GEORGIA DOT RESEARCH PROJECT 18-24 FINAL REPORT ANALYSIS OF THE GEORGIA ADD-ON TO THE 20162017 NATIONAL HOUSEHOLD TRAVEL SURVEY OFFICE OF PERFORMANCE-BASED MANAGEMENT AND RESEARCH 600 WEST PEACHTREE STREET NW ATLANTA, GA 30308 1. Report No.: 2. Government Accession No.: FHWA-GA-21-1824 N/A 4. Title and Subtitle: Analysis of the Georgia Add-On to the 20162017 National Household Travel Survey 7. Author(s): Gwen Kash, Ph.D. (https://orcid.org/0000-0002-6177-3491); Patricia L. Mokhtarian, Ph.D. (https://orcid.org/0000-0003-11190041); Giovanni Circella, Ph.D. (https://orcid.org/0000-0003-1832-396X) 3. Recipient's Catalog No.: N/A 5. Report Date: August 2021 6. Performing Organization Code: N/A 8. Performing Organization Report No.: 18-24 9. Performing Organization Name and Address: 10. Work Unit No.: School of Civil and Environmental Engineering N/A Georgia Institute of Technology 11. Contract or Grant No.: 790 Atlantic Ave NW PI#: 0015718 Atlanta, GA 30318 Phone: (404) 385-1443 Email: patmokh@gatech.edu 12. Sponsoring Agency Name and Address: 13. Type of Report and Period Covered: Georgia Department of Transportation Final; August 2018 August 2021 Office of Performance-based Management and Research 600 West Peachtree Street NW Atlanta, GA 30308 14. Sponsoring Agency Code: N/A 15. Supplementary Notes: Prepared in cooperation with the U.S. Department of Transportation, Federal Highway Administration. 16. Abstract: Through an extensive analysis of the Georgia subsample of the 20162017 National Household Travel Survey, this report provides an in-depth snapshot of the travel behavior of Georgians of all ages. It documents differences in travel needs and behavior by region and between demographic groups, focuses on measurement challenges and improved techniques, and identifies areas where future data collection is needed. In addition to an overview of key travel trends in the state, the report includes chapters on work travel; work flexibility (teleworking and flexible scheduling); new technologies and services, including alternative-fuel vehicles, shared mobility, and online shopping; social inclusion and equity; nonmotorized and access/egress travel; and travel for its own sake. 17. Keywords: Travel behavior, travel patterns, vehicle ownership, commuting, work travel, teleworking, flexible work schedules, online shopping, shared mobility, new technology, alternative-fuel vehicles, nonmotorized travel, social inclusion, equity, travel for its own sake 18. Distribution Statement: No Restriction 19. Security Classification (of this report): Unclassified 20. Security Classification (of this page): Unclassified Form DOT 1700.7 (8-69) 21. No. of Pages: 543 22. Price: Free GDOT Research Project No. 18-24 / T.O. 2014-63 Final Report ANALYSIS OF THE GEORGIA ADD-ON TO THE 20162017 NATIONAL HOUSEHOLD TRAVEL SURVEY By Gwen Kash Postdoctoral Fellow Patricia L. Mokhtarian Susan G. and Christopher D. Pappas Professor and Giovanni Circella Senior Research Engineer Georgia Tech Research Corporation Georgia Institute of Technology Contract with Georgia Department of Transportation in cooperation with U.S. Department of Transportation Federal Highway Administration August 2021 The contents of this report reflect the views of the authors, who are responsible for the facts and the accuracy of the data presented herein. The contents do not necessarily reflect the official views or policies of the Georgia Department of Transportation or the Federal Highway Administration. This report does not constitute a standard, specification, or regulation. ii Symbol in ft yd mi in2 ft2 yd2 ac mi2 fl oz gal ft3 yd3 oz lb T oF fc fl lbf lbf/in2 Symbol mm m m km mm2 m2 m2 ha km2 mL L m3 m3 g kg Mg (or "t") oC lx cd/m2 N kPa SI* (MODERN METRIC) CONVERSION FACTORS APPROXIMATE CONVERSIONS TO SI UNITS When You Know Multiply By To Find LENGTH inches 25.4 millimeters feet 0.305 meters yards 0.914 meters miles 1.61 kilometers AREA square inches 645.2 square millimeters square feet 0.093 square meters square yard 0.836 square meters acres 0.405 hectares square miles 2.59 square kilometers VOLUME fluid ounces 29.57 milliliters gallons 3.785 liters cubic feet 0.028 cubic meters cubic yards 0.765 cubic meters NOTE: volumes greater than 1000 L shall be shown in m3 MASS ounces 28.35 grams pounds 0.454 kilograms short tons (2000 lb) 0.907 megagrams (or "metric ton") TEMPERATURE (exact degrees) Fahrenheit 5 (F-32)/9 Celsius or (F-32)/1.8 ILLUMINATION foot-candles foot-Lamberts 10.76 3.426 lux candela/m2 FORCE and PRESSURE or STRESS poundforce 4.45 newtons poundforce per square inch 6.89 kilopascals APPROXIMATE CONVERSIONS FROM SI UNITS When You Know Multiply By To Find LENGTH millimeters 0.039 inches meters 3.28 feet meters 1.09 yards kilometers 0.621 miles AREA square millimeters 0.0016 square inches square meters 10.764 square feet square meters 1.195 square yards hectares 2.47 acres square kilometers 0.386 square miles VOLUME milliliters 0.034 fluid ounces liters 0.264 gallons cubic meters 35.314 cubic feet cubic meters 1.307 cubic yards MASS grams 0.035 ounces kilograms 2.202 pounds megagrams (or "metric ton") 1.103 short tons (2000 lb) TEMPERATURE (exact degrees) Celsius 1.8C+32 Fahrenheit ILLUMINATION lux candela/m2 0.0929 0.2919 foot-candles foot-Lamberts FORCE and PRESSURE or STRESS newtons 0.225 poundforce kilopascals 0.145 poundforce per square inch Symbol mm m m km mm2 m2 m2 ha km2 mL L m3 m3 g kg Mg (or "t") oC lx cd/m2 N kPa Symbol in ft yd mi in2 ft2 yd2 ac mi2 fl oz gal ft3 yd3 oz lb T oF fc fl lbf lbf/in2 iii TABLE OF CONTENTS EXECUTIVE SUMMARY ...........................................................................................................1 About this Report........................................................................................................................ 2 Key Messages ............................................................................................................................. 4 Challenges and Recommendations ............................................................................................. 7 Leveraging NHTS Data to Understand Complex Travel Patterns ......................................... 7 Mainstreaming Equity............................................................................................................. 8 Assisting Georgians with Mobility Impairments..................................................................... 8 Achieving Sustainable Mobility for Low-income, Carless, and Vehicle-deficit Georgians ... 9 Promoting Transit, Walking, and Biking by Addressing Georgians' Concerns................... 10 Supporting Captive Transit Users ........................................................................................ 10 Improving Pedestrian and Cyclist Safety in Every Neighborhood ....................................... 10 Leveraging New Technologies and Services to Make Transportation More Effective and Equitable ........................................................................................................................... 11 Measuring Work Travel ........................................................................................................ 11 Identifying Data Needs ......................................................................................................... 12 CHAPTER 1. KEY TRAVEL PATTERNS IN THE STATE OF GEORGIA ......................13 Chapter 1 Summary ............................................................................................................... 13 Methodological Notes............................................................................................................... 15 Weighting .............................................................................................................................. 16 Adjustments to Data .............................................................................................................. 17 Assorted Definitions and Notes............................................................................................. 18 Key Terms ............................................................................................................................. 18 Geographic Divisions for Analysis ....................................................................................... 20 Overview................................................................................................................................... 25 Trip Purpose Overview ......................................................................................................... 31 Mode Overview ..................................................................................................................... 34 Trip Patterns by Location of Travel.......................................................................................... 38 Household and Personal Mobility............................................................................................. 40 Trip Purpose.............................................................................................................................. 45 Vehicle Availability and Usage ................................................................................................ 58 Vehicle Ownership ................................................................................................................ 58 Vehicle Usage ....................................................................................................................... 64 Vehicle Fleet Characteristics................................................................................................ 65 Transit Preferences and Use ..................................................................................................... 68 Transit Availability and Use ................................................................................................. 68 iv Transit Service Preferences Among Workers ....................................................................... 69 Summary of Findings on Geographical Differences ................................................................ 77 Work Travel (Chapter 2Chapter 3)..................................................................................... 78 New Technology (Chapter 4) ................................................................................................ 80 Equity (Chapter 5) ................................................................................................................ 80 Nonmotorized Travel and Travel for its Own Sake (Chapter 6Chapter 7)......................... 81 CHAPTER 2. COMMUTE AND WORK PATTERNS ...........................................................83 Chapter 2 Summary ............................................................................................................... 83 Overview................................................................................................................................... 85 Definitions and Technical Notes ........................................................................................... 85 Worker Characteristics ......................................................................................................... 87 Overview of Commuters........................................................................................................ 93 Commute Mode by Person .................................................................................................... 98 Overview of Work Journeys ................................................................................................ 101 Defining the Commute............................................................................................................ 109 Purpose-based and Place-based Anchors........................................................................... 113 Measuring Complex Commutes: The Last-leg and Counterfactual Methods..................... 116 Methods for Definition and Measurement of Commutes .................................................... 119 Results ................................................................................................................................. 130 Summary of Methods Implications for Upcoming Report Sections .................................... 134 Work Travel Distance ............................................................................................................. 136 Region ................................................................................................................................. 138 Peak, Off-peak, and Weekend Travel.................................................................................. 142 Demographic Differences ................................................................................................... 146 Work Travel Distance by Mode.............................................................................................. 150 Commute Duration and Burden .............................................................................................. 152 Commute Times by Time of Day and MPO Tier ................................................................. 154 Commute Times by Mode and Vehicle Ownership ............................................................. 156 Demographic Differences in Commute Duration ............................................................... 158 Total Daily Commute Burden ............................................................................................. 159 Conclusion .............................................................................................................................. 162 CHAPTER 3. TELEWORKING AND WORKER SCHEDULE FLEXIBILITY .............165 Chapter 3 Summary ............................................................................................................. 165 Overview................................................................................................................................. 166 Types of Work Flexibility .................................................................................................... 166 Definitions and Technical Notes ......................................................................................... 168 Overview of Travel Day Telecommuting ............................................................................ 173 v Work Flexibility for Whom? .................................................................................................. 176 Descriptive Statistics on Access to Flexible Schedule and Work Location ........................ 176 Logistic Regression Analysis of Eligibility for Flexible Work Schedule ............................ 186 Logistic Regression Analysis of Eligibility for Flexible Work Location............................. 188 The Effects of Flexible Scheduling ........................................................................................ 192 Models of Work Arrival and Departure Times ................................................................... 200 Frequency of Teleworking...................................................................................................... 202 Telecommuting Frequency among Eligible Workers.......................................................... 202 Telecommuting Frequency and Distance to Work .............................................................. 209 Travel Day Telecommuting ................................................................................................. 213 CHAPTER 4. NEW TECHNOLOGIES AND SERVICES ...................................................219 Chapter 4 Summary ............................................................................................................. 219 Alternative-fuel Vehicles ........................................................................................................ 220 Shared Mobility ...................................................................................................................... 228 Methods ............................................................................................................................... 229 Shared Vehicles: Bikesharing and Carsharing................................................................... 231 Ridehailing .......................................................................................................................... 235 Ridehailing and Vehicle-for-Hire Trips.............................................................................. 239 Online Shopping ..................................................................................................................... 241 Online Shopping and Travel Behavior ............................................................................... 251 CHAPTER 5. SOCIAL INCLUSION AND EQUITY............................................................253 Chapter 5 Summary ............................................................................................................. 253 Overview................................................................................................................................. 255 Mainstreaming Equity......................................................................................................... 258 Key Equity Findings............................................................................................................ 259 Choice of Equity Indicators ................................................................................................ 269 Key Equitable Mobility Indicators ......................................................................................... 270 Vehicle Access........................................................................................................................ 279 Captive Travel..................................................................................................................... 279 Intra-Household Vehicle Allocation: Who Gets the Family Car? ...................................... 286 How Much and What For: The Interrelated Effects of Gender and Age on Mobility and Trip Purpose ........................................................................................................................ 294 Background ......................................................................................................................... 296 Methods ............................................................................................................................... 298 Results ................................................................................................................................. 304 Trip Purpose ....................................................................................................................... 307 vi Discussion ........................................................................................................................... 315 A Compound Gendered Lens .............................................................................................. 321 Health and Disability .............................................................................................................. 322 Health and Physical Activity among Disabled and Nondisabled Georgians ..................... 324 Mobility Differences between Georgians with and without Mobility Impairments ............ 329 Children with Mobility Impairments................................................................................... 332 Variations among Adults with Mobility Impairments ......................................................... 333 Risk Factors for Immobility among Adults with Mobility Impairments.............................. 342 CHAPTER 6. NONMOTORIZED TRANSPORTATION AND ACCESS/EGRESS TRAVEL .....................................................................................................................................347 Chapter 6 Summary ............................................................................................................. 347 Overview................................................................................................................................. 349 Barriers to Walking and Biking More Frequently .............................................................. 352 Physical Activity.................................................................................................................. 355 Access and Egress Travel ....................................................................................................... 361 Measurement of Access/Egress Travel ............................................................................... 362 NHTS Access/Egress Questions and Instructions............................................................... 366 Access/Egress Travel by Georgia Residents ....................................................................... 367 Travel Day Walking and Biking by Georgia Adults .............................................................. 377 Time of Day ......................................................................................................................... 385 Captive and Choice Nonmotorized Travel ............................................................................. 390 Defining Captive Travel...................................................................................................... 390 Model Structure .................................................................................................................. 392 Results ................................................................................................................................. 394 Children's Nonmotorized and School Travel ......................................................................... 401 Identifying School Trips ...................................................................................................... 401 Children's School Travel .................................................................................................... 406 Children's Nonmotorized Travel for All Purposes ............................................................. 409 Frequency of Nonmotorized Travel by Children ................................................................ 414 CHAPTER 7. TRAVEL FOR ITS OWN SAKE ...................................................................421 Chapter 7 Summary ............................................................................................................. 421 Overview and Methods ........................................................................................................... 421 Challenges of Identifying and Measuring TFIOS ............................................................... 423 Measurement of Loop Trips and TFIOS in the 2017 and 2009 NHTS ............................... 426 Trip Purpose for Loop Trips ............................................................................................... 428 Loop Trips in the 2017 Georgia NHTS Subsample................................................................ 431 vii APPENDIX: SAMPLE SIZE TABLES ...................................................................................439 Household and Vehicle Sample Tables .................................................................................. 442 Person Sample Tables ............................................................................................................. 450 Trip, Work Journey, and Access/Egress Travel Sample Tables............................................. 477 ACKNOWLEDGMENTS .........................................................................................................515 REFERENCES ...........................................................................................................................516 viii LIST OF FIGURES Figure 1. Map. Georgia MPO map. ...............................................................................................23 Figure 2. Classification. Urbanicity classifications. ......................................................................24 Figure 3. Stacked bar graph. Person trip purpose by MPO tier .....................................................34 Figure 4. Bar graph. Age of population with mobility impairments. ............................................44 Figure 5. Pie chart. Vehicle ownership among Georgia households. ............................................58 Figure 6. Bar graph. Vehicle sufficiency in Georgia households. .................................................60 Figure 7. Histogram. Distribution of annual household miles driven............................................64 Figure 8. Bar graph. Educational attainment of Georgia workers by gender, MPO tier, household income, and race. ..................................................................................................91 Figure 9. Histogram. Work journeys by time of arrival to or departure from work (weighted). ............................................................................................................................103 Figure 10. Stacked bar graph. Weekday trips in progress by time of day (all ages, weighted)...108 Figure 11. Stacked bar graph. Weekday vehicle trips in progress by time of day (weighted). ...109 Figure 12. Pie chart. Proportions of traditional and nontraditional commutes in Georgia. .........112 Figure 13. Diagram. Trip sequence with multiple potential commute patterns...........................115 Figure 14. Diagram. Alternate methods for identifying commutes within complex work journeys. ...............................................................................................................................117 Figure 15. Diagram. Sequence of trips including a provisional work anchor. ............................122 Figure 16. Diagram. Example work journey classifications of trip sequences............................123 Figure 17. Maps. Example journey with a negative complexity increment. ...............................128 Figure 18. Pie chart. Telework category of Georgia workers and telecommuting frequency, past 30 days. .........................................................................................................................170 Figure 19. Pie chart. Travel day working and teleworking among Georgia adults. ....................172 Figure 20. Bar graphs. Commute schedules of workers with and without flexible working times. ....................................................................................................................................192 Figure 21. Histograms. Arrival times at work for service and nonservice workers by schedule flexibility. ..............................................................................................................195 Figure 22. Histograms. Departure times from work for service and nonservice workers by schedule flexibility. ..............................................................................................................196 Figure 23. Histograms. Arrival times at work for caregivers by schedule flexibility and gender. ..................................................................................................................................198 Figure 24. Histograms. Departure times from work for caregivers by schedule flexibility and gender. ..................................................................................................................................199 ix Figure 25. Histograms. Annual miles driven by AFVs, all vehicles, and recent vehicles...........224 Figure 26. Pie chart. Number of online shoppers per household, past 30 days. ..........................243 Figure 27. Pie chart. Household online shopping by household size, past 30 days.....................247 Figure 28. Bar graph. Percent of adults with annual household income less than $25,000, by race, gender, and mobility impairment.................................................................................257 Figure 29. Histogram. Average marginal effect of being female on probability of being housebound, by age group (percentage points). ...................................................................306 Figure 30. Histogram. Probability of being housebound by gender and age...............................307 Figure 31. Bar graph. Average marginal effect of gender on probability of making trips of various purposes on travel day (in percentage points). ........................................................312 Figure 32. Histograms. Average marginal effects of gender by age (in percentage points)........313 Figure 33. 3D histogram. Predicted probability of making a trip to transport others by gender and age. .................................................................................................................................314 Figure 34. Stacked bar graph. Annual household income of Georgians with and without mobility impairments............................................................................................................324 Figure 35. Bar graph. Health of nondisabled Georgia adults by income and level of physical activity in a typical week. .....................................................................................................328 Figure 36. Bar graph. Mobility aids used by adults with mobility impairments. ........................336 Figure 37. Pie chart. Use of nonmotorized modes by Georgia adults (past 7 days). ...................349 Figure 38. Bar graphs. Gender differences in perceived barriers to walking and biking more frequently among pedestrians and cyclists ages 18+............................................................354 Figure 39. Bar graphs. Weekday and weekend NMT by time of day. ........................................386 Figure 40. Diagram. Example child's travel itinerary with trips to and from school. .................402 Figure 41. Instructions. Excerpt from 2009 NHTS travel diary instructions...............................426 Figure 42. Instructions. Excerpt from 2017 NHTS travel diary instructions...............................427 x LIST OF TABLES Table 1. NHTS MPO classification by county. .............................................................................21 Table 2. Classification of sample households by MPO tier and urbanicity...................................25 Table 3. Summary demographic statistics. ....................................................................................26 Table 4. Major travel indicators by MPO tier................................................................................27 Table 5. Summary demographic and travel statistics by MPO tier. ..............................................28 Table 6. Summary demographic and travel statistics by MPO......................................................28 Table 7. Summary statistics per household, daily..........................................................................29 Table 8. Summary statistics per capita, daily. ...............................................................................30 Table 9. Trip distance and percent of population immobile. .........................................................31 Table 10. Trip purpose by destination: Types and frequencies. ....................................................33 Table 11. Person trips by mode (detailed). ....................................................................................35 Table 12. Mode share and trip distance by MPO tier. ...................................................................36 Table 13. Annual trips and vehicle miles by Georgians and within Georgia. ...............................38 Table 14. Annual person trips within and between MPO tiers (thousands). .................................40 Table 15. Annual vehicle trips within and between MPO tiers (thousands). ................................40 Table 16. Daily household travel by demographic groups. ...........................................................42 Table 17. Daily per capita trips and immobility (persons ages 5+). ..............................................43 Table 18. Trip distance and daily PMT and VMT per person ages 5+..........................................46 Table 19. Annual person trips by purpose and income..................................................................48 Table 20. Gender and racial differences in annual person trips by purpose. .................................51 Table 21. Frequency of trip purposes by age and disability. .........................................................52 Table 22. Annual tripmaking by purpose, gender, and age. ..........................................................54 Table 23. Trip purpose by mode of travel......................................................................................56 Table 24. Mode of travel by trip purpose. .....................................................................................57 Table 25. Vehicle sufficiency by income, race, and number of household vehicles. ....................61 Table 26. Vehicle purchasing behavior by household income. .....................................................62 Table 27. Characteristics of vehicles owned by households of different income levels. ..............63 Table 28. Miles driven per vehicle by household vehicle sufficiency...........................................65 Table 29. Vehicle characteristics by vehicle age cohort................................................................66 Table 30. Characteristics of seldom-used vehicles. .......................................................................67 xi Table 31. Ownership of seldom-used vehicles by annual household income. ..............................68 Table 32. County transit funding and transit usage by MPO.........................................................70 Table 33. County demographics and transit usage by transit funding category. ...........................71 Table 34. Perceived most important factors to make transit a good commute option by geography, vehicle ownership, and transit access/use............................................................73 Table 35. Transit service quality priorities of Georgia workers by demographic group. ..............74 Table 36. Logistic regression: Transit service quality priorities of Georgia workers. ..................76 Table 37. Workforce participation. ................................................................................................88 Table 38. Job categories of Georgia workers.................................................................................89 Table 39. Job categories of Georgia workers by educational attainment. .....................................92 Table 40. Active commuter rates. .................................................................................................94 Table 41. Work journey complexity among active commuters. ...................................................96 Table 42. Number of work journeys per person, active commuters. .............................................97 Table 43. Usual commute mode as reported by NHTS-defined workers. .....................................99 Table 44. Travel day commute mode(s) for active commuters. ..................................................100 Table 45. Use of single-occupancy vehicles, high-occupancy vehicles (HOVs), and alternate modes among active commuters. .........................................................................................102 Table 46. Total annual work journeys in Georgia. ......................................................................104 Table 47. Work journeys by MPO tier (all days).........................................................................105 Table 48. Work journeys by MPO tier (weekdays only). ............................................................106 Table 49. Work journey mode, vehicle occupancy status, and time of day by MPO tier (percentages).........................................................................................................................107 Table 50. Initial anchor classifications based on location and primary activity at location. .......121 Table 51. Duration and purposes of stops in work journeys........................................................125 Table 52. Counterfactual match status of complex work journeys in sample. ............................126 Table 53. Weighted OLS model of work journey complexity increment distance in miles........131 Table 54. Predicted complexity increments.................................................................................132 Table 55. Comparison between full work journey, counterfactual, and last-leg commute measurements. ......................................................................................................................133 Table 56. Distance to work, commute PMT, and work journey PMT.........................................137 Table 57. Average work journey and commute distance by MPO tier........................................138 Table 58. Percent of statewide work-related miles traveled by MPO tier. ..................................140 Table 59. Annual work-related PMT and VMT by MPO tier (millions of miles).......................141 Table 60. Annual work-related PMT and VMT per worker by MPO tier (miles).......................141 xii Table 61. Work travel distance by time of day. ...........................................................................144 Table 62. Work travel distance by time of day and MPO tier. ....................................................145 Table 63. Average work travel distance by demographic and employment characteristics. .......147 Table 64. Average annual work travel distance per worker by demographic and employment characteristics. ......................................................................................................................148 Table 65. Percent of total work travel distance by demographic and employment characteristics. ......................................................................................................................149 Table 66. Work travel distance by mode. ....................................................................................151 Table 67. Work-related travel duration in minutes. .....................................................................154 Table 68. Travel day commute and work journey duration by time of day and MPO tier..........155 Table 69. Work travel duration by mode and vehicle ownership. ...............................................157 Table 70. Work travel duration by MPO tier, gender, education, and age. .................................160 Table 71. Work travel duration by income, race, occupation, and worker type. .........................161 Table 72. Total daily commute burden by MPO tier, gender, education, and age. .....................163 Table 73. Total daily commute burden by income, race, occupation, and worker type. .............164 Table 74. Eligibility for job flexibility among Georgia workers. ................................................167 Table 75. Travel day working and teleworking by MPO tier. .....................................................174 Table 76. Travel day teleworking and conventional commuting among active workers. ...........175 Table 77. Employment flexibility by MPO, job characteristics, educational attainment, and income. .................................................................................................................................177 Table 78. Employment flexibility by sex, caregiver status, age, race, and mobility impairment. ...........................................................................................................................179 Table 79. Employment flexibility by caregiver status. ................................................................180 Table 80. Employment flexibility for workers with a 4-year college degree by location and job characteristics. ................................................................................................................182 Table 81. Employment flexibility for workers without a 4-year college degree by location and job characteristics. .........................................................................................................183 Table 82. Employment flexibility among workers with a 4-year college degree by sex, caregiver status, age, race, and mobility impairment. ..........................................................184 Table 83. Employment flexibility among workers without a 4-year college degree by sex, caregiver status, age, race, and mobility impairment. ..........................................................185 Table 84. Logistic regressions: Eligibility for flexible work schedule among Georgia workers. ................................................................................................................................187 Table 85. Teleworking status of Georgia workers (descriptive statistics)...................................190 Table 86. Multinomial logistic regression: Teleworking status of Georgia workers. ................191 xiii Table 87. Median work arrival and departure times by schedule flexibility, MPO tier, education, and employment characteristics. .........................................................................194 Table 88. Median work arrival and departure times by schedule flexibility, demographic characteristics, and caregiver status. ....................................................................................197 Table 89. Linear regression: Work arrival and departure times for travel day commutes. .........201 Table 90. Average number of days telecommuted in the past 30 days among eligible workers by MPO tier, sex, caregiver status, age, race, and medical condition. .................................204 Table 91. Telecommuting 30-day frequency among eligible workers by MPO tier, sex, caregiver status, age, race, and medical condition. ..............................................................205 Table 92. Average days telecommuted in the past 30 days among eligible workers by employment characteristics. .................................................................................................206 Table 93. Telecommuting 30-day frequency among eligible workers by employment characteristics. ......................................................................................................................207 Table 94. Average days telecommuted in the past 30 days among eligible workers by caregiver status. ....................................................................................................................208 Table 95. Telecommuting 30-day frequency among eligible workers by caregiver status. ........209 Table 96. Distance to work in miles by MPO tier. ......................................................................210 Table 97. Average days telecommuted in the past 30 days among eligible workers by distance to work and MPO tier. ............................................................................................211 Table 98. Telecommuting 30-day frequency among eligible workers by distance to work and MPO tier. ..............................................................................................................................212 Table 99. Travel day commuting and telecommuting among workers ages 18+ by MPO tier, sex, caregiver status, age, race, and medical condition. .......................................................214 Table 100. Travel day commuting and telecommuting among workers ages 18+ by employment characteristics. .................................................................................................216 Table 101. Travel day commuting and telecommuting among workers ages 18+ by caregiver status .....................................................................................................................................217 Table 102. Travel day commuting and telecommuting among workers ages 18+ by distance to work and MPO tier. ..........................................................................................................218 Table 103. Number of alternative-fuel vehicles by MPO tier. ....................................................220 Table 104. Fuel type of alternative- and conventional-fuel vehicles...........................................221 Table 105. Vehicle characteristics: AFVs, all vehicles, and recent vehicles...............................223 Table 106. Annual miles driven by fuel type...............................................................................225 Table 107. Household characteristics for AFVs, all vehicles, and recent vehicles. ....................227 Table 108. Main driver characteristics for AFVs, all vehicles, and recent vehicles....................228 Table 109. Shared mobility questions in the NHTS. ...................................................................230 xiv Table 110. Demographics of bikeshare and carshare users ages 16+..........................................232 Table 111. Frequency of carshare and bikeshare use among users ages 16+, past 30 days. .......234 Table 112. Ridehailing use among Georgians ages 16+..............................................................237 Table 113. Demographics of ridehailing users (column percentages).........................................238 Table 114. Unweighted sample sizes for ridehailing and vehicle-for-hire trips by Georgians ages 16+................................................................................................................................239 Table 115. Monthly and annualized estimates of ridehailing and vehicle-for-hire trips by Georgians ages 16+. .............................................................................................................240 Table 116. Online orders by MPO tier, past 30 days...................................................................241 Table 117. Household online shopping by MPO tier, annual household income, vehicle ownership, and race, past 30 days. .......................................................................................244 Table 118. Household online shopping by household composition. ...........................................246 Table 119. Online shopping by persons ages 16+, past 30 days. ................................................249 Table 120. Which household members placed online orders in the last 30 days for people ages 16+................................................................................................................................250 Table 121. Percent of persons ages 16+ who made a trip to buy goods on the travel day. .........251 Table 122. Demographic characteristics by disability and national origin..................................274 Table 123. Demographics of nonimmigrant and immigrant Georgians by level of education....275 Table 124. Key mobility indicators for Georgia adults by geography, race, education, income, vehicle ownership, gender, and disability...............................................................276 Table 125. Mobility indicators by national origin and educational attainment among Georgia adults.....................................................................................................................................277 Table 126. Driver status and vehicle ownership by Georgia adults of different demographic characteristics. ......................................................................................................................278 Table 127. Purpose by mode for captive, choice, and POV trips (weighted percent and unweighted sample sizes). ....................................................................................................281 Table 128. Mean duration and distance of captive, choice, and POV trips. ................................282 Table 129. Linear regressions: Effect of captive travel on duration of transit and nonmotorized trips................................................................................................................283 Table 130. Linear regression: Duration in minutes of trips by POV and transit. ........................285 Table 131. Weighted sample characteristics of drivers by household vehicle ownership...........287 Table 132. Unweighted sample characteristics of drivers by household vehicle ownership.......288 Table 133. Percent of drivers listed as the "main" driver for a household vehicle......................290 Table 134. Logistic regression: Vehicle allocation within vehicle-deficit households. ..............291 Table 135. Percent of drivers who are listed as main driver for newest household vehicle. .......292 xv Table 136. Logistic regression: Intra-household allocation of new vehicles...............................293 Table 137. Percent of teen drivers who are the main driver for household vehicles, by sex.......294 Table 138. Sample characteristics................................................................................................299 Table 139. Annual trips per capita by purpose, gender, and age (weighted)...............................302 Table 140. Binomial probit model of likelihood of being chronically housebound (having made no trips within the past 7 days). ..................................................................................305 Table 141. Multivariate probit model of likelihood of making trips for various purposes. ........308 Table 142. Summary statistics and indicators for multivariate probit model. .............................310 Table 143. Health and age of Georgia adults with and without mobility impairments. ..............325 Table 144. Driver status and income of Georgians with and without mobility impairments, by health and age. .................................................................................................................326 Table 145. Georgians in poor or fair health by income, age, and disability. ...............................327 Table 146. Walking and physical activity among Georgians with and without mobility impairments, by health and age. ...........................................................................................330 Table 147. Immobility by disability, health, and age. .................................................................331 Table 148. Average daily trips by disability, health, and age......................................................331 Table 149. Behavioral adaptations among disabled and older adults. .........................................332 Table 150. Comparison of children (ages 517) with and without mobility impairments. .........333 Table 151. Demographic characteristics of Georgians with mobility impairments. ...................335 Table 152. Driver status and income of Georgia adults with mobility impairments. ..................339 Table 153. Immobility and average daily trips of Georgians with mobility impairments...........340 Table 154. Walking and physical activity among Georgians with mobility impairments...........341 Table 155. Logistic regression: Immobility among Georgia adults with a mobility impairment. ...........................................................................................................................344 Table 156. Selected average marginal effects on probability of being immobile among adults with a mobility impairment, percentage points. ...................................................................345 Table 157. Walking and biking among Georgia adults (past 7 days). .........................................351 Table 158. Perceived barriers to walking and biking more frequently (pedestrians and cyclists ages 18+). ................................................................................................................353 Table 159. Physical activity among Georgia adults.....................................................................356 Table 160. Usual physical activity of Georgia adults by walking and biking behavior (past 7 days). .................................................................................................................................359 Table 161. Total annual trips and access/egress legs by nonmotorized modes. ..........................361 Table 162. Mode share of walking and biking trips and legs. .....................................................363 Table 163. Instances of walking and biking per person ages 5+. ................................................364 xvi Table 164. Access and egress legs by mode and parent mode (unweighted). .............................365 Table 165. Mode share of legs to access/egress public transit and other modes. ........................368 Table 166. Mean duration in minutes of access/egress legs by mode. ........................................369 Table 167. Mode share of public transit access/egress legs by demographic characteristics (row percentages). ................................................................................................................370 Table 168. Unweighted sample sizes of public transit access/egress legs...................................373 Table 169. Duration in minutes of public transit access/egress legs by demographic characteristics. ......................................................................................................................376 Table 170. Nonmotorized trips/legs per capita and duration of NMT trips/legs. ........................378 Table 171. Mode and purpose of NMT trips/legs by demographic factors. ................................381 Table 172. Mean duration in minutes of leisure and instrumental NMT trips/legs by Georgia adults.....................................................................................................................................383 Table 173. Purpose and duration of NMT trips/legs....................................................................384 Table 174. Unweighted trips/legs by mode and purpose. ............................................................385 Table 175. Weighted NMT trips/legs by time of day ..................................................................387 Table 176. Unweighted NMT trips/legs by time of day. .............................................................388 Table 177. Time of day of NMT trips/legs by day and MPO tier. ..............................................389 Table 178. Vehicle sufficiency of Georgia households and adults..............................................392 Table 179. Duration, quantity, and purpose of nonmotorized travel for captive and choice travelers. ...............................................................................................................................395 Table 180. Linear regression: Daily total minutes of instrumental walking and biking..............398 Table 181. Linear regression: Daily total minutes of leisure walking and biking. .....................399 Table 182. Mode of travel-day trips to and from school by children ages 517 (unweighted)...405 Table 183. Usual mode of travel to and from school by children ages 517 (unweighted). .......405 Table 184. School travel: Mode split of trips and usual mode(s). ...............................................406 Table 185. Walking and biking as percentage of school travel by geography, age, driver status, sex, race, income, and household vehicles. ...............................................................408 Table 186. Purpose of children's nonmotorized trips (weighted and unweighted). ....................411 Table 187. Purpose of children's nonmotorized trips by geography, age, driver status, sex, race, income, and household vehicle ownership. .................................................................413 Table 188. Nonmotorized travel of children ages 517 (past 7 days). ........................................415 Table 189. Percentage of children who have walked or biked in the past 7 days and average number of trips among mode users. .....................................................................................417 Table 190. Nonmotorized travel on the travel day: Percent of trips/legs and percent of children. ................................................................................................................................418 xvii Table 191. Percent of children who walked/biked on the travel day and minutes of nonmotorized travel..............................................................................................................419 Table 192. Destination purpose of loop trips made by Georgians ages 5+ (unweighted). ..........429 Table 193. Loop trips as a percent of total trips by mode and purpose (Georgians ages 5+)......432 Table 194. Loop trips per capita in 2017 (ages 5+). ....................................................................433 Table 195. Mode and purpose of loop trips by Georgians ages 5+. ............................................434 Table 196. Mode, purpose, distance, and duration of loop and nonloop trips.............................436 Table 197. Distance and duration of loop and nonloop trips for fitness and recreation. .............438 Table 198. Overview of Georgia NHTS population types and samples used in this report. .......441 Table 199. Overview of household and vehicle subsamples. ......................................................442 Table 200. Household sample table. ............................................................................................443 Table 201. Sample table: Vehicles owned by Georgia households. ............................................447 Table 202. Overview of person subsamples. ...............................................................................450 Table 203. Person sample table: All persons ages 5+..................................................................451 Table 204. Person sample table: All adults ages 18+. .................................................................457 Table 205. Person sample table: All workers ages 18+ (demographic characteristics)...............463 Table 206. Person sample table: All workers ages 18+ (job and commute characteristics)........469 Table 207. Person sample table: Active commuters ages 18+. ...................................................471 Table 208. Person sample table: Persons ages 16+. ....................................................................473 Table 209. Person sample table: Workers ages 16+. ...................................................................475 Table 210. Summary of traveler populations and instances of travel..........................................478 Table 211. Overview of trip, work journey, and access/egress travel subsamples......................479 Table 212. Trip sample table: All trips by all persons ages 5+ by location.................................480 Table 213. Trip sample table: All trips by all persons ages 5+ by purpose.................................482 Table 214. Trip sample table: All trips by all persons ages 5+ by mode.....................................483 Table 215. Trip sample table: All trips by all persons ages 5+ by demographic factors.............484 Table 216. Trip sample table: All vehicle trips by all persons ages 5+ by location. ...................488 Table 217. Trip sample table: All vehicle trips by all persons ages 5+ by purpose. ...................490 Table 218. Trip sample table: All vehicle trips by all persons ages 5+ by demographic factors. ..................................................................................................................................491 Table 219. Trip sample table: All trips by all persons ages 18+ by location...............................496 Table 220. Trip sample table: All trips by all persons ages 18+ by purpose...............................498 Table 221. Trip sample table: All trips by all persons ages 18+ by mode...................................499 xviii Table 222. Trip sample table: All trips by all persons ages 18+ by demographic factors...........500 Table 223. Trip sample table: All vehicle trips by all persons ages 18+ by location. .................504 Table 224. Trip sample table: All vehicle trips by all persons ages 18+ by purpose. .................506 Table 225. Trip sample table: All vehicle trips by all persons ages 18+ by demographic factors. ..................................................................................................................................507 Table 226. Work journey/commute sample table: Location, distance, and mode. ......................511 Table 227. Work journey/commute sample table: Stops, driving, and time of day. ...................512 Table 228. Work journey/commute sample table: Traveler characteristics. ...............................513 xix EXECUTIVE SUMMARY Travel in Georgia is changing. Atlanta is growing, changing the balances of trips across the state. Advanced technologies are generating a number of new products and services with direct implications for travel demand in Georgia, including ridehailing services (e.g., Uber and Lyft), vehicle sharing (e.g., carsharing services such as Zipcar, bikesharing, and, more recently, electric scooters), and alternative-fuel vehicles (e.g., hybrid and electric cars). Working from home almost doubled between 2000 and 2010, and even before the coronavirus disease 2019 (COVID-19) pandemic, online shopping had become an increasingly significant part of Georgians' retail behavior. These new options are transforming individuals' and households' travel-related decision-making. They will substantially modify the demand for housing, vehicle sales, the amount of travel by private vehicles, and the resulting gasoline tax revenues and emissions of greenhouse gases and criteria pollutants. The impacts of these changes--positive and negative--do not affect all Georgians equally. There is a risk of exacerbating existing inequality, locking Georgia into a two-tiered travel system-- literally and figuratively leaving people behind. To ensure a more equitable future, it is imperative to examine the needs of those who, through age, disability, economic disadvantage, gender, or race, are restricted in their mobility and access to opportunities. Further, climate change poses risks to Georgians' health and economic wellbeing. The state is preparing to invest more than $1 billion to mitigate the effects of sea level rise.1 Additionally, 48 percent of Georgia's population currently lives in areas at elevated risk of wildfires, which are expected to 1 See https://sealevelrise.org/states/georgia/. 1 increase in frequency and intensity as elevated temperatures and droughts become more prevalent.2 It is, therefore, more urgent than ever to identify and enact sustainable transport solutions. Travel demand forecasting models and transportation policy need to be updated to account for the latest trends, improving their accuracy and equity. This report provides a baseline and guidance for such work through extensive analysis of the Georgia add-on to the 20162017 National Household Travel Survey (NHTS). ABOUT THIS REPORT This report provides an in-depth snapshot of the travel behavior of Georgians of all ages. It documents differences in travel needs and behavior by region and between demographic groups, focuses on measurement challenges and improved techniques, and identifies areas where future data collection is needed. More in-depth summaries and a few key findings are included at the beginning of each individual chapter; a brief synopsis of the key messages is provided in Key Messages of this executive summary. Chapter 1 presents greater detail about the NHTS data and methods, provides an overview of general travel patterns in Georgia, and synthesizes findings about regional differences from throughout the report. Chapter 2 focuses on work travel, including more accurate measurement of complex commutes (i.e., commutes including one or more stops between home and work). One in every four 2 See https://statesatrisk.org/georgia. 2 Georgia commutes is complex. The common practice of using the last trip in the chain as a proxy for commute distance undercounts Georgia's annual commute person miles traveled (PMT) by 2.6 billion miles, about 10 percent of total commute PMT. This chapter presents a new, more accurate method of estimating commute distance. Chapter 3 explores flexible work locations (i.e., teleworking) and schedules. It examines which workers' jobs allow for flexible time and/or location and how often workers take advantage of that flexibility. Chapter 4 discusses new technologies and services, including alternative-fuel vehicles, shared mobility, and online shopping. Chapter 5 focuses on social inclusion and equity. The chapter documents the mobility disadvantages among captive mode users and people with mobility impairments; examines the interrelated effects of gender and age on travel behavior; explores how vehicles are allocated within households; and synthesizes equity findings from throughout the report. Chapter 6 provides a portrait of walking and biking in Georgia. Access and egress travel (i.e., travel to reach another mode of transportation such as public transit) account for a substantial portion of nonmotorized travel (NMT); thus, the chapter discusses how to incorporate this travel into the analysis. It also discusses nonmotorized travel by children and children's school travel by all modes. Chapter 7 discusses travel for its own sake, or the intrinsic value of travel, beyond its utilitarian purpose of getting from A to B. Loop trips (i.e., trips with the same start and end location) are an easily identifiable form of travel for its own sake. This chapter discusses changes in the NHTS's 3 methods of soliciting information about loop trips; provides an overview of the frequency, mode, and purposes of loop trips; and reviews continued measuring challenges. KEY MESSAGES Accounting for complexity is critical. Georgians' travel is often composed of chains rather than individual trips. For example, one in four Georgia commutes is complex (i.e., including one or more stops between home and work), making it important not to underestimate the full extent of commute travel. As teleworking rises, some workers are skipping the commute to work; others are still commuting to the office for part of the day and working an additional shift when they get home. The modes used to access public transit are not always included in mode share figures, which undercounts modes like walking and, to a lesser extent, biking. Decisions about whether, where, and how to travel vary by time of day and day of the week. Accurate measurement and modeling of Georgians' travel habits needs to reflect these sources of complexity. New technologies and services are reshaping Georgians' travel. Less than a decade after the founding of Uber, 1 in 10 Georgians used a ridehailing app at least once in the past 30 days. Ridehailing apps accounted for 87 percent of vehicle-for-hire trips in Georgia, even more in small MPO areas and rural areas. Two thirds of Georgia households had likewise purchased something online within the past 30 days, ranging from 72 percent of Atlanta-region households to 52 percent of households in rural areas. Statewide, 1.9 percent of Georgia's vehicles were hybrid, electric, or powered by another alternative fuel. High-speed internet has facilitated the rise of teleworking. These trends have likely continued or accelerated in the years since the NHTS data were collected. 4 Georgians' mobility and travel habits are bifurcated. Residents of the Atlanta region are more mobile than the national average, while residents of smaller metropolitan planning organization (MPO) areas and rural counties are less mobile. Emerging trends such as telecommuting, flexible work scheduling, alternative-fuel vehicles, and online shopping are more pronounced in the Atlanta region compared to the rest of the state. Ridehailing is also more common in Atlanta but has promise for improving mobility in small towns and rural communities. In addition to these geographic differences, there are wide gaps between wealthy and poor Georgians. Transportation-disadvantaged populations such as the elderly and Georgians with mobility impairments, especially those from poor households, are often stuck at home. Three in ten Georgians do not have full vehicle access. Five percent of Georgians live in households with zero vehicles. An additional 26 percent live in vehicle-deficit households (i.e., households with at least one vehicle but fewer vehicles than potential drivers). On average, Georgians in vehicle-deficit households are more mobile than Georgians from zero-vehicle households, but household members who do not have access to the family car face many of the same barriers to mobility as travelers from carless households. Captive travelers pay a double penalty. Georgia's current travel environment constitutes a two-tiered system divided not just by mode, but by the ability to choose between modes. The lowest-income Georgians (<$15,000 annual household income) and Georgians who live in vehicle-deficit households overall walk, bike, and take transit more than their wealthier counterparts. These captive travelers pay a double time penalty. The first penalty is not having the option to drive, which is the fastest mode in many cases. The second penalty is that the trips 5 of captive transit users are also longer than those of choice transit users. The same is true of captive pedestrians and cyclists. Low-income Georgians pay more for less. Low-income Georgians are less mobile than their wealthier counterparts. Only 43 percent of the lowest-income households (those making less than $15,000 per year) are vehicle sufficient, and nearly one third own no vehicles at all. These households purchase older vehicles with an average of 130,000 miles on the odometer at the time of purchase (see chapter 1). Vehicles near the end of their useful lives are financially more accessible to low-income households due to their lower purchase costs. However, these vehicles cost more to maintain and need to be replaced more frequently. Technological improvements in vehicle efficiency are disproportionately benefitting wealthy households. Walking and biking are easy to undercount. In a typical week, 72.6 percent of Georgians will walk, ride a bike, or both, but this nonmotorized travel can be hidden in the data. In addition to the 950,000 nonmotorized trips Georgians make each year, they walk or bike as a way to access/egress another mode of transportation (e.g., public transit) 260,000 times per year. However, this access/egress travel is considered a part of the mode being accessed and is not included in typical mode share calculations. Incorporating these access/egress legs provides a more complete picture of Georgians' walking and biking. Accuracy and equity go hand in hand. Many of the measurement issues identified in this report disproportionately affect the accuracy of the data for one or more marginalized populations. For example, since more women than men make complex commutes, better accounting for complex commutes can: (1) better capture the full extent of work travel in Georgia by identifying 2.6 billion PMT that would not have been included, and (2) improve 6 measurement of women's travel patterns and avoid underestimating their commute distances. Similarly, incorporating transit access and egress trips into estimates of nonmotorized travel improves estimates of walking and biking across the board, but particularly for transit-dependent Georgians. CHALLENGES AND RECOMMENDATIONS Leveraging NHTS Data to Understand Complex Travel Patterns This report provides analysis and examples of how to effectively leverage NHTS data to more accurately analyze complex travel patterns, as follows: The NHTS contains a rich array of data about Georgians' backgrounds and usual travel habits, as well as a snapshot of their travel. Chapter 1, Methodological Notes provides future analysts with an orientation to NHTS data. There is no one "typical" traveler. To understand Georgians' travel behavior, it is necessary to account for diverse needs and experiences. Disaggregating data is vital for uncovering regional variations and demographic differences in mobility. For many topics, the NHTS contains data about "usual" practices and observed travel on the travel day. Comparing the two can identify nuances. For instance, nonmotorized travel is the "usual" commute mode of just 2.3 percent of Georgians, but 3.7 percent of observed commute trips were by NMT, suggesting that a number of Georgians who usually drive to work sometimes walk or bike (see chapter 1, Household and Personal Mobility). When commuters who walk or bike to public transit are included, the total amount of walking and biking is higher than the initial figure would suggest. 7 In addition to the substantive analysis, the techniques presented here, such as our methods of identifying and measuring work travel (see chapter 2, Defining the Commute), school travel (see chapter 6, Identifying School Trips) or more completely documenting NMT (see chapter 6, Access and Egress Travel), are designed to be useful to future analysts. Where appropriate, the report also discusses the limitations of NHTS data. Identifying Data Needs at the end of this section discusses future data needs. Mainstreaming Equity Many forms of social inequality and exclusion affect Georgians' mobility. Low-income people, older adults, and people with disabilities face the strongest barriers; differences by gender and race were also documented. While these problems come from outside the transportation system, if pre-existing inequalities are not taken into account, transportation professionals can inadvertently exacerbate them, leaving disadvantaged communities even farther behind. To prevent this, transportation agencies should mainstream equity by evaluating differential needs or policy outcomes by gender, race, disability, and other sources of social exclusion as a standard part of planning. One critical first step is disaggregating data to ensure that differences between groups are detected. A valuable approach is to partner with organizations outside of the transportation sector, such as senior services centers, healthcare providers, and community groups, for studying and/or implementing projects to help underserved groups. Assisting Georgians with Mobility Impairments Georgians with mobility impairments have a critical need for transportation. Many people with mobility impairments are impoverished. A mobility impairment does not disqualify someone 8 from employment. However, it is worth noting that for some Georgians with disabilities, the biggest obstacle to getting their foot in the door to a new career might be getting their foot to the door (see chapter 1, Trip Purpose). Current paratransit services are inadequate; only 10 percent of adults with disabilities reported using them, while 70 percent reported reducing day-to-day travel. An assessment of the challenges facing local paratransit systems and the experiences of paratransit users is needed. In addition to improving paratransit service, complementary alternatives such as ridehailing should be explored. Fulton County, for example, recently began offering subsidized ridehailing for elderly residents (see chapter 5, Discussion). Children with mobility impairments may be especially vulnerable. However, the NHTS sample of children with mobility impairments is small; more targeted data collection and study are needed (see chapter 5, Children with Mobility Impairments). Achieving Sustainable Mobility for Low-income, Carless, and Vehicle-deficit Georgians The mobility of low-income, carless, and vehicle-deficit Georgians can be improved by improving the level of service for transit and nonmotorized trips. Ridehailing and car-, bike-, and scooter-sharing services can also help carless people, provided they are affordable and technologically accessible. Community carpooling and microtransit could provide more flexible transportation to Georgians without cars. Helping low-income Georgians acquire cars could have consequences for sustainability and congestion, but automobile access improves mobility, quality of life, and economic 9 opportunities. To maximize the social and environmental benefits of new vehicle technology, policies are needed to help low-income families purchase fuel-efficient and electric vehicles. Promoting Transit, Walking, and Biking by Addressing Georgians' Concerns The NHTS provides data on Georgians' preferences for transit service (see chapter 1, Transit Service Preferences Among Workers) and perceived barriers to walking and biking more frequently (see chapter 6, Barriers to Walking and Biking More Frequently). These data provide actionable information by identifying high-priority issues. Captive and choice travelers sometimes have different service priorities; it is important to balance the needs of both groups. Supporting Captive Transit Users Whatever the causes, transit-dependent travelers are experiencing a worse quality of service for commutes and other trips than their wealthier neighbors. To reduce this inequality, it is therefore important to examine potential discrepancies in vehicle frequency and route density, and to examine whether current transit routings match the needs of low-income commuters (see chapter 5, Key Equitable Mobility Indicators). Improving Pedestrian and Cyclist Safety in Every Neighborhood Unsafe pedestrian and cyclist environments discourage travelers with a choice from walking and biking, and leave captive pedestrians and bicyclists to walk in unsafe environments. Improving the quality of the walking and biking environments encourages more Georgians to choose NMT (see chapter 6, Barriers to Walking and Biking More Frequently). Captive pedestrians and cyclists, who are already walking and biking, need safe infrastructure on the routes they are already using to access their homes, work, shops, schools, and transit stations (see chapter 5, Key Equitable Mobility Indicators, and chapter 6, Captive and Choice Nonmotorized Travel). 10 Walking is also an important mode of transportation for people with mobility impairments, making safety and accessible road design even more important (see chapter 6, Travel Day Walking and Biking by Georgia Adults). Leveraging New Technologies and Services to Make Transportation More Effective and Equitable Teleworking, ridehailing, and online shopping are most common among the most mobile and tech-savvy Georgians, centered in the Atlanta region. Room for growth exists with other segments of the state's population. Teleworking, for example, could benefit workers in rural locations and facilitate employment for Georgians with disabilities. Ridehailing services can improve transportation options in rural areas and increase the mobility of disadvantaged groups. Online shopping can provide homebound or busy low-income Georgians better access to goods at an affordable price. Facilitating internet access and technological literacy is key to all three of these trends. Education on using ridehailing apps and smartphones in general is particularly valuable for older adults (see chapter 5, Discussion). Measuring Work Travel One in four Georgia commutes is complex (i.e., including one or more stops between home and work). The common practice of using the last trip in the commute as a proxy for commute distance undercounts Georgia's annual commute PMT by 2.6 billion miles, about 10 percent of the total commute PMT. This report presents a new, more accurate method of estimating commute distance and correctly identifying trip anchors based on a combination of purpose and location (see chapter 2, Defining the Commute). It also provides data on the time of day of work journeys and how work and nonwork travel interact over the course of a day (see chapter 2, Overview of Work Journeys). Commute distances also vary by time of day. Finally, there are 11 differences in the employment characteristics of Atlanta, smaller MPO regions, and rural counties; incorporating employment data can inform our understanding of work travel (see chapter 2, Overview of Work Journeys). Data on telework and mixed telework can also inform projections of travel demand (see chapter 3, The Effects of Flexible Scheduling). Identifying Data Needs Some of the topics covered by the NHTS change more rapidly than others. Given technological changes and the COVID-19 pandemic, more recent data on alternative-fuel vehicles, teleworking, and online shopping would be beneficial. While the NHTS is a strong data source overall, some topics are less robust in the number or format of questions. For example, telecommuting is undermeasured (see chapter 3, Definitions and Technical Notes). The NHTS measurements of physical activity are flawed (see chapter 6, Physical Activity) and while the measurement of loop trips is significantly improved from the 2009 NHTS, the techniques for recording the purposes of loop trips and other forms of travel for its own sake (TFIOS) need improvement (see chapter 7, Challenges of Identifying and Measuring TFIOS). The occupation categories used by the NHTS are overly broad, limiting their predictive ability for modeling work travel. While educational attainment can be used as a partial proxy to subdivide the categories, a more detailed set of categories would be useful for predicting work travel locations and likely schedules. Finally, the 2017 NHTS does not disaggregate ridehailing data from other types of vehicles for hire (i.e., taxis and limos). While this report provides a tentative estimate for what percentage of vehicle-for-hire trips are made through a ridehailing app (see chapter 4, Ridehailing and Vehicle-for-Hire Trips), more direct data would be valuable. Additionally, new mobility services, such as shared scooters, entered the market after the completion of NHTS data collection. 12 CHAPTER 1. KEY TRAVEL PATTERNS IN THE STATE OF GEORGIA CHAPTER 1 SUMMARY This chapter uses analysis of the 2017 National Household Travel Survey (NHTS) to provide an overview of travel patterns in the state of Georgia. It is intended to serve as a reference text. Detailed statistical tables are accompanied by text that provides context and draws attention to notable findings and patterns. Analysis is presented at the state level, and disaggregated by factors such as size of community, sociodemographic characteristics, and vehicle ownership. The chapter is organized in the following sections: Methodological Notes provides the methodological background on the NHTS itself, and notes on how the data have been processed for this report. It discusses the integration of the regular NHTS and the add-on sample commissioned by the Georgia Department of Transportation (GDOT) and instances where the analytic categories used for this report vary from the default categories provided as part of the NHTS's public use dataset. The section also discusses the geographic divisions used for analysis. Overview provides summary statistics at the state level, comparisons with national patterns, and regional differences within the state. Georgians' travel habits are bifurcated. Residents of the Atlanta MPO are more mobile than the national average, while residents of smaller MPOs and rural counties are less mobile. While most of this paper reports travel by residents of Georgia or of MPOs within Georgia, Trip Patterns by Location of Travel examines travel that occurs within 13 Georgia, regardless of the residence of the traveler. Most travel is local; 75 percent of trips stay within a single county, and 95 percent stay within a single MPO. Inter-MPO trips are generally balanced. In other words, the number of trips from A to B is comparable to the number of trips from B to A. However, this does not take time of day into account; there are certainly temporal imbalances, as would be expected on the basis of directional commute and other flows. Household and Personal Mobility examines differences in household and personal mobility by gender, race, age, income, and disability. Elderly Georgians and Georgians with disabilities have strongly reduced mobility compared to their neighbors. Lowincome residents also make fewer trips than other Georgians. Trip Purpose examines trip purpose across different demographic groups and modes. Why people travel gives insight into what utility they derive from their trips; trip purpose can illuminate patterns that may be obscured by raw trip numbers. For example, men who have reached retirement age make somewhat fewer trips than younger adults, but make more discretionary trips. Their overall lower mobility is accompanied by more trips dedicated to personal fulfillment. Senior women do not seem to enjoy this benefit. In fact, the age-related reduction in mobility falls heavily on senior and elderly women. Gendered differences are strong throughout all age groups. Adult and teenage women, and even female children, devote more of their travel to household-serving trips than do their male counterparts. Income also affects trip purpose. Moderate-income households are more mobile than lower-income households, but make more trips to transport someone else than any other group; this suggests moderate-income households' increased mobility over lower-income 14 groups may yield a smaller dividend in terms of utility than would be expected from the raw trip numbers, person miles traveled (PMT), or vehicle miles traveled (VMT). Vehicle Availability and Usage discusses household vehicle ownership. Vehicle ownership differs substantially by household income, not only in terms of the number of household vehicles, but also in terms of the quality. Low-income vehicle owners typically purchase older, high-mileage vehicles that must be replaced more frequently. The section also discusses differences in the quantity of driving between households that are vehiclesufficient and those that have fewer vehicles than potential drivers. Transit Preferences and Use analyzes data about Georgia workers' preferences for transit service using a question included in the add-on commissioned by GDOT. Finally, Summary of Findings on Geographical Differences synthesizes findings on regional differences from throughout the report. METHODOLOGICAL NOTES This report is based on analysis of the Georgia sample from the NHTS, including 8,005 households from the add-on sample commissioned by GDOT and 606 households that were obtained as part of the national sampling design. Households in both subsamples were asked the same survey questions, and survey weights provided by NHTS were calibrated to include both subsamples. Therefore, this report pools the national and add-on subsamples. Because of the add-on module, GDOT also has access to more detailed and disaggregated variables. The dataset used for this report integrates these confidential variables with the updated public use dataset (version 1.1). Extensive technical documentation of NHTS itself can be found 15 on the official website.3 This section will give additional information on the integration of the two datasets, as well as other technical matters. Readers more interested in contents than methodology may wish to proceed to Overview in this chapter. Weighting NHTS uses expansion weights to allow for the production of representative national estimates. Because of the add-on, the publicly available weights applied to Georgia respondents have also been made to produce representative estimates at the state level--something that is not guaranteed for states that did not commission an add-on. Weights are not guaranteed to be representative for small geographical entities or subpopulations. As suggested in the 2017 NHTS Weighting Report (Roth, Dai, and DeMatteis 2017), this report is cautious with analysis of small geographies and notes when a sample size may be too small for a reliable estimate. Unless otherwise noted, all results and statistics are weighted using 7-day weights provided by NHTS. Person weights are used for persons and trips4 and household weights are used for households and vehicles.5 This approach is used in the summary of travel trends commissioned by NHTS itself (McGuckin and Fucci 2018). That report was used as a methodological reference for the present document and is drawn upon for some national comparisons. Unweighted sample tables can be found in the appendix. 3 See https://nhts.ornl.gov/documentation. 4 NHTS-provided trip weights are an annualized version of the person-weights. For normalized figures such as trips per household, the numerator and denominator are each calculated using the appropriate weight (in this case, person weights for trips and household weights for households). 5 When household-level variables are applied to people, person-weights are used. For example, 22.4 percent of Georgia households have an annual income of at least $100,000 (household weights), but those households contain 25.9 percent of Georgia's population (person weights). 16 Adjustments to Data This project integrates the confidential dataset provided to GDOT in 2017 with updates to the public-use dataset released in August 2018.6 Responses of "Other, specify" from the confidential dataset were used to update public variables such as race, homeownership, mode, and internet capability to match NHTS's categories. For example, an "other" response of "iPhone" to a question about what device is used to access the internet would be recoded as "smart phone." Where gender, race, or age are missing, NHTS's imputed values were used. Race was more comprehensively recoded, as follows: To reflect Georgia's racial makeup, the default racial categories are: (1) white nonHispanic, (2) Black and Black multiracial (including Black-Hispanic), and (3) other. For consistency, respondents listed as "multiracial" in the public dataset were reclassified based on their more detailed responses, available in the private dataset. "Other, specify" responses were used to recategorize respondents who did not identify with any of the census-style categories used in the NHTS. For example, a number of Latino respondents selected only "other" and wrote "Latino" or their nationality.7 Others listed countries of origin that they were not sure how to classify, for example, in southwest Asia. 6 Details of what modifications were made from version 1.0 of the dataset are available in the Version 1.1 Release Notes at https://nhts.ornl.gov/documentation. 7 Some disagreement exists over whether to treat Hispanic as an ethnicity or Latino as a race. This report adopts the latter approach. The NHTS questionnaire did not include Latino as an option for race. Accordingly, the write-in "Latino" responses were grouped with respondents who selected white-Hispanic into the created racial category of Latino only. In the three-category definition used for race in this report's analysis, Latinos are included under (3) other. 17 Incorrect or extraneous "other" responses were recoded. For example, a person who selected "white" and wrote "Irish" under "other" is classified by NHTS as multiracial. Cases like this were reclassified in this report as white. Assorted Definitions and Notes NHTS collects trip information only about respondents ages 5 years and older. Following McGuckin and Fucci (2018), reported per capita results are more specifically per person ages 5 years and up, unless otherwise noted. NHTS does not ask about drivers' licensing; rather, they ask "do you/does this person drive?" This creates some ambiguity as to how a respondent might classify a teenager with a learner's permit; 56 respondents under the age of 16 were listed as drivers. Unless otherwise stated, drivers is defined to mean driver of legal driving age (16+).8 At a household level, NHTS defines race based on the race of the person who was the first respondent to the survey. Since 7 percent of households sampled are multiracial, this approach has obvious limitations. This report defines a household's race based on its overall composition, typically dividing between all-white, non-Hispanic households and households with one or more nonwhite and/or Latino members. Key Terms Though terms such as person miles traveled and vehicle trips are commonly used, a brief reminder may be helpful for some readers: 8 The exception is that, to match the methods used to calculate national descriptive statistics by McGuckin and Fucci (2018), drivers of all ages were used to calculate the totals presented in chapter 2, Overview. 18 Trip: A single trip from one origin to one destination, no matter how brief the amount of time spent there. For example, if someone stops for coffee on the way to work, the journey to work would contain two trips. Two types of trips are considered. Person trip: A trip made by a person. Vehicle trip: A trip made by a personal occupancy vehicle (POV). More precisely, the trip is measured as the trip made by the driver of that vehicle. If a father drives his daughter to school, that action would be counted as one vehicle trip (just the father) and two person trips (both the father and the daughter). Vehicle miles traveled (VMT): Similar to vehicle trip, this measure of distance is calculated based just on the distance traveled by the driver of a POV for a trip that occurred in a POV. Person miles traveled (PMT): Miles traveled by a person, regardless of mode or driver status. In the previous example, if the daughter's school was 1 mile away from home, the trip would generate 2 PMT and 1 VMT. A crucial difference is that PMT also captures travel by other modes, such as transit, walking, and biking. However, the method used by NHTS to calculate trip distance, namely the shortest network distance, is likely to understate the mileage of nonmotorized modes, especially cycling, where travelers are likely to choose a more circuitous route. Work journey (WJ): All the trips and intermediate stops a commuter makes between home and work. A work journey is unidirectional (either from home to work, or from work to home). WJs can be simple (proceeding directly between home and work) or complex (containing at least one internal stop). 19 Geographic Divisions for Analysis In addition to statewide figures, this report examines differential patterns between different metropolitan planning organizations (MPOs) and kinds of MPOs within Georgia. Counties are classified into four MPO tiers based on size: Tier 1: Counties within the Atlanta MPO. Tier 2: Counties within MPOs other than Atlanta, with an MPO population over 200,000 (based on American Community Survey [ACS] 2015 5-year estimates). These include Georgia counties that are part of the Tennessee-based Chattanooga MPO. Tier 3: Counties within MPOs with population < 200,000. Tier 4: Counties that are not in any MPO. For analytical purposes, MPO boundaries are defined at the county level, following the practice used for GDOT Research Project 16-31, Impact of Emerging Technologies and Trends on Travel Demand in Georgia (Kim, Mokhtarian, and Circella 2019). All households in a county that is partially within an MPO are considered MPO households. Table 1 shows the full classification for each county, and figure 1 shows a map of MPO boundaries across county lines. The sample size for most individual counties is too small to allow for a county-level breakdown of results; several counties are represented by fewer than 10 households in the survey sample (see the appendix). Most of the comparative geographic analysis in this report focuses instead on types of counties, using the MPO tier system. 20 Table 1. NHTS MPO classification by county. MPO MPO ID MPO Name Tier County Name 1 Albany 3 1 Albany 3 2 Athens 2 2 Athens 2 2 Athens 2 2 Athens 2 3 Atlanta 1 3 Atlanta 1 3 Atlanta 1 3 Atlanta 1 3 Atlanta 1 3 Atlanta 1 3 Atlanta 1 3 Atlanta 1 3 Atlanta 1 3 Atlanta 1 3 Atlanta 1 3 Atlanta 1 3 Atlanta 1 3 Atlanta 1 3 Atlanta 1 3 Atlanta 1 3 Atlanta 1 Table continues on next page. Dougherty Lee Clarke Madison Oconee Oglethorpe Barrow Carroll Cherokee Clayton Cobb Coweta DeKalb Douglas Fayette Forsyth Fulton Gwinnett Henry Newton Paulding Rockdale Spalding County Portion in FIPS MPO NHTS Sample Size, unweighted households 13095 Total 168 13177 Partial 61 13059 Total 318 13195 Partial 63 13219 Partial 90 13221 Partial 32 13013 Partial 26 13045 Partial 46 13057 Total 119 13063 Total 97 13067 Total 378 13077 Total 71 13089 Total 406 13097 Total 59 13113 Total 65 13117 Total 98 13121 Total 536 13135 Total 369 13151 Total 74 13217 Partial 33 13223 Total 53 13247 Total 36 13255 Partial 34 21 Continued from previous page. MPO MPO ID MPO Name Tier Table 1. (Continued). County Name County Portion in FIPS MPO NHTS Sample Size, unweighted households 3 Atlanta 1 Walton 13297 Partial 33 5 Augusta 2 Columbia 13073 Partial 326 5 Augusta 2 Richmond 13245 Total 420 6 Brunswick 3 Glynn 13127 Total 245 7 Cartersville 3 Bartow 13015 Total 167 8 Chattanooga/ Catoosa 2 Catoosa 13047 Total 39 8 Chattanooga/ Catoosa 2 Dade 13083 Partial 6 8 Chattanooga/ Catoosa 2 Walker 13295 Partial 25 Total; 9 Columbus 2 Chattahoochee 13053 military 4 presence 9 Columbus 2 Harris 13145 Partial 88 9 Columbus 2 Muscogee 13215 Partial 411 9 Dalton 3 Whitfield 13313 Total 128 10 Gainesville 2 Hall 13139 Total 384 10 Gainesville 2 Jackson 13157 Partial 128 11 Hinesville 3 Liberty 13179 Total 96 11 Hinesville 3 Long 13183 Partial 19 12 Macon 3 Bibb 13021 Total 292 12 Macon 3 Jones 13169 Partial 58 13 Rome 3 Floyd 13115 Partial 167 14 Savannah 2 Bryan 13029 Partial 75 14 Savannah 2 Chatham 13051 Total 611 14 Savannah 2 Effingham 13103 Partial 125 15 Valdosta 3 Lowndes 13185 Total 202 16 Warner Robins 3 Houston 13153 Total 283 16 Warner Robins 3 Peach 13225 Partial 45 99 Non-MPO 4 Other counties None 931 Because each county contains a wide variety of built environments, a measure of neighborhood type known as urbanicity is sometimes used instead of or in addition to MPO tier. This measure, 22 which was created for the NHTS by consultants at the Claritas company, was released as part of the August 2018 update to the NHTS dataset.9 The urbanicity measure combines built environment measures, such as density, with data about residents' tendencies to travel within or outside of their communities. Claritas describes the community types as in figure 2. Figure 1. Map. Georgia MPO map. Retrieved 9-25-18 from the Georgia Association of Metropolitan Planning Organizations http://www.gampo.org/members.html. The map shows MPO boundaries within the state of Georgia; the Columbus and Chattanooga MPOs extend into neighboring states. 9 Claritas also updated the census-derived built environment variables, e.g., housing density, and percent renters. More details can be found in the 2017 NHTS Version 1.1 Release Notes (NHTS 2018). 23 Figure 2. Classification. Urbanicity classifications. Source: Claritas (2018, pp. 45) As shown in table 2, each MPO tier contains neighborhoods that fall into several different categories. Using urbanicity allows a way to capture these intra-county geographic differences. 24 Table 2. Classification of sample households by MPO tier and urbanicity. OVERVIEW This section compares major travel indicators for the state of Georgia with national figures. (Data will also be compared across different MPO tiers in the Trip Purpose section.) It also summarizes statistics relating to household travel, person travel, trip purpose, and mode choice. Subsequent chapters will examine the relationships between these topics, as well as relationships to demographic traits. The issue of vehicle ownership will also be visited in greater detail. To set the stage for the statistics to follow, table 3 presents the number of households, people, and vehicles in Georgia and nationally. The Atlanta MPO accounts for more than half of the state's census on all indicators, while the non-MPO portions of the state comprise about a fifth of its census. The remaining one quarter is divided between mid-size and small MPO regions. 25 Table 3. Summary demographic statistics. Households National 118,208,251 Persons, all ages 321,419,000 Persons, ages 5+ 301,599,169 Vehicles 222,578,926 Drivers, all ages 223,277,172 Drivers, ages 16+ 222,780,478 Workers 156,988,243 Source: NHTS weighted totals.10 indicates data not available. Georgia (Statewide) 3,651,249 100% 10,204,581 100% 9,555,773 100% 6,997,337 100% 7,036,938 100% 7,000,240 100% 4,778,570 100% Tier 1 Atlanta 1,956,521 53.6% 5,518,955 54.1% 5,181,569 54.2% 3,691,992 52.8% 3,848,238 54.7% 3,832,515 54.7% 2,759,079 57.7% Tier 2 Medium MPOs 593,833 Tier 3 Small MPOs 373,866 Tier 4 Non-MPO Counties 727,029 16.3% 10.2% 19.9% 1,656,452 1,055,347 1,973,827 16.2% 10.3% 19.3% 1,530,993 973,023 1,870,188 16.0% 10.2% 19.6% 1,121,151 707,766 1,476,427 16.0% 10.1% 21.1% 1,096,159 698,309 1,394,231 15.6% 9.9% 19.8% 1,091,691 694,927 1,381,108 15.6% 9.9% 19.7% 752,233 477,924 789,334 15.7% 10.0% 16.5% Table 4 presents major travel indicators. Georgia households are slightly larger than the national average, but they are statistically similar to the national average for most indicators. Following the national pattern (McGuckin and Fucci 2018), there are as many vehicles as drivers. However, vehicle ownership is not universal; 6.9 percent of households own zero vehicles, and 18.9 percent have a vehicle deficit (fewer vehicles than potential drivers). See Vehicle Availability and Usage in this chapter for more details about vehicle ownership. Within Georgia, some differences exist between Atlanta (tier 1) and other areas (tiers 24). Households in Atlanta are somewhat larger, and the ratios of vehicles to households, drivers, and 10 NHTS weights were benchmarked to 2015 American Community Survey 1-year estimates where available and 20112015 5-year estimates otherwise. See Roth, Dai, and DeMatteis (2017) for further details on the weighting process. 26 workers are lower. Rural counties (tier 4) have the highest ratios of vehicles per household, driver, and worker. However, the marked difference in the number of vehicles per worker is driven in part by the lower number of workers per household as compared to the rest of the state. Table 4. Major travel indicators by MPO tier. National Georgia (Statewide) Tier 1 Atlanta Tier 2 Mid-sized MPOs Tier 3 Small MPOs Tier 4 Non-MPO Counties Persons (ages 5+) per HH Workers per HH 2.551 1.328 2.617 1.309 2.648 1.410 2.578 1.267 2.603 1.278 2.572 1.086 Drivers per HH 1.889 1.916 1.887 1.888 1.893 2.031 Vehicles per HH 1.883 1.916 1.887 1.888 1.893 2.031 Vehicles per Driver* Vehicles per Driver Aged 16+ Vehicles per Worker 0.997 0.999 1.418 0.994 1.000 1.464 0.959 0.963 1.338 1.023 1.027 1.490 1.014 1.018 1.481 1.059 1.069 1.870 * NHTS respondents were asked "Do you/does this person drive," and not about licensing. As a result, some people too young to receive a drivers license are recorded as drivers. Unless otherwise specified, this report will be analyzing drivers of driving age (ages 16+). HH = Household. Table 5 gives information about household composition and trip generation at the level of MPO tier, and table 6 gives the same information about individual MPOs. These data are designed to illuminate local differences. However, caution should be employed when examining estimates for MPOs with a small sample size in the NHTS data. In particular, Chattanooga is represented by only 70 households. Hinesville and Rome are represented by fewer than 200 households each (see the appendix). The sample sizes for the remaining small and medium MPOs range from 201 to 811 households (table 1). Due to these sample size constraints, most analysis will proceed at the level of MPO tier. 27 Table 5. Summary demographic and travel statistics by MPO tier. Place Statewide MPO TIER 1. Atlanta 2. Medium-size MPOs 3. Small MPOs 4. Non-MPO Demographics Persons aged Households 5+ 3,651,249 9,555,773 Annual Travel by Residents Vehicle Trips VMT Person Trips Vehicles (Thousands) (Millions) (Thousands) 6,997,336 6,882,190 71,984 11,073,916 PMT (Millions) 131,078 1,956,521 53.6% 593,833 16.3% 373,866 10.2% 727,029 19.9% 5,181,569 54.2% 1,530,993 16.0% 973,023 10.2% 1,870,188 19.6% 3,691,992 52.8% 1,121,151 16.0% 707,766 10.1% 1,476,427 21.1% 3,844,926 55.9% 1,051,526 15.3% 707,440 10.3% 1,278,298 18.6% 40,499 56.3% 9,698 13.5% 6,687 9.3% 15,100 21.0% 6,172,253 55.7% 1,746,745 15.8% 1,115,020 10.1% 2,039,899 18.4% 75,945 57.9% 20,596 15.7% 10,672 8.1% 23,864 18.2% Place 1. Albany 2. Athens 3. Atlanta 4. Augusta 5. Brunswick 6. Cartersville 7. Chattanooga 8. Columbus 9. Dalton 10. Gainesville 11. Hinesville 12. Macon 13. Rome 14. Savannah 15. Valdosta 16. Warner Robins Table 6. Summary demographic and travel statistics by MPO. Demographics Persons aged Households 5+ 43,355 101,328 91,767 229,894 1,956,521 5,181,569 137,340 361,997 40,721 97,161 28,688 75,736 41,322 83,631 98,590 262,221 40,440 110,803 85,636 229,842 24,637 75,667 68,827 185,303 28,072 63,558 139,178 363,409 39,787 107,097 59,340 156,370 Annual Travel by Residents Vehicle Trips VMT Person Trips Vehicles (Thousands) (Millions) (Thousands) 72,507 76,265 558 113,246 176,565 149,989 1,356 262,718 3,691,992 3,844,926 40,499 6,172,253 236,760 257,795 2,061 427,870 75,241 94,922 929 148,967 66,791 61,532 798 93,980 84,019 56,642 560 80,159 173,178 179,854 1,556 285,955 78,853 65,930 501 118,350 190,512 165,397 1,853 268,332 49,648 56,391 851 77,342 123,776 119,437 1,006 198,211 52,625 50,416 407 82,502 260,118 241,850 2,312 421,711 75,688 80,130 506 117,902 112,636 102,416 1,129 164,520 PMT (Millions) 1,052 4,830 75,945 3,502 1,726 1,099 857 2,708 779 3,384 1,015 1,841 704 5,315 882 1,574 28 Compared to the U.S. average, Georgians make fewer trips per household and per capita (table 7 and table 8). The exception is the Atlanta MPO, where tripmaking levels and mileage are above average for Georgia, and close to the U.S. average for most indicators. Additionally, Atlanta residents are less likely to be immobile than residents of other MPO tiers (i.e., make no trips on their travel day; see table 9). Table 7. Summary statistics per household, daily. Person Vehicle Geography Trips PMT Trips Georgia 8.31 98.35 5.16 National 8.60 99.46 5.11 MPO TIER 1. Atlanta 8.64 106.35 5.38 2. Medium-size MPOs 8.06 95.02 4.85 3. Small MPOs 8.17 78.20 5.18 4. Non-MPO 7.69 89.93 4.82 MPO 1. Albany 7.16 66.45 4.82 2. Athens 7.84 144.21 4.48 3. Atlanta 8.64 106.35 5.38 4. Augusta 8.54 69.86 5.14 5. Brunswick 10.02 116.10 6.39 6. Cartersville 8.98 104.94 5.88 7. Chattanooga 5.31 56.84 3.76 8. Columbus 7.95 75.24 5.00 9. Dalton 8.02 52.74 4.47 10. Gainesville 8.58 108.26 5.29 11. Hinesville 8.60 112.90 6.27 12. Macon 7.89 73.28 4.75 13. Rome 8.05 68.75 4.92 14. Savannah 8.30 104.64 4.76 15. Valdosta 8.12 60.77 5.52 16. Warner Robins 7.60 72.69 4.73 From McGuckin and Fucci (2018, 12) MPO represented by small sample VMT 54.01 53.81 56.71 44.74 49.01 56.90 35.28 40.49 56.71 41.12 62.54 76.25 37.10 43.23 33.96 59.29 94.60 40.04 39.77 45.50 34.87 52.15 29 Table 8. Summary statistics per capita, daily. Per Person (ages 5+) Per Driver* Geography Georgia National Person Trips 3.17 3.37 Vehicle PMT Trips 37.58 1.97 38.98 2.00 Vehicle VMT Trips 20.64 2.68 19.13 2.70 MPO TIER 1. Atlanta 3.26 40.16 2.03 21.41 2.74 2. Medium MPOs 3.13 36.86 1.88 17.35 2.63 3. Small MPOs 3.14 30.05 1.99 18.83 2.78 4. Non-MPO 2.99 34.96 1.87 22.12 2.51 MPO 1. Albany 3.06 28.43 2.06 15.10 2.96 2. Athens 3.13 57.56 1.79 16.16 2.42 3. Atlanta 3.26 40.16 2.03 21.41 2.74 4. Augusta 3.24 26.50 1.95 15.60 2.93 5. Brunswick 4.20 48.66 2.68 26.21 3.55 6. Cartersville 7. Chattanooga 3.40 39.75 2.63 28.08 2.23 28.88 2.73 1.86 18.33 2.29 8. Columbus 2.99 28.29 1.88 16.25 2.73 9. Dalton 2.93 19.25 1.63 12.39 2.53 10. Gainesville 11. Hinesville 3.20 40.33 2.80 36.76 1.97 22.09 2.69 2.04 30.81 2.97 12. Macon 13. Rome 2.93 27.22 3.56 30.37 1.77 14.87 2.50 2.17 17.57 3.13 14. Savannah 3.18 40.07 1.82 17.43 2.47 15. Valdosta 3.02 22.57 2.05 12.95 2.82 16. Warner Robins 2.88 27.58 1.79 19.79 2.42 * To match national figures, drivers of all ages are used. From McGuckin and Fucci (2018, p, 13), except for vehicle trips and VMT per person. MPO represented by small sample. VMT 28.03 28.49 28.83 24.24 26.24 29.67 21.66 21.87 28.83 23.46 34.80 35.47 22.60 23.65 19.20 30.11 44.77 21.02 25.27 23.56 17.82 26.64 30 Table 9. Trip distance and percent of population immobile. Average Person Trip Geography Length (mi) Georgia 11.85 MPO TIER 1. Atlanta 12.31 2. Medium-size MPOs 11.80 3. Small MPOs 9.58 4. Non-MPO 11.71 MPO 1. Albany 9.29 2. Athens 18.40 3. Atlanta 12.31 4. Augusta 8.19 5. Brunswick 11.58 6. Cartersville 11.69 7. Chattanooga 10.69 8. Columbus 9.48 9. Dalton 6.59 10. Gainesville 12.62 11. Hinesville 13.13 12. Macon 9.30 13. Rome 8.54 14. Savannah 12.61 15. Valdosta 7.49 16. Warner Robins 9.57 MPO represented by small sample. Average Vehicle Trip Length (mi) 10.46 10.53 9.22 9.45 11.81 7.32 9.04 10.53 8.00 9.79 12.98 9.88 8.65 7.60 11.21 15.09 8.42 8.08 9.56 6.32 11.03 Percent Immobile (0 trips on travel day) 19.4% Percent Immobile, weekdays only 16.1% 18.2% 20.0% 20.7% 21.7% 15.1% 16.0% 17.3% 18.2% 22.8% 24.9% 18.2% 18.4% 12.3% 14.7% 20.5% 22.9% 21.1% 19.3% 25.8% 18.9% 18.1% 16.7% 26.3% 23.9% 21.6% 17.0% 15.1% 14.5% 14.1% 14.7% 12.3% 17.7% 18.9% 19.9% 16.1% 16.1% 19.7% 14.6% 22.1% 12.5% Trip Purpose Overview The NHTS collects detailed information about trip purpose. In most cases, it is desirable, or even necessary (due to small sample sizes), to collapse the NHTS's detailed taxonomy into a smaller number of categories. Table 10 shows descriptive statistics for the most detailed measure of trip purpose--primary activity at destination--and color-coded categories into which these activities are classified. Broadly, trip purpose is classified into three categories for this report. Mandatory trips support required life activities such as working and attending school. Household-serving trips contribute 31 to the maintenance of the household. They include transporting other household members, shopping, errands, and medical services. Household-serving trips benefit the household as a whole but may not directly benefit the traveler (for example, picking up dry cleaning for a family member). Discretionary trips, in contrast, facilitate activities that are intrinsically valuable to the travelers themselves (e.g., socializing, engaging in recreation, and exercising), allow them to participate in a broader community, or allow them to volunteer their time to a good cause. Trips to return home are the final leg of any circuit of trips. Because of this, they are a good proxy for how often someone leaves the house, without the complications of trip chaining. Trip purpose (level 1) will be the primary measure used in this report. When analysis requires the use of a smaller number of categories, the summarized purposes or purpose categories will be used. The classification categories used here are nested and mutually compatible. However, they differ slightly from NHTS's predefined categories.11 The NHTS classifications base purpose determinations on the purpose at destination, not factoring in origin or physical location. As a result, a trip with the purpose of "work" will be a work trip, regardless of whether the destination was the respondent's official work address. Conversely, trips from work to home are classified as "return home" trips rather than "work commute" trips because the purpose at destination will be regular home activities. However, even when trips both to and from work are accounted for, commuting represents a minority of trips: 17.3 percent of person trips and 23.9 percent of vehicle trips. 11 The NHTS combines commuting with other work travel. This study separates them to allow analysis specifically of the work commute versus other related trip purposes. The other difference is the creation of a community activities category. The NHTS classifies religious and community activities with attending school and daycare and volunteer activities with "other," whereas this report combines these three types of community activities into a new category. The categorization of some loop trips were also adjusted (see footnote to table 10). 32 Table 10. Trip purpose by destination: Types and frequencies. Level Zero: Primary Activity at Destination Percent of Percent of Person Vehicle Trips Trips Level 1: Trip Purpose Level 2: Purpose, summarized Category 1: Mandatory Travel Work 11.1% Work-related meeting / trip 1.0% Attend school as a student 3.7% Attend child care 0.3% Attend adult care 0.1% 14.9% Work Commute Other Work-related 1.2% Travel Attend School or 1.1% Day Care Attend School or 0.1% Day Care Attend School or 0.1% Day Care Work Commute Other Work-related Travel Attend School or Day Care Attend School or Day Care Attend School or Day Care Category 2: Household-Serving Travel Drop off /pick up someone 6.7% Buy goods (groceries, clothes, appliances, gas) Buy services (dry cleaners, bank, service a car, pet care) Other general errands (post office, library) 14.1% 1.9% 2.6% Health care visit (medical, dental, therapy) 1.4% 8.4% Transport Someone 15.4% Shopping/Errands 2.3% Shopping/Errands 2.9% Shopping/Errands 1.4% Medical Services Transport Someone Other Household-Serving Other Household-Serving Other Household-Serving Other Household-Serving Category 3: Discretionary Travel Recreational activities (visit parks, movies, bars, museums) 3.0% 1.9% Social/Recreational Personal Enrichment Exercise (go for a jog, walk, walk the dog, go to the gym) 2.9% 2.5% Social/Recreational Personal Enrichment Visit friends or relatives Buy meals (go out for a meal, snack, carry-out) 3.8% 7.9% 3.4% Social/Recreational 7.5% Meals Personal Enrichment Personal Enrichment Religious or other community activities 2.5% 2.1% Community Activities Community Volunteer activities (not paid) 0.5% 0.6% Community Activities Community Other Travel Change type of transportation 1.0% 0.4% Other Other Something else Regular home activities (chores, sleep) 0.3% 34.3% 0.1% Other 32.9% Return Home Other Return Home Work from home (paid) 0.7% 0.9% Return Home Return Home To better capture the purpose of loop trips, loop trips where the primary activity at destination was home or school were reclassified as exercise for nonmotorized and recreation for all other modes. 33 As figure 3 shows, although where, how, and how much Georgians travel may vary from region to region, why they travel is strikingly constant. Statewide, half of all person trips are devoted to the day-to-day business of going to work, school, or daycare, and returning home from all kinds of trips. Just over half of the remainder go to household-serving travel (27 percent of total trips), and 22 percent of trips are discretionary. Figure 3. Stacked bar graph. Person trip purpose by MPO tier (color-coded per table 10). Mode Overview Privately operated vehicles dominate travel in Georgia, receiving 10 times the share of any other mode (table 11). POVs are predominantly privately owned; only 0.2 percent of POV trips were in rental cars. 34 Table 11. Person trips by mode (detailed). Mode Privately Operated Vehicle Pedestrian (walk/wheelchair) Reported Trips, unweighted Estimated Annual Trips, weighted (millions) Percent of Trips, weighted Mode Category 52,675 9467.0 85.5% POV 4,201 887.2 8.0% Nonmotorized Bike 287 68.2 0.6% Nonmotorized School Bus 1,213 324.8 2.9% Bus or train Public Transit (bus, heavy rail, light rail, and streetcar) 515 157.1 1.4% Bus or train Paratransit 56 15.3 0.1% Other ground or water Other Bus 141 22.5 0.2% Bus or train Taxi/Ridehail/Limo 222 63.2 0.6% Other ground or water Air 105 19.0 0.2% Air Other 289 49.5 0.4% Other ground or water POV includes car, van, pickup truck, motorcycle/moped, RV, rental car, and trips in a "company vehicle" originally reported under the "something else" category. "Pedestrian" combines walking with 29 wheelchair trips that were originally reported under the "something else" category. "Other" includes golf cart, Segway, boat, and responses originally reported under the "something else" category (large trucks, agricultural and construction equipment, skateboard/scooter/skates). Two trips by skateboard/scooter/skates were classified as nonmotorized. Pedestrian trips, including both walking and wheelchair use, are the second most common, followed by school buses and public transit. No other single mode accounts for more than 1 percent of trips. As a result, while future targeted analysis will focus specifically on less common modes such as ridehailing, paratransit, and biking, analysis of geographic patterns will focus on the mode categories described in table 11, as shown in table 12. 35 Table 12. Mode share and trip distance by MPO tier. Annual Person Trips (millions) Mode Statewide POV 9,467.0 Nonmotorized (walk, bike, wheelchair) 955.5 Bus or train 504.4 Other ground or water transportation 127.8 Air 19.0 Mode Statewide POV 85.5% Nonmotorized (walk, bike, wheelchair) 8.6% Bus or train 4.6% Other ground or water transportation 1.2% Air 0.17% Mode Statewide POV 103,800 Nonmotorized (walk, bike, wheelchair) 842 Bus or train 4,058 Other ground or water transportation 1,634 Air 20,120 Mode Statewide POV 11.0 Nonmotorized (walk, bike, wheelchair) 0.9 Bus or train 8.1 Other ground or water transportation 12.9 Air 1,059 Tier 1 (Atlanta) Tier 2 (Medium MPOs) Tier 3 (Small MPOs) Tier 4 (Non-MPO) 5,165.6 1,507.5 989.6 1,804.4 562.9 160.2 80.8 151.6 347.7 56.9 29.8 69.9 80.9 19.9 13.7 13.3 15.1 2.03 1.22 0.65 Percent of Person Trips Tier 1 Tier 2 Tier 3 Tier 4 83.7% 86.3% 88.7% 88.5% 9.1% 9.2% 7.2% 7.4% 5.6% 3.3% 2.7% 3.4% 1.3% 1.1% 1.2% 0.65% 0.24% 0.12% 0.11% 0.03% Annual Person Miles (millions) Tier 1 Tier 2 Tier 3 Tier 4 56,869 15,273 9,476 22,181 464 196 58 124 2,637 430 194 797 923 167 293 251 15,052 3,905 651 511 Average Trip Distance (miles) Tier 1 Tier 2 Tier 3 Tier 4 11.0 10.1 9.6 12.3 0.8 1.2 0.7 0.8 7.6 7.6 6.5 11.4 11.5 8.5 21.4 18.8 996 1,927 536 790 POVs account for a smaller proportion of trips in the Atlanta MPO compared to the rest of the state. POV use is offset by comparatively high mode shares for nonmotorized transportation and buses/trains. Nonmotorized mode use is most common in medium and large MPOs (tiers 12). 36 Unsurprisingly, motorized trips comprise the majority of person miles. Motorized trips are the longest in rural, non-MPO counties and the shortest in small MPO counties. Air travel exerts an outsized influence on PMT. Although air travel comprises less than 1 percent of trips made by Georgians, it represents 15 percent of total person miles,12 making it the second largest component of PMT. Interestingly, tier 2 MPO households make proportionally fewer air trips than Atlanta MPO households do (as might be expected), but the ones they do make are about twice as long, on average. A similar comparison holds between tier 4 and tier 3 households. According to NHTS figures, nonmotorized transit accounts for a small share of PMT. The average reported nonmotorized trip distance is less than 1 mile. However, these figures are likely misleading. The NHTS determines trip distance by calculating the shortest path between the origin and destination. This is likely to underestimate the true length of nonmotorized trips; pedestrians and cyclists often follow circuitous routes to avoid safety hazards, find a pleasant walking environment, or to follow a scenic route (Lu, Scott, and Dalumpines 2018; Misra and Watkins 2017). Cyclist trip lengths are particularly prone to being underestimated; 8 percent of trips by adult cyclists have computed speeds of less than 2 mph based on shortest path distance and reported duration. Further, route choices differ by age and gender; on average, as compared to young male cyclists, women and elderly cyclists choose longer trajectories that are perceived to be safer (Misra and Watkins 2018). Therefore, the shortest path distance is likely to be inaccurate for all cyclists, but it is systematically less accurate for female and elderly cyclists than it is for young males. A time-based measure of cycling is likely more accurate. 12 The fact that air trips are extreme outliers in terms of length is the reason air has been left as its own category, despite the small quantity of trips. 37 TRIP PATTERNS BY LOCATION OF TRAVEL This report focuses primarily on travel by residents of Georgia. However, when considering demand for transportation infrastructure, it is important also to account for trips by nonresidents within Georgia and trips that Georgia residents make outside of the state. This section describes trip patterns by location of travel. Table 13 compares trips made by residents of Georgia and trips by nonresidents that took place partly or entirely in Georgia. Table 13. Annual trips and vehicle miles by Georgians and within Georgia. Person Trips (thousands) Entirely Within Georgia Partly Within Georgia Entirely Outside of Georgia Subtotal, Entirely or Partly in Georgia Subtotal, Entirely or Partly Outside of Georgia Total By Georgia Residents By Out-of-State U.S. Residents 10,610,653 162,441 300,822 10,773,094 463,263 11,073,916 205,771 254,385 460,156 Total 10,816,424 416,826 11,233,250 Vehicle Trips (thousands) Entirely Within Georgia Partly Within Georgia Entirely Outside of Georgia Subtotal, Entirely or Partly in Georgia Subtotal, Entirely or Partly Outside of Georgia Total 6,687,867 85,361 108,961 6,773,228 194,322 6,882,189 98,115 137,011 235,126 6,785,982 222,372 7,008,354 VMT (thousands) Entirely Within Georgia Partly Within Georgia Entirely Outside of Georgia Subtotal, Entirely or Partly in Georgia Subtotal, Entirely or Partly Outside of Georgia Total indicates no data. 60,858,290 8,144,777 2,981,012 69,003,067 11,125,789 71,984,079 1,714,764.65 11,062,151 12,776,915 62,573,055 19,206,928 81,779,982 38 Georgians' out-of-state travel is largely offset by trips nonresidents make in Georgia. Georgians make 463 million out-of-state person trips per year, while residents of other U.S. states make 460 million person trips entirely or partially in the state of Georgia.13 Motorized travel is less balanced. After accounting for out-of-state trips made by Georgians elsewhere, Georgia roadways receive 40.8 million net additional trips from out-of-state motorists. However, while Georgia is a net recipient of interstate trips (between Georgia and another location), Georgia is a net donor of trips that are entirely out-of-state. Georgians make 109 million vehicle trips per year that take place entirely outside of Georgia, versus 98 million trips entirely within Georgia made by nonresidents.14 The distinction between trips in a particular place and trips by residents of a particular place is also relevant when examining travel within different parts of Georgia (table 14). Forty percent of person trips by Georgians cross city lines (i.e., the origin and destination are in different cities or towns). Twenty-four percent cross county lines. However, the overwhelming majority of trips stay within a single MPO; only 5.3 percent of trips cross MPO boundaries. Trips that cross between different MPO tiers account for 5 percent of total person trips. The directionality of these trips is balanced, with approximately the same number of trips from A to B as from B to A (table 14 and table 15). Therefore, when discussing trips in spatial terms, this report will classify them by origin rather than double-counting them or apportioning them between origin and destination. 13 Trips by international visitors and residents of U.S. territories are not accounted for in the survey data. 14 The remainder of the report excludes non-Georgians' travel, but includes Georgians' out-of-state travel, unless otherwise specified. 39 Table 14. Annual person trips within and between MPO tiers (thousands). Detail may not sum to totals because of rounding. Table 15. Annual vehicle trips within and between MPO tiers (thousands). Origin Destination 2. Medium 1. Atlanta MPOs 1. Atlanta 3,691,232 31,356 2. Medium MPOs 33,341 938,663 3. Small MPOs 14,287 5,406 4. Non-MPO 37,986 33,337 5. Out of State 7,018 20,215 Total Attracted Trips 3,783,864 1,028,977 Detail may not sum to totals because of rounding. 3. Small MPOs 4. Non-MPO 15,208 40,120 5,062 34,259 660,930 44,669 46,123 1,055,889 5,802 10,152 733,125 1,185,089 5. Out of State 7,200 21,425 3,634 9,916 108,961 151,135 Total Originating Trips 3,785,115 1,032,749 728,926 1,183,251 152,148 6,882,189 While trips between two geographic categories may be balanced on the whole, this does not necessarily apply to trips at a specific time of day. When focusing on rush hour loads or work commutes, a more detailed method for assigning trips to jurisdictions should be considered. HOUSEHOLD AND PERSONAL MOBILITY The Overview section provided an overview of household and personal mobility and discussed regional differences. This section focuses on demographic-based patterns. 40 As shown in table 16, the amount of travel generated by a household differs by gender of household head(s), race,15 income, and vehicle ownership. However, these patterns are complicated by correlations with household size. For a clearer picture of individuals' mobility, it is important to consider person-level travel, as well. Table 17 reveals that many household patterns are also observable at the level of individual household members. Several populations have markedly reduced mobility compared to the state average. While some racial discrepancies exist, the largest differences are by income, age, and disability. The two groups with the most restricted mobility are elderly people (ages 80+) and people with disabilities. Over the course of the year, a Georgia resident with a mobility impairment will make 327 fewer person trips than the average Georgian. An elderly Georgian will make 414 fewer trips, a reduction of more than one trip per day. However, younger seniors are, on average, relatively mobile. The declines in tripmaking associated with aging are mainly seen in the elderly population (ages 80+). Nearly 45 percent of this group made no trips on their travel day. People with disabilities (of all ages) face similar constraints; 40 percent were immobile on their travel day. It should be noted that there is considerable overlap between the elderly and disabled populations; 38.5 percent of elderly respondents reported having a medical condition or handicap that impedes travel (compared to 19.2 percent of younger seniors and 6.8 percent of adults). However, the majority of Georgians with disabilities are younger (figure 4). Therefore, the two groups' needs should not be automatically conflated. 15 See Methodological Notes for more details on how race is defined. 41 Table 16. Daily household travel by demographic groups. 42 Table 17. Daily per capita trips and immobility (persons ages 5+). Subpopulation Person Trips All 3.17 Gender Male 3.16 Female 3.18 Race White non-Hispanic 3.25 Black & Black multiracial 3.10 Other race 3.05 Age Children (515) 2.45 Teens (1617) 2.54 Adults (1864) 3.44 Seniors (6579) 3.08 Elderly (80+) 2.04 Disability No impairment 3.25 Impairment present 2.28 Annual Household Income <$15,000 2.85 $15,000 to $24,999 2.82 $25,000 to $34,999 3.14 $35,000 to $49,999 3.33 $50,000 to $74,999 3.20 $75,000 to $99,999 3.32 $100,000+ 3.42 * Vehicle trips as driver only. Immobile is defined as zero trips on travel day. Vehicle Trips* 1.97 2.02 1.93 2.10 1.88 1.72 0.01 0.76 2.49 2.24 1.16 2.05 1.09 1.26 1.64 2.03 2.21 2.21 2.20 2.14 Percent Immobile 19.4% Percent Immobile, weekdays 16.1% 17.8% 21.0% 13.8% 18.3% 17.8% 21.3% 21.4% 14.5% 18.1% 17.5% 22.7% 22.2% 16.2% 27.6% 44.7% 16.3% 16.1% 13.2% 25.7% 47.8% 17.6% 40.3% 14.1% 37.7% 25.9% 23.2% 20.7% 18.5% 18.7% 15.5% 15.7% 23.5% 18.8% 18.1% 17.3% 14.6% 11.6% 12.1% Calculated using 5-day person weights, excluding households whose travel day fell on a weekend or federal holiday. Vehicle trips for this age group stem from minors learning to drive under the supervision of an adult. Defined by NHTS as "a condition or handicap that makes it difficult to travel outside the home." 43 56.5% 4.0% 1.2% 26.3% 12.0% Children (5-15) Teens (16-17) Adults (18-64) Seniors (65-79) Elderly (80+) Figure 4. Bar graph. Age of population with mobility impairments. Interestingly, the elderly and disabled are the only populations for which weekday immobility was higher than overall immobility. One possible explanation for this phenomenon is a difference in trip purpose (e.g., emphasis on visiting family or attending church; fewer or no work trips on weekdays). However, the difference also suggests that elderly and disabled residents' travel is sometimes contingent on the availability of a friend or family member to assist them; 35 percent of elderly and disabled residents reported asking others for rides. Even if still somewhat independently mobile, some elderly and disabled individuals may curtail their travel during peak periods and, thus, defer some trips to weekends or eliminate them altogether; 63 percent reported reducing their travel.16 On average, men and women make similar numbers of trips (table 17). However, as will be shown in chapter 5, this is not true across all age groups; elderly women's mobility is much lower than that of men of a comparable age. Additionally, trip purposes vary by gender. In addition to making fewer trips, elderly and disabled residents also make shorter trips (table 18). Though women make as many trips as men, their trips are also shorter. Georgians who earn 16 Individuals with disabilities were asked about changing travel behavior as a result of their disability; elderly respondents with no reported disability were asked if they had changed their behavior over the last year. 44 more than $50,000 travel a greater distance than those who earn less, but within the lower income tiers, the relationship does not appear to be linear; there is a spike in trip length and PMT among residents earning between $15,000 and $24,999 as compared to the income groups above and below. It is possible this is linked to employment, e.g., the relative location of low-wage jobs and low-income housing. Average person trip distance is somewhat longer than that of the average vehicle trip. This result likely reflects the outsized influence of long-distance air travel. Note, for example, that the difference diminishes or reverses in households with income of less than $50,000. An additional reason to treat person trip distances with caution is that, with the exception of loop trips, the NHTS determines trip distance by calculating the shortest path between the origin and destination. While this will understate the distance for at least some motorized trips, it is likely to disproportionately underestimate distance for modes that often follow less direct routes (e.g., walking, biking, and transit). TRIP PURPOSE This section examines demographic differences in trip purpose. Specifically, the tables in this section disaggregate by trip purpose the patterns in person trips presented in the previous section, Household and Personal Mobility. This section also examines the relationship between trip purpose and mode. 45 Table 18. Trip distance and daily PMT and VMT per person ages 5+. Subpopulation PMT VMT Average Person Average Vehicle Trip Length (mi) Trip Length (mi) All 37.58 20.64 11.85 10.44 Gender Male 40.53 23.66 12.82 11.71 Female 34.80 17.79 10.93 9.18 Race White non-Hispanic Black & Black multiracial Other race 40.31 34.01 35.54 22.07 20.01 16.54 12.39 10.98 11.68 10.52 10.60 9.64 Age Children (515) Teens (1617) Adults (1864) Seniors (6579) Elderly (80+) Disability 19.36 20.32 44.52 32.77 13.13 0.07 4.56 26.91 20.50 6.91 7.90 8.00 12.96 10.66 6.46 6.61 5.97 10.78 9.14 5.95 No impairment Impairment present 39.49 16.05 21.67 8.98 12.14 7.07 10.57 7.82 Annual Household Income <$15,000 23.17 11.94 8.14 $15,000 to $24,999 28.85 16.47 10.25 $25,000 to $34,999 24.94 17.41 7.95 $35,000 to $49,999 29.68 20.75 8.89 $50,000 to $74,999 43.66 24.62 13.67 $75,000 to $99,999 46.18 23.91 13.91 $100,000+ 46.07 24.09 13.49 Vehicle trips for this age group stem from minors learning to drive under the supervision of an adult. 9.24 10.04 8.56 9.40 11.15 10.89 11.26 Defined by NHTS as "a condition or handicap that makes it difficult to travel outside the home." Table 19 shows person trips by purpose and income. Over all, higher income groups make more trips, but these trips are not always spread proportionally across all types. Discretionary travel is a good example. Dining out and takeout generally increase with income. However, trips for community and religious activities are highest in high- and low-income categories, with a dip in moderate income groups. People in the lowest income bracket make a relatively large number of 46 social and recreational trips; approximately the same number as households earning $50,000$74,999. However, social trips drop markedly among households earning $15,000$34,999. These groups also make more work trips than members of very low-income households, so part of the difference may relate to the availability of time for recreation among the working poor. There is a notable spike in trips to transport someone in the $35,000$49,999 bracket. This likely reflects the travel patterns of households that have acquired at least one car, but may still not have enough vehicles for every potential driver. These households can be characterized as having a vehicle deficit (Blumenberg et al. 2018; see also chapter 5). The presence of trips to transport someone raises the issue of how much utility travelers obtain from their trips, or whether utility for one household member is obtained at a cost of disutility for another. On average, someone in a household earning $35,000$49,999 will make 73 more person trips per year than someone in the next income category down, but 22 of those trips, or 30 percent, will be to transport someone else. 47 Table 19. Annual person trips by purpose and income. Annual Trips per Person (Ages 5+) Household Income Trip Purpose $15,000 to $25,000 to $35,000 to $50,000 to $75,000 to <$15,000 $24,999 $34,999 $49,999 $74,999 $99,999 $100,000+ All Purposes 1,041 1,028 1,145 1,217 1,166 1,211 1,247 Mandatory Travel 120 166 209 198 204 226 205 Work commute 71 104 149 143 144 160 137 Other work-related travel 6 8 9 9 13 15 19 Attend school or daycare 42.7 53.8 51.1 46.3 46.4 51.0 49 Household-Serving Travel 324 288 314 355 306 321 294 Transport someone 69.0 69.7 74.5 96.2 74.4 76.9 84 Shopping or errands 228 197 223 246 219 225 197 Medical/dental services 27 22 16 13 13 18 13 Discretionary Travel 227 212 216 251 250 260 310 Social/recreational 123 87.3 102 134 122 136 158 Dining 72.0 80.0 85.7 88.1 92.5 87.3 115 Community/religious 32.7 44.5 28.9 29.1 35.7 36.0 36.5 Other Travel 10.2 9.1 12.7 13.5 19.0 14.8 19.4 Other 10.2 9.1 12.7 13.5 19.0 14.8 19.4 Return Home 359.9 353.0 392.6 399.6 388.5 390.5 419.5 Return home 360 353 393 400 388 390 420 Table continues on next page. 48 Table 19. (Continued). Continued from previous page. Percent of Trips (Persons Ages 5+) Trip Purpose Mandatory Travel Work commute Other work-related travel Attend school or daycare Household-Serving Travel Transport someone Shopping or errands Medical/dental services Discretionary Travel Social/recreational Dining Community/religious Other Travel Other Return Home Return home $15,000 to $25,000 to $35,000 to $50,000 to $75,000 to <$15,000 $24,999 $34,999 $49,999 $74,999 $99,999 $100,000+ 11.5% 16.1% 18.3% 16.3% 17.5% 18.7% 16.4% 6.8% 10.1% 13.1% 11.7% 12.4% 13.2% 11.0% 0.6% 0.8% 0.8% 0.7% 1.1% 1.2% 1.5% 4.1% 5.2% 4.5% 3.8% 4.0% 4.2% 3.9% 31.1% 28.0% 27.4% 29.2% 26.2% 26.5% 23.6% 6.6% 6.8% 6.5% 7.9% 6.4% 6.4% 6.7% 21.9% 19.1% 19.5% 20.2% 18.7% 18.6% 15.8% 2.6% 2.1% 1.4% 1.0% 1.1% 1.5% 1.1% 21.8% 20.6% 18.9% 20.7% 21.4% 21.4% 24.8% 11.8% 8.5% 8.9% 11.0% 10.4% 11.3% 12.7% 6.9% 7.8% 7.5% 7.2% 7.9% 7.2% 9.2% 3.1% 4.3% 2.5% 2.4% 3.1% 3.0% 2.9% 1.0% 0.9% 1.1% 1.1% 1.6% 1.2% 1.6% 1.0% 0.9% 1.1% 1.1% 1.6% 1.2% 1.6% 34.6% 34.3% 34.3% 32.8% 33.3% 32.2% 33.6% 34.6% 34.3% 34.3% 32.8% 33.3% 32.2% 33.6% 49 As shown in table 20, white respondents make more discretionary trips than do people of color. This tendency holds for all three categories of discretionary travel. Conversely, white respondents make fewer trips to transport someone else. Homebound ("return home") trips provide a proxy for the number of times people leave the house, regardless of any trip chaining that may happen once they leave. Examining the total tripmaking of the elderly (ages 80+) and people with disabilities underscores how people in these groups are accessing fewer destinations, while examining their home trips shows that they also leave the house less frequently (table 21). Unsurprisingly, elderly respondents make very few work-related trips. However, people with mobility impairments also made very few work trips. In fact, people with mobility impairments made nearly 40 percent fewer work trips than did teenagers, most of whom are full-time high school students. NHTS does not contain information about whether participants' disabilities also impose restrictions on their ability to work, so it is not possible to definitively distinguish between unemployment resulting from an inability to work and unemployment resulting from a travel difficulty. However, it is worth noting that for some Georgians with disabilities, the biggest obstacle to getting their foot in the door to a new career might be getting their foot to the door. Chapter 5, Health and Disability will return to the needs of Georgians with mobility impairments. 50 Table 20. Gender and racial differences in annual person trips by purpose. Annual Trips per Person (Ages 5+) Gender Race Trip Purpose All Purposes Mandatory Travel Work commute Other work-related travel Attend school or daycare Household-Serving Travel Transport someone Shopping or errands Medical/dental services Discretionary Travel Social/recreational Dining Community/religious Other Travel Other Return Home Return home All 1,158 188 129 12 47.3 310 77.6 216 17 256 129 91.6 35.2 15.1 15.1 388.7 389 Male 1,155 219 155 14 50.7 269 61.9 196 12 259 132 95.6 30.9 16.2 16.2 391.8 392 White Non- Black & Black Female Hispanic Multiracial 1,162 1,187 1,131 159 184 184 104 132 120 11 14 11 44.1 37.8 52.7 349 301 343 92.4 67.4 96.5 236 218 226 21 15 21 253 288 219 126 144 108 87.8 106 76.7 39.4 38.1 34.0 14.0 14.6 15.4 14.0 14.6 15.4 385.8 399.9 369.6 386 400 370 Percent of Trips Other Race 1,112 217 136 9 71.5 269 71.8 186 12 219 122 69.5 27.3 15.9 15.9 391.1 391 Trip Purpose Mandatory Travel Work commute Other work-related travel Attend school or daycare Household-Serving Travel Transport someone Shopping or errands Medical/dental services Discretionary Travel Social/recreational Dining Community/religious Other Travel Other Return Home All 16.3% 11.1% 1.0% 4.1% 26.8% 6.7% 18.7% 1.4% 22.1% 11.1% 7.9% 3.0% 1.3% 1.3% 33.6% Male Female 19.0% 13.7% 13.4% 9.0% 1.2% 0.9% 4.4% 3.8% 23.3% 30.1% 5.4% 8.0% 16.9% 20.3% 1.0% 1.8% 22.4% 21.8% 11.4% 10.9% 8.3% 7.6% 2.7% 3.4% 1.4% 1.2% 1.4% 1.2% 33.9% 33.2% White Non- Black & Black Hispanic Multiracial 15.5% 16.2% 11.2% 10.6% 1.1% 1.0% 3.2% 4.7% 25.3% 30.4% 5.7% 8.5% 18.4% 19.9% 1.3% 1.9% 24.3% 19.4% 12.1% 9.6% 9.0% 6.8% 3.2% 3.0% 1.2% 1.4% 1.2% 1.4% 33.7% 32.7% Other Race 19.5% 12.2% 0.9% 6.4% 24.2% 6.5% 16.7% 1.0% 19.7% 11.0% 6.2% 2.5% 1.4% 1.4% 35.2% 51 Table 21. Frequency of trip purposes by age and disability. Trip Purpose All Purposes Mandatory Travel Work commute Other work-related travel Attend school or daycare Household-Serving Travel Transport someone Shopping or errands Medical/dental services Discretionary Travel Social/recreational Dining Community/religious Other Travel Other Return Home Return Home Children 515 895 179 2 2 175.1 147 36 103 7 203 117 54.0 31.3 24.8 24.8 341.7 342 Annual Trips per Person (Ages 5+) Age Mobility Impairment Teens Adults Seniors Elderly 1617 1864 6579 80+ Present Absent 928 1,255 1,123 744 831 1,187 200 218 60 12 34 202 31 181 53 9 19 139 0 17 6 1 3 13 168.3 20.0 1.6 1.1 12.5 50.4 155 343 402 277 324 309 43 94.5 57.0 26.6 42.1 80.8 108 232 315 221 228 215 4 17 30 30 54 14 216 269 280 203 181 262 123 133 128 102 84.2 133 68.6 102 100 55.0 65.8 93.8 24.2 33.6 51.5 46.5 31.5 35.4 14.3 13.9 10.0 6.8 8.3 15.7 14.3 13.9 10.0 6.8 8.3 15.7 342.5 410.7 370.6 244.6 282.9 398.1 342 411 371 245 283 398 Percent of Trips Trip Purpose Mandatory Travel Work commute Other work-related travel Attend school or daycare Household-Serving Travel Transport someone Shopping or errands Medical/dental services Discretionary Travel Social/recreational Dining Community/religious Other Travel Other Return Home Return Home Children 515 20.0% 0.2% 0.2% 19.6% 16.4% 4.1% 11.6% 0.8% 22.7% 13.1% 6.0% 3.5% 2.8% 2.8% 38.2% 38.2% Teens 1617 21.5% 3.3% 0.0% 18.1% 16.8% 4.7% 11.6% 0.5% 23.3% 13.2% 7.4% 2.6% 1.5% 1.5% 36.9% 36.9% Adults Seniors Elderly 1864 6579 80+ Present 17.4% 14.4% 1.3% 1.6% 27.4% 7.5% 18.5% 1.4% 21.4% 10.6% 8.1% 2.7% 1.1% 1.1% 5.3% 4.7% 0.5% 0.1% 35.8% 5.1% 28.1% 2.7% 24.9% 11.4% 8.9% 4.6% 0.9% 0.9% 1.6% 1.3% 0.2% 0.2% 37.3% 3.6% 29.7% 4.0% 27.3% 13.7% 7.4% 6.3% 0.9% 0.9% 4.1% 2.2% 0.4% 1.5% 39.0% 5.1% 27.5% 6.5% 21.8% 10.1% 7.9% 3.8% 1.0% 1.0% 32.7% 33.0% 32.9% 32.7% 33.0% 32.9% 34.0% 34.0% Absent 17.0% 11.7% 1.1% 4.2% 26.1% 6.8% 18.1% 1.1% 22.1% 11.2% 7.9% 3.0% 1.3% 1.3% 33.5% 33.5% 52 As shown in table 22, age-related patterns in mobility are highly gendered. For example, women's mobility decreases at age 65. Men ages 6579, on the other hand, experience little reduction in their overall mobility; a decrease in work travel is largely offset by increases in discretionary and household-serving travel. The differences are even more stark among the elderly (ages 80+). Elderly men make 77 percent as many trips as younger adult men (ages 1864). Elderly women, on the other hand, make only half as many trips as younger adult women, making them substantially less mobile than children.17 Elderly women with disabilities fare even worse, making just 371 trips per year (versus 673 for elderly men with disabilities). Gendered differences in travel manifest at earlier ages, as well. Among adults, women make 62 more trips per year than men overall. However, an even larger gap occurs for household-serving travel in particular; compared to adult men, adult women make 113 more household-serving trips per year, or 2.2 per week. The overall gap shrinks from 113 to 62 largely because adult women make 66 fewer commute trips than adult men do. 17 Some of this difference may come from the fact that women live longer, and a higher percentage of elderly women report having a medical condition that interferes with their mobility (43 percent versus 29 percent). However, the gender gap persists even after accounting for disability. Nondisabled elderly men make 1,062 trips, versus 839 for women. 53 Table 22. Annual tripmaking by purpose, gender, and age. Children 5-15 Trip Purpose Male Female All Purposes 925 862 Mandatory Travel 177 181 Work commute 3 1 Other work-related travel 3 Attend school or daycare 170 180 Household-Serving Travel 144 150 Transport someone 33 40 Shopping or errands 104 102 Medical/dental services 7 8 Discretionary Travel 222 182 Social/recreational 134 100 Dining 60 48 Community/religious 28 35 Other Travel 32 17 Other 32 17 Return Home 350 332 Return Home 350 332 Note: detail may not sum to totals because of rounding. Teens 16-17 Male Female 945 912 225 176 36 26 0 0 188 150 114 194 18 67 91 123 4 5 234 198 137 110 72 65 25 23 18 11 18 11 354 332 354 332 Adults 18-64 Male Female 1,223 1,285 256 182 215 149 18 15 23 18 286 399 72 116 203 260 11 23 262 275 130 136 103 100 29 38 13 15 13 15 407 414 407 414 Seniors 65-79 Male Female 1,196 1,062 83 41 73 35 8 4 1 2 405 399 61 53 318 312 26 33 306 259 139 119 117 86 49 53 14 7 14 7 389 355 389 355 Elderly 80+ Male Female 948 636 28 4 26 1 2 1 2 345 242 23 28 284 187 37 26 256 175 133 85 74 45 49 45 7 7 7 7 312 209 312 209 54 This gendered division of household-serving travel begins early; teenaged girls (ages 1617) make 1.7 times as many household-serving trips as teenaged boys. It would be a mistake to attribute this difference entirely to social outings to the local mall. Teen girls make 1.3 times as many shopping and errands trips as teen boys, but they make 3.6 times as many trips to transport someone else as do teen boys.18 In a typical week, a teen boy will make 0.4 trip to transport someone, while a teen girl will make 1.4 trips to transport someone. At the same time, teen girls make markedly fewer recreational trips than do teen boys. Chapter 5, How Much and What For: The Interrelated Effects of Gender and Age on Mobility and Trip Purpose will revisit these issues. Table 23 and table 24 respectively show trip purpose by mode of travel and mode of travel by trip purpose. So, for example, 11.8 percent of POV trips have a purpose of work commute (table 23), and 90.7 percent of work commute trips are by POV (table 24). Some uncommon modes have a relatively small number of trips in the NHTS Georgia data. For greater transparency when interpreting estimates, sample sizes by mode are included in table 23. In particular, the reader should be extremely cautious when interpreting statistics about the purpose of trips by paratransit and bus transportation (aside from school bus and public transit). 18 Under Georgia's graduated driver's license program, teenagers of all genders face limitations on transporting nonfamily members. Teens are prohibited from transporting nonfamily members for the first 6 months after receiving their license, then limited to one nonfamily passenger for the next 6 months. While the data do not have licensing information about the teen drivers surveyed, these restrictions by definition apply to all 16-year-old drivers and some portion of 17-year-old drivers. Teen drivers who have had their license for at least a year are still limited to three nonfamily passengers under age 21, and no driver under 18 is permitted to drive between midnight and 5 a.m. For details, see: https://dds.georgia.gov/teen-drivers. 55 Table 23. Trip purpose by mode of travel. Purpose Privately Pedestrian Operated (walk or Vehicle wheelchair) Bike School Bus Public Transit Taxi, Other Other Ridehail, or Ground or Bus Limo Paratransit Water Air Work commute Other work-related travel 11.8% 1.0% 6.2% 0.8% 6.1% 0.6% 0.1% 21.0% 1.0% 3.8% 5.6% 11.4% 4.1% 9.0% 20.2% 3.0% 5.2% 17.4% Attend school or daycare Transport someone 2.7% 7.3% 3.9% 2.1% 1.8% 44.8% 6.7% 7.4% 4.2% 4.3% 4.1% 1.9% 3.8% 0.1% 1.9% 3.6% 5.1% Shopping or errands 20.3% 11.3% 15.3% 11.8% 1.5% 6.9% 13.6% 10.5% Medical/dental services 1.5% 0.4% 0.1% 4.6% 6.9% 2.6% 23.6% 0.7% Social/recreational 8.8% 38.3% 26.2% 3.4% 6.5% 36.8% 9.3% 3.3% 14.4% 4.9% Dining 8.3% 7.4% 6.8% 0.0% 4.5% 7.5% 9.8% 5.7% 0.9% Community/religious 3.3% 1.4% 1.0% 0.3% 1.4% 4.3% 10.8% 6.1% 0.1% Other 0.8% 2.9% 2.9% 3.6% 2.8% 6.2% 7.1% 4.7% 3.3% 64.6% Return home 34.1% 25.3% 39.8% 40.5% 34.8% 19.0% 36.1% 35.8% 36.7% 1.9% Total annual trips, in thousands 9,466,980 887,175 68,190 324,798 157,103 22,487 63,151 15,302 49,478 19,005 85.5% 8.0% 0.6% 2.9% 1.4% 0.2% 0.6% 0.1% 0.4% 0.2% Sample size (unweighted trips) 52,675 4,201 287 1,213 515 141 222 56 289 105 Including rental cars. The most common trip purpose for air was to change to a different mode of transportation. This is likely because most air travelers will take a different form of transportation from the airport to their final destination. Blank cells indicate combinations that were not found in the dataset (e.g., no participant reported using a school bus to go shopping). 56 Table 24. Mode of travel by trip purpose. Purpose Privately Operated Vehicle Pedestrian (walk or wheelchair) Bike School Bus Public Transit Other Bus Taxi, Other Ridehail, or Ground or Limo Paratransit Water Work commute Other work-related travel 90.7% 84.6% 4.5% 6.3% 0.3% 0.2% 0.3% 2.7% 1.4% 0.1% 1.1% 0.6% 2.2% 0.1% 0.8% 1.3% Attend school or daycare 56.3% 7.7% 0.3% 32.2% 2.6% 0.2% 0.3% 0.1% 0.2% Transport someone 93.4% 2.5% 2.9% 0.9% 0.1% <0.1% 0.2% Shopping or errands 93.2% 4.9% 0.5% 0.9% <0.1% 0.2% 0.1% 0.3% Medical/dental services 88.5% 2.5% <0.1% 4.5% 1.0% 1.0% 2.3% 0.2% Social/recreational 67.4% 27.6% 1.4% 0.9% 0.8% 0.7% 0.5% <0.1% 0.6% Dining 89.9% 7.5% 0.5% <0.1% 0.8% 0.2% 0.7% 0.3% Community/religious 92.6% 3.7% 0.2% 0.3% 0.7% 0.3% 2.0% 0.3% 0.0% Other 55.6% 17.6% 1.4% 8.2% 3.0% 1.0% 3.1% 0.5% 1.1% Return home 86.8% 6.0% 0.7% 3.5% 1.5% 0.1% Including rental cars. Blank cells indicate combinations that were not found in the dataset (e.g., no participant reported using a school bus to go shopping). 0.6% 0.1% 0.5% Air 0.1% 2.8% 0.2% 0.1% 0.0% 8.5% <0.1% 57 POVs account for the majority of trips for every purpose. They account for more than 80 percent of trips for every purpose except attending school or daycare, social and recreational trips, and trips whose purposes could not be classified. Non-POV trips to school were primarily by school bus. A further 8 percent of trips to school were by walking and biking, and 2.6 percent were by transit (table 24). Aside from home trips, social and recreational trips make up the largest share of nonmotorized modes, with shopping and errands in second place. The largest proportion of transit trips are work trips, followed by shopping and errands. VEHICLE AVAILABILITY AND USAGE Vehicle Ownership Georgia households own an average of 1.92 vehicles. As figure 5 shows, 93 percent of households own at least one vehicle. Number of Household Vehicles Three or more, 24.9% Zero, 6.9% One, 33.6% Two, 34.6% Figure 5. Pie chart. Vehicle ownership among Georgia households. 58 In addition to raw numbers of vehicles, it is important to consider the number of vehicles available in relationship to the number of potential drivers. As shown in figure 6, 19 percent of Georgia households have at least one vehicle, but not as many vehicles as potential drivers. These households can be described as having a vehicle deficit (Blumenberg et al. 2018). Approximately half of vehicle-deficit households have what this report refers to as a hard deficit, where the number of vehicles is smaller than the number of household members listed as drivers. The other half have a soft deficit, with enough vehicles available for each listed driver, but not enough for other household members of driving age.19 Nondeficit households include vehicle-sufficient households (i.e., an equal number of vehicles as potential drivers), and vehicle-surplus households (i.e., more vehicles than potential drivers). About 5 percent of vehicle-sufficient households have a "soft" surplus because one or more people of driving age is listed as a nondriver. 19 Some writers, including Blumenberg et al., (2018) calculate vehicle deficit based purely on the number of drivers. This report considers nondriving adults and teenagers for several reasons. First, the NHTS's determination of "driver" is somewhat ambiguous with regard to whether teenagers with learner's permits and elderly people who are licensed to drive but refrain from doing so should be considered drivers. Basing vehicle sufficiency measures on the number of people of driving age provides a more consistent measure. Second, nondriving adults present additional, competing travel needs that must be met with available vehicles. Third, some adults and teenagers may fail to learn to drive if they do not expect to have access to a car; it is more consistent with the purpose of these classifications to consider households with such individuals as being vehicle-deficit rather than vehicle-sufficient. 59 Figure 6. Bar graph. Vehicle sufficiency in Georgia households. Researchers debate whether households with zero vehicles should be referred to as "carless" or "car free," the latter to suggest that not owning a vehicle is a valid lifestyle choice.20 However, in Georgia, the demographics of zero-vehicle and vehicle-deficit households suggest that these households have reduced vehicle ownership out of economic necessity (table 25). Nearly one third of households in the lowest income bracket are carless, while less than 1 percent of households earning $75,000 or above are car-free. The proportion of nondeficit households increases steadily at higher income levels. There are also notable racial discrepancies; households with nonwhite members are more likely to have a vehicle deficit or be carless. Lower-income households were less likely than higher-income households to have purchased a vehicle within the past year (table 26). However, among only vehicle-owning households, the 20 For the same reason, this research team was reluctant to imply that households with fewer vehicles than driving-age members are "deficient," but use the admittedly similar term "deficit" in keeping with Blumenberg et al. (2018). 60 lowest-income households are nearly as likely as the wealthiest households to have purchased a vehicle within the past year. Table 25. Vehicle sufficiency by income, race, and number of household vehicles. All Households Annual Household Income 1. <$15,000 2. $15,000 to $24,999 3. $25,000 to $34,999 4. $35,000 to $49,999 5. $50,000 to $74,999 6. $75,000 to $99,999 7. $100,000+ Number of Vehicles in Household 1 2 3+ Race of Household Members White non-Hispanic Some or all nonwhite Zero-Vehicle 6.9% 32.5% 6.9% 2.9% 1.1% 1.1% 0.4% 0.1% n/a n/a n/a 2.9% 11.6% Deficit Nondeficit 18.9% 74.1% 24.3% 30.5% 23.8% 22.0% 17.4% 12.9% 9.2% 43.2% 62.6% 73.3% 76.8% 81.5% 86.7% 90.7% 36.4% 14.4% 6.9% 63.6% 85.6% 93.1% 13.0% 25.9% 84.1% 62.5% It is likely that low-income households' frequent vehicle purchases are by obligation rather than choice; the average vehicle purchased by the lowest-income households already has 133,973 miles on the odometer, nearly 40,000 more miles than any other income group. Although vehicles nearer the end of their life spans are cheaper to purchase, maintenance costs may become prohibitive, leading to faster turnover. Nearly one quarter of vehicles being driven by the lowest-income households were purchased in the past 12 months (table 26). Table 27 also shows the condition of vehicles that were not recently purchased. 61 Table 26. Vehicle purchasing behavior by household income. Annual Household Income Percent of Households that Purchased a Vehicle Recently All Households Vehicle-Owning Households Only Condition of Newly Purchased Vehicles Mean Age of Vehicle (years) Mean Mileage of Vehicle All income levels 26.1% 28.0% 7.3 69,333 1. <$15,000 20.7% 30.6% 11.4 133,973 2. $15,000 to $24,999 17.4% 18.7% 9.2 86,610 3. $25,000 to $34,999 23.6% 24.3% 8.8 94,393 4. $35,000 to $49,999 25.1% 25.4% 8.3 74,867 5. $50,000 to $74,999 27.0% 27.3% 6.6 61,638 6. $75,000 to $99,999 31.5% 31.7% 6.4 55,073 7. $100,000+ 33.8% Odometer reading at time of survey, 011 months after date of purchase. 33.8% 5.0 43,019 62 Table 27. Characteristics of vehicles owned by households of different income levels. Mean Vehicle Annual Household Income Age (years) Average Vehicle Odometer Reading <$15,000 14.2 138,001 $15,000 to $24,999 12.7 124,183 $25,000 to $34,999 11.8 118,761 $35,000 to $49,999 11.1 111,126 $50,000 to $74,999 11.0 110,765 $75,000 to $99,999 10.4 100,295 $100,000+ 9.1 87,436 Odometer reading at time of survey, 011 months after date of purchase. Percent of Vehicles that are Newly Purchased Mean Age of Newly Purchased Vehicles Mean Mileage of Newly Purchased Vehicles 24.1% 11.4 133,973 13.2% 9.2 86,610 17.0% 8.8 94,393 15.5% 8.3 74,867 14.5% 6.6 61,638 15.5% 6.4 55,073 15.2% 5.0 43,019 63 Vehicle Usage On average, each vehicle in a household was driven 11,939 miles over the course of the year. In total, each household used their vehicles to drive an average of 22,472 miles, as shown in figure 7. The median was 18,048 miles. The figure is top-coded at 100,000 miles for clarity; the 99th percentile for miles driven was 102,166 miles. Figure 7. Histogram. Distribution of annual household miles driven. The relationship between number of vehicles and household VMT is not constant across household type. Vehicle-deficit households, in particular, drive each vehicle more than other households do (table 28). 64 Table 28. Miles driven per vehicle by household vehicle sufficiency. Vehicle Ownership Status All vehicle-owning households Nondeficit, single potential driver Nondeficit, multiple potential drivers Vehicle-deficit households Percent of Households 93.1% 26.4% 47.7% 18.9% Annual Miles Driven Per Household Vehicle 12,169 11,145 11,889 14,304 Vehicle Fleet Characteristics Georgia's fleet of personal occupancy vehicles is diverse and evolving. The median age of vehicles is 10 years, but 5.6 percent are less than 2 years old and 1.3 percent have been on the road for 40 years or more. Table 29 divides vehicles into cohorts based on the passage of California's Low Emissions Vehicle standards, which serve as a proxy for tightening emissions standards. In addition to differences in age and mileage, body types vary between age cohorts. Contemporary vehicles include a larger proportion of SUVs and fewer pickup trucks and motorcycles. New vans are also somewhat less common than vans from the two preceding decades. Alternative-fuel vehicles now account for 4.2 percent of contemporary vehicles and 1.8 percent of the overall fleet. Of these, 63 percent are hybrid vehicles, 29 percent are electric, and 5.0 percent are plug-in hybrids. Alternative-fuel Vehicles in chapter 4 further discusses alternative-fuel vehicles. 65 Table 29. Vehicle characteristics by vehicle age cohort. Number of vehicles Percent of fleet Average age Newly purchased by household (within past 12 months) Vehicle Mileage Mean odometer reading 049,999 mi 50,00099,999 mi 100,000149,999 mi 150,000199,999 mi 200,000+ mi Vehicle Type Auto/wagon SUV Pickup truck Van or minivan Motorcycle Other Alternative fuel (any body type) Pre-LEV All Vehicles (before 1993) 6,997,337 319,663 100% 4.6% 10.9 34.7 LEV I (19932003) 1,854,254 26.5% 17.3 LEV II Contemporary (20042014) (20152017) 3,909,762 856,277 55.9% 12.2% 7.9 1.5 15.8% 8.6% 10.2% 11.8% 49.3% 105,434 28.6% 24.1% 20.6% 14.4% 12.3% 49.5% 23.6% 17.5% 5.4% 2.4% 1.6% 1.8% 154,431 13.0% 15.8% 20.6% 19.9% 30.6% 42.2% 8.1% 40.1% 2.5% 3.3% 3.8% 0.8% 166,628 6.1% 10.6% 23.8% 28.0% 31.5% 44.3% 20.2% 24.1% 5.6% 3.1% 2.7% 0.2% 94,991 24.2% 35.3% 23.9% 11.4% 5.3% 52.7% 24.8% 13.6% 5.8% 2.1% 0.9% 2.2% 19,453 95.1% 3.9% 0.4% 0.6% 49.9% 32.3% 12.0% 3.7% 0.9% 1.2% 4.2% 66 In addition to "normal" vehicles used in day-to-day travel, some residents have seldom-used vehicles. These vehicles are not frequently driven because their owners have primary vehicles that are used for most household trips (as differentiated from owners who have a single vehicle but drive infrequently). These secondary, seldom-used vehicles may be hobby or leisure vehicles (e.g., motorcycles, campers, antique cars, etc.). Others are special-use vehicles, such as a pickup truck kept on hand for occasional freight hauling. These seldom-used vehicles account for approximately 4 percent of vehicles in Georgia. As shown in table 30, these vehicles tend to be older, and are more likely to include leisure vehicles such as RVs, motorcycles, and miscellaneous vehicles such as campers. Table 30. Characteristics of seldom-used vehicles. Type of Vehicle Regular Seldom Used 1. Car/Wagon 50.8% 33.7% 2. Van 5.5% 2.9% 3. SUV 24.6% 9.2% 4. Pickup 17.7% 18.6% 5. Other Truck 0.3% 0.7% 6. RV 0.3% 4.3% 7. Motorcycle 0.8% 27.2% 97. Something Else 0.0% 3.4% Age Mean Age 10.56 16.64 Age Cohort 0. Pre-LEV (pre-1993) 4.1% 14.2% 1. LEV1 (19932003) 25.6% 45.1% 2. LEV2 (20042014) 57.4% 36.8% 3. Contemporary (20152017) 12.8% 3.9% A seldom-used vehicle is defined as a vehicle with fewer than 1,050 annual miles that is driven less than half as many miles as would be expected given the household annual miles driven and number of vehicles. 67 Seldom-used vehicles are predominantly owned by higher-income households (table 31). Households earning less than $50,000 annually account for half of all households and 33.6 percent of seldom-used vehicles. Those earning more than $100,000 account for 22.4 percent of all households and 34.7 percent of the vehicles. Table 31. Ownership of seldom-used vehicles by annual household income. Annual Household Income 1. <$15,000 2. $15,000 to $24,999 3. $25,000 to $34,999 4. $35,000 to $49,999 5. $50,000 to $74,999 6. $75,000 to $99,999 7. $100,000+ Percent of Households 16.8% 10.4% 11.0% 12.1% 16.4% 10.9% 22.4% Percent of Regular Vehicles 8.2% 7.6% 9.4% 12.0% 18.5% 13.7% 30.6% Percent of Seldom-Used Vehicles 6.3% 7.3% 7.3% 12.7% 17.2% 14.5% 34.7% TRANSIT PREFERENCES AND USE While personal occupancy vehicles are the dominant form of transportation in Georgia, there is a societal interest in promoting the use of public transit. This section summarizes transit availability and then analyzes Georgia workers' preferences about transit quality. Transit Availability and Use Table 32 summarizes county-level transit funding within each MPO.21 Twelve MPOs receive funding to offer fixed-route public transit in at least one county. Four MPOs offer only rural 21 Data on transit funding generously provided by Garrow et al. (2018). Preliminary report is available at http://garrowlab.ce.gatech.edu/sites/default/files/20191027%20Rural%20Transit%20in%20Georgia.pdf 68 (on-demand) service. Thirty-day transit usage also varies, from a low of 3.4 percent of the population ages 5+ (Hinesville) to 17.3 percent (Athens). Transit access can vary considerably within an MPO. Table 33 classifies counties by their level of access to transit. The level of access is based on transit offerings in the county, and in the MPO of which the county is part. The majority of the population resides in counties with a full fixed-route transit system (though this does not take into account quality of the system or proximity to a transit station). The majority of counties, which are home to 36.5 percent of the population, have partial access to transit. Some of these counties are rural counties that provide on-demand service. The remainder are counties in MPOs that do not offer fixed-route service. However, residents of these counties can theoretically access transit in other counties within the MPO. Unsurprisingly, transit use is highest in counties with full access. Nevertheless, 89 percent of users in counties with partial access have used transit within the past 30 days. Transit Service Preferences Among Workers As part of the NHTS add-on module, workers in Georgia were asked an additional question about transit preferences. From a list of seven possibilities, they were asked to select the "three most important factors that would make [your/their]22 public transit system a good option for [your/their] commute." Participants' selections were not ranked against each other--each was a value of either yes or no. 22 This is one of the few opinion questions in the NHTS that was still asked of subjects whose responses were recorded by a proxy. 69 Table 32. County transit funding and transit usage by MPO. MPO Counties in MPO Number of Counties in Each Transit Funding Category* None Rural (OnDemand) Urban Urban & Rural City Only Fixed Route Transit Albany 2 0 1 0 1 0 Athens 4 3 0 1 0 0 Atlanta 18 4 5 6 2 1 Augusta 2 0 1 0 1 0 Cartersville 1 0 0 0 1 0 Chattanooga 3 0 3 0 0 0 Columbus 3 2 0 1 0 0 Gainesville 2 0 1 0 1 0 Hinesville 2 0 1 0 1 0 Macon 2 0 1 1 0 0 Rome 1 0 0 1 0 0 Savannah 3 0 2 0 1 0 Rural (On-Demand) Service Only Brunswick 1 0 1 0 0 0 Dalton 2 0 2 0 0 0 Valdosta 1 0 1 0 0 0 Warner Robins 2 1 1 0 0 0 Non-MPO Counties Non-MPO 110 27 78 0 0 5 * As reported by Garrow et al. (2018). Fixed route service for an entire county (as compared to city, which covers only part of a county). Percent of population ages 5+, weighted. Usage Transit Use, Past 30 Days 14.7% 17.3% 15.7% 10.3% 7.9% 11.0% 11.4% 8.3% 3.4% 13.0% 8.1% 10.7% 4.3% 9.2% 12.1% 8.0% 6.9% 70 Table 33. County demographics and transit usage by transit funding category. County Information Transit* Use (past 30 days) Transit Funding Status County MPO* Demographics Number of Percent of Counties Population** Statewide Population Workers Ages 5+ Only MPO Tier 1 (Atlanta) Population Workers Ages 5+ Only MPO Tiers 24 Population Workers Ages 5+ Only No Access None None 27 4.3% 2.9% 0.3% 2.9% 0.3% None None Partial Access Rural Fixed-route 1 1.4% 8.5% 6.2% 8.5% 6.2% 9 5.3% 8.5% 5.6% 9.5% 6.9% 5.6% 2.4% Rural Rural 83 18.1% 8.3% 3.1% 8.3% 3.1% Rural Fixed-route Full Access 15 11.7% 9.2% 8.5% 11.3% 11.5% 6.5% 4.5% Fixed-route Fixed-route 24 59.2% 15.8% 15.6% 16.9% 17.5% 12.8% * For counties not in an MPO, county status is used for both columns. ** Defined by the NHTS as public or commuter bus, rail (Amtrak, commuter, elevated, light rail), and street car; does not include paratransit/dial-a-ride. Funding for rural (on-demand) transit only. Funding for fixed-route service in part or all of county jurisdiction, with or without funding for rural on-demand service. Indicates access to fixed-route transit service at county level; does not consider individual distance to transit or level of service. 9.4% 71 In general, proximity and cost were the two most common responses, though the relative frequencies varied by geography, vehicle ownership, and current transit access and use (table 34). Proximity and other convenience measures were more highly valued by residents of MPO tiers 1 and 2 and counties with full transit access. Residents of small MPOs and non-MPO counties selected cost more frequently. The most common response for zero-vehicle households was cost; this was the only group for whom proximity was not either the first or second choice. Occasional transit users were more likely to focus on both proximity and cost than either nonusers or moderate/frequent users; these groups' preferences were more heterogeneous. Workers whose "usual commute mode" was transit, paratransit, or some other bus represent 3.4 percent of the workforce. The top concern of this group, by 9 percentage points, was cost. The second- and third-most common concerns for these commuters were proximity and consistent on-time performance. For most other groups, the third-most common concern was compatibility with their schedule. While both characteristics are time-based, nontransit commuters were more likely to focus on frequency, whereas transit commuters were more focused on reliability. Preferences also varied across demographic groups. Table 35 shows responses by income, sex, age, disability, and race. 72 Table 34. Perceived most important factors to make transit a good commute option by geography, vehicle ownership, and transit access/use. Close to Work and Home Fits Faster than Reasonable Consistently Schedule Driving Cost on Time Avoids Travel Stress Safety All Workers MPO Tier 1. Atlanta 2. Medium MPOs 3. Small MPOs 55% 58% 55% 47% 45% 43% 55% 45% 38% 43% 33% 27% 48% 45% 48% 54% 36% 35% 36% 41% 29% 32% 22% 23% 27% 25% 27% 35% 4. Non-MPO County Access to Transit None 47% 61% 41% 54% 27% 20% 59% 64% 36% 23% 25% 25% 27% 24% Partial 51% Full 57% Household Vehicle Ownership Zero-vehicle 42% Vehicle-deficit 54% Nondeficit 56% 44% 46% 32% 42% 47% 34% 40% 21% 29% 41% 52% 46% 51% 50% 48% 38% 36% 45% 40% 34% 25% 31% 19% 26% 30% 30% 26% 28% 32% 25% Actual Transit Usage (past 30 days) None 55% 15 days 62% 46% 43% 38% 47% 49% 34% 35% 34% 28% 31% 28% 21% 6+ days Usual Commute Mode 54% 41% 39% 47% 41% 37% 22% Public transit, other bus, or paratransit 44% 39% 40% 53% 42% 39% 20% Something else 56% 46% 37% 48% 37% 27% 26% 73 Table 35. Transit service quality priorities of Georgia workers by demographic group. Close to Work and Home Fits Faster than Reasonable Consistently Schedule Driving Cost on Time Avoids Travel Stress Safety All workers 55% 45% 38% 48% 36% 29% 27% Annual Household Income <$15,000 41% $15,000 to $24,999 53% $25,000 to $34,999 50% $35,000 to $49,999 53% $50,000 to $74,999 54% $75,000 to $99,999 56% $100,000+ 62% 37% 40% 42% 38% 43% 48% 52% 19% 30% 32% 32% 32% 41% 51% 55% 48% 55% 52% 53% 47% 42% 42% 43% 36% 39% 36% 34% 33% 30% 25% 31% 32% 33% 27% 27% 37% 33% 29% 29% 27% 26% 21% Sex Male Female 56% 54% 46% 44% 40% 36% 48% 49% 36% 36% 30% 28% 22% 32% Age Teens 1617 Adults 1864 Seniors 6579 Elderly 80+ 52% 56% 52% 61% 68% 45% 42% 9% 22% 39% 32% 12% 67% 48% 52% 53% 12% 36% 39% 37% 6% 29% 40% 35% 49% 27% 22% 23% Presence of Mobility Impairment Disability 47% 33% 29% 50% 37% 34% 35% No disability 56% 45% 38% 48% 36% 29% 27% Race White non-Hispanic 59% 50% 43% 47% 34% 28% 21% Black and Black multiracial Other race 50% 52% 40% 38% 32% 32% 51% 50% 41% 34% 30% 32% 33% 36% To isolate the effects of these interrelated factors, responses were modeled using logistic regression, with a separate model estimated for each possible response. Table 36 presents the 74 odds ratios23 associated with each variable. The models highlight differential priorities that generally conform to the ideas of captive and choice riders. Compared to low-income workers, wealthier workers were more likely to prioritize proximity, schedule, and speed, and less likely to cite cost as an important factor. A similar split was evident along racial lines, with white, non-Hispanic workers more likely than workers of color to emphasize proximity, schedule, and speed. Black workers were more likely than white workers to emphasize reliability, and all nonwhite workers were more likely to cite safety. It is worth noting that once income is controlled for, there is no evidence of racial differences in the likelihood of prioritizing affordability. The idea of transit as a means for avoiding travel stress was most popular in the highly congested Atlanta metropolitan area. Workers from households with cars were more likely to express interest in avoiding travel stress than workers from zero-vehicle households. Interestingly, lowincome workers were comparatively more likely than other groups to select "avoid travel stress," though it was still one of the least commonly chosen options for these respondents. 23 Odds ratios (ORs) are the factors by which the odds of choosing each response differ for people with the associated characteristic (e.g., living in a medium MPO area), compared to the base or reference group (for the continuous variable age, the OR is the factor by which the odds differ with a 1-year increase in age). An OR of 1 means that the characteristic has no impact on the odds of choosing that response. An OR greater than 1 means that people with that characteristic are more likely to choose the associated response than those in the base group, and conversely for an OR less than 1. 75 Table 36. Logistic regression: Transit service quality priorities of Georgia workers. 1. Close to Work & 2. Fits Schedule Home 3. Faster than Driving Covariates OR P-Value OR P-Value OR P-Value MPO Tier (base: Tier 1 Atlanta) 2 - Medium MPOs 0.922 0.375 1.662 <0.001 *** 0.674 <0.001 *** 3 - Small MPOs 0.703 0.002 *** 1.169 0.169 0.538 <0.001 *** 4 - Non-MPO 0.600 0.001 *** Transit use, past 30 days (base: not used) 15 days 1.190 0.457 0.847 0.873 0.295 0.560 0.584 0.002 *** 1.208 0.424 6+ days 0.930 0.600 0.884 0.371 1.004 0.977 Annual Household Income (base: <$15,000) $15,000 to $24,999 1.569 0.058 * 1.231 0.388 1.649 0.064 * $25,000 to $34,999 1.503 0.061 * 1.265 0.279 1.819 0.015 ** $35,000 to $49,999 1.561 0.032 ** 0.982 0.930 1.496 0.090 * $50,000 to $74,999 1.616 0.016 ** 1.241 0.284 1.434 0.114 $75,000 to $99,999 $100,000+ 1.689 0.011 ** 1.527 0.038 ** 2.075 0.002 *** 2.051 <0.001 *** 1.732 0.005 *** 2.809 <0.001 *** Household vehicle ownership (base: zero vehicles) Vehicle-deficit 1.144 0.650 1.164 0.633 1.102 0.795 Nondeficit 1.019 0.948 1.178 0.602 1.459 0.305 Race (base: white non-Hispanic) Black or Black multiracial 0.744 0.002 *** 0.764 0.006 *** 0.763 0.007 *** Other race 0.765 0.042 ** 0.657 0.002 *** 0.685 0.006 *** Other Female 0.984 0.846 0.981 0.812 0.890 0.162 Age in years 0.998 0.547 0.997 0.293 0.995 0.104 Has mobility impairment 0.855 0.621 0.713 0.270 0.923 0.826 Constant 0.977 0.943 0.638 0.198 0.414 0.026 ** Note: Because weighted logistic regression was used, pseudo-R 2 statistics were not provided. Odds ratio. 4. Reasonable 5. Consistently on 6. Avoids Travel Cost Time Stress OR P-Value OR P-Value OR P-Value 1.039 1.304 1.515 0.674 0.020 ** 0.009 *** 1.056 1.291 1.198 0.560 0.030 ** 0.254 0.596 <0.001 *** 0.600 <0.001 *** 0.657 0.018 ** 0.576 0.020 ** 0.968 0.814 0.972 0.909 1.173 0.247 1.094 0.702 1.666 <0.001 *** 0.839 1.035 0.935 0.993 0.786 0.650 0.449 0.871 0.741 0.973 0.231 0.023 ** 1.048 0.796 0.915 0.856 0.829 0.845 0.843 0.300 0.677 0.444 0.376 0.398 0.637 0.852 0.857 0.870 0.587 0.552 0.081 * 0.503 0.498 0.515 0.019 ** 0.005 *** 0.956 0.880 1.175 0.587 1.042 0.894 0.820 0.523 2.662 0.009 *** 3.127 0.003 *** 1.147 0.151 1.075 0.584 1.283 0.012 ** 1.023 0.869 1.007 0.947 1.247 0.124 0.962 0.996 1.014 1.060 0.634 0.234 0.966 0.860 0.958 1.000 0.923 0.671 0.610 0.870 0.809 0.244 0.879 1.012 1.122 0.139 0.150 <0.001 *** 0.724 <0.001 *** 7. Safety OR P-Value 1.055 1.394 1.103 0.618 0.011 ** 0.578 0.732 0.248 0.724 0.058 * 0.842 0.666 0.785 0.756 0.725 0.622 0.488 0.087 * 0.286 0.199 0.156 0.028 ** 1.383 0.358 1.393 0.352 1.572 <0.001 *** 2.032 <0.001 *** 1.650 0.992 1.395 0.295 <0.001 *** 0.025 0.327 0.002 *** 76 Gender was not strongly affiliated with preferences with one notable exception: the odds of women citing safety concerns were 65 percent greater than those of men. This finding echoes many other studies on transit security and gender, and it is worth noting that concerns about safety can depress transit use among women (Keane 1998, Loukaitou-Sideris 2014, Clark et al. 2016). Once other differences are controlled for, recent transit use has relatively few effects on expressed preferences.24 Occasional users are less concerned about cost than both frequent users and nonusers. Frequent users are more likely to mention avoiding travel stress. Both types of user may be less concerned about safety than non-users, but the effect is insignificant for occasional users and only borderline significant for frequent users. SUMMARY OF FINDINGS ON GEOGRAPHICAL DIFFERENCES The first section of this chapter (see Geographic Divisions for Analysis in Methodological Notes) has described the division of Georgia's counties into MPO tiers used throughout the report.25 The second section (Overview) provided a basic overview of mobility indicators by MPO size and, where possible, individual MPOs. The third section (Trip Patterns by Location of Travel) analyzed trip location (as opposed to residence of the traveler, which is the classification used for the rest of this report), and the seventh section (Transit Preferences and Use) discussed 24 In addition to including transit use as a covariate, a series of models fully interacted by transit ridership was estimated. These models were not an improvement over the more parsimonious models presented here. Models fully interacted by Atlanta vs. other MPO tiers were also tried and rejected. 25 Tier 1 comprises counties that partially or wholly fall within the Atlanta MPO. Tier 2 consists of counties that partially or wholly fall within a medium MPO region, and tier 3 consists of counties in small MPO regions. Tier 4 consists of counties that do not fall within any MPO. See table 1 for classifications of individual counties. 77 differences in transit availability, use, and preferences among Georgians from different MPOs and MPO tiers. This section summarizes key findings about geographical differences from subsequent chapters. Work Travel (Chapter 2Chapter 3) At a state level, a plurality of jobs (46 percent) is in the professional/managerial/technical sector, followed by sales and service (27 percent), blue-collar (i.e., manufacturing, construction, maintenance, and farming, 18 percent) and clerical (9 percent) (see chapter 2, Worker Characteristics). There is, however, significant geographic variation. Blue-collar jobs account for 29 percent of total jobs in non-MPO counties (tier 4), which is more than twice the share in Atlanta (tier 1). Tiers 2 and 3 fall somewhere in the middle. The inverse is true of professional jobs, which are most common in tier 1 and least common in tier 4. These industry differences have implications for commuter schedules, since blue-collar workers tend to have more atypical schedules (e.g., working nights, weekends, holidays, or unpredictable hours). Perhaps relatedly, the annual total work journeys per worker is higher in non-MPO counties than elsewhere in the state. Private occupancy vehicles, whether singly-occupied or carrying passengers, are the dominant commute mode (see chapter 2, Commute Mode by Person). In Atlanta, where transit offerings are comparatively robust, 91 percent of workers usually commute by POV, compared to 94 percent of workers in medium MPO counties and 97 percent of workers in small-MPO and non-MPO counties. 78 Commute durations are longest in the Atlanta MPO area (see chapter 2, Commute Duration and Burden ). The region also has the largest variability in commute duration by time of day; the average PM peak commute is 42.2 minutes, which is 13.6 minutes longer than the average overnight commute. Non-MPO counties have the second-longest commute times, but the least variability by time of day. Small MPO counties have the shortest commute durations on average. On a typical workday, 12 percent of Georgia commuters spend 2 hours or more traveling to and from work (see chapter 2, Commute Duration and Burden). These heavy commute burdens are not equally distributed; 2-hour commute burdens are more than twice as common in the Atlanta region as elsewhere in the state (16.6 percent in Atlanta versus 6.27.1 percent elsewhere). Complex commutes (those involving at least one stop) are equally common across MPO tiers (see chapter 2, Overview of Commuters). Additional stops typically add distance to commutes. However, idiosyncratic local geography--such as Atlanta's complex highways or the rivers bisecting downtown Macon, Columbus, and Albany--resulted in some commutes being made geographically shorter by adding stops that encouraged the traveler to follow the geographically shortest route rather than the fastest one (see chapter 2, Demographic Differences). At the time of data collection, 46 percent of Georgia workers had a flexible work schedule, location, or both (see chapter 3, Overview). Both kinds of flexibility are more commonly available to workers in Atlanta, where 54 percent have one or both kinds of work flexibility, compared to 37 percent of workers elsewhere in the state. Flexible work locations (including teleworking and home-based work) are most common in Atlanta, where commute durations are greater than in the rest of the state, and in non-MPO counties, where the distance between home and work is, on average, higher than elsewhere in Georgia. 79 New Technology (Chapter 4) Adoption of alternative-fuel vehicles, ridehailing services such as Uber and Lyft, and online shopping are all more pronounced in Atlanta compared to the rest of the state (see chapter 4, sections on Alternative-fuel Vehicles, Shared Mobility, and Online Shopping). However, while ridehailing usage is lowest in non-MPO counties, users from these counties make more ridehailing trips than users from other types of counties (see chapter 4, Ridehailing). Ridehailing accounts for an estimated 87 percent of all vehicle-for-hire trips in Georgia, with the remainder conducted by traditional taxi and limo services (see chapter 4, Ridehailing and Vehicle-for-Hire Trips). However, in small MPO regions and non-MPO counties, ridehailing accounts for 95 percent of all vehicle-for-hire trips. Equity (Chapter 5) Per capita tripmaking is the highest in Atlanta MPO counties and the lowest in non-MPO counties (see chapter 5, Key Equitable Mobility Indicators). Similarly, immobility is lowest in Atlanta and highest in non-MPO counties. However, despite the higher average mobility in Atlanta, the region is still home to a number of mobility-disadvantaged populations, especially low-income and transit-dependent people. Neighborhoods matching the urban category, which in Georgia are found only in Atlanta, appear to confer some mobility benefits, including supporting mobility among people with mobility impairments (see chapter 5, Risk Factors for Immobility among Adults with Mobility Impairments) After controlling for other factors, Atlanta, Columbus (medium), and Brunswick (small) have the longest transit trips, followed by Savannah and Gainesville (see chapter 5, Key Equitable Mobility Indicators). Non-MPO counties have the shortest transit trips. In all types of counties, 80 captive transit users face significantly longer travel times than choice users (see chapter 5, Vehicle Access). Nonmotorized Travel and Travel for its Own Sake (Chapter 6Chapter 7) While there are some regional differences in nonmotorized travel (NMT), the larger differences are by neighborhood type, which can vary substantially within an MPO (see chapter 6, Overview). Neighborhood type is therefore a more useful way of examining differences in walking and biking. This report examines walking and biking trips as well as nonmotorized legs to access and egress public transit. NMT is most common in the densest urban neighborhood type (which, in Georgia, is only found in Atlanta). On average, residents of urban neighborhoods make more than three times as many NMT trips and transit access/egress legs than residents of second-city and suburban neighborhoods, and more than seven times as many trips/legs as residents of small towns and rural areas (see chapter 6, Travel Day Walking and Biking by Georgia Adults). The average urban resident spends more than 20 minutes per day walking and/or biking, compared to less than 8 minutes in all other neighborhood types. The purpose of NMT also varies by neighborhood type (see chapter 6, Travel Day Walking and Biking by Georgia Adults). The majority of NMT in all neighborhood types is for instrumental purposes. However, leisure travel accounts for more than a third of NMT in small-town and rural neighborhoods versus just 17 percent in urban neighborhoods. Loop trips, which are predominantly NMT and represent a form of travel for its own sake, are most common in Atlanta and non-MPO counties and lower in small and medium MPO regions (see chapter 7, Loop Trips in the 2017 Georgia NHTS Subsample). 81 Finally, among children ages 517, leisure and socialization account for more than half of nonmotorized trips in rural and small-town neighborhoods (51.9 percent and 51.5 percent respectively), as well as in non-MPO counties more generally (51.5 percent) (see chapter 6, Children's Nonmotorized Travel for All Purposes). Elsewhere in the state, these trips for "fun" purposes are outnumbered by instrumental trips. In the aggregate, 8 percent of children in medium MPO counties typically walk or bike to/from school, followed by Atlanta MPO counties (see chapter 6, Children's School Travel). In nonMPO counties, just 3.5 percent walk or bike to/from school. However, the differences are stronger by neighborhood type; 22 percent of children in urban neighborhoods walk or bike to/from school, more than double the percentage in second-city neighborhoods. Just 0.6 percent of children in rural neighborhoods walk or bike to/from school. 82 CHAPTER 2. COMMUTE AND WORK PATTERNS CHAPTER 2 SUMMARY This chapter analyzes work-related travel, particularly commuting. Overview includes technical and vocabulary notes and provides an overview of workforce and commute characteristics. In addition to mode choice and carpooling behavior, the section describes the prevalence of complex work journeys (i.e., trips to or from work with at least one intermediate stop). Commutes to sites other than a respondent's "official" work location are also discussed. Together, complex commutes and commutes involving a nontraditional work location account for one third of work journeys in Georgia, highlighting the importance of effectively measuring these journeys. Defining the Commute discusses several methods of dealing with complex work journeys when measuring work travel. The section focuses primarily on the methods used to define commute distances and durations for the purposes of analysis in this chapter. Readers primarily interested in the results of the analysis may choose to quickly advance to the fifth section (Commute Duration and Burden), which summarizes definitions to be used in the third section, Work Travel Distance, and the fourth section, Work Travel Distance by Mode. Work Travel Distance examines commute distance, person miles traveled, and vehicle miles traveled. Georgia residents travel a total of 26.4 billion miles each year in their journeys to and from work, and 23.6 billion of those miles are as a driver of a private vehicle. Two thirds of commute PMT and VMT occur during weekday peak hours. There 83 is considerable regional variation. Commute distances are longest in the most and least populous areas of the state (i.e., the Atlanta MPO and non-MPO counties). Work Travel Distance by Mode focuses on work travel duration. Workers' median "usual" commute duration is 25 minutes. However, some workers must travel much longer. On a typical day, 12.3 percent of active commuters will spend a total of 2 hours or more traveling to and from work. Commute durations are greatest in the Atlanta MPO. The Atlanta MPO also experiences the greatest variability by time of day. Commute Duration and Burden discusses the amount of time spent on individual commutes and the total amount of time spent on commuting per day (commute burden). Commute durations differ sharply by mode. The average auto commute lasts 31 minutes, versus 17 minutes for nonmotorized commutes. The average public transit commute lasts 70 minutes. However, captive riders' commutes are even slower than those of choice transit riders (i.e., commuters from vehicle-sufficient households). The average transit commute for a captive rider is 20.4 minutes longer than the average choice transit commute, despite the fact that the average distance of captive riders' commutes is 3.1 miles shorter than a choice rider's transit commute (14.2 miles vs. 17.3 miles). The data also suggest the presence of "captive" walkers and cyclists: 8.3 percent of nonmotorized commutes by workers without full access to vehicles last an hour or longer, versus just 0.7 percent of choice nonmotorized commutes. These findings suggest that Georgia's current commuting environment constitutes a two-tiered system divided not just by mode, but by the ability to choose between modes. 84 OVERVIEW Definitions and Technical Notes In general, a commute is travel for the purpose of getting to or from work. However, since some commuters make stops in between their home and workplace, more specific vocabulary is needed. A commuter is a person who travels from a home location to a different location for the purposes of paid employment. This chapter uses general data about the "usual" habits of commuters, and also travel diary data from a single day of travel. It refers to respondents who reported traveling to a workplace on their travel day as active commuters. The phrase "work journey" refers to all the trips and intermediate stops a commuter makes between home and work.26 A work journey is unidirectional (i.e., either from home to work, or from work to home). As a result, most active commuters will have two or more work journeys in a single day. "Circuit" refers to the full sequence of trips from home to work and back. The home and work locations are anchors of the work journey, and any destinations between the two are considered stops. A simple work journey proceeds directly between home and work. A complex work journey contains at least one internal stop. For example, a commuter might stop to pick up a child from school, buy coffee, or shop for groceries on the way to or from work. A work journey might also include longer stops, such as a professional who is pursuing an advanced degree by taking classes after work.27 26 The phrase "work journey" is used rather than "tour" to avoid confusion with the alternate definition used by the NHTS in the tour files it provides to researchers. 27 In contrast to this classification, the NHTS classifies any stop of 30 minutes or more as an anchor in its own right. The reasons for departing from the NHTS classification are discussed further in chapter 2, Defining the Commute. 85 A commute is the portion of the work journey that should be considered work-related travel. For a direct (simple) work journey, the work journey and the commute are identical. For a complex work journey, it is necessary to determine which portions of the journey should be attributed to the commute, and which should be considered as related to the purposes of the intermediate stops. The next section, Defining the Commute, will compare alternate methods of apportioning person miles traveled, vehicle miles traveled, and travel time between work and nonwork purposes. A technique the research team refers to as the counterfactual method is well-suited for a realistic portrayal of complex commutes. In this technique, the portion of a work journey assigned to the commute is based on what the commute would have looked like had the commuter gone directly from home to work with no intermediate stops. The NHTS contains two types of data about work-related travel. The first consists of questions asked as part of the general questionnaire, such as "usual" commute mode, industry of employment, and flexibility of work schedule. Questions it also contains about telecommuting will be analyzed in chapter 3. The second type of data is work trips made by respondents on their travel days. These travel diary data are used to provide more accurate mode shares, distances, and many other measures. Travel diary data are the basis for most analyses in this chapter. It should be noted that travel times are self-reported by survey respondents, but travel distances were calculated by the NHTS as a shortest path distance, rather than based on the route chosen by the traveler. Additionally, travel days ran from 4:00 a.m. through 3:59 a.m., with the result that night shift workers might actually begin or end their day at work. More details about how this report identifies work journeys and commutes within the trip diaries are available in chapter 2, Defining the Commute. 86 Worker Characteristics Table 37 shows levels of workforce participation.28 Statewide, 61.5 percent of adults ages 18 and older are classified as workers by NHTS.29 Of these, 80.4 percent work full-time. Labor force participation is higher among men than women. Female workers are more likely to work parttime, and are also more likely to hold multiple jobs. Regional variations are also present. Workforce participation is highest in the Atlanta MPO region, and lowest in non-MPO counties. Workforce participation is highest for members of Generation X (who were 3752 years old at the time of the survey). However, many Georgians (18.1 percent) continue to work past retirement age; 21.4 percent of younger seniors (ages 6579) are still working, as are 3.6 percent of Georgians ages 80 and older. The dominant industry varies by region, gender, race, age, and income (table 38). Professional, managerial, and technical positions are the most common, both among workers overall and among all subpopulations except the lowest-income workers and those without a college degree. Sales and service is the largest sector for low-income and noncollege-educated workers, and the second largest for other groups. For men, the second-most common category is blue-collar jobs such as manufacturing, construction, maintenance, and farming. Blue-collar jobs are most common in non-MPO counties (tier 4), followed by small MPOs (tier 3). They are the least common in Atlanta MPO counties (tier 1). 28 For details on sample sizes for this and subsequent tables, see the appendix. As a reminder, unless otherwise stated, all statistics presented are weighted. 29 NHTS defines a worker as someone who worked for pay or profit, or was temporarily absent from paid employment, in the week before completing the travel survey ("last week"). All references to "workers" in this section refer to NHTS-defined workers. 87 Table 37. Workforce participation. Adults Workers Only All adults Gender Male Female MPO Tier 1. Atlanta MPO 2. Medium MPOs 3. Small MPOs 4. Non-MPO Age Cohort Millennial and Gen Z (1836) Generation X (3752) Pre-retirement age Baby Boomer (5364) Retirement age (65+) Annual Household Income <$35,000 $35,000 to $49,999 $50,000 to $74,999 $75,000 to $99,999 Percent Workers* 61.5% Full-time 80.4% More Than Part-time One Job 19.6% 9.1% 69.0% 85.4% 14.6% 8.4% 54.6% 74.4% 25.6% 9.8% 65.6% 81.5% 18.5% 8.8% 60.8% 76.4% 23.6% 8.9% 61.0% 79.3% 20.7% 9.4% 51.0% 80.7% 19.3% 9.9% 68.6% 74.5% 25.5% 8.9% 79.0% 88.8% 11.2% 9.7% 60.8% 83.4% 16.6% 9.1% 18.1% 50.2% 49.8% 5.3% 46.9% 63.2% 66.3% 69.1% 68.7% 77.6% 84.9% 87.3% 31.3% 22.4% 15.1% 12.7% 11.3% 12.8% 9.8% 7.1% $100,000+ 75.0% 85.7% 14.3% Race White non-Hispanic only 62.5% 80.7% 19.3% Black and Black multiracial 59.5% 79.7% 20.3% Other 61.9% 80.3% 19.7% * Percentage of all cases in row that work. See the appendix for unweighted sample sizes. As defined by NHTS, a worker is someone who worked for pay or profit or was temporarily absent from paid employment in the week before completing the travel survey. Both full-time and part-time workers may have more than one job. 6.0% 8.2% 11.7% 6.6% 88 Table 38. Job categories of Georgia workers. Occupation* Manufacturing, Clerical or Construction, Professional, Sales or Administrative Maintenance, or Managerial, or Service Support Farming Technical All workers ages 18+ 26.9% 9.3% 17.7% 46.0% Gender Male 24.9% 2.8% 27.7% 44.6% Female 29.2% 16.9% 6.1% 47.6% MPO Tier 1. Atlanta MPO 26.8% 9.6% 14.1% 49.4% 2. Medium-size MPOs 27.9% 9.7% 17.1% 45.0% 3. Small MPOs 27.5% 9.3% 21.6% 41.5% 4. Non-MPO 25.8% 7.9% 28.7% 37.4% Age Cohort Millennial and Gen Z (1836) 34.8% 8.0% 19.1% 38.1% Generation X (3752) 20.5% 9.5% 16.8% 53.1% Pre-retirement age Baby Boomer (5364) 22.7% 9.7% 18.0% 49.3% Retirement age (65+) 30.3% 16.0% 12.9% 40.7% Annual Household Income <$35,000 38.8% 10.6% 27.3% 23.1% $35,000 to $49,999 30.2% 10.8% 23.8% 34.7% $50,000 to $74,999 24.5% 12.3% 19.2% 43.9% $75,000 to $99,999 24.8% 9.1% 12.7% 53.4% $100,000+ 17.9% 5.9% 8.0% 68.2% Race White non-Hispanic only 24.9% 8.5% 15.3% 51.3% Black and Black multiracial 30.6% 11.5% 20.5% 37.2% Other 26.7% 7.9% 21.7% 43.7% Education Level High school or less 38.3% 6.7% 38.4% 16.4% Associates degree or some college 32.1% 14.2% 19.8% 33.7% Bachelor's degree or higher 16.9% 7.4% 4.9% 70.8% * All workers were limited to a single self-reported occupation. Of workers, 0.1 percent reported their occupation as "other." Note: NHTS defines a worker as someone who worked for pay or profit or was temporarily absent from paid employment in the week before completing the travel survey ("last week"). 89 Workers in professional, technical, and managerial professions tend to be the wealthiest, with a median household income of $75,000$99,999.30 Clerical and administrative workers have a median household income of $50,000$74,999. Workers in the remaining two categories (sales/service and blue-collar) are the least well off, with a median household income of $35,000$49,999. Much of the variation in industry across gender, region, and race is related to differences in educational attainment. As shown in figure 8 and table 38, 71 percent of workers in the highest income households hold at least a bachelor's degree, and 68 percent of workers in these households work in the professional sector. In contrast, only 20 percent of workers in the lowestincome households hold a bachelor's degree, and only 23 percent work in the professional sector. As shown in table 39, differences in occupation category by gender, region, and race are much smaller among college-educated workers. For example, the proportion of college-educated men who work in blue-collar positions is 5 percentage points higher than the proportion of collegeeducated women who do so. For noncollege graduates, this difference is 32 percentage points. The occupational categories of college graduates are broadly similar across all MPO tiers, though blue-collar professions are slightly less common in Atlanta, and sales and service jobs are more common. 30 No individual-level salary data are available, so these figures represent the combined income of all household members. Additionally, income is provided as a categorical variable, so the median is also presented as a category. 90 Percent of Workers With a Bachelor's Degree or Higher 80% 70% 60% 50% 40% 48% 42% 30% 20% 10% 0% 53% 42% 31% 26% 71% 52% 39% 32% 20% 50% 43% 36% Men Women MPO Tier 1 MPO Tier 2 MPO Tier 3 MPO Tier 4 <$35,000 $35,000 to $49,999 $50,000 to $74,999 $75,000 to $99,999 $100,000+ White non-Hispanic Black and black multiracial Other Figure 8. Bar graph. Educational attainment of Georgia workers by gender, MPO tier, household income, and race. However, a college degree notably fails to erase a racial difference in employment category: the proportion of Black31 college-educated workers in professional, managerial, and technical positions is 10 percentage points lower than the proportion of white non-Hispanic college educated workers, and 11 percentage points lower than college-educated workers of another race. This difference is larger than the difference between white and Black workers without a college degree. 31 This report uses "Black" as an umbrella term to include African Americans and Black immigrants. 91 Table 39. Job categories of Georgia workers by educational attainment. College Graduate Workers (Bachelor's Degree or Higher) Manufacturing, Clerical or Construction, Professional, Sales or Administrative Maintenance, Managerial, or Service Support or Farming Technical All college-educated workers 16.9% 7.4% 4.9% 70.8% Gender Male 17.4% 3.6% 7.4% 71.6% Female 16.4% 11.3% 2.3% 70.0% MPO Tier 1. Atlanta MPO 18.2% 7.3% 3.8% 70.7% 2. Medium MPOs 14.4% 7.6% 7.6% 70.3% 3. Small MPOs 11.7% 8.0% 6.0% 74.3% 4. Non-MPO 15.1% 7.4% 7.7% 69.8% Race White non-Hispanic only 16.9% 5.8% 4.0% 73.3% Black and Black multiracial 18.4% 11.7% 7.1% 62.7% Other 13.8% 7.1% 5.0% 74.1% Workers Without 4-Year College Degree Manufacturing, Clerical or Construction, Professional, Sales or Administrative Maintenance, Managerial, or Service Support or Farming Technical All workers without 4-year college degree 34.8% 10.9% 28.0% 26.1% Gender Male 30.2% 2.2% 42.3% 25.0% Female 40.7% 22.0% 9.6% 27.5% MPO Tier 1. Atlanta MPO 36.3% 12.3% 25.5% 25.7% 2. Medium MPOs 37.9% 11.2% 24.1% 26.3% 3. Small MPOs 34.6% 9.9% 28.7% 26.8% 4. Non-MPO 29.5% 8.0% 35.8% 26.4% Race White non-Hispanic only 32.8% 11.2% 26.5% 29.5% Black and Black multiracial 37.3% 11.3% 28.0% 22.9% Other 35.9% 8.6% 33.9% 21.5% Note: NHTS defines a worker as someone who worked for pay or profit or was temporarily absent from paid employment in the week before completing the travel survey ("last week"). 92 Overview of Commuters Table 40 shows the percentage of workers who reported at least one work journey on their travel day.32 This report defines these workers as active commuters. On an average weekday, 70 percent of NHTS-defined workers and 44 percent of the total adult population reported making one or more work journeys. The weekday proportion of active commuters for full-time workers was 77 percent, compared to 54 percent of part-time workers. Approximately one in five workers surveyed on a weekend or holiday commuted on that day. Weekend commuting was more common among service and blue-collar workers than clerical and professional workers, and among groups that are disproportionately employed in the service and blue-collar sectors (e.g., Black and low-income workers). It was also more common among part-time workers. Relatedly, weekend commutes are more common in non-MPO counties, which have a relatively high concentration of blue-collar jobs, than in other parts of the state. 32 The travel day ran from 4:00 a.m. through 3:59 a.m. the next day. A small number of people who were not classified as workers also reported work journeys; this group made up 1.3 percent of active commuters. Some of these are likely due to temporary employment, and others may represent a change in worker status for the participant between completing the main questionnaire and the travel diary. These journeys are not included in the percentage of workers reporting work journeys, but they will be included in per-capita and per-worker trip generation rates (for example, table 41 and table 42). 93 Table 40. Active commuter rates. All Adults Workers Only* Weekends Weekends Weekdays & Holidays Weekdays & Holidays All Days Only Only All Days Only Only All adults 34.4% 43.8% 13.0% 55.2% 70.0% 21.0% Gender Male 40.3% 50.4% 16.2% 57.8% 72.1% 23.1% Female 28.9% 37.5% 10.3% 52.2% 67.6% 18.8% MPO Tier 1. Atlanta MPO 36.3% 46.3% 11.6% 54.8% 69.0% 18.3% 2. Medium-size MPOs 32.3% 43.3% 13.7% 52.3% 70.7% 21.6% 3. Small MPOs 33.4% 42.8% 16.7% 53.9% 71.8% 24.7% 4. Non-MPO 31.3% 37.8% 13.7% 60.1% 72.2% 27.0% Age Cohort Millennial and Gen Z (1836) 40.4% 48.3% 20.4% 58.4% 69.5% 29.7% Generation X (3752) 43.0% 58.5% 11.4% 53.9% 72.0% 14.7% Pre-retirement age Baby Boomer (5364) 33.4% 43.4% 11.5% 54.1% 70.2% 18.8% Retirement age (65+) 8.6% 10.6% 3.8% 44.7% 59.3% 17.7% Annual Household Income <$35,000 28.5% 33.8% 15.8% 59.3% 69.5% 33.6% $35,000 to $49,999 36.9% 46.7% 15.1% 58.1% 75.4% 22.6% $50,000 to $74,999 38.4% 46.9% 20.1% 57.8% 71.0% 29.9% $75,000 to $99,999 41.9% 53.9% 8.3% 59.3% 74.8% 12.4% $100,000+ 36.5% 50.8% 6.6% 48.3% 66.8% 8.9% Race White non-Hispanic only 33.3% 44.1% 10.4% 52.7% 69.8% 16.6% Black and Black multiracial 35.5% 42.1% 18.3% 58.6% 69.3% 30.1% Other 36.2% 46.3% 13.1% 58.1% 73.0% 21.9% Occupational Category (Workers Only)* Sales or service 57.7% 66.7% 37.0% Clerical or administrative support 55.6% 74.7% 11.4% Manufacturing, construction, maintenance, or farming 64.0% 80.5% 23.6% Professional, managerial, or technical 54.1% 71.9% 14.3% Worker Type (Workers Only)* Full time 60.0% 76.8% 21.0% Part time 44.6% 53.9% 24.2% * Excludes the 1.3% of active commuters who were not defined as workers by NHTS. NHTS defines a worker as someone who worked for pay or profit or was temporarily absent from paid employment in the week before completing the travel survey ("last week"). 94 Table 41 shows the complexity of commutes made by various groups of workers. Overall, 39.6 percent of active commuters made at least one complex commute (with an interim stop). Three quarters of those who made a complex commute also made at least one simple commute, e.g., a simple commute in one direction and a complex commute in the other. Trip chaining is less common on weekends: 68.4 percent of weekend commuters made only simple work journeys, as opposed to 59.4 percent of weekday commuters.33 The prevalence of complex commutes is comparable for all MPO tiers, and across income levels. There are some differences across age groups, with younger workers being more likely to make only simple work journeys. Consistent with previous findings about gendered travel patterns (McQuaid and Chen 2012, Loukaitou-Sideris 2016), female commuters are more likely to have trip chained. Black commuters are also more likely to make complex work journeys than commuters of other races. Of active commuters, 88 percent reported exactly two work journeys (table 42). Close to 7 percent reported one work journey. Participants who reported an odd number of work journeys typically started or ended their travel day away from home; 74 percent of these started or ended at work. The next most common nonhome location to start or end the day was visiting friends/relatives. These "odd" commuters, then, are mainly people who work night shifts or extended shifts, with a minority of people who stay out late after work and the occasional air traveler who is midtrip when the travel day ends. 33 The sample size of the weekend commuters category was too small to disaggregate further, so that group is not separately cross-tabulated in the tables presented here. For unweighted sample sizes, see the appendix. 95 Table 41. Work journey complexity among active commuters. All Days Weekdays All active commuters Gender Male Female MPO Tier 1. Atlanta MPO 2. Medium-size MPOs 3. Small MPOs 4. Non-MPO Age Cohort Millennial and Gen Z (1836) Generation X (3752) Pre-retirement age Baby Boomer (5364) Retirement age (65+) Annual Household Income <$35,000 $35,000 to $49,999 $50,000 to $74,999 $75,000 to $99,999 $100,000+ Race White non-Hispanic only Black and Black multiracial Other Simple Only Both Complex Simple & Only Complex Simple Only Both Complex Simple & Only Complex 60.4% 10.1% 29.5% 59.4% 10.5% 30.1% 65.1% 54.4% 8.5% 12.1% 26.4% 33.5% 64.6% 52.8% 8.7% 12.7% 26.7% 34.5% 60.5% 61.1% 60.7% 59.5% 10.3% 9.5% 8.8% 10.6% 29.2% 29.4% 30.4% 29.9% 60.0% 58.5% 59.5% 58.1% 10.6% 10.1% 10.5% 10.3% 29.4% 31.4% 30.0% 31.6% 63.7% 59.1% 56.3% 59.7% 9.1% 10.6% 11.1% 10.5% 27.1% 30.3% 32.6% 29.8% 62.5% 58.4% 55.7% 57.0% 10.5% 10.2% 10.6% 11.8% 27.0% 31.3% 33.7% 31.2% 60.9% 58.0% 60.9% 60.9% 60.3% 10.6% 9.0% 10.8% 8.5% 10.0% 28.6% 33.0% 28.3% 30.6% 29.7% 61.0% 57.0% 57.4% 59.6% 59.9% 11.2% 8.8% 11.6% 8.7% 10.4% 27.9% 34.2% 30.9% 31.7% 29.7% 62.2% 55.5% 65.3% 9.0% 13.9% 5.4% 28.9% 30.5% 29.3% 61.3% 54.2% 63.7% 9.5% 14.5% 5.3% 29.2% 31.3% 31.1% 96 Table 42. Number of work journeys per person, active commuters. All Days Weekdays More Than More Than One WJ Two WJs Two WJs One WJ Two WJs Two WJs All adults 6.9% 88.1% 5.0% 6.6% 88.2% 5.3% Gender Male 8.1% 86.3% 5.6% 7.5% 86.6% 5.9% Female 5.4% 90.4% 4.3% 5.4% 90.1% 4.4% MPO Tier 1. Atlanta MPO 5.8% 89.8% 4.3% 5.5% 90.1% 4.4% 2. Medium MPOs 7.3% 85.8% 6.9% 7.1% 85.2% 7.7% 3. Small MPOs 8.3% 85.4% 6.4% 8.3% 84.9% 6.8% 4. Non-MPO 9.0% 85.9% 5.1% 8.8% 86.0% 5.3% Age Cohort Millennial and Gen Z (1836) 5.9% 88.9% 5.3% 5.7% 88.6% 5.7% Generation X (3752) 7.6% 88.5% 3.9% 7.4% 88.6% 4.1% Pre-retirement age Baby Boomer (5364) 7.7% 86.2% 6.1% 6.8% 87.1% 6.0% Retirement age (65+) 6.5% 86.0% 7.4% 6.4% 85.0% 8.5% Annual Household Income <$35,000 8.0% 86.2% 5.8% 8.5% 85.2% 6.2% $35,000 to $49,999 4.6% 87.7% 7.7% 4.3% 86.9% 8.8% $50,000 to $74,999 8.8% 86.7% 4.4% 7.6% 88.0% 4.5% $75,000 to $99,999 5.4% 89.0% 5.6% 4.8% 89.5% 5.7% $100,000+ 6.1% 90.6% 3.3% 5.8% 91.0% 3.2% Race White non-Hispanic only 6.5% 88.7% 4.8% 6.0% 88.9% 5.1% Black and Black multiracial 8.7% 86.9% 4.4% 8.8% 87.1% 4.1% Other 4.2% 88.4% 7.4% 3.8% 88.0% 8.3% 97 Commute Mode by Person As shown in table 43, personal occupancy vehicles (whether singly-occupied or carrying passengers) are the self-reported usual commute mode for 93.2 percent of workers statewide.34 There is considerable regional variation; "only" 91.1 percent of Atlanta MPO residents commute by car, while in non-MPO counties, the use of private autos is almost universal. Unsurprisingly, transit usage is highest in Atlanta, where there are more robust transit offerings. Transit usage is also higher among women than men, and among low-income and Black workers. A comparison of these figures with the modes actually used for work journeys on the travel day (table 44) suggests that Georgia residents' commute patterns are, unsurprisingly, slightly more complicated than reporting a single "usual" commute mode suggests. For instance, the percent of commuters reporting making a nonmotorized trip as part of a work journey is 1.6 times higher than the percent reporting nonmotorized travel as their "usual" work mode.35 Interestingly, more men than women reported nonmotorized means as their usual mode, but when trips actually made on the travel day are examined, this apparent gender gap disappears. 34 Specifically, participants were asked to report how they "usually" got to their main job last week. If multiple modes were used, they were instructed to select the one that they used for a longer distance. This question was not asked of people who reported that they usually telecommute, though it was asked of workers who telecommute some days. Most periodic telecommuters, like most workers in general, selected private auto as their usual commute mode. This question will be revisited in chapter 3. 35 Although the "usual" modes were provided by all commuters while the travel day modes were reported only by active commuters, there was no appreciable difference between the "usual" commute mode of active commuters and inactive commuters. 98 Table 43. Usual commute mode as reported by NHTS-defined workers. All workers POV* Nonmotorized (Walking, Biking) Public Transit or Other Bus or Train Other Ground or Water 93.2% 2.3% 3.8% 0.4% Gender Male 93.8% 2.7% 2.6% 0.5% Female 92.5% 1.8% 5.3% 0.3% MPO Tier 1. Atlanta MPO 91.1% 2.2% 5.9% 0.4% 2. Medium MPOs 94.2% 3.4% 1.8% 0.5% 3. Small MPOs 96.8% 1.5% 1.4% 0.3% 4. Non-MPO 97.0% 2.0% 0.4% 0.3% Age Cohort Millennial and Gen Z (1836) 91.9% 2.8% 4.3% 0.5% Generation X (3752) 94.2% 2.1% 3.2% 0.3% Pre-retirement age Baby Boomer (5364) 93.6% 1.5% 4.1% 0.3% Retirement age (65+) 95.3% 2.2% 2.5% Annual Household Income <$35,000 88.6% 4.1% 7.1% 0.3% $35,000 to $49,999 94.2% 2.2% 3.1% 0.5% $50,000 to $74,999 96.4% 1.4% 1.5% 0.5% $75,000 to $99,999 95.3% 1.4% 2.9% 0.0% $100,000+ 94.3% 1.8% 2.9% 0.3% Race White non-Hispanic only 95.3% 2.4% 1.7% 0.2% Black and Black multiracial 90.1% 1.5% 7.8% 0.4% Other 91.8% 3.4% 3.4% 1.0% * Including privately owned, rental, and company vehicles. Includes all workers, whether or not they reported a work journey on their travel day. Air 0.3% 0.4% 0.2% 0.4% 0.2% 0.1% 0.3% 0.3% 0.2% 0.5% 0.3% 0.5% 0.7% 0.4% 0.1% 0.5% 99 Table 44. Travel day commute mode(s) for active commuters. Non- Public motorized Transit or Other (Walking, Other Bus Ground or POV* Biking) or Train Water Air All active commuters 95.0% 3.7% 3.8% 1.4% 0.2% Gender Male 95.1% 3.6% 2.9% 1.5% 0.2% Female 94.9% 3.8% 5.0% 1.2% 0.4% MPO Tier 1. Atlanta MPO 93.6% 4.4% 6.1% 1.5% 0.4% 2. Medium MPOs 96.4% 4.2% 1.5% 0.9% 3. Small MPOs 98.3% 2.3% 0.6% 0.6% 0.0% 4. Non-MPO 96.5% 1.8% 0.4% 1.8% 0.2% Age Cohort Millennial and Gen Z (1836) 94.0% 4.5% 4.8% 1.9% 0.1% Generation X (3752) 96.1% 3.1% 2.8% 1.1% 0.5% Pre-retirement age Baby Boomer (5364) 95.0% 3.1% 3.9% 1.3% 0.1% Retirement age (65+) 95.6% 3.6% 3.1% Annual Household Income <$35,000 90.8% 6.7% 6.2% 2.1% 0.7% $35,000 to $49,999 96.6% 2.6% 4.0% 1.6% 0.9% $50,000 to $74,999 97.6% 1.5% 1.5% 1.1% 0.6% $75,000 to $99,999 96.4% 2.4% 3.2% 0.6% 0.2% $100,000+ 96.4% 3.2% 3.5% 1.0% 0.4% Race White non-Hispanic only 96.6% 3.0% 1.8% 0.8% 0.1% Black and Black multiracial 92.6% 3.7% 7.4% 2.5% 0.3% Other 94.8% 6.1% 3.4% 1.2% 0.6% * Including privately owned, rental, and company vehicles. Excludes taxis and ridehailing, which are included under Other Ground or Water. Note: Of participants, 3.6 percent reported using multiple modes, and are included in totals for all relevant modes. 100 Table 45 shows rates of driving alone (single-occupancy vehicle [SOV] trips) among active commuters. Of active commuters, 64 percent drove alone for all the trips that made up the work journeys on their travel day. Only 15 percent drove alone for some trips, and the remainder either were in a car with multiple people or used a different mode. Twenty-two percent of active commuters did not spend any portion of their work journeys driving alone. Driving alone is more common among men than women and is most common in small MPOs. Younger workers drive alone less than older ones, as do lower-income and nonwhite workers. Overview of Work Journeys In chapter 1, the sections on Trip Patterns by Location of Travel and Household and Personal Mobility provided statistics summarized by commuter. This section provides summary statistics about the individual work journeys. Georgia residents make nearly 2 billion work journeys each year, or an average of 415 per adult worker.36 Figure 9 shows the temporal distribution of these work journeys based on the time of arrival at or departure from the work anchor. Seventy percent of all work journeys take place between 6:009:59 a.m. or 3:006:59 p.m.; this report defines these two time windows as peak periods. 36 Per-worker statistics are calculated based on total work journeys from all adults divided by total NHTSdefined workers. The numerator includes the 1.2 percent of work journeys that are made by "nonworkers," likely people who are working short-term or sporadic jobs (for example, some farm work). 101 Table 45. Use of single-occupancy vehicles, high-occupancy vehicles (HOVs), and alternate modes among active commuters. All SOV Some SOV Trips No SOV Trips All active commuters Gender Male Female MPO Tier 1. Atlanta MPO 2. Medium MPOs 3. Small MPOs 4. Non-MPO Age Cohort Millennial and Gen Z (1836) Generation X (3752) Pre-retirement age Baby Boomer (5364) Retirement age (65+) Annual Household Income <$35,000 $35,000 to $49,999 $50,000 to $74,999 $75,000 to $99,999 $100,000+ Race White non-Hispanic only Black and Black multiracial Other * Transit, walking, etc. All WJ Segments by SOV SOV + HOV or HOV or HOV + SOV + Alt. Mode* Alt. Mode All Trips by Alt. Mode 63.6% 14.6% 16.6% 5.2% 65.7% 60.9% 12.8% 16.9% 16.4% 5.0% 16.7% 5.5% 61.8% 63.8% 69.4% 65.9% 15.5% 14.9% 14.2% 11.7% 15.9% 6.7% 17.7% 3.6% 14.7% 1.7% 18.8% 3.7% 61.9% 62.6% 68.2% 66.9% 13.0% 18.5% 11.0% 14.1% 19.1% 6.0% 14.5% 4.4% 15.6% 5.1% 14.7% 4.4% 53.2% 63.5% 71.7% 67.4% 66.9% 14.0% 13.1% 11.8% 14.2% 18.8% 23.7% 9.2% 20.0% 3.4% 13.5% 2.9% 14.5% 3.9% 10.3% 4.0% 68.3% 58.0% 58.5% 14.0% 14.5% 17.3% 14.1% 3.6% 19.8% 7.7% 18.5% 5.7% 102 12.5% 10.1% 12.0% 10.1% 6.7% 2.7% 0.6%0.2%0.3%0.1%0.9% 6.1% 7.0% 4.1% 3.1%2.4%2.7%3.2% 6.1% 3.2% 1.6%1.4%1.7%1.4% Dark purple bars indicate peak periods. Figure 9. Histogram. Work journeys by time of arrival to or departure from work (weighted). As table 46 shows, 92.7 percent of work journeys are by POV, and in 68.5 percent of cases, the commuter drove alone for the entire work journey. Table 47, table 48, and table 49 show the same information broken down by MPO tier. Interestingly, the annual number of work journeys per worker is substantially higher in non-MPO counties than in any other region (i.e., 451 work journeys per worker versus 404410 in tiers 13). One likely contributing factor to this difference is that multiple-job holding is highest in non-MPO counties (table 37). Additionally, adults who are not workers produce an average of 9.9 work journeys per nonworker in tier 4 counties, versus 7.0 work journeys per nonworker in tiers 1 and 8.6 in tiers 2 and 3 (not tabulated). 103 Table 46. Total annual work journeys in Georgia. Total (Millions) All Days Percent of Total WJs Per Worker* Weekdays Only Total Percent of Per (Millions) Total WJs Worker* Total work journeys 1,967.3 415.4 1,746.4 368.7 Simple work journeys 1,490.3 75.8% 314.7 1,311.9 75.1% 277.0 Complex work journeys 477.0 24.2% 100.7 434.4 24.9% 91.7 Mode POV 1,824.5 92.7% 385.2 1,618.4 92.7% 341.7 Multimodal with POV 16.6 0.8% 3.5 16.5 0.9% 3.5 Other 125.9 6.4% 26.6 111.2 6.4% 23.5 Nonmotorized 45.3 2.3% 9.6 42.2 2.4% 8.9 Public transit or other bus/train 50.7 2.6% 10.7 43.4 2.5% 9.2 Other ground or water 18.4 0.9% 3.9 15.0 0.9% 3.2 Multimodal without POV or air 9.1 0.5% 1.9 8.1 0.5% 1.7 Air or air multimodal 2.4 0.1% 0.5 2.4 0.1% 0.5 Driver Status (POV and Multimodal POV WJs) Driver for entire WJ 1,687.8 85.8% 356.4 1,502.1 86.0% 317.2 Driver for part of WJ 22.0 1.1% 4.6 21.1 1.2% 4.5 Passenger 133.8 6.8% 28.3 114.1 6.5% 24.1 Vehicle Occupancy Status (POV and Multimodal POV WJs) Drive alone entire WJ 1,347.6 68.5% Family sharing: drive with household passenger for 1+ legs 165.7 8.4% Carpool driver: drive with nonhousehold passenger for 1+ legs Time of Day 212.8 10.8% 284.5 35.0 44.9 1,188.7 158.8 189.5 68.1% 251.0 9.1% 33.5 10.9% 40.0 AM peak (6 am9:59 am) 693.0 35.2% 146.3 632.1 36.2% 133.5 Midday (10 am2:59 pm) 310.3 15.8% 65.5 264.8 15.2% 55.9 PM peak (3 pm6:59 pm) 690.4 35.1% 145.8 623.3 35.7% 131.6 Overnight (7 pm6:59 am) 273.6 13.9% 57.8 226.2 * Based on total population of workers ages 18+. Excludes multimodal trips. Based on time of arrival at or departure from work anchor. 13.0% 47.8 104 Table 47. Work journeys by MPO tier (all days). Annual Total, Millions Per Worker* Tier 1 Tier 2 Tier 3 Tier 4 Tier 1 Tier 2 Tier 3 Tier 4 Total work journeys 1,119.8 299.3 194.5 353.7 409.4 403.5 410.0 450.7 Simple work journeys 845.1 229.8 148.7 266.6 309.0 309.7 313.6 339.8 Complex work journeys 274.6 69.6 45.7 87.1 100.4 93.8 96.4 111.0 Mode POV 1,010.9 284.9 189.7 339.1 369.6 384.1 399.9 432.1 Multimodal with POV 12.4 2.9 0.9 0.4 4.6 3.9 1.9 0.5 Other 96.5 11.3 3.9 14.3 35.3 15.2 8.2 18.2 Nonmotorized 28.9 7.2 2.1 7.1 10.6 9.7 4.4 9.1 Public transit or other bus/train 46.4 2.2 0.8 1.3 17.0 2.9 1.7 1.6 Other ground or water 10.5 1.8 0.5 5.6 3.8 2.4 1.1 7.1 Multimodal without POV or air 8.5 0.1 0.4 3.1 0.2 0.9 Air or air multimodal 2.0 0.3 0.7 0.4 Driver Status (POV and Multimodal POV WJs) Driver for entire WJ 945.5 256.3 174.9 311.1 345.7 345.5 368.7 396.4 Driver for part of WJ 12.8 5.8 1.7 1.6 4.7 7.9 3.5 2.0 Passenger 67.0 25.6 14.1 27.0 24.5 34.6 29.7 34.5 Vehicle Occupancy Status (POV and Multimodal POV WJs) Drive alone entire WJ 747.3 203.8 145.4 251.0 Family sharing: drive with household passenger for 1+ legs 89.8 24.8 16.2 34.8 Carpool driver: drive with non- household passenger for 1+ legs 127.7 36.6 16.6 32.0 Time of Day 273.2 32.8 46.7 274.7 33.5 49.3 306.6 34.2 34.9 319.8 44.4 40.7 AM peak (6 am9:59 am) 406.8 101.4 65.6 119.1 148.7 136.7 138.4 151.8 Midday (10 am2:59 pm) 170.7 52.6 31.0 56.0 62.4 70.9 65.5 71.4 PM peak (3 pm6:59 pm) 394.2 102.8 67.1 126.3 144.1 138.5 141.5 161.0 Overnight (7 pm6:59 am) 148.2 42.5 30.7 52.2 54.2 57.3 64.7 66.6 * Based on total population of workers ages 18+. Excludes multimodal trips. Based on time of arrival at or departure from work anchor. 105 Table 48. Work journeys by MPO tier (weekdays only). Annual Total, Millions Per Worker* Tier 1 Tier 2 Tier 3 Tier 4 Tier 1 Tier 2 Tier 3 Tier 4 Total work journeys 1,019.5 254.8 159.1 313.0 372.8 343.5 335.4 398.8 Simple work journeys 765.1 191.9 119.4 235.6 279.7 258.7 251.6 300.2 Complex work journeys 254.4 62.9 39.7 77.4 93.0 84.8 83.7 98.7 Mode POV 919.9 243.8 154.5 300.3 336.3 328.7 325.7 382.6 Multimodal with POV 12.4 2.8 0.9 0.4 4.6 3.8 1.9 0.5 Other 87.2 7.9 3.7 12.4 31.9 10.6 7.8 15.8 Nonmotorized 28.9 4.3 1.9 7.1 10.6 5.8 4.0 9.1 Public transit or other bus/train 39.1 2.2 0.8 1.3 14.3 2.9 1.7 1.6 Other ground or water 9.6 1.3 0.5 3.6 3.5 1.7 1.1 4.6 Multimodal without POV or air 7.6 0.1 0.4 2.8 0.2 0.9 Air or air multimodal 2.0 0.0 0.3 0.7 0.4 Driver Status (POV and Multimodal POV WJs) Driver for entire WJ 866.0 221.6 140.4 274.1 316.6 298.7 296.0 349.3 Driver for part of WJ 12.8 5.2 1.5 1.6 4.7 7.0 3.2 2.0 Passenger 55.5 19.9 13.5 25.2 20.3 26.9 28.4 32.1 Vehicle Occupancy Status (POV and Multimodal POV WJs) Drive alone entire WJ 679.7 175.3 116.2 217.6 Family sharing: drive with household passenger for 1+ legs 86.6 22.9 14.4 34.8 Carpool driver: drive with non- household passenger for 1+ legs 117.8 31.1 12.2 28.4 Time of Day 248.5 31.7 43.1 236.2 30.8 42.0 244.9 30.4 25.7 277.2 44.4 36.2 AM peak (6 am9:59 am) 377.4 88.9 56.5 109.3 138.0 119.8 119.2 139.3 Midday (10 am2:59 pm) 149.7 42.9 24.4 47.7 54.7 57.9 51.5 60.8 PM peak (3 pm6:59 pm) 364.9 88.5 56.7 113.2 133.4 119.3 119.5 144.2 Overnight (7 pm6:59 am) 127.5 34.5 21.4 42.8 46.6 46.5 45.2 54.5 * Based on total population of workers ages 18+. Excludes multimodal trips. Based on time of arrival at or departure from work anchor. 106 Table 49. Work journey mode, vehicle occupancy status, and time of day by MPO tier (percentages). Work travel shares the roads (and sidewalks and rails) with nonwork travel. Figure 10 and figure 11 show the total number of trips and vehicle trips in progress at each hour of the day. A 107 trip from 9:3010:30 a.m., for example, would be in progress during both the 9:009:59 a.m. and the 10:0010:59 a.m. periods. The figures show that commute trips dominate peak-hour travel while noncommute trips dominate mid-day trips. Figure 10. Stacked bar graph. Weekday trips in progress by time of day (all ages, weighted). 108 Figure 11. Stacked bar graph. Weekday vehicle trips in progress by time of day (weighted). DEFINING THE COMMUTE What is a commute? When the U.S. Department of Transportation (USDOT) began to conduct large national surveys with the 1969 Nationwide Personal Transportation Survey, the answer to this question may have seemed self-evident. More than 50 years later, American's journeys to and from work have become more complicated. An influx of women to the workforce and the growing dominance of the dual-worker family has increased the numbers of workers on the road; many working parents (especially mothers) incorporate dropping children off at school into their daily commute (Gimenez-Nadal and Molina 2016; McGuckin, Zmud, and Nakamoto 2005). 109 More people work multiple jobs, or balance employment and higher education (McFarland et al. 2019). Less-seismic shifts, such as the so-called "Starbucks effect," have added a morning coffee stop to many commutes (McGuckin, Zmud, and Nakamoto 2005). These complex journeys to work, sometimes referred to as chained or tour commutes, have become more common over the past few decades (ibid.). Commute complexity has been analyzed in terms of mode choice, congestion, sustainability, and demographic differences such as gender (Concas and Winters 2007; McGuckin, Zmud, and Nakamoto 2005; Paleti, Bhat, and Pendyala 2013; Zhu et al. 2018). As commuters' journeys to and from work become more complex, it is important to make sure that the tools used to measure work travel can provide accurate, meaningful data about work journeys that do not fit the traditional mold of a single, uninterrupted trip between the home and the workplace. However, there is a lack of consensus on how to measure complex commutes. Researchers focusing on complexity often consider the full distance traveled between home and work (Paleti, Bhat, and Pendyala 2013; Zhu et al. 2018). The default measurement of work travel provided by the NHTS, in contrast, defines work travel based on the last trip in the chain (McGuckin and Fucci 2018). Researchers would benefit from having a more systematic measure of what portion of the travel between home and work should be considered work-related. This section identifies some critical definitional and measurement issues associated with commute travel, especially for complex work journeys. It then compares different techniques for 110 resolving these issues, and identifies best practices that will be incorporated into the analysis for the upcoming sections on Work Travel Distance and Work Travel Distance by Mode. First, we consider the entire work journey, including all intermediate stops. We discuss the importance of defining commute anchors by both purpose and location, rather than just purpose (which is the default for NHTS). Having identified whole work journeys, we then compare two alternate measures for determining what portion of each journey should be counted as commute distance (measured in PMT): (1) using the last leg of the journey, the default method when analyzing the NHTS trip file; and (2) modeling a counterfactual simple commute to estimate the distance that would have been traveled had no stops been made. Georgians make 2 billion journeys to work each year. As shown in figure 12, two thirds of these follow the pattern of a traditional commute: straight from the respondent's home location to the respondent's work location (or vice versa) with no stops in between.37 The remaining work journeys include at least one intermediate stop, involve a work or home site that differs from the respondent's home or work address, or both. 37 This figure does not include trips with the stated purpose of working from home. 111 Figure 12. Pie chart. Proportions of traditional and nontraditional commutes in Georgia. Given the complexity of Georgians' actual work-related travel, the first step in measuring commuting habits is defining what a commute is. If an office worker stops to shop for groceries and pick up a child on the way home from work, what portion of those trips should be considered part of the work commute? If a manager at a construction company travels from her home to a construction site where her employees are working and then proceeds to her company's central office, where did her commute end? If a teleworker goes to a coffee shop or a co-working space, is his trip a commute at all? These philosophical issues are at the heart of two practical questions that must be answered in order to measure work travel. First, what destinations and purposes should count as "home" and "work" anchors for a commute? Second, when a work journey between a home and work anchor includes intermediate stops, what portion of this travel should be considered as part of the commute? We will argue that the NHTS's traditional answers to both these questions need to be adjusted to more accurately reflect the modern commute. We will first examine alternate methods of identifying the home and work anchors that begin or end a work journey. We will 112 then compare methods for carving out the work-related portion of complex work journeys in order to define a commute. Purpose-based and Place-based Anchors Work journey anchors can be identified by type of activity at the location (trip purpose), by the type of location itself, or a combination of the two. A purpose-based approach to anchors has several advantages, including that it: Identifies commuting trips to conduct paid employment at locations that differ from the commuter's official work address. Since the NHTS limits each respondent to a single "work" location, this is especially important for people with more than one job or job site. Differentiates between a trip to work and a weekend trip to use the gym at work. Correctly identifies commutes by people who spent the night somewhere other than their address of record (e.g., someone working on extended assignment in another city, or those who spent the night with a romantic partner with whom they do not cohabitate fulltime). In short, considering purpose is important because people do not always work or sleep at their official work or home addresses, and some trips to the work address may be for nonwork purposes. NHTS's method of identifying work trips is based entirely on purpose and does not consider location. The tours are also location-blind; a homework tour is defined by a purpose of "regular home activities" at one end and a purpose of "work," not including working from home, at the 113 other. This location-blind approach, which originated in an era of lower computer processing power, is an efficient way of capturing simple commutes. The approach will also capture most complex work journeys because they will all end, eventually, with a "work" or "home" purpose.38 However, by NHTS's definition, a commute tour is considered to end at home only when the purpose of the final trip is "regular home activities" or "work from home (paid)." When any of several other activities that might take place at the home location (e.g., dropping off or picking up a family member, changing type of transportation, exercise) are conducted there, the location is not recognized as a "home" anchor, and trips following a stop at home for such purposes will be included as part of the tour. Consider the following series of four trips: Home Work Transport Someone Transport Someone Home This sequence of trip purposes can describe multiple scenarios (figure 13): A father went to work and picked up his children from school and daycare on the way home. A father went to work, drove straight home to pick up his daughter and took her to karate class before returning home himself. In scenario I, the commuter makes one simple work journey and one complex work journey. In scenario II, he makes two simple work journeys, followed by additional trips that are unrelated to 38 NHTS tours also consider nonwork, nonhome stops to be anchors if they are longer than 30 minutes. However, these can simply be joined back together (i.e., a HomeOther + OtherWork tour constitutes a Home Work work journey). 114 work travel. However, a location-blind method cannot distinguish between scenarios I and II. The father's first stop at home in case II would not register as a "home" trip because the purpose was not "regular home activities"; it was to pick up his daughter and transport her somewhere else. Those nonwork trips would erroneously be considered part of his commute. Figure 13. Diagram. Trip sequence with multiple potential commute patterns. A location-blind approach also makes it difficult to distinguish a trip to the gym on the way to work (a complex work journey) from a walk around the block followed by a drive to work (a simple work journey), since both initial trips depart from home and have a purpose of "fitness." Similarly, a walk to happy hour with coworkers at the end of the work day, followed by a return to the office to change mode of transportation (i.e., pick up the car), would register the two walk trips as part of a complex commute, whether or not the person made further stops on the way home. Conversely, stopping on the way home from work to do housework for an elderly relative 115 would register as a "home" anchor because chores are a regular home activity. The internal stop at a relative's house would incorrectly be identified as the end of the work journey. To avoid these errors, the most effective definition of work journey anchors should consider both purpose and place. Unfortunately, the public use NHTS dataset does not contain any variables that would allow researchers to incorporate location into work journey identification. We are able to correct for this problem by using variables from the confidential version of the data made available to GDOT (and thence to us) by virtue of its commissioning the add-on sample. Analysts using the public-use version would not be able to incorporate location into their definition of anchors; however, the proposed method does not require identifiable data such as specific GIS coordinates. Thus, the anonymized location categories (i.e., "home," "work," and "other") and ID numbers used could arguably be released publicly without imperiling confidentiality; doing so would make the public-use dataset more powerful for researchers who are not able to access the confidential data. Measuring Complex Commutes: The Last-leg and Counterfactual Methods Once full work journeys are identified, it is necessary to identify which portions of those observed work journeys are considered commutes. If a commuter stops for gas at a filling station that does not require deviation from the shortest route between home and work, it may be best to consider the entire work journey to be the commute. If, in contrast, a commuter goes to night classes at a university in a nearby city after work, less of the distance traveled from workplace to home is actually related to the commute. Methodologically, a uniform rule is needed that can accommodate different scenarios such as these. 116 Figure 14 shows an example of an observed work journey and two methods of extracting the "real" commute from the larger work journey. In the NHTS data, the most obvious way to apportion work travel is to analyze commutes at the trip level. This report refers to this as the "last-leg" method, because work trips in the trip file are the last leg of the work journey.39 Figure 14. Diagram. Alternate methods for identifying commutes within complex work journeys. In this example, the commuter has stopped for coffee on the way to her office. The coffee shop is very close to her workplace. As a result, the last-leg method (dotted line) identifies a commute that is much shorter than the distance between home and work. This is a key shortcoming of the last-leg method. 39 NHTS classifies trips based on purpose at destination, so even when just using the "last-leg" method, commute trips in the "home" direction need to be identified based on the trips that preceded them. 117 An alternate method is to identify the commute based on a counterfactual work journey: how long would the trip have been if the commuter had proceeded directly from home to work (dashed line)? Using the counterfactual distance as the commute distance gives a more accurate picture of what portion of the work journey is work-related, and how much distance was added by the additional stops. The ideal method of computing the counterfactual distance would be to use the same process the NHTS uses to calculate the distance of reported trips: calculate the shortest-path distance using the Google application programming interface (API). However, this solution is impractical for several reasons. To begin with, it requires exact coordinates of all travel locations, which are unavailable to users of the public dataset. Second, the Google API only provides estimates for current or future conditions; no functionality exists to calculate what conditions were like when the data were collected in 2017. Many commuters make a simple work journey in one direction and a complex work journey in the other. In these cases, the observed simple commute can be used as a counterfactual for a complex commute between the same two anchors. However, for commuters who made no simple work journeys, no obvious counterfactual is available.40 As will be described, this study imputes counterfactuals for these commutes by first modeling commute distance for cases where an 40 NHTS contains a "distance to work" variable, but even when considering only simple commutes between the home location and the normal work location, the "distance to work" variable differed from the observed commute distance by at least 0.2 mile in 34.9 percent of cases (unweighted, N=6,881). In 1 percent of these cases, the difference was 4.3 miles or greater, and the largest observed discrepancy was 111.6 miles. It is likely that the issue is due in part to the fact that distance to work was calculated based on an address provided during an earlier stage of the survey than the travel diary, sometimes by a different person than provided the trip information. In addition, some respondents may have listed the "home office" or human resources (HR) address rather than their actual work site. For these reasons, the distance to work variable is of limited utility as a measure of counterfactual commutes. 118 observed counterfactual is available, and then using that model to predict the distance for cases where it is not available. Methods for Definition and Measurement of Commutes The identification of commutes, and measurement of their key characteristics, was completed through the following steps: 1. Identify valid anchors. 2. Use anchors to group travel day trips into journeys. 3. Validate work journeys by classifying all journeys as either "work" or "other". 4. Calculate distances and times at the work journey level, and identify modes and stop characteristics. 5. Identify potential counterfactual commutes for complex commutes. 6. For complex work journeys with an observed counterfactual, use weighted ordinary least squares (OLS) regression to model the "complexity increment," or the incremental contribution of internal stops to the total work journey PMT. 7. For complex work journeys with no observed counterfactual available, use the model of complexity increments to impute counterfactual distances for unmatched complex work journeys. 8. Use observed and predicted complexity increments to calculate a work journey's commute PMT. Identifying Anchors The travel day origin site and all destinations were evaluated for their potential to be anchors to a work journey based on their purpose and location. The purpose of a location was defined as the 119 primary activity conducted there, either as the origin of the first trip of the day, or as a destination otherwise.41 Location was analyzed using two variables from the confidential dataset. Location type classifies all places as home, work, school, or other (regardless of purposes at those locations). For this analysis, school and other were combined into a single category, "other." Location number is a unique identifier for each location visited by a household member on the travel day. Location number was used to resolve ambiguities and questions (such as home activity at a nonhome location), and to screen out journey-level loops. Neither of these variables contain any geographic identifiers that would pose a risk to participant confidentiality if released publicly. Table 50 shows the criteria used in this study to define anchors. In contrast to the current NHTS approach, the home location was always considered to be a home anchor, regardless of trip purpose. This includes a number of trips where respondents, perhaps mistakenly, recorded their purpose at the home location to be "work for pay" rather than "work from home (paid)."42 The NHTS considers the former purpose (but not the latter) to signify the occurrence of a commute tour, so if using only the purpose indicator, this would cause the home to be mislabeled as a "work" destination. This error is especially important to correct because it tends to create 41 From a data management standpoint, it is important to remember that NHTS's file is organized by trips rather than locations. Each trip involves two locations (an origin and a destination). The destination of one trip is the origin of the next. Because the last destination at the end of the day does not form the start of a new trip, the number of locations is one greater than the number of trips. In the data, each location (except the start and end points for the day) will be listed twice in the trip-file: first as the destination of one trip, and then as the origin of the subsequent trip. Using destination characteristics will yield the needed data about all locations except the location where the traveler started the day. 42 It is possible that some respondents were reporting working from a separate office at their home address, such as a therapist or accountant with an office having its own separate entrance. A smaller number of respondents reported "working from home" at their work location; these were recoded as simply "work." 120 imaginary commutes if a person leaves home with an origin purpose of "home activities," and then returns home to an erroneously tagged "work" destination. Except for home, this report considers any destination with a purpose of work to be a work anchor, regardless of location. We include "other" location types in this rule because each participant was limited to providing a single work address. This does not match the reality of multiple jobs or job sites; 16 percent of work activity occurred at locations categorized as "other." Table 50. Initial anchor classifications based on location and primary activity at location. PRIMARY ACTIVITY Regular Home Work for Pay Other Purpose* Activities Home Home Home Home LOCATION TYPE Work Provisional Work Work Provisional Work Other Provisional Home Work Not an Anchor * Including work from home (paid). In only 25 cases out of nearly 60,000, participants reported working from home at a location other than home or work. The research team identified two types of provisional anchors (i.e., destinations that needed further examination to determine whether or not they were anchors). Provisional home anchors consisted of home activity at a nonhome location; these locations could reflect either a situation such as spending the night at a friend's home or hotel (making the location a legitimate, albeit temporary, home anchor), or one such as helping with housework at another person's home (for 121 which the location would not be a home anchor). In a second pass through the data, provisional home anchors were considered anchors if either of the following two criteria were true: (1) the traveler began or ended the day at that location, or (2) the traveler never visited the "home" location on the travel day. Trips made to a validated provisional home destination were included whether or not that trip specifically was the first or last trip of the day. Provisional home anchors that did not meet either criterion were reclassified as non-anchors. Provisional work anchors were identified when nonwork purposes occurred at the work location, addressing situations as shown in figure 15. In the example in the figure, the traveler went out for dinner after work, then returned to the office to pick up her car. Trip 3 ends at a provisional work anchor. If the provisional anchor is not identified, the work journey home would appear to include trips 2, 3, and 4, whereas just considering trip 4 to be a simple work journey might be more appropriate. Since the validity of a provisional work anchor depends on the sequence of trips before and after it, these provisional anchors were evaluated during the validation of work journeys. Figure 15. Diagram. Sequence of trips including a provisional work anchor. 122 Defining Work Journeys The itineraries of respondents with at least one valid work anchor and one valid home anchor on the travel day were evaluated for the presence of work journeys. The process began by dividing all of the day's trips into journeys, with a new journey beginning each time an anchor was encountered. Figure 16 shows several hypothetical sequences of trips and how they would be parsed as journeys and work journeys. Figure 16. Diagram. Example work journey classifications of trip sequences. As scenarios I and II show, HW and WH journeys were considered work journeys regardless of how many non-anchor destinations were visited over the course of the journey or 123 the duration of those non-anchor stops. HH and WW journeys were discarded.43 Some respondents had more than two work journeys (scenario III). Scenario IV illustrates how provisional work anchors (involving nonwork activity at a work location) were evaluated. The first provisional anchor is not considered part of a work journey because the preceding and following anchors are both home; the traveler was at the work location, but only there for nonwork activities. The second provisional anchor is considered a work anchor because it follows a definitive work anchor (and thus is a return to the workplace after a "side trip") and precedes a home anchor. Many commuters made fitness walks at the beginning of the travel day like the one illustrated in scenario V. In the NHTS's location-blind tour file, these loop trips were inappropriately classified as part of the HW journey. As table 51 shows, 55 percent of complex work journeys in Georgia involved a single, short stop. The most common purposes were shopping/errands, transporting someone, and dining. Social, recreational, and work-related stops had much longer internal dwell times. As we will see, these longer dwell times were associated with larger complexity increments than stops with shorter dwell times, with the exception of stops to transport someone else. 43 WW journeys might have the same beginning and ending destination, such as if the participant went out for a quick afternoon coffee. A WW journey might also be between two different locations at which the participant worked for pay. Both types of WW journeys were treated the same when classifying them, and neither are considered to constitute part of a commute. This classification has one potential limitation with respect to multiple job-holders: if the two ends of a WW journey were jobs with two different employers, it might be more accurate to consider that journey part of the participant's commute. Unfortunately, the NHTS does not contain enough detail to differentiate a journey between two locations associated with a single job from a journey between a first and second job. 124 Table 51. Duration and purposes of stops in work journeys. Stop Categories Percent of all WJs Percent of Complex Mean Internal Dwell (N=10,490) WJs (N=2,618) Time (Minutes)* Itinerary (Number and Duration of Stops) No stops (simple) 75.8% - 0.0 Single short stop (<30 min) 13.3% 54.7% 9.3 Single long stop (30+ min) 5.0% 20.5% 97.7 Multiple short 2.5% 10.5% 18.9 Short + long or multiple long 3.5% 14.4% 132.5 Total Purpose of Stop(s) 100.0% 100.0% 46.1 Shopping/errands 10.7% 44.3% 22.0 Transport someone 6.9% 28.3% 9.2 Dining 5.0% 20.6% 26.4 Social/recreational/fitness 2.6% 10.8% 85.9 Work-related 0.7% 3.0% 81.4 Other 3.3% 13.7% 85.9 Multiple purposes 3.9% 16.3% * For itineraries, total dwell time of all internal stops is shown. For the purpose of this report, dwell time shown is the average for a single stop of that type. If a journey contained two 10-minute stops for shopping, the stops would be counted separately rather than as a single, 20-minute stop. Work journeys with stops for multiple purposes are listed under "Multiple purposes" and also under each relevant individual purpose. Identifying Counterfactuals and Complexity Increments Once work journeys were identified, journey-level statistics were calculated based on the trips included in each journey. These include work journeylevel mode(s); total PMT and VMT; the number, duration, and purpose of stops; and total travel time and internal dwell time (duration of all stops). For complex work journeys, the data were examined for the presence of a trip that could serve as an observed counterfactual, which is defined as a simple work journey made by the participant 125 between the same origindestination pair44 at some other point on the travel day. As shown in table 52, matches were identified for 60.5 percent of complex work journeys. Same-direction matches were used where available, but in most cases, the matched simple work journey was in the opposite direction as the complex work journey (e.g., a commuter's evening journey from work that includes a stop is matched with the simple morning journey to work). These matched complex work journeys were used to build a model to predict a counterfactual for the 37.5 percent of unmatched complex work journeys where no observed counterfactual was available. Table 52. Counterfactual match status of complex work journeys in sample. Category Number of Cases (Unweighted) Percent Total complex work journeys 2,618 Matched with opposite-direction simple commute 1,484 56.7% Matched with same-direction simple commute 101 3.9% No observed counterfactual 983 37.5% Excluded from analysis* 50 1.9% * Reasons for exclusion: supercommute >100 mi (N=35), invalid distance provided (N = 10), commute includes air travel (N = 5). The weighted median distance of the complex work journeys was 16.2 miles and the 95th percentile was 52.6 miles. The small number of work journeys longer than 100 miles was found to exert a disproportionate influence in the model; we therefore excluded them to provide more accurate estimates for shorter work journeys. 44 Location numbers were used in matching to avoid errors when matching trips of respondents with multiple home or work locations. 126 Conceptually, we define a complexity increment as the incremental distance contributed by the internal stops in a complex work journey (compared to the mileage of a simple commute with the same origin and destination). Mathematically, the complexity increment for a complex work journey that has been matched with an observed counterfactual is the observed total PMT for the complex work journey minus the observed PMT for its simple counterfactual. For the set of matched cases, the weighted median complexity increment was 1.9 miles. The mean was 4.8 miles, with a maximum of 79.7 miles. For some work journeys, the added stops did not add miles; in 3.8 percent of cases, the complexity increment was zero or near-zero.45 In 7.8 percent of cases, the complexity increment was below zero. This is likely due to the fact that the recorded distances calculated by the Google API optimize for shortest travel time rather than distance, and thus recorded distances (in particular, those for the simple commutes being used for the counterfactual match) tend to favor highways over surface streets. If an added stop causes the commuter to use local roads instead of a more circuitous highway, the stop may actually reduce the mileage of the work journey compared to its simple commute counterfactual, such as the example in figure 17.46 45 The distances for the two directions can vary due to one-way roads, traffic conditions at the time the Google API was queried, etc. To decide what constitutes "near-zero", the researchers compared opposite-direction matched simple commutes from participants with no complex work journeys. This report defines near-zero as a complexity increment with an absolute value of less than 0.2 mile. 46 The median negative complexity increment was -0.85 miles, the 95th percentile was -4.3 miles, and the longest was -8.9 miles. Negative complexity increments were most commonly found in Atlanta, which has a complex highway network, and in smaller cities with rivers running through the downtown (e.g., Macon, Columbus, and Albany). 127 Figure 17. Maps. Example journey with a negative complexity increment. Using weighted OLS regression on the complex work journeys for which an observed counterfactual was available, the research team modeled the complexity increment as a function of the distance of the observed complex work journey, stops made, and other sociodemographic variables. The fitted model was then used to predict complexity increments for the unmatched complex work journeys (i.e., those with no observed counterfactual). Complexity increments are a measure of change; the results of the model tell how many miles stops of various types and lengths would be expected to add to (or sometimes subtract from47) a 47 For unmatched complex work journeys, the predicted complexity increment was near-zero but negative in 1.4 percent of cases, and nonnegligibly negative in 9.0 percent of cases (unweighted). 128 commute. These data can be used to examine issues like how total commute PMT might change if complex commutes became more common (i.e., replacing some simple commutes with complex commutes). The model also allows for prediction of the counterfactual (simple) work journey distance for unmatched work journeys, defined as complex work journey PMT minus the complexity increment. This reveals how long these commutes would have been without internal stops. The next step is to calculate the commute portion of total observed PMT for complex work journeys. Initially we computed this as the observed (where available) or predicted counterfactual distance. However, the commute PMT represents the actual mileage traveled by the worker that should be attributed to his or her commute. Therefore, the commute PMT is capped at 100 percent of observed work journey PMT. This cap accounts for the minority of cases with a negative complexity increment, making sure that more miles are not allocated to a worker's commute than that worker actually traveled. The next section presents the results of these estimations and compares them with the results obtained from looking at the entire work journey and using the last-leg method. 129 Results Model of Complexity Increments Table 53 shows the complexity increment distance (in miles) of matched commutes as a function of trip distance, stops, and demographic and environmental information.48 The full model explains more than half the variation in complexity increments. Most of this explanatory power comes from characteristics of the work journey itself: a model based only on distance and stops has an R2 of 0.486; the addition of demographic variables results in a final R2 of 0.544. Longer observed work journeys are associated with larger complexity increments; these effects differ between rural and nonrural areas. A stop of any length adds to the complexity increment, but stops of a longer duration are associated with larger increases. Stops to transport someone, which tend to be relatively brief, exert an outsized effect on the complexity increment given their duration. Demographic variables such as gender, age, income, and vehicle ownership are also significant. Mode was insignificant, perhaps due to the high private auto mode share. 48 Mode is included in the model, but was not found to be significant either in the single category of "alternate mode" shown here or in more disaggregated forms. This may be because 97 percent of complex work journeys were entirely or partially by personal vehicle. 130 Table 53. Weighted OLS model of work journey complexity increment distance in miles. Full Model Trip Characteristics Only Model Covariate Coef P-Value Coef P-Value WJ PMT WJ PMT2 Interaction: rural x WJ PMT Interaction: rural x WJ PMT2 -0.0673 0.00674 0.175 -0.00615 0.372 <0.001 *** 0.016 ** 0.005 *** 0.0198 0.00432 0.855 0.060 * WJ internal dwelltime, minutes 0.0174 0.013 ** 0.0207 0.015 ** Number of stops by duration Short (<30 minutes) 1.349 <0.001 *** 1.238 <0.001 *** Medium (3059 minutes) 2.460 <0.001 *** 2.358 <0.001 *** Long (60+ minutes) 3.451 <0.001 *** 3.516 0.001 *** Purpose: transport someone (yes/no) Alternate mode used for part or all of trip 2.889 -0.422 <0.001 *** 0.795 2.923 <0.001 *** Household size & vehicle sufficiency (Reference: 1-driver, 1-vehicle households) Vehicle-sufficient with 2+ drivers -2.262 <0.001 *** Vehicle-deficit or zero-vehicle -1.557 0.046 ** Female -0.993 0.023 ** Age Age2 -0.271 0.00290 0.009 *** 0.010 ** Household income (reference: <$35,000) $35,000$49,999 -0.885 0.119 $50,000$74.999 0.938 0.101 $75,000$99,999 1.238 0.148 $100,000+ 2.190 0.004 *** Income missing (dummy variable) 1.176 0.200 Constant Model R2 5.582 0.544 0.015 ** -2.042 0.486 0.036 Rural is defined by urbanicity of participant's home address rather than census designation. Defined as any mode besides personal occupancy vehicle. Vehicle-sufficient is defined as having a number of vehicles equal to or greater than the number of household members ages 16+. Vehicle-deficit is defined as having at least one vehicle, but fewer vehicles than household members ages 16+. Table 54 shows predicted complexity increments based on the most common itineraries and average work journey distance. The average predicted complexity increment is 5 miles; the median is 3.5 miles. Trips with a single short stop, the most common type of trip, have a 131 predicted complexity increment of 3.2 miles. A single long stop, on average, is associated with a complexity increment of 7.4 miles. Table 54. Predicted complexity increments. Mean Predicted Complexity Increment (miles) Percent of Complex Work Journeys (N=2,568) Average total predicted complexity increment* 5.01 Complexity Increments for Example Itineraries Single short stop 3.22 54.7% Single medium stop 4.86 8.3% Single long stop 7.37 12.2% Two short stops 4.73 8.7% Three short stops 6.23 1.4% Short stop + medium stop 6.36 3.4% Short stop + long stop 8.87 3.2% Medium stop + long stop 10.50 1.0% * The figure here is the average for all complex WJs under 100 miles in the NHTS GA sample, including WJs with negative complexity increments. For consistency, in this table, predicted values are used for both matched and unmatched WJs. This is different from table 55, where predicted values are used for unmatched WJs and observed values are used for matched WJs. Complex Commute Distance Having discussed how internal stops change commute distance, we now estimate the share of total observed work journey PMT that pertains to the commute (i.e., commute PMT), and compare the estimates with those obtained by the last-leg method, and with total observed work journey PMT (table 55). The average complex work journey is 19.4 miles. Among complex commutes, the average distance assigned to the commute purpose, based on the counterfactual approach (i.e., subtracting the predicted complexity increment from the complex work journey length, where observed counterfactuals are not available), is 14.3 miles. When using the last-leg method, the average distance assigned to the commute purpose would be just 8.8 miles. Across all complex work journeys, about three quarters of the total journey PMT is allocated to the 132 commute under the counterfactual approach, compared to only about two fifths under the last-leg approach. Of course, since 76 percent of work journeys are simple, the differences between the two methods diminish when taking both simple and complex work journeys into consideration. In that case, the commute purpose accounts for more than 90 percent of the distance using the counterfactual method, or more than 80 percent using the last-leg method. Even so, the differences between the two methods remain nontrivial. Table 55. Comparison between full work journey, counterfactual, and last-leg commute measurements. WJ Length Category All complex WJs Percent of Complex WJs Mean WJ PMT: Full WJ 19.4 Mean Commute PMT (Counterfactual Method)* 14.3 (73.9%) Mean Commute PMT (Last-Leg Method) 8.8 (45.1%) <5 mi 10.2% 3.4 2.3 (69.4%) 1.7 (49.7%) 59.99 mi 18.9% 7.5 5.6 (75.1%) 3.4 (45.3%) 1024.99 mi 44.3% 16.7 12.8 (76.8%) 7.9 (47.1%) 2549.99 mi 22.6% 33.9 25.9 (76.2%) 14.8 (43.7%) 5074.99 mi 3.2% 58.0 35.8 (61.7%) 26.8 (46.2%) 7599.99 mi 0.9% 84.3 43.4 (51.5%) 29.0 (34.4%) All WJs (simple and complex) 14.8 13.6 (91.8%) 12.3 (82.7%) * Defined as the counterfactual (simple) WJ PMT, capped at 100% of observed (complex) WJ PMT. Observed counterfactuals are used where available. When there is no observed counterfactual, the counterfactual is calculated by subtracting the predicted complexity increment from the observed complex WJ PMT. Defined as the PMT of the final trip in a complex WJ. In a home-to-work WJ, the last leg is the trip that ends at a work anchor; in a work-to-home WJ, the final trip ends at a home anchor. Percent of average row-category WJ PMT in parentheses. At the state level, using the last-leg method would underestimate Georgia's commute PMT by 2.6 billion miles per year compared to using the counterfactual method, an amount that constitutes about 10 percent of the total commute PMT. As such, the counterfactual model is a clear improvement over the last-leg technique. 133 Summary of Methods Implications for Upcoming Report Sections Which method of measuring work travel is preferable? This depends on the question to be answered. For example, when considering commute burden on individual travelers, using the full work journey is likely to be a better proxy for the full amount of travel required of each worker, particularly if the stops involved are household-serving travel such as dropping off children at school. Because women are disproportionately responsible for such trips, looking only at the commute portion of a complex work journey may understate the true toll of commutes that might otherwise appear short (Gimenez-Nadal and Molina 2016). However, this must be balanced with the fact that male commuters are more likely to make nonessential stops such as going for coffee (McGuckin, Zmud, and Nakamoto 2005), in which cases using the full work journey would arguably offer an inflated view of the commute burden. For estimates of PMT and VMT generated by commuting, it might be preferable to isolate the portion of each work journey that should be considered work-related. For this purpose, the counterfactual method is a clear improvement over the last-leg method, which underestimates the true amount of work travel. To provide data to answer a broad range of questions, this report uses several different measurements of commute distance and duration in its upcoming chapters. Henceforward, the definitions of the measurements are as follows: Work journey PMT: Full distance traveled by the commuter, including all stops. Work journey VMT: Full distance driven by the commuter over the course of the work journey, excluding miles as a passenger in a POV as well as miles traveled by walking, biking, and public transportation. 134 Work journey duration: Travel time, in minutes, of the whole work journey (excluding dwell time for intermediate activities). Commute PMT: Work-related portion of the whole work journey, calculated using the counterfactual method as described in this section. Commute VMT: Work-related VMT of the whole work journey. Commute VMT will be scaled proportionally based on the ratio of commute PMT to whole work journey PMT. For example, if a commuter's commute PMT were 80 percent of the whole work journey PMT, the commute VMT would be considered to comprise 80 percent of the work journey VMT. Commute duration: Duration, in minutes, of the work-related portion of the whole work journey. Like commute VMT, commute duration will be scaled proportionally to work journey duration.49 Supercommute: A work journey in excess of 100 miles. The data included 85 such commutes, or 0.8 percent of all work journeys. Six of these were by air. The groundbased supercommutes ranged between 100 and 447 miles. The longest of these are likely not daily commutes, but rather capture the small percent of the population that will be making longer work-related trips on any given day. Unless otherwise stated, to provide more representative estimates of "typical" work travel, supercommutes are excluded from the analysis. This report provides alternate estimates of total PMT and VMT (with and without the supercommutes). Note: Since commute PMT was not modeled for work 49 Since traffic speeds may vary at different stages of the commute, this method for estimating commute travel time is likely somewhat noisier than the estimations of VMT. However, it will provide a better measure than other available alternatives, such as using the last-leg travel time. 135 journeys longer than 100 miles, for complex supercommutes, the commute PMT is defined as the full work journey PMT. WORK TRAVEL DISTANCE Table 56 compares workers' distance to work (i.e., the shortest-path distance between their home and work addresses) with the work journeys observed on the travel day. While "distance to work" has the advantage of having been asked of all workers, rather than just those who happened to travel to work on the day they filled out their diary, it is likely that some workers' reported work address does not match the location where they actually conduct their work (see Defining the Commute in this chapter for further discussion). While just 0.8 percent of observed work journeys exceeded 100 miles, 2.3 percent of workers listed a work address more than 100 miles from their home. The supercommutes implied by the "distance to work" variable are, on average, 142 miles longer than the observed travel day supercommutes. Even after excluding supercommutes, the mean distance to work still exceeds the mean commute PMT. For these reasons, the research team suggests caution when analyzing the reported distance to work, and focuses primarily on the observed travel day work journeys. A few workers reported air as their usual commute mode (0.3 percent), or as their travel day mode (0.1 percent). Travel diary air commutes were nearly twice as long as the average distance to work of commuters who described air as their usual mode of travel, suggesting that the air work journeys reported in the travel diary data largely reflect unusual work travel (i.e., business trips). However, the unweighted sample sizes for both kinds of air commute are very small, and extreme caution should be used when interpreting these findings. Because it is an extreme outlier 136 in terms of distance, air travel will be excluded from further analysis. Ground supercommutes will also be excluded unless otherwise stated. Table 56. Distance to work, commute PMT, and work journey PMT. Workers* Observed WJs Distance to Commute Work PMT Work Journey PMT Median 11.3 10.3 11.3 95th percentile 46.2 38.5 42.4 Proportion of supercommutes (>100 miles) 2.3% 0.8% 0.8% Mean Commute Distances All modes and distances 25.2 16.2 17.4 Regular commutes (<100 miles) 14.4 13.6 14.8 Supercommutes (>100 miles) 481.1 339.1 339.1 Terrestrial supercommutes Air supercommutes 599.8 660.0 191.1 1146.4 191.1 1146.4 Mean excluding air 23.5 14.8 16.0 * This question was asked of all workers; only 6,978 out of 8,363 (unweighted) responded. Weighted N=3.98 million workers (out of a total population of 4.78 million workers). The sample included 10,463 work journeys made by 5,101 unique commuters (unweighted). Weighted N=1.96 billion WJs by 2.65 million active commuters (annually). Note: These figures should not be used to calculate perworker rates. Rather, per-worker rates should be calculated using a denominator of all workers, including those who did not work on their travel day. As shown in table 46, Georgians make an average of 415 work journeys per worker annually. Since commute PMT was not calculated for complex work journeys greater than 100 miles, WJ PMT is used for 36 complex supercommutes. Personal occupancy vehicle or other form of ground transportation. No commutes by boat were reported. Mode for the "distance to work" column is determined by the usual commute mode. There are an estimated 2.4 million work journeys by air each year and 12,500 regular air commuters (weighted). However, due to extremely low unweighted sample sizes (N=11 for distance to work and N=6 for observed WJs), these estimates should be treated with caution. 137 Region Table 57 shows average work journey and commute distances by MPO tier. Average commute distances are longest in the most and least populous areas of the state, averaging 15.5 miles for residents of non-MPO counties and 14.1 miles for residents of the Atlanta MPO. Commutes are noticeably shorter in small and medium MPO counties (10.9 and 11.2 miles, respectively). In all parts of the state, the average full work journey is 1.01.3 miles longer than the average commute. Table 57. Average work journey and commute distance by MPO tier. Percent Commute Commute of WJs PMT VMT WJ PMT WJ VMT Supercommutes Excluded* (weighted N=1.94 billion WJs per year) Statewide 100.0% 13.6 12.1 14.8 13.2 MPO Tier (Worker Residence) 1. Atlanta MPO 57.1% 14.1 12.3 15.4 13.4 2. Medium MPOs 15.3% 11.2 10.0 12.4 11.1 3. Small MPOs 9.9% 10.9 10.1 11.9 11.1 4. Non-MPO 17.7% 15.5 14.7 16.7 15.8 Terrestrial Supercommutes Included (weighted N=1.96 billion WJs per year) Statewide 100.0% 14.8 13.1 16.0 14.2 MPO Tier (Worker Residence) 1. Atlanta MPO 57.0% 15.2 13.2 16.5 14.3 2. Medium MPOs 15.3% 12.1 10.9 13.3 12.0 3. Small MPOs 9.9% 12.2 11.3 13.2 12.3 4. Non-MPO 17.8% 17.1 15.9 18.2 17.0 * A supercommute is a one-way work journey longer than 100 miles. Commutes by air excluded. No supercommutes by water were reported. The lengthier commutes in non-MPO areas reflect the lower employment prospects in more rural parts of the state. Whether by necessity or lifestyle choice, 22.9 percent of commutes by residents of non-MPO counties were to a job site in an MPO, requiring the commuter to travel to a different community. In comparison, only 16.2 percent of commutes by residents of small and 138 medium MPOs involved travel to a different MPO or non-MPO county. While Atlanta MPO residents were the least likely to travel outside of their home MPO (2.3 percent of commutes), the PMT of commutes that stay entirely within the Atlanta MPO is higher than commute PMT for commutes within medium and small MPOs (13.4 miles versus 7.9 miles). As a result, average commute distances are longest in the Atlanta MPO. Table 58 shows the percent of statewide work travel originating from residents of each MPO tier; table 59 shows the total miles on which table 58 is based. In keeping with the longer commutes, residents of non-MPO counties account for 16.7 percent of workers but 20.3 percent of commute PMT. Atlanta also accounts for a slightly disproportionate share of commute mileage, while small- and medium-MPO counties account for a smaller portion of commute mileage than they do of workers. A comparison of PMT and VMT shows that Atlanta MPO residents, on average, spend a smaller portion of their commutes behind a steering wheel and a larger portion either as a passenger or using alternate modes of transportation. Residents of small-MPO and non-MPO counties spend a higher share of their commute travel driving. Residents of medium-MPO counties have identical shares of commute PMT and VMT, indicating that the average percent of commute mileage as a driver matches the state average. Table 60 shows annual commute miles per worker. As with the average distance for an individual commute, per-worker mileage is higher in the Atlanta MPO and non-MPO counties than in small and medium MPOs. In non-MPO counties, the higher per-worker total is also influenced by the increased number of work trips per worker, increased levels of multiple job- 139 holding, and a higher number of work trips by people who may be irregularly employed but are not classified as workers (see Overview of Work Journeys in this chapter). Table 58. Percent of statewide work-related miles traveled by MPO tier. Commute PMT Commute VMT WJ PMT Percent of WJ VMT Workers Supercommutes Excluded* (weighted N=1.94 billion WJs per year) Statewide miles (millions) 26,417 23,609 28,777 MPO Tier 1. Atlanta MPO 59.2% 57.7% 59.2% 2. Medium MPOs 12.6% 12.6% 12.8% 3. Small MPOs 7.9% 8.2% 8.0% 4. Non-MPO 20.3% 21.4% 20.0% Terrestrial Supercommutes Included (weighted N=1.96 billion WJs per year) Statewide 28,908 25,717 31,268 MPO Tier 1. Atlanta MPO 58.8% 57.3% 54.9% 2. Medium MPOs 12.5% 12.7% 11.8% 3. Small MPOs 8.2% 8.5% 7.4% 4. Non-MPO 20.5% 21.5% 17.6% * A supercommute is a one-way work journey longer than 100 miles. Commutes by air excluded. No supercommutes by water were reported. 25,684 57.8% 12.8% 8.3% 21.2% 27,792 57.4% 12.9% 8.6% 21.2% 100% 57.8% 15.5% 10.1% 16.7% 140 Table 59. Annual work-related PMT and VMT by MPO tier (millions of miles). Commute PMT Commute VMT WJ PMT Supercommutes Excluded* (weighted N=1.94 billion WJs per year) Statewide 26,417 23,609 MPO Tier (Worker Residence) 1. Atlanta MPO 15,636 13,633 2. Medium MPOs 3,323 2,979 3. Small MPOs 2,097 1,942 4. Non-MPO 5,361 5,055 Terrestrial Supercommutes Included (weighted N=1.96 billion WJs per year) Statewide 28,908 25,717 MPO Tier (Worker Residence) 1. Atlanta MPO 16,987 14,736 2. Medium MPOs 3,621 3,271 3. Small MPOs 2,370 2,193 4. Non-MPO 5,930 5,517 * A supercommute is a one-way work journey longer than 100 miles. Commutes by air excluded. No supercommutes by water were reported. 28,777 17,045 3,679 2,295 5,758 31,268 17,180 3,696 2,311 5,514 WJ VMT 25,684 14,839 3,285 2,127 5,433 27,792 15,942 3,576 2,378 5,896 Table 60. Annual work-related PMT and VMT per worker by MPO tier (miles). Commute PMT Commute VMT WJ PMT Supercommutes Excluded* (weighted N=1.94 billion WJs per year) Statewide 5,578 4,985 MPO Tier (Worker Residence) 1. Atlanta MPO 5,717 4,984 2. Medium MPOs 4,479 4,016 3. Small MPOs 4,421 4,094 4. Non-MPO 6,831 6,441 Terrestrial Supercommutes Included (weighted N=1.96 billion WJs per year) Statewide 6,104 5,430 MPO Tier (Worker Residence) 1. Atlanta MPO 6,211 5,388 2. Medium MPOs 4,882 4,409 3. Small MPOs 4,996 4,623 4. Non-MPO 7,556 7,031 * A supercommute is a one-way work journey longer than 100 miles. Commutes by air excluded. No supercommutes by water were reported. 6,076 6,232 4,959 4,840 7,337 6,602 6,281 4,983 4,872 7,027 WJ VMT 5,423 5,425 4,428 4,484 6,923 5,868 5,829 4,821 5,013 7,513 141 Peak, Off-peak, and Weekend Travel As shown in table 61, 88.7 percent of work journeys and 88.2 percent of commute PMT and VMT occur on weekdays, and 66.6 percent of PMT occurs specifically during the weekday peak hours. The average commute PMT of off-peak weekday commutes is lower than the average weekday peak commute or weekend commute. This is especially true of mid-day commutes (10.9 miles, as compared to 14.2 miles for average weekday peak and average weekend). NHTS does not factor in time of day when calculating trip distances, so the observed differences cannot be attributed to congestion.50 They more likely relate to characteristics of the work or the workers. White-collar workers (clerical/administrative and professional/managerial/technical sectors) make 72.1 percent of their commutes during weekday peak hours, versus 63.4 percent of commutes by blue-collar workers and just 47.4 percent of commutes by service-sector workers (not tabulated). In addition to requiring more off-peak commuting than white-collar jobs, service jobs tend to be more geographically dispersed. Further, service-sector salaries are often lower, and low-income job seekers sometimes have a smaller job search radius due to transportation challenges (Blumenberg and Pierce 2014, Shen and Sanchez 2005).51 Parenthood may be an additional contributor to shorter midday commutes. Heterosexual couples have been found to make career and housing decisions that allow one parent (usually the mother) to work close to home in order to be able to pick up children after school (Jun and Kwon 2015, McQuaid and Chen 2012, Craig and van Tienoven 2019). Relatedly, an off-peak commute is 50 Congestion may, in fact, create differences in the distance of peak versus off-peak commutes, but NHTS does not provide data appropriate for examining the issue. 51 See also table 63. 142 more likely to reflect part-time work, and part-time workers' commutes tend to be shorter than those of full-time workers. Table 62 shows peak and nonpeak work travel miles by MPO tier. The proportion of commute miles that occurs during weekday peak hours is highest in the Atlanta MPO, which also has the highest proportion of white-collar workers (see table 38). Small MPOs, with the lowest proportion of weekday peak miles, have the second-lowest proportion of white-collar workers.52 52 Non-MPO areas have lower white-collar employment. However, the "blue-collar" category, which is especially common in non-MPO counties, contains a wide range of job types. It may be that agricultural work is more likely to have a daytime schedule than other forms of blue-collar work (e.g., maintenance, manufacturing). NHTS only includes the job category rather than the more specific occupation. 143 Table 61. Work travel distance by time of day. Percent of Statewide Totals Percent Commute Commute of WJs PMT VMT All work journeys (millions of miles) 26,417 23,609 Weekday 88.7% 88.2% 88.2% AM peak (6 am9:59 am) 32.3% 34.2% 33.9% Midday (10 am2:59 pm) 13.2% 10.6% 10.8% PM peak (3 pm6:59 pm) 31.7% 32.4% 32.5% Overnight (7 pm6:59 am) 11.6% 11.0% 11.0% Weekend or Holiday 11.3% 11.8% 11.8% AM peak (6 am9:59 am) 3.1% 3.1% 3.0% Midday (10 am2:59 pm) 2.3% 2.4% 2.4% PM peak (3 pm6:59 pm) 3.4% 3.7% 3.9% Overnight (7 pm6:59 am) All work journeys 2.4% 2.6% 2.5% 13.6 12.1 Avg. Commute and Work Journey Distance Com. PMT Com. VMT Weekday 88.7% 13.5 12.1 AM peak (6 am9:59 am) 32.3% 14.4 12.8 Midday (10 am2:59 pm) 13.2% 10.9 9.9 PM peak (3 pm6:59 pm) 31.7% 13.9 12.5 Overnight (7 pm6:59 am) 11.6% 12.9 11.6 Weekend or Holiday 11.3% 14.2 12.7 AM peak (6 am9:59 am) 3.1% 13.3 11.7 Midday (10 am2:59 pm) 2.3% 14.0 12.8 PM peak (3 pm6:59 pm) 3.4% 14.9 13.8 Overnight (7 pm6:59 am) 2.4% 14.6 12.5 Annual Miles Per Worker Com. PMT Com. VMT All work journeys 5,528 4,941 Weekday 88.7% 4,922 4,397 AM peak (6 am9:59 am) 32.3% 1,909 1,692 Midday (10 am2:59 pm) 13.2% 593 536 PM peak (3 pm6:59 pm) 31.7% 1,807 1,619 Overnight (7 pm6:59 am) 11.6% 613 551 Weekend or holiday 11.3% 656 588 AM peak (6 am9:59 am) 3.1% 171 151 Midday (10 am2:59 pm) 2.3% 132 120 PM peak (3 pm6:59 pm) 3.4% 208 192 Overnight (7 pm6:59 am) 2.4% 146 125 Excludes commutes greater than 100 miles. Weighted N=1.96 billion per year. Based on time of arrival at or departure from work anchor. WJ PMT 28,777 88.5% 33.1% 11.6% 33.4% 10.5% 11.5% 2.9% 2.3% 3.9% 2.4% 14.8 WJ PMT 14.8 15.2 13.0 15.6 13.4 15.2 13.6 15.1 16.9 14.9 WJ PMT 6,022 5,375 2,008 702 2,027 637 701 175 142 235 149 WJ VMT 25,684 88.7% 32.9% 11.7% 33.6% 10.5% 11.3% 2.9% 2.4% 3.7% 2.4% 13.2 WJ VMT 13.2 13.5 11.7 14.0 12.0 13.3 12.0 13.7 14.5 12.7 WJ VMT 5,375 4,810 1,782 634 1,821 572 614 155 129 202 128 144 Table 62. Work travel distance by time of day and MPO tier. Commute VMT (millions)* Work Journey VMT (millions)* Statewide 23,609 25,684 Weekdays AM peak (6 am9:59 am) 20,824 (88.2%) 8,013 (33.9%) 22,778 (88.7%) 8,439 (32.9%) Midday (10 am2:59 pm) 2,538 (10.8%) 3,004 (11.7%) PM peak (3 pm6:59 pm) 7,666 (32.5%) 8,626 (33.6%) Overnight (7 pm6:59 am) 2,607 (11.0%) 2,709 (10.5%) Weekend or holiday 2,785 (11.8%) 2,906 (11.3%) Tier 1. Atlanta MPO 13,633 14,839 Weekdays 12,387 (90.9%) 13,539 (91.2%) AM peak (6 am9:59 am) 4,774 (35.0%) 5,018 (33.8%) Midday (10 am2:59 pm) 1,636 (12.0%) 1,920 (12.9%) PM peak (3 pm6:59 pm) 4,446 (32.6%) 5,008 (33.7%) Overnight (7 pm6:59 am) 1,531 (11.2%) 1,592 (10.7%) Weekend or holiday 1,245 (9.1%) 1,301 (8.8%) Tier 2. Medium MPOs 2,979 3,285 Weekdays 2,577 (86.5%) 2,861 (87.1%) AM peak (6 am9:59 am) 986 (33.1%) 1,059 (32.2%) Midday (10 am2:59 pm) 280 (9.4%) 360 (11.0%) PM peak (3 pm6:59 pm) 957 (32.1%) 1,071 (32.6%) Overnight (7 pm6:59 am) 355 (11.9%) 370 (11.3%) Weekend or holiday 402 (13.5%) 424 (12.9%) Tier 3. Small MPOs 1,942 2,127 Weekdays 1,539 (79.3%) 1,699 (79.9%) AM peak (6 am9:59 am) 583 (30.0%) 627 (29.5%) Midday (10 am2:59 pm) 167 (8.6%) 203 (9.6%) PM peak (3 pm6:59 pm) 583 (30.0%) 652 (30.7%) Overnight (7 pm6:59 am) 206 (10.6%) 216 (10.2%) Weekend or holiday 403 (20.7%) 428 (20.1%) Tier 4. Non-MPO 5,055 5,433 Weekdays 4,321 (85.5%) 4,680 (86.1%) AM peak (6 am9:59 am) 1,670 (33.0%) 1,735 (31.9%) Midday (10 am2:59 pm) 456 (9.0%) 521 (9.6%) PM peak (3 pm6:59 pm) 1,679 (33.2%) 1,894 (34.9%) Overnight (7 pm6:59 am) 515 (10.2%) 531 (9.8%) Weekend or holiday 734 (14.5%) 753 (13.9%) Classifications are based on commuter's home address. Excludes supercommutes (>100 miles). * Percent of regional total in parentheses. Based on time of arrival at or departure from work anchor. 145 Demographic Differences Table 63 shows demographic differences in average commute and work journey distance. As noted in the previous section in this chapter on Peak, Off-peak, and Weekend Travel, lowincome workers have the shortest commutes, in terms of distance. Upper-middle income households (earning $50,000$75,000 annually) have the longest commutes. Agewise, Generation X workers have the longest commutes, and workers who are above retirement age have the shortest. This is likely related to the fact that the majority of employed seniors work part-time, and part-time workers' commutes are shorter than those of full-time workers. Men's commute distances are longer than those of women, but a closer look reveals that this is only actually true among workers without a college degree; the commute distances for collegeeducated men and women are comparable. As will be presented in the next section in this chapter on Work Travel Distance by Mode, the differences between men's and women's commute durations are smaller than the differences in distance. These same patterns can be observed in annual per-worker miles of work travel (table 64) and percent of miles of work travel (table 65), with some variations. For example, while collegeeducated women's day-to-day commute distances are on par with those of college-educated men's, their annual commute PMT is lower because a larger share of women work part-time. 146 Table 63. Average work travel distance by demographic and employment characteristics. 147 Table 64. Average annual work travel distance per worker by demographic and employment characteristics. 148 Table 65. Percent of total work travel distance by demographic and employment characteristics. 149 WORK TRAVEL DISTANCE BY MODE Table 66 shows differences in commute length by mode. Including multimodal commutes (but excluding supercommutes), 95 percent of commute PMT comes from POVs (as driver or passenger), 89 percent as a driver, and 75 percent as a driver with no passengers in the vehicle. Table 66 is a reminder that PMT may overstate the prominence of some modes and understate others. Nonmotorized travel, for instance, represents just 0.3 percent of annual commute and work journey miles, but its mode share in terms of number of journeys is more than 10 times higher (3.2 percent). At the opposite end of the spectrum, if supercommutes were included, air travel would account for 0.1 percent of all work journeys but 8.1 percent of total work journey PMT.53 Ground supercommutes would comprise 0.7 percent of work journeys and 7.3 percent of total work journey PMT (not tabulated). The average distance for an individual commute is 13.6 miles (or 14.8 miles for the full work journey). Excluding supercommutes, the mode with the longest average commute distance is transit, followed by multimodal commutes. For full work journeys, multimodal work journeys are the longest. 53 This is the reason for generally excluding supercommutes from the analysis. 150 Table 66. Work travel distance by mode. 151 The average commute and work journey distances for a nonmotorized work journey are both 1.5 miles, making it the only mode for which full work journeys are not longer than commutes. However, this is not because nonmotorized commuters do not make stops. Auto users are the most likely to make stops on their commutes (24.2 percent of unimodal commutes by auto are complex). While nonmotorized complex work journeys are less common, nonmotorized commutes are slightly more likely to include a stop than transit work journeys (3.7 percent versus 3.3 percent). What this does suggest, however, is that nonmotorized commuters who make stops are likely to visit destinations that are directly along their normal route rather than making detours. COMMUTE DURATION AND BURDEN Having discussed the distance of work travel, we turn now to the duration. Both distance and duration are measurements of a commute's length, but because speeds vary by mode and region, distances do not, by themselves, indicate how much of a worker's day is devoted to traveling to and from the job. This report presents three measures of commute duration. 1. Self-reported "usual" commute duration, asked of all workers. 2. Travel day commute duration: The duration of commutes observed on the travel day. For complex commutes (those containing one or more stops), the commute duration is calculated using the methods described previously in Defining the Commute. 3. Travel day work journey duration: The full travel duration of all trips involved in the work journey (as opposed to the commute-only portion presented for measure 2). 152 Time spent at stops is not included in commute or work journey duration. Supercommutes (longer than 100 miles) are excluded unless otherwise stated. As shown in table 67, reported "usual" commute times were slightly shorter than observed travel day commute times. The average "usual" commute was 30.9 minutes, the average travel day commute was 32.5 minutes, and the average travel day work journey lasted 35.9 minutes. Depending on which indicator is consulted, 1013 percent of work travel lasts at least 1 hour. This figure excludes supercommutes, which last an average of 4 hours and 11 minutes. The bottom half of table 67 shows the difference between an individual commuter's reported usual commute duration and that commuter's travel day commute and work journey durations. In a plurality of cases, the "usual" and travel day commute times differ by less than 5 minutes. In 16.8 percent of cases, the travel day commute is shorter than the "usual" commute, and in 37.6 percent of cases, the travel day commute is longer. The fact that observed commutes are more likely to be longer than the "usual" duration suggests that respondents report their "usual" commute times under good traffic conditions. With respect to supercommutes, 92.3 percent are longer than the corresponding "usual" times, and 82.5 percent exceed the duration of the usual commute by at least 30 minutes. This suggests that most supercommutes are unusual work travel or business trips and do not reflect that person's typical commute times. For this reason, supercommutes are excluded from further analysis in this chapter. 153 Table 67. Work-related travel duration in minutes. Workers* Observed Work Journeys "Usual" Commute Duration Travel Day: Commute Travel Day: WJ Travel Day: Supercommute Mean (minutes) 30.9 32.5 35.9 250.4 Percentiles (minutes) 5th 50th 95th 5 5 5 106 25 26 30 222 75 81 90 490 Percent lasting 60+ minutes 11.9% 10.3% 13.1% 100% Compared to the "usual" commute, the travel day duration was... Shorter by at least 5 minutes 16.8% 13.7% 6.4% The same (04 minutes difference) 45.6% 42.9% 1.3% Longer by at least 5 minutes 37.6% 43.4% 92.3% * This question was asked of all workers; 6,961 out of 8,363 (unweighted) responded. Weighted N=3.97 million workers. Based on all work journeys 0100 miles reported in travel diaries. The sample included 10,378 work journeys made by 5,067 unique commuters (unweighted). Weighted N=1.94 billion WJs by 2.61 million active commuters. Based on all work journeys greater than 100 miles reported in travel diaries. The sample included 85 supercommutes made by 67 unique commuters (unweighted). Weighted N=15.42 million supercommutes by 36,366 active commuters. (Travel day duration minus usual duration) -5. -5 < (Travel day duration minus usual duration) < 5. (Travel day duration minus usual duration) 5. Commute Times by Time of Day and MPO Tier Commute times are longest during weekday evening peak hours (table 68). The Atlanta MPO has the longest commute times and the most variability; the average weekday PM peak commute is 42.2 minutes, which is 13.6 minutes longer than the average overnight commute. Non-MPO counties have the second-longest commute times, but the least variability by time of day. Small MPO counties have the shortest commute durations on average. 154 Table 68. Travel day commute and work journey duration by time of day and MPO tier. Commute Duration Work Journey Duration Mean Median Percent Mean Median Percent (Minutes) (Minutes) 1 Hour (Minutes) (Minutes) 1 Hour Statewide Weekdays 32.4 27 13.6% 35.8 30 17.1% AM peak (6 am9:59 am) 32.1 27 13.3% 33.9 30 15.1% Midday (10 am2:59 pm) 27.3 20 10.1% 33.1 24 15.3% PM peak (3 pm6:59 pm) 36.7 30 18.1% 41.4 33 23.1% Overnight (7 pm6:59 am) 27.1 24 6.4% 28.9 25 8.2% Weekend or holiday 32.7 25 13.3% 35.9 26 16.4% Tier 1. Atlanta MPO Weekdays 36.4 30 17.4% 39.9 31 21.1% AM peak (6 am9:59 am) 35.9 30 16.4% 37.7 30 18.2% Midday (10 am2:59 pm) 29.9 24 12.5% 36.1 29 17.3% PM peak (3 pm6:59 pm) 42.2 35 24.2% 47.2 40 29.7% Overnight (7 pm6:59 am) 28.6 25 6.8% 30.2 25 9.6% Weekend or holiday 38.6 28 19.9% 42.4 30 23.3% Tier 2. Medium MPOs Weekdays 26.8 21 6.7% 30.7 25 10.1% AM peak (6 am9:59 am) 25.9 20 6.4% 28.0 23 9.0% Midday (10 am2:59 pm) 23.6 20 5.5% 31.1 20 11.2% PM peak (3 pm6:59 pm) 29.5 25 8.0% 33.5 29 11.9% Overnight (7 pm6:59 am) 26.4 20 5.8% 29.7 20 6.6% Weekend or holiday 28.4 20 9.0% 32.7 21 12.6% Tier 3. Small MPOs Weekdays 24.6 20 8.0% 27.6 20 10.9% AM peak (6 am9:59 am) 24.0 20 7.5% 25.8 20 8.8% Midday (10 am2:59 pm) 21.3 15 6.8% 26.3 16 12.0% PM peak (3 pm6:59 pm) 27.6 23 10.8% 31.6 25 14.8% Overnight (7 pm6:59 am) 21.9 20 3.3% 23.3 20 4.7% Weekend or holiday 24.7 20 4.5% 27.3 22 7.0% Tier 4. Non-MPO Weekdays 27.9 20 9.7% 30.5 25 12.9% AM peak (6 am9:59 am) 28.4 20 11.4% 29.6 20 12.9% Midday (10 am2:59 pm) 25.2 18 8.4% 28.6 20 14.5% PM peak (3 pm6:59 pm) 29.2 22 9.8% 33.4 28 14.4% Overnight (7 pm6:59 am) 25.8 24 7.0% 27.3 25 7.0% Weekend or holiday 30.0 25 9.7% 31.1 27 11.9% Classifications are based on commuter's home address. Excludes supercommutes (>100 miles one way). Based on time of arrival at or departure from work anchor. 155 Commute Times by Mode and Vehicle Ownership As shown in table 69, nonmotorized commutes are the shortest in terms of time, and public transit commutes are the longest. However, table 69 suggests a two-tiered commuting system divided not just by mode, but by the ability to choose between modes. Unsurprisingly, "choice" commuters from vehicle-sufficient households (i.e., those with vehicles available for every potential driver) are more likely to commute by POV. However, choice commuters enjoy commutes of a shorter duration even when they choose an alternative mode. The average transit commute for a captive rider54 is 20.4 minutes longer than the average choice transit commute (table 69), despite the fact that the average distance of captive riders' commutes is 3.1 miles shorter than a choice rider's transit commute (14.2 miles versus 17.3 miles) (not tabulated). In other words, on average, commuters who use transit out of necessity receive worse service than commuters who take transit by choice. There are many potential explanations for this inequality. Choice riders can opt out of the transit system if the available routes do not align with their needs, while captive riders may have to use transit regardless of travel times. Service jobs, a prominent source of low-income employment, may be more geographically dispersed than white-collar jobs, rather than being concentrated in higher-density employment nodes that can be more efficiently served by transit. It is possible 54 A "captive" rider is a commuter from a zero-vehicle or vehicle-deficit household (i.e., a household which owns at least one car, but has fewer cars than potential drivers). The transit mode share for commutes made by residents of low-income vehicle-deficit households (<$35,000 per year) is more than double the rate of transit use by vehicle-sufficient households at the same income level (5.0 percent versus 2.0 percent). This suggests that when residents of vehicle-deficit households use transit, they are likely to be doing so because they were not able to drive or get a ride from a household member. 156 that choice riders may be more likely to complain to transit operators, increasing the likelihood that their issues will be addressed. Table 69. Work travel duration by mode and vehicle ownership. Commute Duration Mean Median (Minutes) (Minutes) Work Journey Duration Percent Mean Median Percent 1 Hour (Minutes) (Minutes) 1 Hour All Work Journeys 0100 Miles Household Vehicle Ownership* Zero-vehicle Vehicle-deficit Vehicle-sufficient 39.2 32.8 32.1 25 24.4% 42.5 25 13.3% 35.2 27 13.3% 35.7 30 27.8% 27 15.3% 30 17.2% Mode POV Nonmotorized Public transit or other bus/train Other ground or water Multiple modes 31.6 17.0 69.5 17.5 56.5 26 12.3% 34.8 15 5.1% 17.7 63 58.0% 70.8 15 0.9% 17.6 52 42.3% 81.0 30 15.7% 15 5.7% 63 58.0% 15 0.9% 79 62.4% Vehicle-sufficient Household Work Travel ("Choice" Commuters) POV 31.9 27 12.8% 35.3 Nonmotorized 12.5 5 0.7% 13.6 Public transit or other bus/train 58.2 60 50.7% 58.2 Multiple modes 43.1 38 27.8% 77.2 30 16.5% 5 2.1% 60 50.7% 61 57.3% Vehicle-deficit and Zero-vehicle Household Work Travel POV 30.5 25 10.5% 33.0 26 12.8% Nonmotorized 20.2 15 8.3% 20.7 15 8.3% Public transit or other bus/train 78.6 72 63.8% 80.9 72 63.8% Multiple modes 71.6 79 58.8% 85.3 85 68.2% * Vehicle-sufficient households have at least one vehicle for each potential driver. Vehicle-deficit households have at least one vehicle, but fewer vehicles than potential drivers. Due to inadequate sample size, trips with the mode "other" are not subdivided by vehicle ownership. Whatever the causes, the effect is that transit-dependent commuters are experiencing worse transit service than their choice-rider neighbors. To reduce this inequality, it is therefore important to examine potential discrepancies in vehicle frequency and route density, and to examine whether current transit routings match the needs of low-income commuters. 157 Vehicle ownership also affects nonmotorized commutes; 8.3 percent of nonmotorized commutes by captive pedestrians and cyclists lasted an hour or more, versus just 0.7 percent of nonmotorized choice commutes. Inspiring more commuters to choose walking and cycling is a laudable goal, but this finding suggests that captive pedestrians and cyclists need a different kind of assistance. In addition to facilitating access to other transportation options, it is critical to ensure that the locations where these obligatory nonmotorized commuters walk and bike have adequate and safe infrastructure. Transit service improvements and subsidies for obligatory nonmotorized commuters could help improve transportation for commuters without vehicle access. However, it should be noted that the time discrepancy between commuters in vehicle-sufficient and vehicle-insufficient households vanishes when examining commutes made by cars (as either driver or passenger). Helping all these workers acquire cars could have serious consequences for sustainability and congestion, but there is no denying that it would be a substantial quality of life improvement for many, not just in terms of shortening an existing commute but also in terms of expanding future employment opportunities (O'Regan and Quigley 1998, Shen and Sanchez 2005, Blumenberg and Pierce 2014, Blumenberg 2016, Loukaitou-Sideris 2016, Smart and Klein 2018). It may be beneficial to explore ways of promoting community carpooling or carsharing and providing microtransit service to a wider variety of routes. Demographic Differences in Commute Duration Table 70 and table 71 show work travel duration broken down by demographic characteristics. The usual commute, travel day commute, and travel day work journey are shown. As mentioned in the section on Work Travel Distance, in terms of distance, men's commutes are, on average, 158 longer than women's by 1.8 miles. One notable finding here is that this difference does not extend to work travel duration (table 70). While men's median "usual" commute time is 5 minutes longer than that of women, the difference for observed commutes on the travel day is only 2 minutes. When the full work journey is considered, men's and women's commute times differ by just 30 seconds on average (and the medians are identical). In other words, while female workers do not produce as much VMT as male workers do, from a quality of life standpoint, female workers spend just as much time commuting as their male colleagues do. Similarly, while Black commuters' commute distance was almost identical to that of white workers, Black workers' commutes last, on average, 7.4 minutes longer than the commutes of white workers (table 71). This was the largest intergroup difference observed. It may relate to racial differences in mode choice and residential patterns (e.g., work journey durations are longer in Atlanta and when using transit). In terms of economic and occupational characteristics, the longest commutes were found among blue-collar workers. Total Daily Commute Burden To understand the cumulative effects of work travel on Georgians' time use, in addition to examining the duration of individual one-way commutes, it is worth examining the total amount of time spent commuting per day (i.e., the commute burden). Table 72 and table 73 show the total amount of time spent traveling to and from work on a typical work day, broken down by various demographic characteristics. As a point of comparison, total daily commute distance is also shown. 159 Table 70. Work travel duration by MPO tier, gender, education, and age. "Usual" Commute Travel Day Commute Travel Day Work Journey Mean Median Percent Mean Median Percent Mean Median Percent (Minutes) (Minutes) 1 Hour (Minutes) (Minutes) 1 Hour (Minutes) (Minutes) 1 Hour Statewide* 30.9 25 11.9% 32.5 MPO Tier (Worker Residence) 1. Atlanta MPO 34.6 30 15.7% 36.6 2. Medium MPOs 23.4 20 5.0% 27.0 3. Small MPOs 23.0 20 5.9% 24.6 4. Non-MPO 27.3 20 10.9% 28.1 Gender Male 31.4 25 13.0% 32.8 Female College-educated 29.2 20 11.3% 32.0 29.7 25 11.5% 32.0 College-educated men 29.5 25 11.5% 31.7 College-educated women 29.8 25 11.4% 32.4 Without 4-year College Degree 31.0 25 12.8% 32.7 Men without college degree 32.6 25 14.0% 33.5 Women without college degree 28.8 20 11.2% 31.6 Age Cohort Millennial and Gen Z (1836) 28.8 20 10.7% 31.8 Generation X (3752) 32.2 25 13.4% 33.6 Pre-retirement age Baby Boomer (5364) 31.2 25 13.2% 32.6 Retirement age (65+) 27.9 20 13.2% 27.9 * Excludes commutes greater than 100 miles one way. Weighted N=1.96 billion per year. Defined as bachelor's degree or higher. 26 10.3% 35.9 30 17.6% 40.2 20 7.0% 31.0 20 7.4% 27.5 21 9.7% 30.6 27 14.1% 36.0 25 13.0% 35.5 28 12.2% 35.6 27 12.0% 35.6 28 12.5% 35.7 25 14.6% 35.9 28 15.4% 36.4 25 13.4% 35.3 25 13.1% 35.1 30 13.8% 36.8 27 15.1% 36.3 21 9.9% 32.0 30 13.1% 30 21.3% 25 10.4% 20 10.2% 25 12.8% 30 17.4% 30 16.6% 30 16.2% 30 16.2% 30 16.1% 30 17.6% 30 18.1% 28 17.0% 28 16.2% 30 17.0% 30 18.8% 25 17.1% 160 Table 71. Work travel duration by income, race, occupation, and worker type. "Usual" Commute Mean Median (Minutes) (Minutes) Travel Day Commute Travel Day Work Journey Percent Mean Median Percent Mean Median Percent 1 Hour (Minutes) (Minutes) 1 Hour (Minutes) (Minutes) 1 Hour Statewide* 30.9 25 11.9% 32.5 Annual Household Income <$35,000 30.0 20 13.4% 32.3 $35,000 to $49,999 28.2 20 9.0% 30.0 $50,000 to $74,999 31.5 25 12.5% 33.5 $75,000 to $99,999 31.3 25 12.2% 33.0 $100,000+ 30.5 25 11.9% 32.5 Race White non-Hispanic only 28.0 20 10.5% 30.5 Black and Black multiracial 35.4 30 15.8% 36.6 Other 28.7 25 11.0% 30.5 Occupational Category (NHTS-designated Workers Only) Sales or service 28.8 20 11.8% 31.0 Clerical or administrative support Blue collar 30.9 25 13.9% 33.1 34.6 25 15.4% 34.5 Professional, managerial, or technical 29.6 25 10.8% 32.7 Worker Type (NHTS-designated Workers Only) Part-time 28.6 20 11.2% 30.4 Full-time 30.8 25 12.5% 32.9 * Excludes commutes greater than 100 miles one way. Weighted N=1.96 billion per year. Blue collar refers to manufacturing, construction, maintenance, or farming. 26 10.3% 35.9 25 14.6% 35.9 25 11.3% 32.4 28 13.6% 36.3 27 14.3% 36.1 30 13.2% 36.6 25 11.5% 33.3 30 17.6% 40.8 25 12.5% 33.6 25 12.5% 34.5 28 14.0% 36.9 30 16.8% 36.8 27 13.0% 36.1 20 12.0% 34.0 28 14.0% 36.2 30 13.1% 27 17.7% 25 13.8% 30 16.7% 30 17.8% 30 17.7% 28 14.5% 30 22.2% 28 15.0% 27 15.8% 30 16.6% 30 18.6% 30 17.2% 25 16.1% 30 17.2% 161 The median daily commute duration is 55 minutes (or 60 minutes if the full work journey is included). However, 12.3 percent of commuters spend 2 hours or more traveling to and from work. These heavy commute burdens are not equally distributed; 2-hour commute burdens are more than twice as common in the Atlanta MPO as elsewhere in the state (16.6 percent in Atlanta versus 6.27.1 percent elsewhere). If the whole work journey is included, more than one in five Atlanta-MPO employees has a commute burden of 2 or more hours. As with individual commute durations, Black residents also have the longest total average commute burden (73.6 minutes, versus 62.1 and 64.5 for white non-Hispanic and other race, respectively). Retirement-age commuters generally have somewhat smaller median commute burdens than other age groups. They also have a lower proportion of 2-hour commute burdens. As a result, the age difference in average commute burden is larger than the difference in median commute burden. CONCLUSION This chapter has considered Georgians' travel to and from job sites. As telecommuting becomes more common, many Georgia workers "travel" to work without leaving their houses. Some workers also benefit from flexible work schedules, enabling them to avoid peak-hour work travel, and/or better balance work and life demands. The next chapter will focus on patterns associated with telecommuting and work schedule flexibility. 162 Table 72. Total daily commute burden by MPO tier, gender, education, and age. Daily Commute Totals Minutes Minutes Percent Miles (Mean) (Median) 2 Hours (Mean) Daily Work Journey Totals Minutes Minutes Percent Miles (Mean) (Median) 2 Hours (Mean) Statewide* 66.2 55 12.3% 27.8 73.0 MPO Tier (Worker Residence) 1. Atlanta MPO 74.4 63 16.6% 28.6 81.7 2. Medium MPOs 56.5 47 7.1% 23.4 64.7 3. Small MPOs 50.6 42 6.2% 22.4 56.6 4. Non-MPO 56.5 45 6.4% 31.5 61.4 Gender Male 66.9 56 12.9% 29.5 73.5 Female 65.3 54 11.6% 25.6 72.4 College-educated 66.1 58 11.7% 27.2 73.4 College-educated men 65.8 57 12.2% 27.2 73.9 College-educated women 66.3 60 11.0% 27.2 72.9 Without 4-year College Degree 66.2 53 12.7% 28.1 72.7 Men without college degree 67.5 55 13.3% 30.9 73.3 Women without college degree 64.4 50 12.0% 24.4 71.9 Age Cohort Millennial and Gen Z (1836) 65.4 55 11.6% 26.5 72.1 Generation X (3752) 67.5 57 12.9% 29.8 74.0 Pre-retirement age Baby Boomer (5364) 66.9 52 13.8% 27.4 74.4 Retirement age (65+) 58.2 50 7.7% 23.7 66.8 * Includes all active commuters with no supercommutes on travel day (weighted N=2.6 million). Defined as bachelor's degree or higher. 60 16.3% 30.2 70 21.0% 31.2 53 10.8% 25.9 45 9.3% 24.5 50 9.9% 33.8 60 16.9% 32.0 60 15.5% 28.0 65 16.1% 29.9 65 17.3% 30.4 65 14.9% 29.4 60 16.4% 30.5 60 16.6% 33.0 59 16.0% 26.9 60 15.0% 29.0 63 16.7% 32.2 60 18.0% 30.0 54 17.4% 26.4 163 Table 73. Total daily commute burden by income, race, occupation, and worker type. Daily Commute Totals Daily Work Journey Totals Minutes Minutes Percent Miles Minutes Minutes Percent Miles (Mean) (Median) 2 Hours (Mean) (Mean) (Median) 2 Hours (Mean) Statewide* 66.2 55 12.3% 27.8 73.0 60 16.3% 30.2 Annual Household Income <$35,000 65.9 52 12.4% 23.5 73.3 60 16.8% 25.6 $35,000 to $49,999 63.5 53 10.4% 29.6 68.5 58 12.9% 31.5 $50,000 to $74,999 67.4 55 12.9% 31.4 73.1 60 15.9% 33.8 $75,000 to $99,999 67.8 57 13.2% 29.4 74.2 63 16.4% 31.9 $100,000+ 65.7 60 12.2% 28.2 74.0 65 17.8% 31.4 Race White non-Hispanic only 62.1 52 10.0% 28.1 68.0 60 13.4% 30.4 Black and Black multiracial 73.6 60 16.4% 28.1 82.2 65 21.4% 31.0 Other 64.5 52 11.5% 25.5 71.0 60 15.3% 27.7 Occupational Category (NHTS-designated Workers Only) Sales or service 63.2 52 10.6% 26.3 70.4 59 15.0% 28.7 Clerical or administrative support Blue collar 66.7 53 14.2% 25.8 74.4 60 17.4% 28.6 69.4 55 14.3% 31.0 74.1 60 16.8% 32.9 Professional, managerial, or technical 67.1 59 12.5% 28.1 74.2 65 16.8% 30.8 Worker Type (NHTS-designated Workers Only) Part-time 61.7 49 9.4% 22.8 69.5 52 15.1% 25.3 Full-time 67.3 57 13.0% 28.9 73.8 60 16.6% 31.4 * Includes all active commuters with no supercommutes on travel day (weighted N=2.6 million). Blue collar refers to manufacturing, construction, maintenance, or farming. 164 CHAPTER 3. TELEWORKING AND WORKER SCHEDULE FLEXIBILITY CHAPTER 3 SUMMARY This chapter analyzes workplace forces that afford workers more choice in how, when, or if they travel to work. In particular, its focus is on teleworking and schedule flexibility. Overview includes technical and vocabulary notes and provides an overview of schedule flexibility and teleworking as an overall proportion of worker activity in Georgia. It also provides summary statistics about the number of telecommuters each day by MPO tier. Work Flexibility for Whom? examines access to flexible work conditions. It considers demographic differences between Georgians who have the ability to flex their schedule or work location and those whose job is less flexible. Some of the differences in work flexibility are related to job function or prestige; workers in high-income households and with college degrees are more likely to have a flexible worksite or schedule. However, using regression analysis, we document that workers' likelihood of being allowed to telework or set their own schedule is also influenced by where they live and demographic factors that are not related to the nature of the work in which they engage. The Effects of Flexible Scheduling examines the effects of schedule flexibility on arrival times to and departure times from work. In general, workers with flexible schedules tend to arrive at and depart from work later than workers with inflexible schedules. While schedule flexibility influences departure and arrival times, it was not found to have a significant effect on commute duration. 165 Frequency of Teleworking examines the incidence of teleworking by looking at: (1) the monthly frequency of teleworking by telecommute-eligible workers, and (2) commuting and telecommuting on the travel day. In particular, we focus on differences by region, gender and caregiver status, mobility impairment, and distance between the home and work location. Work flexibility, particularly teleworking, has taken on new significance with the onset of the coronavirus disease 2019 (COVID-19) pandemic in 2020. The data analyzed in this report obviously reflect pre-COVID-19 conditions. With respect to teleworking prevalence (i.e., adoption and frequency; Frequency of Teleworking), it remains to be seen what the postCOVID-19 "new normals" will be; the 2017 NHTS data will provide an important benchmark against which to gauge future levels of teleworking. With respect to telework eligibility (see sections on Descriptive Statistics on Access to Flexible Schedule and Work Location and Logistic Regression Analysis of Eligibility for Flexible Work Location), changes may be more uneven. The pandemic has made it clear that many kinds of essential work cannot be accomplished remotely. On the other hand, the pandemic may also expand the number and types of job functions that are considered possible to accomplish remotely, and in many instances transform teleworking from a perk to a public health measure seen to be in the interests of employers as well as employees. OVERVIEW Types of Work Flexibility Employers provide Georgia workers with two kinds of flexibility: scheduling flexibility, or the ability to adjust start and end times, and location flexibility, or the ability to work from home or 166 a "third place" such as a coffee shop or coworking space (Oldenburg 1989).55 As shown in table 74, 40.9 percent of Georgia workers have work schedule flexibility. Twenty-seven percent have flexible work locations (either telecommuting periodically56 or working from home on a regular basis). There is substantial overlap between workers with schedule flexibility and those with location flexibility; 22.3 percent of workers have both schedule and location flexibility, while 18.7 percent have schedule flexibility only. Most workers who have work location flexibility also have schedule flexibility; just 4.9 percent of workers have location flexibility without schedule flexibility. Table 74. Eligibility for job flexibility among Georgia workers. Flexibility Type Schedule Schedule Location Only Only* and Location* All workers 18.7% 4.9% 22.3% MPO Tier 1. Atlanta MPO 19.2% 5.7% 27.2% 2. Medium MPOs 18.2% 4.0% 15.1% 3. Small MPOs 18.5% 3.1% 15.1% 4. Non-MPO counties 17.4% 3.9% 15.9% * Includes home-based workers and telecommute-eligible workers. Totals Schedule Location* No With or Without With or Without Flexibility Location Schedule 54.2% 40.9% 27.1% 47.9% 62.7% 63.2% 62.8% 46.4% 33.3% 33.6% 33.3% 32.9% 19.1% 18.2% 19.8% A slight majority of workers in the Atlanta MPO area have flexibility with regard to schedule, location, or both. Elsewhere in the state, 63 percent of workers have neither schedule nor location flexibility. 55 Flexible scheduling, or flextime, is assessed by asking participants, "At your primary job, do you have the ability to set or change your own start time?" If the response is yes, then for simplicity of exposition, workers are considered to have a flexible schedule (or "schedule flexibility"), whether or not they choose to take advantage of the option to vary their work schedule. Location flexibility is assessed with a series of questions; see Worker Telework Eligibility Categories. 56 Table 74 and the accompanying text include the 3 percent of workers, shown in figure 18, who are eligible to telecommute but have not done so in the past 30 days. 167 Definitions and Technical Notes Teleworking refers to working and interacting remotely with an employer, colleagues, and clients using the internet and/or telephone. Typically, a teleworker works from home. However, some may work out of nearby coffee shops, coworking spaces, or similar facilities--a type of "third place," to use the term popularized by Oldenburg (1989) and often applied in teleworking contexts (e.g., Venolia et al. 2014). Teleworking and telecommuting are often used interchangeably. In this report, however, telecommuting refers to a specific case of teleworking, in which workers whose primary workplace is outside the home substitute teleworking for their usual conventional commute. Teleworking is an umbrella term that comprises telecommuting, but also employment where home is the "usual" work site (i.e., home-based work). Worker Telework Eligibility Categories For the purposes of this report, with respect to teleworking, there are three types of workers: Home-based workers, or workers who usually or always work from home. Home-based workers include people who work for remote employers (such as online tutoring companies) and self-employed people who operate a business out of their home office. Telecommute-eligible workers, or workers whose primary workplace is outside the home, but "have the option of working from home or an alternate location."57 In other words, with telecommuting, teleworking is substituting for a conventional commute. Although being an eligible telecommuter does not require taking advantage of this option, most 57 This is the wording used in the NHTS questionnaire; see p. 44 of the Retrieval Questionnaire found at https://nhts.ornl.gov/assets/2016/NHTS_Retrieval_Instrument_20180228.pdf. 168 (78 percent) telecommute-eligible workers report telecommuting at least once in the past 30 days (figure 18). Telecommute-ineligible workers, or nonhome-based workers who do not have the option of telecommuting. Home-based workers and telecommute-eligible workers are considered to have a flexible work location. These categories are measured by the NHTS with a series of questions. Workers are first asked if they "usually" work from home. If so, they are considered home-based workers. Nonhome-based workers are asked a follow-up question about whether they "have the option of working from home or an alternate location instead of going into your/their primary workplace." Workers who answer "yes" to this second question are considered telecommute-eligible; those who answered "no" are considered telecommute-ineligible.58 Telecommute-eligible workers are asked about the number of days they worked from home over the past 30 days; this information is not asked of home-based or telecommute-ineligible workers. However, the travel day telework categories used in this report (see Travel Day Work and Telework Categories) do cover teleworking by all workers, regardless of their official telework status. 58 This question wording may result in the self-misclassification as "telecommute-eligible" of some selfemployed workers whose jobs routinely take them to "alternate locations" (e.g., plumbers, electricians, etc.), because they do not "usually" work from home (which would have classified them as a "home-based worker"). 169 Figure 18. Pie chart. Telework category of Georgia workers and telecommuting frequency, past 30 days. Travel Day Work and Telework Categories In addition to ascertaining the telework eligibility status of individuals, it is also of interest to analyze whether they actually teleworked on the day the travel diary was completed. We define an active worker as someone who reported working at a home or nonhome location on the travel diary day. Because respondents are not asked directly about travel day teleworking, that classification was made based on respondents' stated trip purposes (or, if no trips were made, stated reason for not making any trips). Specifically, respondents have engaged in a conventional commute if they meet either of the following criteria: Reported working for pay, engaging in work-related business, or working "from home" at their work address. Reported working for pay or engaging in work-related business at any location besides their home and work addresses. 170 This definition diverges from the definition used in chapter 2 on commuting, by including trips for the purposes of off-site work-related meetings. The reason for this methodological difference is that the most relevant question for the present analysis is whether or not the worker leaves the house (versus accomplishing the same work from home), whereas for previous analysis the focus was on travel to and from the workplace. For the purposes of analyzing teleworking, the relevant distinction is whether or not work causes a worker to leave the home. Participants have teleworked on the travel day if they meet any of the following criteria: Reported no trips on the travel day with the stated reason that they worked from home. Described primary activity at any location visited on the travel day as "work from home (paid)."59 Reported working for pay or engaging in a work-related meeting at their home address.60 Travel day teleworking is subdivided into exclusive telework (i.e., the participant did not work or conduct work business outside of the home on the travel day), and mixed telework, where the participant reported both a conventional commute and teleworking.61 Across the 7-day week, on average, 5.2 percent of Georgia adults, or 13.1 percent of active workers, report teleworking on a given day (figure 19). 59 Twenty-two participants reported working specifically "from home" at a location other than home or work; since these may reflect telecommuting from a friend or relative's house or a secondary address, this report defers to the participant's judgment and classifies them as travel-day telecommuters. The exception to this rule is that if respondents reported working from "home" at their work location, it was recoded as working outside the home. 60 Unfortunately, the data do not provide a way to distinguish between off-site work meetings and telework at a "third place" (such as a caf), so this definition of telework does not include third-place teleworking. 61 These two categories are sometimes described as "full day" and "partial day" telecommuting; this report generally avoids that terminology because its classification does not depend on the number of hours worked. 171 Figure 19. Pie chart. Travel day working and teleworking among Georgia adults. It should be noted that participants can report teleworking on the travel day even if their job does not usually allow telecommuting. This may reflect a second job, or an exception granted by an employer in the case of an illness, weather emergency, or other circumstance.62 While these data were collected before the outbreak of COVID-19, recent events have illustrated that some workers who are not normally considered telework-eligible can sometimes be given the option of working from home in extraordinary circumstances. The travel diary data unfortunately does not allow for distinguishing between nonhome teleworking sites (such as coworking spaces) and other work locations (such as a second job, alternate office address, or visit to a client). To be conservative, therefore, in this analysis travel day telecommuting refers only to working from home, or to reporting a nonhome activity as such; any other work-related nonhome activity is not assumed to be teleworking. In that respect, this analysis will underestimate the amount of teleworking occurring. 62 In addition, a small percentage of respondents not designated as workers by the NHTS reported teleworking on the travel day. They are included in figure 19 for completeness. See Overview of Travel Day Telecommuting for further discussion. 172 In addition, the travel day telecommuting metric likely (even further) underestimates the true incidence of telecommuting because it is based on the reported primary activity at a location, and each location is limited to a single listed activity. If individuals are working from home for an extended period of time, it is likely that they conduct a mix of work and regular home activities, and which purpose participants choose to list may be somewhat idiosyncratic. Even with this caveat, however, measuring travel day telecommuting provides valuable information to this analysis. Overview of Travel Day Telecommuting As shown in table 75, on a typical weekday, nearly 498,000 Georgia adults will telecommute, representing 6.5 percent of the total adult population. This is equivalent to 13 percent of people who worked on the travel day ("active workers") (not tabulated). There are clear regional differences, with telecommuting being most common in Atlanta. As with the analysis of commuting travel in chapter 2, both commuting and telecommuting are observed by Georgians who are not officially categorized as workers.63 These "nonworkers" are included in table 75 for completeness. They represent a comparatively small group (just 151 respondents). Because the other survey items analyzed for this chapter were only asked of NHTS-defined workers, for consistency, throughout the remainder of the chapter, the researchers restrict the sample to NHTS-defined workers. Specifically, telecommuting by "nonworkers" is excluded from all analysis in this chapter aside from table 75 and figure 19. 63 The NHTS defines workers based on their workforce participation in the week leading up to the survey; economic activity on the travel day by "nonworkers" might reflect a change in employment status, irregular work, or an error. 173 Table 75. Travel day working and teleworking by MPO tier. Active Workers Teleworked on Travel Day Weekdays Total Population (Worked for Pay on Travel Day) (Exclusively or Partially) All adults* ages 18+ 7,704,068 3,854,775 (50.0%) 497,627 (6.5%) MPO Tier 1. Atlanta MPO 4,167,843 2,290,234 (55.0%) 354,766 (8.5%) 2. Medium MPOs 1,219,439 577,655 (47.4%) 48,334 (4.0%) 3. Small MPOs 777,264 351,886 (45.3%) 25,024 (3.2%) 4. Non-MPO 1,539,522 630,659 (41.0%) 64,090 (4.2%) All workers* ages 18+ 4,736,051 3,761,714 (79.4%) 471,779 (10.0%) MPO Tier 1. Atlanta MPO 2,735,122 2,222,401 (81.3%) 335,322 (12.3%) 2. Medium MPOs 741,844 567,367 (76.5%) 44,741 (6.0%) 3. Small MPOs 474,303 358,100 (75.5%) 23,874 (5.0%) 4. Non-MPO 784,782 608,566 (77.5%) 60,964 (7.8%) Weekends and Holidays All adults ages 18+ 7,704,068 1,201,200 (15.6%) 169,268 (2.2%) All workers ages 18+ 4,736,051 1,155,736 (24.4%) 139,343 (2.9%) All Days All adults ages 18+ 7,704,068 3,044,842 (39.5%) 397,404 (5.2%) MPO Tier 1. Atlanta MPO 4,167,843 1,811,751 (43.5%) 285,910 (6.9%) 2. Medium MPOs 1,219,439 430,139 (35.3%) 37,264 (3.1%) 3. Small MPOs 777,264 278,552 (35.8%) 20,105 (2.6%) 4. Non-MPO 1,539,522 524,399 (34.1%) 54,126 (3.5%) All workers ages 18+ 4,736,051 2,972,715 (62.8%) 371,129 (7.8%) MPO Tier 1. Atlanta MPO 2,735,122 1,772,912 (64.8%) 268,167 (9.8%) 2. Medium MPOs 741,844 418,858 (56.5%) 34,295 (4.6%) 3. Small MPOs 474,303 273,055 (57.6%) 18,368 (3.9%) 4. Non-MPO 784,782 507,891 (64.7%) 50,299 (6.4%) * NHTS defines a worker as someone who worked for pay or profit or was temporarily absent from paid employment in the week before completing their travel survey ("last week"). Of adults who were not classified as workers, 2.4 percent reported working on the travel day, including 0.9 percent who reported teleworking. These "nonworker" workers are included in totals for all adults, but not totals for workers. Note: Percents shown are percent of row subpopulation (adults or workers). However, it is worth noting that more than one third of these irregular workers--36 percent-- reported teleworking. This is a much higher rate than was observed among regular workers. This suggests that teleworking may be more common among people whose employment situations are 174 not easily captured by the standard worker/nonworker binary measure. The issue of teleworking among irregular workers may therefore merit further study. As table 75 shows, on a typical weekday, approximately 80 percent of workers will be actively working for pay. On weekends, this figure is 24 percent, and there is some variation across MPO tiers. Table 76 shows teleworking among these active workers. Table 76. Travel day teleworking and conventional commuting among active workers. Weekdays All workers* ages 18+ Exclusive Telework (Worked from Home Only) Mixed Telework (Worked from Home & Outside of Home) Conventional Commute (Worked Outside of Home Only) All Teleworking (Exclusive + Mixed) 9.0% 3.6% 87.5% 12.5% MPO Tier 1. Atlanta MPO 11.5% 3.6% 84.9% 15.1% 2. Medium MPOs 5.1% 2.8% 92.1% 7.9% 3. Small MPOs 3.0% 3.6% 93.3% 6.7% 4. Non-MPO Weekends and Holidays All workers ages 18+ All Days 5.8% 7.9% 4.2% 4.1% 90.0% 87.9% 10.0% 12.1% All workers ages 18+ 8.8% 3.7% 87.5% 12.5% MPO Tier 1. Atlanta MPO 11.3% 3.8% 84.9% 15.1% 2. Medium MPOs 5.1% 3.1% 91.8% 8.2% 3. Small MPOs 3.6% 3.1% 93.3% 6.7% 4. Non-MPO 6.2% 3.7% 90.1% 9.9% Note: Percentages shown are row percentages based on the population of active workers (workers who reported working for pay or working from home for pay on the travel day). * NHTS defines a worker as someone who worked for pay or profit or was temporarily absent from paid employment in the week before completing their travel survey ("last week"). As table 76 shows, at least 12.5 percent of active workers on a typical day will be working from home for some or all of their work day. Teleworking, particularly exclusive teleworking, is much more common in the Atlanta region than it is in the rest of the state. The area with the secondhighest rate of teleworking is non-MPO counties, where nearby work opportunities may be less 175 available. Rates of teleworking are similar on weekdays and weekends (12.5 percent and 12.1 percent, respectively). Due to sample size concerns, weekend telecommuting is not disaggregated by MPO tier. WORK FLEXIBILITY FOR WHOM? The opportunity to work from home or have a flexible schedule is not evenly distributed among Georgia workers. This section examines differences by region, employment characteristics, and demographic characteristics in who has the ability to choose their own work location or schedule. Workers are considered to have a flexible schedule if they have some leeway over what time they start and stop their work day. For this study, the category of flexible location includes: (1) workers who usually work outside the home but have the ability to telecommute (even if they have not done so within the past 30 days), and (2) home-based workers. Workers who fit neither category do not have location flexibility. Descriptive Statistics on Access to Flexible Schedule and Work Location As shown in table 77, 45.9 percent of workers have access to some form of flexibility, whether that be schedule (18.7 percent), location (4.9 percent), or both (22.3 percent). Work flexibility is more common in the Atlanta MPO region, where a slight majority of workers have some form of flexibility, than in the rest of the state. Workers with higher education levels and incomes are more likely to have flexible schedules and locations. Work flexibility is also more common among professional, managerial, and technical jobs than other occupation categories. 176 Table 77. Employment flexibility by MPO, job characteristics, educational attainment, and income. 177 However, access to employment flexibility is also correlated with a number of demographic characteristics that are not directly related to job type or function (table 78). White workers are more likely to have flexible jobs than workers of other races. Workers with a mobility impairment are slightly more likely than nondisabled workers to have a flexible schedule, and much more likely to have a flexible location (38.5 percent versus 26.9 percent). It is unclear whether this is an accommodation offered by employers to workers with disabilities, or if mobility-impaired workers unable to find flexible employment simply exit the labor market. Men are more likely to have a flexible schedule and location than women. This gap is more pronounced among caregivers for children under the age of 16 than it is among noncaregivers (table 79), despite the fact that the burden of childcare falls disproportionately on women (McQuaid and Chen 2012, Loukaitou-Sideris 2016). As the next section (Logistic Regression Analysis of Eligibility for Flexible Work Schedule) will show, many of these demographic differences in employment flexibility persist after controlling for other factors. 178 Table 78. Employment flexibility by sex, caregiver status, age, race, and mobility impairment. Flexibility Type Totals Flexible Schedule Flexible Location* Schedule Only Location* Schedule & No Only Location Flexibility With or Without Location With or Without Schedule All workers 18.7% 4.9% 22.3% 54.2% 40.9% 27.1% Sex Male Female Caregiver Status 19.9% 17.2% 4.4% 5.4% 23.7% 20.6% 52.0% 56.8% 43.6% 37.8% 28.1% 26.0% Noncaregiver Caregiver, youngest child ages 015 18.4% 19.1% 5.3% 4.1% 21.2% 24.0% 55.1% 52.7% 39.6% 43.1% 26.5% 28.2% Age Cohort Millennial and Gen Z (1836) Generation X (3752) Pre-retirement age Baby Boomer (5364) Retirement age (65+) 15.5% 19.7% 20.9% 25.5% 4.9% 4.6% 5.9% 1.9% 16.6% 25.4% 24.5% 33.4% 63.0% 50.2% 48.7% 39.1% 32.2% 45.1% 45.3% 58.9% 21.5% 30.0% 30.4% 35.4% Race White non-Hispanic only Black and Black multiracial Other 20.9% 14.4% 18.8% 3.5% 7.5% 4.6% 26.0% 17.0% 18.8% 49.6% 61.1% 57.8% 46.9% 31.3% 37.6% 29.4% 24.5% 23.4% Mobility Impairment Absent Present 18.8% 7.8% 4.9% 2.5% 22.0% 36.0% 54.2% 53.6% 40.9% 43.8% 26.9% 38.5% * Includes home-based workers and telecommute-eligible workers. A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child 515 years old. 179 Table 79. Employment flexibility by caregiver status. Flexibility Type Totals Flexible Schedule Flexible Location* Schedule Only Location* Schedule & No Only Location Flexibility With or Without Location With or Without Schedule Male noncaregiver 18.9% 5.0% 22.4% 53.8% 41.3% 27.3% Female noncaregiver Male caregiver 17.8% 21.6% 5.6% 3.4% 19.9% 26.0% 56.7% 49.1% 37.7% 47.6% 25.5% 29.4% Female caregiver 16.2% 5.1% 21.7% 57.1% 37.9% 26.8% By Age of Youngest Child Youngest, ages 04 18.7% 4.1% 23.5% 53.7% 42.2% 27.6% Youngest, ages 515 19.5% 4.1% 24.5% 51.9% 44.0% 28.7% Household Type Male co-caregiver 21.6% 3.4% 26.1% 48.8% 47.7% 29.6% Female co-caregiver 16.3% 4.7% 22.9% 56.1% 39.2% 27.6% Male single caregiver 21.6% 0.0% 20.0% 58.4% 41.6% 20.0% Female single caregiver 15.7% 6.5% 17.1% 60.7% 32.8% 23.6% Household Type and Age of Youngest Child Male co-caregiver, youngest ages 04 21.4% 3.4% 24.8% 50.5% 46.1% 28.2% Female co-caregiver, youngest ages 04 14.2% 5.5% 23.3% 57.1% 37.4% 28.8% Female single caregiver, youngest ages 04 22.1% 2.7% 12.0% 63.2% 34.1% 14.6% Male co-caregiver, youngest ages 515 21.8% 3.5% 27.5% 47.2% 49.3% 31.0% Female co-caregiver, youngest ages 515 18.3% 4.0% 22.5% 55.3% 40.7% 26.4% Male single caregiver, youngest ages 515 23.1% 0.0% 16.0% 60.9% 39.1% 16.0% Female single caregiver, youngest ages 515 12.5% 8.4% 19.7% 59.4% 32.1% 28.1% * Includes home-based workers and telecommute-eligible workers. A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child 515 years old. Male single caregivers with youngest child ages 04 are omitted due to small sample size (5 individuals). 180 Sixty percent of college-educated workers have a flexible schedule, location, or both (table 80), whereas only 34 percent of workers without a 4-year college degree have location or schedule flexibility (table 81). In addition to the direct difference in flexibility between workers with and without a college degree, the relationship between some other demographic / employment characteristics and flexibility differs between those two groups. For example, for low-education workers, sales and service jobs are among the least likely to have flexibility (along with blue-collar jobs), whereas for high-education workers, sales and service jobs are among the most likely to have flexibility (along with professional, managerial, and technical jobs). Similarly, the gender gap in flexibility is much larger among college-educated workers. Among high-education workers, there is a 14 percentage-point difference between genders: 67 percent of college-educated male workers have some form of flexibility, versus 53 percent of similarly educated women (table 82). Approximately one third of low-education workers have some kind of employment flexibility, regardless of gender (table 83). A number of demographic and employment characteristics are correlated with each other quite apart from flexibility considerations, and thus, in addition to examining the relationship to flexibility of one or two other variables at a time, it is desirable to be able to assess the relationship of one variable to flexibility while simultaneously controlling for as many other variables as possible. In the next two subsections, we use logistic regression to isolate the effects of different demographic variables and job characteristics on work schedule and location flexibility, respectively. 181 Table 80. Employment flexibility for workers with a 4-year college degree by location and job characteristics. Flexibility Type Schedule Only Location* Only All workers with a 4-year college degree 20.6% 4.7% MPO Tier 1. Atlanta MPO 2. Medium MPOs 3. Small MPOs 4. Non-MPO counties 19.8% 23.4% 26.4% 17.5% 5.3% 4.3% 3.8% 1.7% Worker Type Full-time Part-time 21.6% 20.4% 4.6% 4.7% Occupational Category Sales or service Clerical or administrative support Blue collar 14.7% 23.3% 19.3% 5.2% 7.5% 1.6% Professional, managerial, or technical 21.8% 4.5% Annual Household Income <$35,000 $35,000 to $49,999 $50,000 to $74,999 $75,000 to $99,999 $100,000+ 17.1% 25.2% 15.4% 22.8% 21.6% 6.3% 3.0% 4.6% 4.4% 4.7% * Includes home-based workers and telecommute-eligible workers. Blue collar refers to manufacturing, construction, maintenance, or farming. Schedule & Location 35.0% No Flexibility 39.7% Totals Flexible Schedule Flexible Location* With or Without Location With or Without Schedule 55.6% 39.7% 40.2% 21.5% 28.6% 23.0% 34.7% 50.7% 41.1% 57.8% 60.0% 45.0% 55.0% 40.5% 45.5% 25.8% 32.5% 24.6% 34.9% 35.0% 38.9% 39.8% 56.6% 55.4% 39.5% 39.7% 41.2% 13.3% 20.8% 36.7% 38.9% 55.9% 58.3% 36.9% 55.9% 36.6% 40.0% 58.6% 46.4% 20.8% 22.4% 41.2% 19.6% 18.2% 28.0% 31.6% 44.7% 57.0% 53.6% 52.1% 41.2% 29.0% 36.6% 43.4% 43.3% 54.4% 66.3% 25.9% 21.2% 32.6% 36.0% 49.4% 182 Table 81. Employment flexibility for workers without a 4-year college degree by location and job characteristics. Flexibility Type Schedule Location* Only Only All workers without a 4-year college degree 17.0% 5.0% MPO Tier 1. Atlanta MPO 18.3% 6.1% 2. Medium MPOs 14.3% 3.7% 3. Small MPOs 15.0% 2.8% 4. Non-MPO counties 17.3% 4.7% Worker Type Full-time 14.6% 4.3% Part-time 17.8% 5.2% Occupational Category Sales or service 12.5% 6.4% Clerical or administrative support Blue collar 20.4% 17.6% 3.1% 4.0% Professional, managerial, or technical 21.2% 4.8% Annual Household Income <$35,000 14.5% 6.1% $35,000 to $49,999 11.5% 3.5% $50,000 to $74,999 20.7% 4.9% $75,000 to $99,999 23.6% 4.1% $100,000+ 20.5% 4.8% * Includes home-based workers and telecommute-eligible workers. Blue collar refers to manufacturing, construction, maintenance, or farming. Schedule & Location 12.1% 12.8% 10.4% 9.1% 13.5% 12.5% 11.9% 11.1% 11.9% 7.8% 17.9% 8.3% 9.5% 11.0% 13.9% 21.3% No Flexibility 65.9% 62.8% 71.7% 73.1% 64.5% 68.6% 65.1% 70.0% 64.6% 70.6% 56.0% 71.1% 75.4% 63.4% 58.4% 53.4% Totals Flexible Schedule With or Without 29.1% 31.1% 24.6% 24.0% 30.8% 27.1% 29.7% 23.6% 32.3% 25.4% 39.2% 22.8% 21.1% 31.7% 37.5% 41.8% Flexible Location* With or Without 17.1% 18.9% 14.1% 11.9% 18.2% 16.8% 17.2% 17.5% 15.0% 11.9% 22.8% 14.4% 13.0% 15.8% 18.1% 26.1% 183 Table 82. Employment flexibility among workers with a 4-year college degree by sex, caregiver status, age, race, and mobility impairment. Flexibility Type Totals Schedule Location* Schedule & No Flexible Schedule Flexible Location* Only Only Location Flexibility With or Without With or Without All workers with a 4-year college degree 20.6% 4.7% 35.0% 39.7% 55.6% 39.7% Sex Male 23.1% 4.0% 39.9% 33.0% 63.0% 43.9% Female 18.0% 5.5% 29.9% 46.6% 47.9% 35.4% Caregiver Status Noncaregiver 20.4% 5.0% 33.6% 40.9% 54.1% 38.7% Caregiver, youngest child ages 015 20.9% 4.2% 37.0% 37.9% 57.9% 41.2% Age Cohort Millennial and Gen Z (1836) 17.3% 5.5% 32.5% 44.7% 49.8% 37.9% Generation X (3752) 21.5% 3.9% 35.9% 38.6% 57.5% 39.9% Pre-retirement age Baby Boomer (5364) 23.6% 5.5% 34.9% 36.0% 58.5% 40.4% Retirement age (65+) 23.4% 2.8% 45.1% 28.7% 68.5% 47.9% Race White non-Hispanic only 21.0% 3.9% 36.9% 38.1% 58.0% 40.9% Black and Black multiracial 16.5% 6.1% 30.7% 46.6% 47.3% 36.9% Other 26.7% 5.9% 33.6% 33.9% 60.2% 39.4% Mobility Impairment Absent 20.8% 4.8% 34.7% 39.8% 55.4% 39.4% Present 10.0% 1.5% 55.5% 33.0% 65.5% 57.0% * Includes home-based workers and telecommute-eligible workers. A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. 184 Table 83. Employment flexibility among workers without a 4-year college degree by sex, caregiver status, age, race, and mobility impairment. Flexibility Type Totals Schedule Location* Schedule & No Flexible Schedule Flexible Location* All workers without a 4-year college degree Only 17.0% Only 5.0% Location 12.1% Flexibility 65.9% With or Without 29.1% With or Without 17.1% Sex Male Female Caregiver Status 17.5% 16.5% 4.7% 5.4% 12.0% 12.2% 65.8% 66.0% 29.5% 28.6% 16.7% 17.5% Noncaregiver Caregiver, youngest child ages 015 16.8% 17.5% 5.5% 4.1% 12.0% 12.2% 65.8% 66.2% 28.8% 29.7% 17.5% 16.3% Age Cohort Millennial and Gen Z (1836) Generation X (3752) Pre-retirement age Baby Boomer (5364) Retirement age (65+) 14.3% 17.8% 18.9% 27.1% 4.5% 5.4% 6.3% 1.3% 6.2% 14.9% 16.5% 24.3% 75.0% 62.0% 58.4% 47.3% 20.5% 32.6% 35.4% 51.4% 10.7% 20.2% 22.8% 25.6% Race White non-Hispanic only Black and Black multiracial Other 20.9% 13.2% 12.6% 3.0% 8.3% 3.8% 15.1% 9.3% 8.0% 61.0% 69.2% 75.6% 35.9% 22.4% 20.6% 18.1% 17.6% 11.8% Mobility Impairment Absent Present 17.2% 6.4% 5.0% 3.3% 11.8% 22.7% 65.9% 67.6% 29.1% 29.1% 16.9% 26.0% * Includes home-based workers and telecommute-eligible workers. A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. 185 Logistic Regression Analysis of Eligibility for Flexible Work Schedule Some of the demographic differences discussed in the previous section (Descriptive Statistics on Access to Flexible Schedule and Work Location) are likely attributable to intergroup differences in education attainment and job type. However, using logistic regression analysis, the research team found that even after controlling for relevant characteristics of the jobs themselves, eligibility for a flexible work schedule is associated with characteristics unrelated to the employee's qualifications or work function (table 84). Being female and nonwhite are both associated with a statistically significant reduction in the likelihood of having a flexible work schedule among both home-based and nonhome-based workers. Among nonhome-based workers, being a caregiver for a child is associated with an increased likelihood of having a flexible schedule, but this effect only applies to male caregivers once the children reach school age. The likelihood of having schedule flexibility increases with age. Where a worker lives impacts the likelihood that she will have a flexible schedule, but only for nonhome-based workers. For those who work primarily from home, MPO tier does not have a significant effect. An additional set of models incorporating interaction terms with education level was estimated. The models are not shown because they did not represent an improvement to the overall goodness of fit. However, a few key insights are worth noting. In line with patterns in the descriptive statistics in the previous section, the gender gap in schedule flexibility is greater among college-educated workers as compared to the gap between men and women without a four-year college degree. Conversely, race-based inequality in schedule flexibility is smaller among college-educated workers than for workers without a four-year degree. 186 Table 84. Logistic regressions: Eligibility for flexible work schedule among Georgia workers. All Workers Nonhome-based Home-based Workers Only Workers Only Female Caregiver, youngest child 04 years old -0.3593 *** 0.2418 ** -0.3420 *** 0.2446 ** -0.4727 ** 0.0464 Caregiver, youngest child 515 years old 0.3126 *** 0.3011 *** 0.4706 Female x caregiver, youngest 04 years old -0.0380 -0.1008 0.5872 Female x caregiver, youngest 515 years old -0.3925 *** -0.4863 *** 0.3251 Education level (reference: high school or less) Some college or associate degree 0.2901 *** 0.2274 *** 0.7466 *** Bachelor's (4-year) degree 0.7758 *** 0.7343 *** 1.0166 *** Postgraduate or professional degree 0.8919 *** 0.8300 *** 1.4035 *** Race (reference: white non-Hispanic) Black and Black multiracial -0.5917 *** -0.5136 *** -1.1412 *** Other -0.1898 ** -0.1274 -0.7870 *** MPO Tier (reference: Atlanta) 2. Medium MPOs -0.3350 *** -0.3611 *** -0.0684 3. Small MPOs -0.3247 *** -0.3253 *** -0.4075 4. Non-MPO -0.3421 *** -0.3653 *** -0.1077 Age 0.0209 *** 0.0200 *** 0.0272 *** Mobility impairment 0.0025 -0.0968 0.5338 Occupational Category (reference: sales and service) Clerical or administrative support Blue collar 0.0275 -0.1324 0.0634 -0.1506 -0.2739 0.1720 Professional, managerial, or technical 0.3304 *** 0.3705 *** 0.0327 Full-time worker 0.0498 0.0942 -0.1484 Home-based worker 1.9913 *** Constant -1.6710 *** -1.6515 *** 0.0482 Model Indicators Number of cases, N 7960 6890 1070 Final log likelihood, LL() -4692.4 -4239.9 -432.7 Market share log likelihood, LL(MS) McFadden's pseudo-R2: 1-LL()/LL(MS) -5463.5 0.141 -4582.6 0.075 -494.2 0.125 Coefficients shown. *** denotes significance for =.01, ** for =.05, and * for =.10. A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. Blue collar refers to manufacturing, construction, maintenance, or farming. 187 Logistic Regression Analysis of Eligibility for Flexible Work Location So far, this report has treated location flexibility as a binary variable. However, flexible work location is actually subdivided between home-based workers, who usually work from home, and telecommute-eligible workers, who work outside the home but have the option to telework sometimes instead of going to their usual work location. The remaining workers, telecommuteineligible workers, are nonhome-based workers without the opportunity to telecommute. Table 85 shows the distribution of telecommute eligibility and home-based work by various demographic characteristics. We constructed a multinomial logistic regression to model the likelihood of being a telecommute-eligible or home-based worker, as compared to the base category of telecommute-ineligible worker (table 86). As with schedule flexibility, being female reduces the likelihood of being telecommute-eligible and home-based. Being a caregiver influences telecommute eligibility and being a home-based worker differently. Being a caregiver of children ages 515 is associated with an increased likelihood of being telecommute-eligible, but only for men. However, for home-based work, female caregivers are more likely to be home-based; there is no effect for male caregivers. People of color are less likely to be telecommute-eligible; race does not have a significant effect on home-based work. Having a mobility impairment does not have a significant effect on the likelihood of being telecommute-eligible, but it does increase the likelihood of being a home-based worker. This suggests that many disabled workers are selecting home-based work rather than being offered accommodations by employers. Workers in Atlanta are more likely to have the option to 188 telecommute and to be home-based workers. While being a full-time worker did not have a significant effect on schedule flexibility, it is associated with an increased likelihood of being telecommute-eligible and a decreased likelihood of being a home-based worker. 189 Table 85. Teleworking status of Georgia workers (descriptive statistics). Home-based All workers 13.7% MPO Tier 1. Atlanta MPO 16.0% 2. Medium MPOs 8.8% 3. Small MPOs 8.8% 4. Non-MPO 13.0% Sex Male 12.2% Female 15.4% Age Cohort Millennial and Gen Z (1836) 10.3% Generation X (3752) 14.0% Pre-retirement age Baby Boomer (5364) 17.2% Retirement age (65+) 22.6% Race White non-Hispanic only 13.7% Black and Black multiracial 13.6% Other 13.4% Medical Condition Absent 13.5% Present 23.7% Occupation Category Sales or service 16.4% Clerical or administrative support 9.8% Blue collar 6.9% Professional, managerial, or technical 15.4% Education HS or less 9.3% Some college or associate degree 10.6% Bachelor's degree or higher 18.2% Work Schedule Part-time 19.1% Full-time 12.3% Annual Household Income <$35,000 11.6% $35,000 to $49,999 10.7% $50,000 to $74,999 12.9% $75,000 to $99,999 9.5% $100,000+ 18.6% Schedule Flexibility (Flextime) Flexible 25.7% Not flexible 5.2% Blue collar refers to manufacturing, construction, maintenance, or farming. Telecommute-eligible 13.5% 17.0% 10.3% 9.5% 6.9% 15.9% 10.7% 11.2% 16.1% 13.1% 12.9% 15.8% 10.8% 10.0% 13.5% 15.0% 9.2% 7.4% 6.3% 20.1% 5.3% 8.4% 21.5% 5.3% 15.5% 5.0% 4.9% 9.5% 17.9% 23.9% 28.7% 3.0% Telecommuteineligible 72.8% 67.1% 80.8% 81.7% 80.2% 71.9% 74.0% 78.5% 69.9% 69.6% 64.5% 70.5% 75.5% 76.6% 73.1% 61.3% 74.4% 82.8% 86.8% 64.6% 85.4% 81.0% 60.2% 75.7% 72.2% 83.4% 84.4% 77.6% 72.6% 57.5% 45.6% 91.8% 190 Table 86. Multinomial logistic regression: Teleworking status of Georgia workers. Base Category: Telecommute-ineligible workers Telecommute-eligible Worker Odds Ratio P-value Female Caregiver, youngest child 04 years old 0.746 1.133 0.001 *** 0.345 Caregiver, youngest child 515 years old 1.379 0.008 *** Female x caregiver, youngest 04 0.969 0.878 Female x caregiver, youngest 515 0.625 0.014 ** Education level (reference: high school or less) Some college or associate degree 1.531 0.002 *** Bachelor's (4-year) degree 3.394 0.000 *** Postgraduate or professional degree 3.437 0.000 *** Race (reference: white non-Hispanic) Black and Black multiracial 0.690 0.000 *** Other 0.656 0.002 *** MPO Tier (reference: Atlanta) 2. Medium MPOs 0.553 0.000 *** 3. Small MPOs 0.469 0.000 *** 4. Non-MPO 0.492 0.000 *** Age 1.009 0.002 *** Mobility impairment 1.478 0.175 Occupational Category (reference: sales and service) Clerical or administrative support Blue collar Professional, managerial, or technical 0.731 0.630 1.181 0.053 * 0.002 *** 0.088 * Full-time worker Constant 1.797 0.058 0.000 *** 0.000 *** Home-based Worker Odds Ratio P-Value 0.768 0.002 *** 0.754 0.883 2.302 1.527 0.102 0.397 0.000 *** 0.030 ** 1.408 2.780 2.506 0.003 *** 0.000 *** 0.000 *** 0.876 0.810 0.136 0.120 0.554 0.503 0.686 1.027 1.681 0.000 *** 0.000 *** 0.004 *** 0.000 *** 0.023 ** 0.461 0.502 0.656 0.492 0.112 0.000 *** 0.000 *** 0.000 *** 0.000 *** 0.000 *** Model Indicators Number of cases, N 7,972 Final log likelihood, LL() -5617.21 Market share log likelihood, LL(MS) McFadden's pseudo-R2: 1-LL()/LL(MS) -6140.95 0.085 Base category: Telecommute-ineligible workers (those who work outside the home without the option of teleworking). A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. Telecommute-eligible workers usually work outside of the home but have the option of telecommuting. Home-based workers usually work from home. Blue collar refers to manufacturing, construction, maintenance, or farming. 191 THE EFFECTS OF FLEXIBLE SCHEDULING Having examined which workers are allowed a degree of flexibility with regard to their work starting time and location in the previous section, this section discusses the effects of that flexibility on workers' behavior. Figure 20 shows weighted histograms of the time of arrival at and departure from work for workers' travel-day commutes. Workers with flexible schedules tend to both arrive at and depart from work later than those with inflexible schedules. Figure 20. Bar graphs. Commute schedules of workers with and without flexible working times. 192 The difference is most pronounced with respect to arrival times at work. The median arrival time for a worker with a flexible schedule is 8:34 a.m., 34 minutes later than the median arrival time for workers with inflexible schedules (table 87). For departure from work, the difference is just 15 minutes (5:00 p.m. vs. 4:45 p.m.). Figure 20 also shows a small secondary peak in the departure times of workers with flexible schedules during the afternoon, though it is unknown if these early departures are followed by teleworking after the worker arrives at home. Flexible work scheduling is also associated with less dispersion of arrival and departure times. The interquartile range of work arrival times for workers with inflexible schedules is 200 minutes (7:0010:20 a.m.) (not included in table). For workers with flexible schedules, it is just 125 minutes (7:409:45 a.m.), a time period that is more than an hour shorter. For departures, the difference is again somewhat smaller. Workers with inflexible schedules have an interquartile range of 180 minutes (3:006:00 p.m.), versus 150 minutes for workers with flexible schedules (3:306:00 p.m.). Table 87 also shows information about arrival and departure times disaggregated by MPO tier, occupational category, educational attainment, and worker type. Flexible scheduling is associated with wider differences in arrival times in small MPO areas and non-MPO counties. However, for departure times, the differences are larger in Atlanta and medium MPOs. Contrary to the general trend, part-time workers with flexible schedules tend to arrive at and depart from work earlier than those with inflexible schedules. 193 Table 87. Median work arrival and departure times by schedule flexibility, MPO tier, education, and employment characteristics. Median Work Arrival Time (am) Median Work Departure Time (pm) Inflexible Flexible Difference Inflexible Flexible Difference Schedule Schedule (minutes) Schedule Schedule (minutes) All workers 8:00 8:34 34 4:45 5:00 15 MPO Tier 1. Atlanta MPO 8:05 8:38 33 4:45 5:00 15 2. Medium MPOs 7:55 8:20 25 4:30 4:50 20 3. Small MPOs 8:00 9:00 60 4:55 4:45 -10 4. Non-MPO 7:40 8:28 48 5:00 5:00 0 Occupational Category Sales or service 9:06 9:10 4 5:00 5:00 0 Clerical or administrative support 7:55 8:30 35 4:41 4:30 -11 Blue collar* 7:01 8:00 59 4:05 5:00 55 Professional, managerial, or technical 7:54 8:30 36 4:55 5:00 5 Educational Attainment High school or less 8:00 8:26 26 4:30 5:00 30 Some college or associate degree 8:00 8:37 37 4:45 5:00 15 Bachelor's (4-year) degree 7:57 8:40 43 5:00 5:00 0 Postgraduate or professional degree 7:55 8:40 45 5:00 5:00 0 Worker Type Part-time 10:03 9:35 -28 4:30 3:50 -40 Full-time 7:50 8:30 40 4:50 5:00 10 * Blue collar refers to manufacturing, construction, maintenance, or farming. The difference in median arrival times between workers with and without flexible scheduling is largest among blue-collar workers (59 minutes) and smallest among sales and service workers (4 minutes). The difference in median departure times between workers with and without flexible schedules is also largest among blue-collar workers (55 minutes). Clerical and administrative workers with flexible schedules are the only occupation category to leave work earlier than their counterparts with nonflexible schedules (a difference of 11 minutes). However, as visualized in figure 21 and figure 22, while flexible scheduling is not associated with substantial changes in median arrival and departure times for sales and service jobs, it is associated with a dramatic reduction in the interquartile range. For example, the interquartile 194 range of arrival times for service jobs with inflexible schedules stretches across 6 hours (7:30 a.m.1:25 p.m.), more than double the interquartile range for service jobs with flexible schedules (8:1511:00 a.m.). However, the NHTS occupation categories are broad, and one possible interpretation of this finding is that, since jobs that allow for flexible schedules tend to be higher status, the difference in schedules may more accurately reflect a difference between higher-status service jobs (e.g., call center manager) versus lower-status jobs (e.g., retail employee). Figure 21. Histograms. Arrival times at work for service and nonservice workers by schedule flexibility. 195 Figure 22. Histograms. Departure times from work for service and nonservice workers by schedule flexibility. Table 88 shows median work arrival and departure times by schedule flexibility, disaggregated by demographic characteristics. Schedule flexibility is associated with larger differences between median arrival times and between median departure times for women and older workers. The differences between the median arrival times of flexible and inflexible workers are similar for caregivers and childless adults. However, caregivers show a larger difference in departures at the end of the work day. Flex time is associated with a larger difference in schedule among caregivers for older children (ages 515) than among those caring for younger children. 196 Figure 23 and figure 24 visually depict the variability of schedules among caregivers for young children by gender. Table 88. Median work arrival and departure times by schedule flexibility, demographic characteristics, and caregiver status. Median Work Arrival Time (am) Median Work Departure Time (pm) Inflexible Flexible Difference Inflexible Flexible Difference Schedule Schedule (minutes) Schedule Schedule (minutes) All workers 8:00 8:34 34 4:45 5:00 15 Sex Male 7:55 8:20 25 5:00 5:00 0 Female 8:00 8:55 55 4:35 5:00 25 Caregiver Status Noncaregiver 8:00 8:40 40 5:00 5:00 0 Caregiver, youngest child ages 015 7:55 8:30 35 4:30 5:00 30 Age Cohort Millennial and Gen Z (1836) 8:20 8:50 30 5:00 5:00 0 Generation X (3752) 7:45 8:30 45 4:30 5:00 30 Pre-retirement age Baby Boomer (5364) 7:50 8:30 40 4:30 5:00 30 Retirement age (65+) 8:00 9:00 60 5:00 4:15 -45 Annual Household Income <$35,000 8:00 9:00 60 4:30 5:00 30 $35,000 to $49,999 8:00 8:30 30 4:45 4:30 -15 $50,000 to $74,999 8:00 8:30 30 4:45 4:30 -15 $75,000 to $99,999 8:00 8:45 45 5:00 5:00 0 $100,000+ 7:53 8:30 37 5:00 5:00 0 Race White non-Hispanic only 7:55 8:30 35 5:00 4:55 -5 Black and Black multiracial 8:00 8:30 30 4:30 5:00 30 Other 8:15 9:00 45 4:30 5:30 60 Mobility Impairment Absent 8:00 8:35 35 4:45 5:00 15 Present 10:45 8:10 -155 3:30 3:10 -20 Caregiver Status by Gender Male noncaregiver 8:00 8:25 25 5:00 5:00 0 Male caregiver 7:46 8:17 31 4:40 5:00 20 Female noncaregiver 8:03 8:55 52 5:00 5:00 0 Female caregiver 7:58 8:53 55 4:25 4:45 20 Caregiver Status by Age of Youngest Child Youngest child 04 years 8:00 8:30 30 4:50 5:00 10 Youngest child 515 years 7:45 8:30 45 4:25 5:00 35 197 Figure 23. Histograms. Arrival times at work for caregivers by schedule flexibility and gender. 198 Figure 24. Histograms. Departure times from work for caregivers by schedule flexibility and gender. 199 Models of Work Arrival and Departure Times Linear regression was used to isolate the relationship between flexible scheduling and work arrival and departure times (table 89). All else held equal, a worker with a flexible schedule will arrive at work an average of 22 minutes later than a comparable worker with an inflexible schedule, and they will depart from work an average of 6 minutes later. Unsurprisingly, occupation category and full-time status exert a strong influence on work arrival and departure times. Once other factors are accounted for, MPO tier is not associated with significant differences in either departure or arrival time. As measured by R2, the goodness of fit is not high for either model (especially the departure time model), signifying that arrival and departure times are largely determined by factors unavailable in the data. The researchers also modeled commute duration as a function of schedule flexibility and found no practically significant effect.64 64 When the model was estimated, the flexibility coefficient was small in magnitude, statistically significant, but with a counterintuitive positive sign (indicating that, after controlling for commute distance and other factors, those with schedule flexibility had commutes that were 0.96 minutes longer on average). This result likely reflects the opposite direction of causality (if the commute is longer duration--which, after controlling for distance and mode, means that it is more congested--the respondent is more likely to seek schedule flexibility). In consideration of all these factors the research team concludes that flexibility is not a good predictor of commute duration, perhaps because of counteracting effects reflecting both directions of causality. 200 Table 89. Linear regression: Work arrival and departure times for travel day commutes. Arrival Time at Work Coefficient P-value Flexible schedule 22.017 0.000 *** Log commute distance (miles) -18.129 0.000 *** Work Journey Mode (reference: private auto) Nonmotorized (pedestrian or cyclist) -23.472 0.207 Public transit or other bus/rail -10.483 0.607 Multimodal and other 52.761 0.028 ** Full-time worker -81.372 0.000 *** Occupational Category (reference: sales and service) Clerical or administrative support Blue collar -49.854 -49.265 0.000 *** 0.000 *** Professional, managerial, or technical -39.647 0.000 *** Education level (reference: high school or less) Some college or associate degree 7.599 0.364 Bachelor's (4-year) degree 2.334 0.787 Postgraduate or professional degree 4.074 0.653 MPO Tier (reference: Atlanta) 2. Medium MPOs -7.804 0.197 3. Small MPOs -2.388 0.739 4. Non-MPO -9.447 0.362 Race (reference: white non-Hispanic) Black and Black multiracial 21.122 0.004 *** Other 8.301 0.346 Age Age2 -5.692 0.053 0.000 *** 0.001 *** Female 5.017 0.361 Caregiver, youngest child 04 years old -8.206 0.328 Caregiver, youngest child 515 years old -2.367 0.749 Day of week (reference: Sunday) Monday -104.764 0.000 *** Tuesday -107.431 0.000 *** W ednesday -109.528 0.000 *** Thursday -115.699 0.000 *** Friday -105.665 0.000 *** Saturday -70.653 0.023 ** Constant 912.744 0.000 *** Departure Time from Work Coefficient P-value 5.915 0.000 *** 8.429 0.000 *** 4.270 2.922 -12.146 27.561 0.207 0.607 0.028 ** 0.000 *** -11.491 -49.278 -18.382 0.000 *** 0.000 *** 0.000 *** 2.835 13.220 23.422 0.364 0.787 0.653 -11.103 -8.073 -16.550 0.197 0.739 0.362 -20.785 17.582 2.162 -0.030 -4.937 -34.201 -23.706 0.004 *** 0.346 0.000 *** 0.001 *** 0.361 0.328 0.749 -5.575 -18.079 -18.085 -6.525 -32.719 -4.919 947.020 0.000 *** 0.000 *** 0.000 *** 0.000 *** 0.000 *** 0.023 ** 0.000 *** R2 0.089 Number of cases, N 5,061 Departure and arrival times are given in minutes past midnight. Coefficients are in minutes. Blue collar refers to manufacturing, construction, maintenance, or farming. 0.024 5,091 201 FREQUENCY OF TELEWORKING How often do workers who have the option of working from home or a third place do so? This report approaches this question using two types of data. First, it examines the frequency of telecommuting among telecommute-eligible workers. Second, it looks at travel-day teleworking among Georgia workers, regardless of their stated eligibility for telework. Telecommuting Frequency among Eligible Workers When workers have the ability to telecommute, the overwhelming majority choose to do so; 78 percent of telecommute-eligible workers reported telecommuting at least once in the past 30 days (table 90). Telecommuting is least common in small MPO areas, where nevertheless 62 percent of eligible workers telecommuted. Workers in rural non-MPO counties are no more likely than average to telecommute. However, those who do telecommute do so for a higher number of days; the average telecommuter from a non-MPO area telecommutes 9.3 days per month, compared to the state average of 6.7 days. Preretirement age Baby Boomers and Generation X workers are the most likely to telework; the Baby Boomers do so for more days on average. While white workers are more likely to telecommute than Black workers, among those who do telecommute, Black workers do so for more days on average. Workers with disabilities are slightly less likely than average to take advantage of the ability to telecommute, but among those who did so, the mean days telecommuted for disabled workers (11.7) is the highest of any subgroup studied. 202 For a finer-grained breakdown, table 91 presents the distribution across telecommuting frequency categories for the same variables as in table 90. There are not strong differences in the percentage of workers in different occupation categories who telecommute at least once (table 92, table 93), but among those who telecommute at least once, blue-collar workers telecommute the most days.65 Part-time workers and low-income workers are less likely to telecommute but do so for more days. 65 One possible explanation for this phenomenon is that there is a larger difference in job function between telecommute-eligible and -ineligible blue-collar jobs than in other industries. Some portion of this finding may also reflect a misclassification as telecommuters of self-employed workers such as movers, plumbers, and other professions that are based primarily on house calls. 203 Table 90. Average number of days telecommuted in the past 30 days among eligible workers by MPO tier, sex, caregiver status, age, race, and medical condition. Percent of Eligible Workers who Telecommuted Mean Days, All Eligible Workers Mean Days, Workers who Telecommuted All workers 77.5% 5.2 6.7 MPO Tier 1. Atlanta MPO 79.9% 5.0 6.2 2. Medium MPOs 72.8% 5.7 7.8 3. Small MPOs 62.3% 4.3 7.0 4. Non-MPO 76.1% 7.1 9.3 Sex Male 76.7% 5.3 6.9 Female Caregiver Status 79.0% 5.1 6.4 Noncaregiver 75.0% 4.6 6.1 Caregiver, youngest child ages 015 81.1% 6.1 7.5 Age Cohort Millennial and Gen Z (1836) 73.5% 4.7 6.5 Generation X (3752) 79.7% 4.9 6.1 Pre-retirement age Baby Boomer (5364) 80.8% 6.7 8.3 Retirement age (65+) 70.6% 5.1 7.2 Race White non-Hispanic only 80.1% 5.1 6.3 Black and Black multiracial 71.0% 5.8 8.1 Other 76.6% 4.6 6.1 Medical Condition Absent 77.5% 5.1 6.6 Present 74.0% 8.6 11.7 Based on telecommute-eligible workers. Excludes home-based workers. A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. 204 Table 91. Telecommuting 30-day frequency among eligible workers by MPO tier, sex, caregiver status, age, race, and medical condition. Did Not Telecommute (0 Days) Occasional (14 Days) Moderate (510 Days) Frequent (11+ Days) All workers 22.5% 37.7% 27.9% 11.9% MPO Tier 1. Atlanta MPO 20.1% 39.6% 31.1% 9.3% 2. Medium MPOs 27.2% 33.1% 21.2% 18.5% 3. Small MPOs 37.7% 34.4% 17.3% 10.5% 4. Non-MPO 23.9% 31.0% 19.0% 26.1% Sex Male 23.3% 38.4% 25.9% 12.4% Female Caregiver Status 21.0% 36.5% 31.4% 11.1% Noncaregiver 25.0% 37.8% 27.9% 9.3% Caregiver, youngest child ages 015 18.9% 37.5% 28.0% 15.7% Age Cohort Millennial and Gen Z (1836) 26.5% 36.7% 26.7% 10.1% Generation X (3752) 20.3% 39.7% 30.8% 9.2% Pre-retirement age Baby Boomer (5364) 19.2% 37.3% 22.5% 21.0% Retirement age (65+) 29.4% 27.6% 31.4% 11.5% Race White non-Hispanic only 19.9% 39.8% 29.3% 11.0% Black and Black multiracial 29.0% 30.9% 25.4% 14.7% Other 23.4% 40.7% 24.9% 11.0% Medical Condition Absent 22.5% 38.3% 27.6% 11.6% Present 26.0% 8.8% 49.2% 16.0% Based on telecommute-eligible workers. Excludes home-based workers. A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. 205 Table 92. Average days telecommuted in the past 30 days among eligible workers by employment characteristics. Percent of Eligible Workers who Telecommuted Mean Days, All Eligible Workers All workers 77.5% 5.2 Occupation Category Sales or service 75.6% 7.4 Clerical or administrative support 74.6% 3.9 Blue collar 72.5% 8.0 Professional, managerial, or technical 78.8% 4.4 Education HS or less 68.9% 7.3 Some college or associate degree 72.2% 6.3 Bachelor's degree or higher 80.1% 4.6 Work Schedule Part-time 66.8% 6.9 Full-time 78.4% 5.1 Annual Household Income <$35,000 64.3% 6.7 $35,000 to $49,999 69.4% 5.2 $50,000 to $74,999 77.0% 5.5 $75,000 to $99,999 78.3% 4.6 $100,000+ 79.6% 4.7 Schedule Flexibility (Flextime) Not flexible 78.1% 5.2 Flexible 73.9% 4.9 Based on telecommute-eligible workers. Excludes home-based workers. Blue collar refers to manufacturing, construction, maintenance, or farming. Mean Days, Workers who Telecommuted 6.7 9.8 5.2 11.1 5.5 10.7 8.7 5.7 10.3 6.4 10.4 7.4 7.2 5.9 5.9 6.6 6.7 206 Table 93. Telecommuting 30-day frequency among eligible workers by employment characteristics. Did Not Telecommute (0 Days) Occasional (14 Days) All workers 22.5% 37.7% Occupation Category Sales or service 24.4% 28.2% Clerical or administrative support Blue collar 25.4% 27.5% 41.2% 30.4% Professional, managerial, or technical 21.2% 40.8% Education HS or less 31.1% 26.1% Some college or associate degree 27.8% 29.6% Bachelor's degree or higher 19.9% 41.5% Work Schedule Part-time 33.2% 19.1% Full-time 21.6% 39.2% Annual Household Income <$35,000 35.7% 22.5% $35,000 to $49,999 30.6% 23.8% $50,000 to $74,999 23.0% 30.4% $75,000 to $99,999 21.7% 45.4% $100,000+ 20.4% 41.5% Schedule Flexibility (Flextime) Not flexible 21.9% 38.3% Flexible 26.1% 33.9% Based on telecommute-eligible workers. Excludes home-based workers. Blue collar refers to manufacturing, construction, maintenance, or farming. Moderate (510 Days) 27.9% 23.1% 25.6% 14.9% 31.0% 18.1% 25.5% 29.9% 22.5% 28.4% 21.8% 35.9% 31.2% 22.5% 29.4% 27.7% 29.3% Frequent (11+ Days) 11.9% 24.4% 7.7% 27.2% 7.0% 24.8% 17.1% 8.7% 25.3% 10.8% 19.9% 9.7% 15.4% 10.5% 8.7% 12.1% 10.7% The ability to telecommute can be of particular benefit to caregivers. As shown in table 94 and table 95, when caregivers are eligible for telecommuting, they are more likely than noncaregivers to take advantage of the option. The difference between caregivers and noncaregivers is greater for women than for men. Caregivers for young children (ages 04) are more likely to use the option to telecommute than caregivers for older children. 207 Table 94. Average days telecommuted in the past 30 days among eligible workers by caregiver status. Percent of Eligible Workers who Telecommuted Mean Days, All Eligible Workers Mean Days, Workers who Telecommuted All workers Caregiver Status 77.5% 5.2 6.7 Noncaregiver 75.0% 4.6 6.1 Caregiver, youngest child ages 015 years 81.1% 6.1 7.5 Caregiver Status by Sex Male noncaregiver 74.3% 4.5 6.0 Female noncaregiver 79.7% 6.2 7.8 Male caregiver 76.1% 4.7 6.2 Female caregiver 84.5% 5.9 6.9 By Age of Youngest Child Youngest ages 04 years 84.2% 6.4 7.6 Youngest ages 515 years 78.4% 5.9 7.5 Household Type Male co-caregiver 79.8% 6.3 7.9 Female co-caregiver 86.2% 5.9 6.8 Male single caregiver 69.5% 2.4 3.5 Female single caregiver 77.3% 5.8 7.5 Based on telecommute-eligible workers. Excludes home-based workers. A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. 208 Table 95. Telecommuting 30-day frequency among eligible workers by caregiver status. Did Not Telecommute (0 Days) Occasional (14 Days) Moderate (510 Days) Frequent (11+ Days) All workers Caregiver Status 22.5% 37.7% 27.9% 11.9% Noncaregiver 25.0% 37.8% 27.9% 9.3% Caregiver, youngest child ages 015 years 18.9% 37.5% 28.0% 15.7% Caregiver Status by Sex Male noncaregiver 25.7% 38.2% 26.8% 9.3% Female noncaregiver 20.3% 38.6% 24.9% 16.2% Male caregiver 23.9% 37.3% 29.4% 9.3% Female caregiver 15.5% 34.8% 35.3% 14.5% By Age of Youngest Child Youngest ages 04 years 15.8% 42.1% 26.8% 15.2% Youngest ages 515 years 21.6% 33.4% 29.0% 16.0% Household Type Male co-caregiver 20.2% 38.2% 25.2% 16.5% Female co-caregiver 13.8% 36.9% 36.0% 13.3% Male single caregiver 30.5% 63.2% 6.3% 0.0% Female single caregiver 22.7% 25.8% 32.3% 19.2% Based on telecommute-eligible workers. Excludes home-based workers. A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. Telecommuting Frequency and Distance to Work To analyze the relationship between distance to work and telecommuting frequency, the researchers began by examining key percentiles for distance to work by MPO tier (table 96). Then, after trying several alternate specifications, they divided respondents into five categories based on round-number distances nearest to these key percentiles. 209 Table 96. Distance to work in miles by MPO tier. 25th Percentile All Workers 5.1 MPO Tier 1. Atlanta MPO 5.9 2. Medium MPOs 4.3 3. Small MPOs 3.9 4. Non-MPO 4.5 Excluding Atlanta (Tiers 24) 4.3 50th Percentile 11.4 12.8 9.3 7.9 12.9 9.5 75th Percentile 21.1 22.3 15.7 13.9 25.2 18.7 95th Percentile 46.3 42.0 46.7 48.0 75.2 60.7 Unsurprisingly, workers who live the farthest away are the most likely to telecommute and do so for more days (table 97 and table 98). Workers who live more than 45 miles from their workplace telecommute an average of 10.3 days, versus the state average of 6.7 days. Somewhat more surprisingly, the group who telecommutes the second-highest number of days is workers who live the closest to work (5 miles or less). This pattern holds for counties in MPOs of all sizes. The sole exception to this pattern is non-MPO counties, where in general workers who live more than 10 miles from work telecommute more often than workers who live 10 miles or fewer from their workplace. 210 Table 97. Average days telecommuted in the past 30 days among eligible workers by distance to work and MPO tier. Percent of Eligible Workers who Telecommuted All Workers 77.5% By Distance to Work 05 miles 5.110 mi 10.120 mi 20.145 mi >45 mi 72.2% 70.5% 82.2% 80.0% 86.9% Tier 1: Atlanta MPO Only 05 miles 5.110 mi 10.120 mi 20.145 mi >45 mi 74.7% 72.2% 83.2% 84.0% 79.5% Tier 2: Medium MPOs Only 05 miles 5.110 mi 10.120 mi 20.145 mi >45 mi 58.2% 80.7% 78.5% 77.1% 91.6% Tier 3: Small MPOs Only 05 miles 5.110 mi 10.120 mi 20.145 mi >45 mi 72.1% 52.2% 75.6% 32.3% 85.2% Tier 4: Non-MPO Counties 05 miles 5.110 mi 10.120 mi 20.145 mi >45 mi 84.1% 68.0% 71.9% 66.9% 98.3% All Workers Excluding Atlanta (Tiers 24) 05 miles 5.110 mi 10.120 mi 20.145 mi >45 mi 68.1% 67.5% 76.6% 62.3% 93.2% Mean Days, All Eligible Workers 5.2 5.8 3.9 4.4 5.6 8.9 6.0 4.0 4.3 5.3 7.5 5.0 5.2 5.3 6.1 9.7 6.0 3.1 3.1 1.9 10.0 5.7 2.6 7.3 9.3 10.8 5.4 3.8 5.1 7.2 10.1 Mean Days, Workers who Telecommuted 6.7 8.0 5.6 5.3 7.0 10.3 8.1 5.5 5.1 6.3 9.5 8.6 6.4 6.7 7.9 10.6 8.3 6.0 4.2 6.0 11.7 6.8 3.9 10.1 13.9 11.0 8.0 5.6 6.7 11.5 10.9 211 Table 98. Telecommuting 30-day frequency among eligible workers by distance to work and MPO tier. Did Not Telecommute (0 Days) All workers 22.5% By Distance to Work 05 miles 27.8% 5.110 mi 29.5% 10.120 mi 17.8% 20.145 mi 20.0% >45 mi 13.1% Tier 1: Atlanta MPO Only 05 miles 25.3% 5.110 mi 27.8% 10.120 mi 16.8% 20.145 mi 16.0% >45 mi 20.5% Tier 2: Medium MPOs Only 05 miles 41.8% 5.110 mi 19.3% 10.120 mi 21.5% 20.145 mi 22.9% >45 mi 8.4% Tier 3: Small MPOs Only 05 miles 27.9% 5.110 mi 47.8% 10.120 mi 24.4% 20.145 mi 67.7% >45 mi 14.8% Tier 4: Non-MPO Counties 05 miles 15.9% 5.110 mi 32.0% 10.120 mi 28.1% 20.145 mi 33.1% >45 mi 1.7% All Workers Excluding Atlanta (Tiers 24) 05 miles 31.9% 5.110 mi 32.5% 10.120 mi 23.4% 20.145 mi 37.7% >45 mi 6.8% Occasional (14 Days) 37.7% 31.7% 32.0% 48.4% 38.1% 22.7% 29.5% 31.6% 49.5% 41.3% 23.4% 20.3% 40.0% 45.7% 40.9% 31.5% 46.0% 26.5% 51.7% 7.0% 18.3% 54.6% 30.1% 20.0% 23.2% 12.1% 35.4% 32.7% 42.6% 23.8% 22.2% Moderate (510 Days) 27.9% 24.8% 32.4% 27.1% 28.8% 28.6% 29.8% 35.3% 28.8% 32.7% 28.3% 23.1% 28.1% 17.2% 19.2% 11.9% 8.6% 18.9% 18.5% 23.2% 42.9% 10.8% 37.9% 18.4% 4.2% 44.6% 16.3% 27.4% 17.7% 11.2% 28.9% Frequent (11+ Days) 11.9% 15.7% 6.1% 6.7% 13.2% 35.5% 15.3% 5.4% 4.9% 10.0% 27.8% 14.8% 12.7% 15.5% 17.0% 48.2% 17.5% 6.9% 5.3% 2.1% 24.0% 18.7% 0.0% 33.5% 39.4% 41.7% 16.4% 7.4% 16.3% 27.3% 42.1% 212 Travel Day Telecommuting In addition to looking at behavior over the course of a month, it is useful to examine telecommuting on the travel day itself. This analysis takes as its base all workers ages 18+, regardless of whether they report being eligible for telework. It includes telework-ineligible workers because sometimes employees who are usually ineligible for teleworking are granted exceptions due to illness, inclement weather, natural disasters, or, more recently, pandemics. On an average day, 8.8 percent of people who work will exclusively telework, and a further 3.7 percent will work both from home and outside the home (mixed telework) (table 99). In other words, 12.5 percent of workers will telework for at least part of their work day.66 Teleworking is most common in the Atlanta MPO, and second-most common in non-MPO counties. Women are more likely to exclusively telework than men, though a higher percentage of men engage in mixed telework. One in five people with mobility impairments who are working on any given day will do so exclusively from home, which is more than double the statewide average. 66 As discussed in Overview in this chapter, the true total is likely higher because this figure only includes the primary activity at each location. 213 Table 99. Travel day commuting and telecommuting among workers ages 18+ by MPO tier, sex, caregiver status, age, race, and medical condition. All workers Exclusive Telework (Worked from Home Only) 8.8% Mixed Telework (Worked from Home & Outside of Home) 3.7% Conventional Commute (Worked Outside of Home Only) 87.5% Total Teleworking (Exclusive and Mixed) 12.5% MPO Tier 1. Atlanta MPO 11.3% 3.9% 84.9% 15.1% 2. Medium MPOs 5.1% 3.1% 91.8% 8.2% 3. Small MPOs 3.0% 3.1% 93.9% 6.1% 4. Non-MPO 6.3% 3.7% 90.0% 10.0% Sex Male 7.9% 4.1% 88.0% 12.0% Female Caregiver Status Noncaregiver Caregiver, youngest child ages 015 years 9.9% 8.3% 9.6% 3.1% 3.3% 4.4% 87.0% 88.5% 85.9% 13.0% 11.5% 14.1% Age Cohort Millennial and Gen Z (1836) Generation X (3752) Pre-retirement age Baby Boomer (5364) Retirement age (65+) 7.4% 8.5% 10.2% 16.8% 3.6% 3.4% 4.1% 5.2% 89.0% 88.1% 85.6% 78.1% 11.0% 11.9% 14.4% 21.9% Race White non-Hispanic only 10.2% 3.7% 86.1% 13.9% Black and Black multiracial 7.3% 3.1% 89.5% 10.5% Other 6.5% 4.9% 88.6% 11.4% Medical Condition Absent 8.6% 3.6% 87.7% 12.3% Present 20.4% 6.1% 73.6% 26.4% Based on workers who reported working outside the home or working from home for pay on the travel day. A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. Workers with a college degree and from high-income households are more likely to telework on the travel day (table 100). A strong relationship also exists with time flexibility: of workers who 214 can set their own schedule, one in four work from home for at least part of the day, versus just 4 percent of workers with inflexible schedules. Caregivers, especially female caregivers, are more likely than noncaregivers to work from home (table 101). The group most likely to telework is female co-caregivers (those living in a household with multiple caregivers). Single caregivers are less likely to work from home than noncaregivers, likely reflecting not a difference in preferences, but a difference in access to teleworking. In particular, single caregivers are often in lower education or lower income groups that have less access to teleworking. The relationship between distance to work and travel-day teleworking appears to be different when measuring monthly frequency versus daily behavior. In the monthly data, those with the smallest and largest distances to work were more likely to telecommute (table 97). When looking at travel-day telecommuting (table 102), aside from Atlanta, any elevation in telecommuting among workers who live 5 miles or less from their workplace is decidedly less pronounced. 215 Table 100. Travel day commuting and telecommuting among workers ages 18+ by employment characteristics. Exclusive Telework (Worked from Home Only) Mixed Telework (Worked from Home & Outside of Home) Conventional Commute (Worked Outside of Home Only) All workers 8.8% 3.7% 87.5% Occupation Category Sales or service 7.0% 3.9% 89.1% Clerical or administrative support 5.3% 1.8% 92.8% Blue collar Professional, managerial, or technical 2.4% 13.0% 3.0% 4.2% 94.6% 82.8% Education HS or less 2.8% 2.7% 94.5% Some college or associate 6.9% 2.4% 90.6% Bachelor's or higher 13.5% 5.1% 81.4% Work Schedule Part-time 10.0% 4.1% 85.9% Full-time 8.5% 3.6% 87.9% Annual Household Income <$35,000 5.3% 3.5% 91.2% $35,000 to $49,999 5.2% 3.0% 91.8% $50,000 to $74,999 7.5% 2.6% 89.9% $75,000 to $99,999 7.1% 4.1% 88.9% $100,000+ 14.7% 4.6% 80.6% Schedule Flexibility (Flextime) Not flexible 2.0% 2.4% 95.7% Flexible 19.0% 5.7% 75.4% Based on workers who reported working outside the home or working from home for pay on the travel day. Total Teleworking (Exclusive and Mixed) 12.5% 10.9% 7.2% 5.4% 17.2% 5.5% 9.4% 18.6% 14.1% 12.1% 8.8% 8.2% 10.1% 11.1% 19.4% 4.3% 24.6% 216 Table 101. Travel day commuting and telecommuting among workers ages 18+ by caregiver status All workers Caregiver Status Noncaregiver Caregiver, youngest child ages 015 Caregiver Status by Sex Male noncaregiver Female noncaregiver Male caregiver Female caregiver By Age of Youngest Child Youngest ages 04 Youngest ages 515 Household Type Male co-caregiver Female co-caregiver Male single caregiver Female single caregiver Exclusive Telework (Worked from Home Only) 8.8% Mixed Telework (Worked from Home & Outside of Home) 3.7% Conventional Commute (Worked Outside of Home Only) 87.5% 8.3% 9.6% 3.3% 4.4% 88.5% 85.9% 7.6% 8.5% 9.2% 11.1% 9.3% 9.9% 8.5% 12.0% 5.8% 7.9% 3.5% 5.3% 3.0% 3.3% 5.6% 3.4% 5.4% 3.8% 0.0% 1.9% 88.9% 86.2% 87.8% 85.6% 85.1% 86.6% 86.1% 84.3% 94.2% 90.2% Total Teleworking (Exclusive and Mixedl) 12.5% 11.5% 14.1% 11.1% 13.8% 12.2% 14.4% 14.9% 13.4% 13.9% 15.7% 5.8% 9.8% Based on workers who reported working outside the home or working from home for pay on the travel day. A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. 217 Table 102. Travel day commuting and telecommuting among workers ages 18+ by distance to work and MPO tier. Exclusive Conventional Teleworking on Telework Mixed Telework Commute Travel Day (Worked from Home Only) (Worked from (Worked Outside of Home & Outside of Home Only) Home) (Exclusive and Mixed) All workers By Distance to Work 05 miles 5.110 mi 10.120 mi 20.145 mi >45 mi 1.8% 1.6% 1.8% 2.3% 2.8% 2.4% 1.9% 1.9% 4.5% 3.9% 95.9% 96.6% 96.3% 93.2% 93.3% 4.1% 3.4% 3.7% 6.8% 6.7% Tier 1: Atlanta MPO Only 05 miles 5.110 mi 10.120 mi 20.145 mi >45 mi 2.2% 1.7% 2.6% 3.1% 3.6% 2.4% 2.0% 1.6% 3.9% 6.8% 95.4% 96.3% 95.9% 93.0% 89.7% 4.6% 3.7% 4.1% 7.0% 10.3% Tier 2: Medium MPOs Only 05 miles 5.110 mi 10.120 mi 20.145 mi >45 mi 0.3% 0.5% 0.3% 2.1% 8.7% 0.9% 1.9% 2.6% 4.7% 3.1% 98.8% 97.6% 97.1% 93.2% 88.1% 1.2% 2.4% 2.9% 6.8% 11.9% Tier 3: Small MPOs Only 05 miles 5.110 mi 10.120 mi 20.145 mi >45 mi 0.7% 1.1% 0.9% 1.1% 1.9% 4.3% 3.4% 0.9% 0.8% 2.5% 95.0% 95.6% 98.2% 98.1% 95.6% 5.0% 4.4% 1.8% 1.9% 4.4% Tier 4: Non-MPO Counties 05 miles 5.110 mi 10.120 mi 20.145 mi >45 mi 2.5% 3.0% 0.0% 0.0% 0.0% 2.1% 0.0% 3.2% 7.5% 2.1% 95.4% 97.0% 96.8% 92.5% 97.9% 4.6% 3.0% 3.2% 7.5% 2.1% Atlanta (Tiers 2-4) 05 miles 5.110 mi 10.120 mi 20.145 mi >45 mi 1.3% 1.4% 0.3% 0.7% 2.4% 2.3% 1.7% 2.5% 5.7% 2.4% 96.4% 96.9% 97.2% 93.6% 95.2% 3.6% 3.1% 2.8% 6.4% 4.8% Based on workers who reported working outside the home or working from home for pay on the travel day. 218 CHAPTER 4. NEW TECHNOLOGIES AND SERVICES CHAPTER 4 SUMMARY This chapter examines three emerging trends shaping Georgians' mobility in the following sections: Alternative-fuel Vehicles describes characteristics of the 130,000 alternative-fuel vehicles owned by Georgians. Hybrid cars are the most common AFV, followed by electric vehicles (EVs). AFVs are disproportionately owned by Atlanta households and high-income households. AFVs are driven somewhat fewer miles than other vehicles of a similar age. Shared Mobility discusses shared mobility, focusing on carsharing, bikesharing, and ridehailing. The section compares demographic characteristics of the users of these services to those of the general public. It also estimates the total monthly trips using these modes. While carsharing and bikesharing use are still relatively uncommon, 1 in 10 Georgians has used a ridehailing app in the past 30 days. This analysis estimates that ridehailing now accounts for 87 percent of all vehicle-for-hire trips, with the remainder comprising trips in conventional taxi and limo services. Ridehailing accounts for an even higher percentage of VFH trips in small MPOs and non-MPO counties. Online Shopping discusses online shopping. In an average month, more than half of Georgians ages 16+, and more than two thirds of Georgia households, place at least one online order for delivery. This section discusses demographic differences in who is shopping online and how frequently. Although online shopping is most common among 219 adults ages 1852, two in five seniors (6579) and one in five elderly adults (80+) have placed an order in the past 30 days. Online shopping is much more common among the wealthy and white, and uncommon among low-income people and carless households that could presumably benefit from the convenience of having goods delivered. ALTERNATIVE-FUEL VEHICLES As shown in table 103, in 2017, 1.9 percent of Georgia's vehicle fleet was composed of vehicles using alternative-fuel sources. These alternative-fuel vehicles comprise a larger percentage (2.5 percent) of the vehicle fleet in Atlanta than elsewhere in the state. Table 103. Number of alternative-fuel vehicles by MPO tier. All Vehicles* (N=16,921) Alternative-fuel Vehicles (AFVs) (N=313) Statewide 6,982,773 130,216 (1.9%) MPO Tier 1. Atlanta MPO 3,679,778 92,511 (2.5%) 2. Medium MPOs 1,120,485 18,619 (1.7%) 3. Small MPOs 707,626 7,001 (1.0%) 4. Non-MPO counties 1,474,885 12,085 (0.8%) * Includes gas, diesel, and AFVs. AFVs include hybrids, electric, plug-in hybrids, flex fuel, ethanol, and bifuel engines. As shown in table 104, AFVs were dominated by hybrid vehicles, which accounted for 61.7 percent of the fleet. Electric vehicles were the second-most common, accounting for 29.5 percent of the fleet. Plug-in hybrids comprised only 4.9 percent of the fleet, but national sales of plug-in hybrids increased substantially in 2018 and 2019, after the close of survey data 220 collection.67 Conventional-fuel vehicles are dominated by gasoline vehicles, with a small percentage of diesel engines. Table 104. Fuel type of alternative- and conventional-fuel vehicles. Weighted Percent All Vehicles All Alternative-Fuel Vehicles (AFVs)* Hybrid 61.71% Electric 29.48% Plug-in hybrid 4.89% Flex fuel or E85 3.16% Other AFV (bifuel and unspecified) All Conventional-Fuel Vehicles (CFVs) 0.75% Gas 97.79% Diesel Other CFV 2.19% 0.02% Note: Percentages shown are percent of category (AFV or CFV). Unweighted Sample Size 16,546 313 228 64 11 8 2 16,606 16,233 367 6 Unweighted Percent 72.84% 20.45% 3.51% 2.56% 0.64% 97.75% 2.21% 0.04% * All vehicle years (19872016). When the sample of AFVs is limited to recent vehicles (2004 or later), it contains 222 hybrids (61.8% weighted), 58 electric (29.5% weighted), 10 plug-in hybrid (5.0%), 7 flex fuel/E85 (3.2%), and 1 other (0.5%). Electric and plug-in hybrids are more common among recent vehicles, which correlates with a decrease in hybrid vehicles as a percentage of the overall fleet. All vehicle years (1900-2017). When the sample of CFVs is limited to recent vehicles (2004 or later), it contains 11,052 gas vehicles (98.1% weighted), 232 diesel vehicles (1.9%), and 1 "high tech" (.003%). Other CFV includes three "high-tech" and one each of nitro burner, C16 racing fuel, and unspecified mixture. Table 105 shows more details on AFVs, in comparison with all vehicles in Georgia. Because 96 percent of AFVs are from 2004 or later, this report provides statistics for all recent vehicles (from 2004 onward) to control for vintage when comparing AFVs to the fleet at large. 67 See https://afdc.energy.gov/data/10567. 221 The median age of AFVs is 4 years, compared to 10 years for all vehicles in Georgia. AFVs are also slightly newer than the typical recent vehicle. More than 80 percent of AFVs are cars or wagons, versus around half of all vehicles and recent vehicles. The complete lack of certain body types among AFVs, notably vans, indicates that some body types are more readily available as AFVs. Mean annual miles driven is lower for AFVs (11,778) than the state average (11,940), and also lower than the average for recent vehicles (13,109). 222 Table 105. Vehicle characteristics: AFVs, all vehicles, and recent vehicles. Vehicle Characteristics Alternative-fuel Vehicles (AFV) All Vehicles Recent Vehicles (20042017) Vehicle Age (Years) Mean 5.5 10.9 6.7 Median 4 10 6 Annual Miles Driven in Vehicle Mean 11,778 11,940 13,109 Median 10,915 9,585 10,913 Odometer Mileage Mean 57,725 105,440 80,804 Median 39,000 94,214 68,932 Vehicle Type (Column Percentage) Car/wagon 81.7% 49.5% 52.2% Van 0.0% 5.4% 5.5% SUV 6.5% 23.6% 26.2% Pickup 4.3% 17.5% 13.3% Other truck 0.0% 0.6% 0.4% RV 0.6% 0.7% 0.3% Motorcycle 0.0% 2.4% 1.9% Something else 6.9% 0.3% 0.3% Vehicle Age Cohort* (Column Percentage) Pre-LEV (pre-1993) 1.9% 4.6% 0.0% LEV1 (19932003) 2.5% 26.7% 0.0% LEV2 (20042014) 67.7% 56.3% 82.0% New vehicles (20152017) 27.9% 12.3% 18.0% Newly-purchased (Past 12 Months) Not newly purchased 81.3% 84.2% 81.5% Newly purchased 18.7% 15.8% 18.5% * Cohorts are based on the passage of California's Low Emissions Vehicle (LEV) standards, which serve as a proxy for tightening emissions standards. Figure 25 shows the weighted distribution of annual miles driven. Compared to all vehicles or all recent vehicles, the AFV distribution has a thicker left tail (more vehicles driven less than a thousand miles per year) and a thinner right tail. This dynamic illustrates why AFVs have a higher median and lower mean mileage compared to all vehicles (table 105). 223 Figure 25. Histograms. Annual miles driven by AFVs, all vehicles, and recent vehicles. 224 Table 106 provides greater details on the average annual miles driven by fuel type for all vehicle years and for recent vehicles (from 2004 onward). On average, AFVs are driven fewer miles than CFVs (11,778 miles versus 11,943 miles). As shown by the lower standard deviation, there is also less variation in how far AFVs are driven. The distribution of AFVs shown in figure 25 is not as strongly right-skewed as the distribution of vehicles in general; 1.1 percent of CFVs are driven more than 50,000 miles annually, versus just 0.05 percent of AFVs. Table 106. Annual miles driven by fuel type. All vehicles All conventional-fuel vehicles (CFV) Mean Annual Miles Driven in Vehicle 11,940 11,943 Standard Deviation* 12,388 12,455 Gas 11,900 12,301 Diesel All alternative-fuel vehicles (AFV) 14,004 11,778 18,177 8,079 Hybrid and plug-in hybrid 12,315 7,930 Electric 9,652 7,529 All recent vehicles (2004 or later) All recent CFV 13,109 13,134 12,841 12,942 Gas 13,076 12,717 Diesel All recent AFV 16,277 12,131 21,720 8,071 Hybrid and plug-in hybrid 12,580 7,971 Electric 10,124 7,442 * Based on an estimate of simple random sampling (SRS) variance within each subpopulation. Includes gas, diesel, and other CFV. Includes hybrid, plug-in hybrid, electric, flex fuel/E85, and other. Table 106 also shows heterogeneity within AFVs and CFVs. Among CFVs, diesel vehicles have a higher mean annual miles driven than gasoline vehicles. Among AFVs, hybrids and plug-in hybrids have a larger mean annual miles driven than fully electric vehicles. It is likely this 225 difference is due in part to technological limitations on the range of the electric vehicles available in 2016 (the most recent model year of EVs in the sample) and before. The difference in annual miles driven between AFVs and CFVs cannot be assumed to be a product of the vehicles themselves; they may reflect differences in the lifestyles and preferences of the people who chose to purchase AFVs versus CFVs. This section turns now to the demographic characteristics of the households and main drivers of AFVs as compared to vehicle owners in general. Table 107 shows the distribution of AFVs and other vehicles among different types of households. Of AFVs, 71 percent are owned by households in Atlanta, compared to 56 percent of recent vehicles and 53 percent of all vehicles. AFVs are mostly owned by wealthier households that own, on average, 2.7 vehicles. One-driver households are less common than the average among AFVs. Table 106 showed that AFVs are driven less than CFVs on average. Table 107 shows that this is still the case when annual miles are normalized by total household vehicles, household drivers, or household members of driving age. 226 Table 107. Household characteristics for AFVs, all vehicles, and recent vehicles. Household Characteristics Alternative-fuel Vehicles (AFV) All Vehicles Recent Vehicles (20042017) MPO Tier 1. Atlanta MPO 71.0% 52.7% 55.9% 2. Medium MPOs 14.3% 16.0% 16.1% 3. Small MPOs 5.4% 10.1% 9.6% 4. Non-MPO counties 9.3% 21.1% 18.3% Annual Household Income <$35,000 9.4% 25.3% 20.1% $35,000 to $49,999 6.1% 12.0% 11.6% $50,000 to $74,999 10.2% 18.4% 18.7% $75,000 to $99,999 17.6% 13.7% 14.3% $100,000+ 56.7% 30.6% 35.3% Number of Household Vehicles Mean 2.74 2.66 2.54 Median 2 2 2 Number of Household Drivers (Ages 16+)* Mean 2.23 2.10 2.10 One driver 14.3% 22.3% 20.9% Two drivers 57.4% 53.4% 55.6% Three or more drivers 28.3% Annual Miles Driven in All Household Vehicles (Mean) 24.1% 23.3% Total 30,476 29,743 29,908 Per vehicle Per household driver 11,314 13,589 11,743 14,621 12,356 14,693 Per household member ages 16+ 13,271 13,766 13,945 * 45 vehicles (unweighted) were in households that reported having zero drivers, representing 0.2 percent of all vehicles (weighted) and 0.1 percent of recent vehicles (weighted). No AFVs were owned by zero-driver households. The sum of all miles driven in all household vehicles, whether by a household driver or someone else. Does not include miles driven by household drivers in rental cars or other non-household vehicles. Households with zero drivers recoded as one driver for the purposes of this statistic. Table 108 shows driver characteristics for AFVs, all vehicles, and recent vehicles. The main driver for an AFV is more likely to be male, white, highly educated, and a worker. These 227 demographic factors may correlate with AFV ownership due to demographic differences in attitudes, disposable income, and travel needs, or a combination of these factors. Table 108. Main driver characteristics for AFVs, all vehicles, and recent vehicles. Driver Characteristics Sex Male Female Age Cohort Under 18 Millennial and Gen Z (1636) Gen X (3752) Pre-retirement age Boomer (5364) Seniors (6579) Elderly (80+) Race White non-Hispanic only Black, Black multiracial, and Black Hispanic Other Employment Status Nonworker Worker Educational Attainment High school or less Some college or associate degree Bachelor's degree Graduate or professional degree Alternative-fuel Vehicles (AFV) 53.0% 47.0% 2.8% 16.9% 37.4% 27.8% 13.6% 1.5% 78.3% 10.8% 10.9% 27.4% 72.6% 8.8% 21.3% 35.9% 34.0% All Vehicles 50.2% 49.8% 1.3% 22.9% 31.5% 25.4% 16.6% 2.5% 64.3% 24.7% 10.9% 33.8% 66.2% 25.9% 32.0% 23.2% 18.9% Recent Vehicles (20042017) 44.4% 55.6% 1.2% 23.7% 33.4% 24.0% 15.6% 2.0% 65.0% 23.8% 11.2% 31.4% 68.6% 21.0% 30.8% 26.3% 21.9% SHARED MOBILITY Shared mobility, as defined by the Shared Use Mobility Center, broadly encompasses "transportation services and resources that are shared among users, either concurrently or one 228 after another."68 Some forms of shared mobility, such as carpooling and public transit, are not reliant on emerging technologies. However, the proliferation of location-enabled smartphones and other new technologies have facilitated the evolution of several forms of shared mobility. These include bike- and carsharing systems, as well as the rise of ridehailing apps such as Uber and Lyft. More recently, scooter sharing has become a part of the transportation landscape, particularly in the Atlanta region. However, this recent development is not reflected in the 2017 NHTS data. This section focuses on bikesharing, carsharing, and ridehailing, the latter of which is sometimes referred to as ridesharing.69 It provides a portrait of shared mobility use from when the data were collected in 20162017. With the increased availability of many types of shared mobility, it is likely that usage of all of these services has risen in the intervening years. Methods NHTS assesses use of shared mobility services with the questions shown in table 109. As table 109 shows, 10.2 percent of Georgians ages 16 and up report having used a ridehailing app at least once in the past 30 days, and 1.0 percent report using a carsharing app. Because the question about bikesharing was only asked of respondents who had bicycled within the past 7 days but asked about bikesharing use within the past 30 days, it is not possible to get a precise estimate of bikesharing use among the general population.70 However, 8 percent of recent 68 https://sharedusemobilitycenter.org/what-is-shared-mobility/. 69 The NHTS uses the word "ridesharing." The researchers prefer to use "ridehailing" in this report because it avoids ambiguity and confusion with other forms of shared rides, such as carpooling. 70 For example, someone who used a bikesharing service 14 days previously and had not biked since would not have been asked whether she had used a bikesharing service. 229 cyclists (those who rode a bike within the past 7 days) reported having used a bikesharing service at some point in the past 30 days. Table 109. Shared mobility questions in the NHTS. Mobility Type Question Wording Who Was Asked? Ridehailing In the past 30 days, how many All respondents times have you purchased a ride ages 16+ with a smartphone rideshare app (e.g., Uber, Lyft, Sidecar)? Affirmative Responses (Ages 16+) Weighted Unweighted 10.2% 1,176 (7.54%) Carsharing In the past 30 days, how many All respondents times did you use a carsharing ages 16+ service where a car can be rented by the hour (e.g., Zipcar or Car2Go)? 1.0% 104 (0.67%) Bikesharing In the past 30 days, how many times did you use a bikeshare program (e.g., Bikeshare, Zagster, or CycleHop)? All recent cyclists* ages 5+ 8.0% 51 (5.70%) * Reported at least one cycling trip in the past 7 days. Recent cyclists represent 5.8 percent of respondents (unweighted). In addition to these questions, when respondents filled out their travel diary, one of the choices of mode for trips was "Taxi/limo (including Uber/Lyft)." Because ridehailing services are combined with more traditional vehicle-for-hire services, travel diaries do not provide a precise estimate of ridehailing usage. However, those data are included here to provide a baseline for future analysis. To provide an exploratory estimation of what proportion of these trips are using a ridehailing app, we also compare them with estimates of total trips based on the question in table 109. 230 As table 109 also shows, because bikeshare and carshare services are utilized by a small fraction of the population, the sample sizes for users of these services are quite small. Survey weights are unreliable with small sample sizes; thus, the analysis of these two user groups will be based on unweighted data and may not be representative. The results should therefore be interpreted with caution. In addition to the 51 adults who reported using bikeshare systems, 26 children between ages 5 and 15 were recorded as bikeshare users. Ten of these were under the age of 10. Bikesharing apps require a smart phone and a credit card. Most, including Relay and Zagster, two services available in Georgia, require riders to be at least 18 years of age. Some require a governmentissued ID for verification. Further, the available bicycles are sized for adult riders. One potential explanation for this finding is that some children were riding bicycles that had been unlocked on their behalf by adults. Relay, for example, allows members to unlock up to four bicycles at a time (though they still legally require all riders to be ages 18+). Another potential explanation is that parents filling out the survey on their child's behalf interpreted the question differently than intended, for example including borrowing a bicycle from a neighbor or some other kind of bicycle-sharing program. More targeted data collection would provide better information about this finding, and about bikesharing in general. Shared Vehicles: Bikesharing and Carsharing Table 110 shows demographic breakdowns of bikeshare users, recent cyclists who have not used a bikeshare, and carshare users. Weighted statistics for the population as a whole are included for comparison. Results should be interpreted with caution due to small sample sizes (included in the table). 231 Table 110. Demographics of bikeshare and carshare users ages 16+. Bikeshare Users Past 30 Days (Unweighted) Other Recent Cyclists* (Unweighted) Carshare Users All Persons Past 30 Days Ages 16+ (Unweighted) (Weighted) Total 51 (100%) 842 (100%) 104 (100%) MPO Tier 1. Atlanta MPO 13 (25%) 230 (27%) 38 (37%) 54.4% 2. Medium MPOs 23 (45%) 348 (41%) 38 (37%) 15.8% 3. Small MPOs 12 (24%) 191 (23%) 19 (18%) 10.1% 4. Non-MPO counties 3 (6%) 73 (9%) 9 (9%) 19.8% Sex Male 27 (53%) 503 (60%) 42 (40%) 47.9% Female 24 (47%) 339 (40%) 62 (60%) 52.1% Age Cohort Millennial and Gen Z (1636) 23 (45%) 237 (28%) 35 (34%) 36.7% Gen X (3752) 10 (20%) 237 (28%) 22 (21%) 27.6% Pre-retirement age Boomer (5364) 8 (16%) 222 (26%) 27 (26%) 19.5% Retirement age (65+) 10 (20%) 146 (17%) 20 (19%) 16.2% Race White non-Hispanic only 18 (35%) 634 (75%) 41 (39%) 54.8% Black and Black multiracial 23 (45%) 146 (17%) 51 (49%) 32.1% Other 10 (20%) 62 (7%) 12 (12%) 13.1% Driver Status Nondriver 9 (18%) 69 (8%) 11 (11%) 12.6% Driver 42 (82%) 773 (92%) 93 (89%) 87.4% Mobility Impairment Absent 45 (88%) 811 (96%) 88 (85%) 90.7% Present 5 (10%) 31 (4%) 15 (14%) 9.3% Annual Household Income <$15,000 12 (24%) 104 (12%) 23 (22%) 14.5% $15,000 to $34,999 15 (29%) 86 (10%) 21 (20%) 19.6% $35,000 to $74,999 9 (18%) 207 (25%) 28 (27%) 28.4% $75,000+ 13 (25%) 421 (50%) 28 (27%) 37.4% Vehicle Deficit Category of Household Zero-vehicle 6 (12%) 44 (5%) 13 (13%) 5.0% Deficit (hard or soft) 18 (35%) 133 (16%) 22 (21%) 27.0% Nondeficit (sufficient/surplus) 27 (53%) 665 (79%) 69 (66%) 68.0% Transit Use, Past 30 Days No 36 (71%) 714 (85%) 66 (63%) 88.9% Yes 15 (29%) 127 (15%) 38 (37%) 11.1% Walking, Past 30 Days No 4 (8%) 79 (9%) 18 (17%) 28.0% Yes 47 (92%) 763 (91%) 86 (83%) 72.0% Note: Because survey weights may not produce accurate estimates for small subsamples, unweighted percentages are shown for carshare and bikeshare users, as well as other recent cyclists. * Respondents who reported at least one cycling trip within the past 7 days but did not report using a bikeshare within the past 30 days. 232 While the majority of Georgia's population is in Atlanta, the plurality of bikeshare users, as well as of other recent cyclists, live in medium MPO areas. Carshare users are divided evenly between Atlanta and medium MPO areas, with a smaller proportion of users in small MPO areas and non-MPO counties. There is a well-documented gender gap in cycling (Emond et al. 2009). Men dominate among recent cyclists who are not bikeshare users, but bikeshare users are more evenly divided by gender. The majority of carshare users surveyed are female. Compared to bikeshare nonusers and the general population, a larger proportion of bikeshare users are Black. Black residents also make up a larger proportion of carshare users than the general population. Low-income people make up a larger share of carshare users than of the general population. They are also disproportionately represented among bikeshare users compared to other recent cyclists. About half of bikeshare users live in vehicle-deficit or zero-vehicle households, compared to just one third of the general population. Interestingly, the percent of carshare users from vehiclenondeficit households is approximately equal to that of the general population. With respect to walking and transit use, compared to the general population, a higher proportion of bikeshare and carshare users have used these modes. In the case of walking, there is little difference between bikeshare users and other recent cyclists. However, bikeshare users are more likely to be transit users than other recent cyclists are. As shown in table 111, Georgians who had used a bikeshare at least once in the past 30 days reported an average of 5.6 uses, versus 3.8 carshare uses among users. This is equivalent to 233 497,000 carshare uses per month (258,000) and 165,000 bikeshare uses (77,000). The bikeshare estimate should be taken with further caution because it does not include users who rode a bike within the past 30 days but not within the past 7 days. Table 111. Frequency of carshare and bikeshare use among users ages 16+, past 30 days. Bikeshare Users All Users 5.6 MPO Tier 1. Atlanta MPO 5.3 2. Medium MPOs 6.5 34. Small MPOs and non-MPO counties 4.5 Sex Male 5.7 Female 5.5 Age Cohort 1652 (Gen X, Millennial, and Gen Z) 5.0 53+ (Baby Boomer and retirement age) 6.7 Race White non-Hispanic only 7.1 Black and other 4.8 Annual Household Income <$35,000 5.4 $35,000+ 5.9 Vehicle Deficit Category of Household Deficit or zero-vehicle 5.1 Nondeficit (sufficient/surplus) 6.0 Carshare Users 3.8 3.8 2.9 5.1 3.0 4.4 4.4 3.1 2.6 4.6 4.5 3.3 3.8 3.8 Even after accounting for the small sample size, there were statistical differences in frequency of bikeshare use by race, income, and vehicle-deficit category (=.01).71 Although, as shown in table 110, Blacks are more likely to have used bikesharing at all in the previous 30 days, table 111 shows that the amount of bikeshare use during that time is greater among users who are 71 Calculated by a t-test for two-group categories and chi-squared test for categories with 3+ groups. 234 white non-Hispanic than among users of color. It is lower among the lowest-income users (those in households with an annual income of less than $35,000) than among moderate- and highincome users. It is higher among users in vehicle-nondeficit households than those from zerovehicle or vehicle-deficit households. Taken together, table 110 and table 111 suggest that although bikeshare adoption is associated, to some extent, with markers of necessity such as lack of transportation alternatives or financial resources, bikeshare usage is not. Among carshare users, there were marginally-significant differences by race (=.10), with people of color reporting two more usage occasions in the past 30 days (4.6), on average, than non-Hispanic whites (2.6). Ridehailing As shown in table 112, 10.2 percent of Georgians ages 16 and older reported using a ridehailing app within the past 30 days. This is comparable with the national figure of 9.8 percent found by Conway, Salon, and King (2018). Ridehailing adoption was most common in Atlanta and least common in non-MPO counties. However, users in non-MPO counties made substantially more trips by ridehailing in that timeframe. There were not pronounced differences in the percent of people who used ridehailing by gender or race. However, among ridehailing users, Black riders made more trips than white riders and people of other races. Millennials and members of Gen Z were more likely to use ridehailing apps and made more trips than older users. Two groups with elevated transportation needs, nondrivers and people with mobility impairments, were less likely to use ridehailing. However, those nondrivers and people with mobility impairments who did use a ridehailing app made more trips than drivers and people without mobility impairments. 235 The percentage of people who use ridehailing is highest among people living in households with an annual income of $75,000 or more per year. However, among ridehailing users, the lowestincome users made more trips per person. These results contrast with those for bikesharing (see Shared Vehicles: Bikesharing and Carsharing), in which adoption was higher for lower-income people, but trip frequency was lower. Georgians who had used transit were also more likely to report using a ridehailing app (31.6 percent versus 7.5 percent) and made more trips (5.3 versus 4.7). Georgians who had reported one or more walking trips were also more likely to report using ridehailing but reported fewer ridehailing trips than those who had not made any pedestrian trips. Table 113 presents similar information as table 112. However, instead of showing the percentage of, for example, Atlanta residents who use ridehailing, it shows the percentage of ridehailing users who are from Atlanta. 236 Table 112. Ridehailing use among Georgians ages 16+. All persons ages 16+ MPO Tier 1. Atlanta MPO 2. Medium MPOs 3. Small MPOs 4. Non-MPO counties Sex Male Female Age Cohort Millennial and Gen Z (1636) Gen X (3752) Pre-retirement age Boomer (5364) Retirement age (65+) Race White non-Hispanic only Black and Black multiracial (incl. Black Hispanic) Other Driver Status Nondriver Driver Mobility Impairment Absent Present Annual Household Income <$15,000 $15,000 to $24,999 $25,000 to $34,999 $35,000 to $49,999 $50,000 to $74,999 $75,000 to $99,999 $100,000+ Vehicle Deficit Category of Household Zero-vehicle Deficit (hard or soft) Nondeficit (sufficient/surplus) Transit Use, Past 30 Days No Yes Walking, Past 30 Days No Yes Percent Who Have Used Ridehailing (past 30 days) 10.2% 14.6% 7.6% 4.3% 3.0% 10.9% 9.5% 15.3% 12.2% 4.9% 1.5% 10.3% 10.1% 9.7% 8.9% 10.4% 10.8% 4.4% 7.2% 6.6% 5.2% 7.8% 7.8% 11.0% 17.8% 15.6% 7.2% 11.0% 7.5% 31.6% 7.0% 11.4% Mean Trips Among Ridehailing Users 4.9 4.6 4.1 2.6 13.0 4.9 5.0 5.4 4.6 3.6 2.7 3.9 7.1 3.6 6.2 4.8 4.8 7.8 11.1 6.1 4.9 3.7 3.7 5.2 3.9 8.5 4.2 4.7 4.7 5.3 7.6 4.3 237 Table 113. Demographics of ridehailing users (column percentages). MPO Tier 1. Atlanta MPO 2. Medium MPOs 3. Small MPOs 4. Non-MPO counties Sex Male Female Age Cohort (adults only) Millennial and Gen Z (1636) Gen X (3752) Pre-retirement age Boomer (5364) Retirement age (65+) Race White non-Hispanic only Black and Black multiracial (incl. Black Hispanic) Other Driver Status Nondriver Driver Mobility Impairment Absent Present Annual Household Income <$15,000 $15,000 to $24,999 $25,000 to $34,999 $35,000 to $49,999 $50,000 to $74,999 $75,000 to $99,999 $100,000+ Vehicle Deficit Category of Household Zero-vehicle Deficit (hard or soft) Nondeficit (sufficient/surplus) Transit Use, Past 30 Days No Yes Walking, Past 30 Days No Yes Ridehailing Users (Past 30 Days) 78.2% 11.8% 4.3% 5.8% 51.5% 48.5% 55.0% 33.2% 9.4% 2.5% 55.5% 32.0% 12.5% 11.0% 89.0% 96.0% 4.0% 10.2% 6.1% 5.2% 9.0% 12.5% 12.6% 44.4% 7.7% 19.0% 73.3% 65.6% 34.4% 19.4% 80.6% All Georgians Ages 16+ 54.4% 15.8% 10.1% 19.8% 47.9% 52.1% 36.7% 27.6% 19.5% 16.2% 54.8% 32.1% 13.1% 12.6% 87.4% 90.7% 9.3% 14.5% 9.5% 10.1% 11.9% 16.6% 11.7% 25.7% 5.0% 27.0% 68.0% 88.9% 11.1% 28.0% 72.0% 238 Ridehailing and Vehicle-for-Hire Trips This section estimates total annual ridehailing trips and compares them with the estimated total of trips by vehicle-for-hire. VFH trips include ridehailing trips, as well as more traditional taxi and limo services. This is done by comparing responses to the ridehailing question analyzed in the previous section (Ridehailing) with reported trips by VFH on the travel day (table 114). Table 114. Unweighted sample sizes for ridehailing and vehicle-for-hire trips by Georgians ages 16+ Rideshare (Past 30 Days)* All Vehicle-for-Hire Trips (Travel Day Users Times Used Users Trips All Georgia 1176 4741 109 205 MPO Tier 1. Atlanta MPO 689 2920 62 111 2. Medium MPOs 349 1209 25 45 3. Small MPOs 108 326 18 38 4. Non-MPO counties 35 286 4 11 * Based on the following question: "In the past 30 days, how many times have you purchased a ride with a smartphone rideshare app (e.g., Uber, Lyft, Sidecar)?" Based on trips on the travel day with mode recorded as "Taxi/Limo (including Uber/Lyft)." While 10 percent of Georgians (ages 16 and older) reported using a ridehailing app within the past 30 days (table 112), VFH trips only accounted for 0.6 percent of reported trips on the travel day itself (weighted), or 165 total trips (unweighted). Because of the small sample sizes, the researchers combine MPO tiers in two different ways: (1) the Atlanta region versus tiers 24 combined, and (2) tiers 1 (Atlanta), 2, and 3+4 combined. As shown in table 115, Georgians ages 16+ take an estimated 56 million trips by taxi, limo, and ridehailing per year (based on the travel diaries), and 49 million trips by ridehailing alone (based 239 on the ridehailing question). Therefore, an estimated 87.0 percent of vehicle-for-hire trips took place by ridehailing, and the remaining 13.0 percent by traditional taxis and limos.72 Table 115. Monthly and annualized estimates of ridehailing and vehicle-for-hire trips by Georgians ages 16+. All Georgia Region Atlanta MPO Rest of Georgia MPO Tier Atlanta MPO Medium MPOs Small MPOs & non-MPO counties Ridehailing Trips* All Vehicle-for-Hire (VFH) Trips Percent Ridehailing Ridehailing All Per 30 Days Annual Per 30 Days Annual VFH 4,001,029 48,679,186 4,597,320 55,934,059 87.0% 2,903,628 1,097,400 35,327,474 13,351,705 3,324,791 1,272,529 40,451,620 15,482,439 87.3% 86.2% 2,903,628 388,362 709,039 35,327,474 4,725,070 8,626,635 3,324,791 525,472 747,058 40,451,620 6,393,238 9,089,201 87.3% 73.9% 94.9% * Based on the following question: "In the past 30 days, how many times have you purchased a ride with a smartphone rideshare app (e.g., Uber, Lyft, Sidecar)?" Thirty-day estimates are converted to annual estimates by multiplying by Based on trips on the travel day with mode recorded as "Taxi/Limo (including Uber/Lyft)." By default, trip weights produce annual estimates, which have been converted to 30-day estimates by multiplying by 30/365. Ridehailing accounts for a larger percentage of VFH trips in Atlanta (87.3 percent) compared to the rest of Georgia (86.2 percent). However, when the data are further disaggregated, ridehailing appears to make up a larger portion of VFH trips in small MPOs and non-MPO counties (94.9 percent) compared to medium MPOs (73.9 percent). While this may be an artifact of small sample sizes, one possible explanation is that traditional taxi services are not as widely available in small towns and rural communities. 72 Due to the small sample of VFH trips in the travel diary, the 95 percent confidence interval of VFH trips is between 52.2 million and 74.1 million. This would mean that the percentage of ridehailing trips might range between 66 and 93 percent, before accounting for additional uncertainty in the number of ridehailing trips. 240 ONLINE SHOPPING As online shopping becomes more prominent, it may change travel behavior in the form of fewer or shorter shopping trips, and may lead to longer-term changes by remaking the kinds and quantity of retail businesses in Georgia communities. Online shopping has become more prominent during the COVID-19 pandemic, but in 2017, it was already common. The NHTS asks respondents ages 16+ how many times in the past 30 days they "purchase[d] something online and had it delivered." As shown in table 116, 52 percent of Georgians reported placing at least one online order, and 66 percent of households contained at least one person who placed an online order. Table 116. Online orders by MPO tier, past 30 days. Percent Placing 1+ Orders Persons Ages 16+ Households Number of Orders Total Orders Per Household All 51.9% 66.0% 12,332,000 16,725,000 3.4 4.6 MPO Tier 1. Atlanta MPO 57.1% 72.4% 7,297,000 9,805,000 3.7 5.0 2. Medium MPOs 51.1% 64.4% 2,047,000 2,828,000 3.4 4.8 3. Small MPOs 47.6% 62.3% 1,111,000 1,547,000 3.0 4.1 4. Non-MPO counties 40.4% 52.3% 1,878,000 2,546,000 2.6 3.5 Note: An order consists of instances when a household member "order[ed] something online and had it delivered" in the past 30 days. Because orders may be split into multiple shipments by the companies fulfilling them, the number of deliveries may exceed the number of orders. Number of household orders is derived from responses of individual household members. In some cases, multiple household members may have listed the same order (such as a kitchen item to be used by household members generally). Therefore, the estimated number of orders is reported as a range. The high estimate assumes no overlap between household members' purchases, and the low estimate assumes total overlap. The number of orders was asked of individual household members, and there is some possibility that multiple household members may have taken "credit" for the same purchase, for instance a 241 piece of furniture that will be used by the entire family. The researchers therefore report total orders as a range. Per table 116, we estimate that Georgians placed between 12.3 and 16.7 million orders per year, or 3.44.6 per household (including those with zero orders).73 Online shopping was most common in Atlanta, where 72.4 percent of households placed one or more orders, and least common in non-MPO counties, where just 52.3 percent of households placed an online order. The average Georgia household has precisely one online shopper. However, when households with zero online purchases are excluded, the average is 1.6 online shoppers per household. Figure 26 shows the number of online shoppers in Georgia households. 73 There is not a 1:1 correlation between orders and resulting deliveries. A single order may contain multiple items, and companies that fulfill those orders may divide them into multiple shipments. 242 Figure 26. Pie chart. Number of online shoppers per household, past 30 days. As shown in table 117, online shopping is nearly universal among high-income Georgians, with 94 percent of households that earn more than $100,000 a year making at least one online purchase. It is substantially less common among low-income households (32 percent of those earning less than $15,000). Among online shopping households, households earning $100,000 or more place an average of 710 online orders per month, compared to approximately 34 for all income categories below $35,000. 243 Table 117. Household online shopping by MPO tier, annual household income, vehicle ownership, and race, past 30 days. All households MPO Tier 1. Atlanta MPO 2. Medium MPOs 3. Small MPOs 4. Non-MPO counties Annual Household Income <$15,000 $15,000 to $24,999 $25,000 to $34,999 $35,000 to $49,999 $50,000 to $74,999 $75,000 to $99,999 $100,000+ Vehicle Deficit Category of Household Zero-vehicle Deficit (hard or soft) Nondeficit (sufficient/surplus) Household Race All white non-Hispanic Nonwhite/mixed race (some or all household members are nonwhite and/or Hispanic) Orders per Household Households Percent of Households All with 1+ with 1+ Online Order(s) Households Order(s) 66.0% 3.4 4.6 5.1 6.9 72.4% 64.4% 62.3% 52.3% 3.7 5.0 3.4 4.8 3.0 4.1 2.6 3.5 5.2 6.9 5.4 7.4 4.8 6.6 4.9 6.7 31.5% 47.0% 56.9% 66.0% 76.6% 80.6% 93.8% 1.1 1.3 1.7 2.0 2.0 2.5 2.7 3.5 3.4 4.6 4.4 5.9 6.6 9.5 3.4 4.2 3.7 4.3 3.6 4.4 4.1 5.3 4.5 6.0 5.4 7.4 7.0 10.1 25.2% 63.1% 70.6% 0.9 1.0 2.6 3.8 3.8 5.1 3.5 4.0 4.1 6.0 5.4 7.2 72.2% 4.0 5.5 58.9% 2.6 3.5 5.6 7.7 4.4 5.9 Note: An order consists of an instance when a household member "order[ed] something online and had it delivered" in the past 30 days. Number of household orders is derived from responses of individual household members. In some cases, multiple household members may have listed the same order (such as a kitchen item to be used by household members generally). Therefore, the estimated number of orders is reported as a range. The high estimate assumes no overlap between household members' purchases, and the low estimate assumes total overlap. Similarly, white households are more likely to be online shoppers than nonwhite/mixed race households, and among online shopping households make more purchases on average. Vehiclesufficient households are more likely to be online shoppers and place more orders than households with fewer available vehicles. 244 Table 118 shows online orders by household composition. Households with children are more likely to place online orders. The average number of orders is also larger for online shopping households with children compared to those without children, and largest for households with children ages five or younger. However, when this is disaggregated by number of adults in the household, it is clear that this increase in online ordering is limited to households with two or more adults. Just 41.5 percent of single-caregiver households with children ages five or younger placed an online order, versus 57.3 percent of single adults without children and 75.9 percent of households with a young child and two or more adults. This is likely related to income; singleparent households face elevated economic challenges compared to other types of households. Whether or not a household is an online shopping household is strongly correlated with the number of members ages 16 or older. More than three quarters of households with three or more people ages 16+ made at least one online purchase, compared to half of households with one person age 16+. Residents of larger households are therefore more likely to live in an online shopping household, regardless of whether they personally make a purchase. Figure 27 shows the proportion of households where none, some, or all members ages 16+ placed an online order, subdivided by household size. 245 Table 118. Household online shopping by household composition. Orders per Household Households Percent of Households All with 1+ with 1+ Online Order(s) Households Order(s) All households 66.0% 3.4 4.6 5.1 6.9 Number of Household Members Ages 16+ 1 50.0% 1.9 1.9 3.8 3.8 2 71.3% 4.0 5.5 5.6 7.7 3+ 78.0% 4.1 6.5 5.2 8.3 Household Composition (Categories) Working-age adult(s), no children under 16 69.0% 3.4 4.5 4.9 6.5 Youngest child 05 72.5% 5.0 7.0 6.8 9.7 Youngest child 615 Retired, no children under 16 73.9% 52.0% 4.2 5.8 2.0 2.6 5.7 7.9 3.9 5.0 Household Composition (Detailed) One adult, no children 2+ adults, no children 57.3% 75.9% 2.3 2.3 4.0 5.8 4.0 4.0 5.3 7.6 One adult, youngest child 05 41.5% 1.8 1.8 4.3 4.3 2+ adults, youngest child 05 One adult, youngest child 6-15 78.3% 57.1% 5.6 8.0 2.4 2.7 7.1 10.3 4.3 4.6 2+ adults, youngest child 615 79.2% 4.8 6.9 6.0 8.7 One adult, youngest child 1621 2+ adults, youngest child 1621 62.4% 85.0% 3.3 4.4 4.7 7.3 5.2 7.1 5.6 8.6 One adult, retired, no children 34.1% 1.0 1.0 2.9 2.9 2+ adults, retired, no children 61.1% 2.5 3.4 4.1 5.6 Note: An order consists of an instance when a household member "order[ed] something online and had it delivered" in the past 30 days. Number of household orders is derived from responses of individual household members. In some cases, multiple household members may have listed the same order (such as a kitchen item to be used by household members generally). Therefore, the estimated number of orders are reported as a range. The high estimate assumes no overlap between household members' purchases, and the low estimate assumes total overlap. 246 Figure 27. Pie chart. Household online shopping by household size, past 30 days. This report now turns from examining households to examining the individuals who live in those households (table 119). Just over half of people ages 16+ personally placed an online order, but 71 percent live in an online shopping household (whether or not they personally placed an order).74 74 This is higher than the 66 percent of households that placed at least one order because larger households, which represent a higher share of the population than of households, are more likely to contain one or more online shoppers. 247 As shown in table 119, elderly Georgians are the least likely to personally place an online order (20.4 percent) and also the least likely to live in an online shopping household (42.5 percent). Teens ages 1617 have the second-lowest percentage of online shoppers (29.5 percent), perhaps because of limited economic autonomy. However, unlike elderly Georgians, teens are about as likely as adults ages 1852 to live in an online shopping household. Groups with mobility challenges (i.e., nondrivers, people with mobility impairments, and residents of zero-vehicle and vehicle-deficit households) are all less likely to personally place online orders and to live in online shopping households than their nondisadvantaged counterparts. Table 120 provides more detail about which household members are making purchases. For example, 45 percent of teens (ages 1617) live in an online shopping household without having personally placed an order, more than twice the rate for any other group. 248 Table 119. Online shopping by persons ages 16+, past 30 days. Online Shoppers Orders/Person Online In Online All Online Shopper Shopping Persons Shoppers (Personally Household 16+ Only Placed (1+ Order was 1+ Order) Placed in HH) All 51.9% 70.8% 2.3 4.5 MPO Tier 1. Atlanta MPO 57.1% 75.5% 2.5 4.4 2. Medium MPOs 51.1% 71.4% 2.4 4.8 3. Small MPOs 47.6% 68.2% 2.1 4.4 4. Non-MPO counties 40.4% 58.8% 1.8 4.4 Sex Male 49.2% 72.5% 2.1 4.3 Female 54.4% 69.3% 2.5 4.6 Age Cohort Teens (1617) 29.5% 74.5% 0.8 2.7 Millennial and Gen Z (1836) 55.1% 75.9% 2.6 4.6 Gen X (3752) 61.0% 76.4% 3.1 5.0 Pre-retirement age Boomer (5364) 50.6% 66.9% 2.0 4.0 Senior (6579) 40.3% 57.8% 1.4 3.6 Elderly (80+) 20.4% 42.5% 0.5 2.4 Race White non-Hispanic only 59.6% 77.5% 2.9 4.9 Black and Black multiracial 39.1% 58.2% 1.4 3.5 Other 51.0% 73.8% 2.2 4.4 Driver Status Nondriver 23.8% 50.8% 0.9 3.6 Driver 55.9% 73.7% 2.5 4.5 Mobility Impairment Absent 53.9% 72.6% 2.4 4.5 Present 32.1% 53.5% 1.3 4.1 Annual Household Income <$15,000 24.8% 38.3% 0.8 3.3 $15,000 to $24,999 32.7% 52.6% 1.1 3.5 $25,000 to $34,999 42.5% 59.5% 1.4 3.4 $35,000 to $49,999 50.8% 69.5% 1.8 3.6 $50,000 to $74,999 55.9% 78.2% 2.3 4.1 $75,000 to $99,999 60.3% 80.2% 2.8 4.6 $100,000+ 74.4% 94.3% 4.1 5.6 Vehicle Deficit Category of Household Zero-vehicle 23.1% 29.9% 0.8 3.3 Deficit (hard or soft) 37.9% 66.1% 1.4 3.6 Nondeficit (sufficient/surplus) 59.6% 75.7% 2.8 4.7 Note: An order consists of an instance when a household member "order[ed] something online and had it delivered" in the past 30 days. 249 Table 120. Which household members placed online orders in the last 30 days for people ages 16+. Self Only Other Household Member(s) Only Self and Other Household Member(s) Nobody in Household All 17.4% 18.9% 34.5% 29.2% MPO Tier 1. Atlanta MPO 18.7% 18.4% 38.4% 24.5% 2. Medium MPOs 16.7% 20.3% 34.4% 28.6% 3. Small MPOs 16.9% 20.6% 30.7% 31.8% 4. Non-MPO counties 14.7% 18.4% 25.7% 41.2% Sex Male 12.6% 23.3% 36.6% 27.5% Female 21.8% 15.0% 32.6% 30.7% Age Cohort Teens (1617) 4.0% 45.0% 25.5% 25.5% Millennial and Gen Z (1836) 15.7% 20.7% 39.5% 24.1% Gen X (3752) 20.1% 15.4% 40.9% 23.6% Pre-retirement age Boomer (5364) 18.5% 16.4% 32.1% 33.1% Senior (6579) 19.8% 17.5% 20.5% 42.2% Elderly (80+) 10.5% 22.2% 9.8% 57.5% Race White non-Hispanic only 17.2% 17.9% 42.4% 22.5% Black and Black multiracial 19.1% 19.1% 20.0% 41.8% Other 14.0% 22.8% 36.9% 26.2% Driver Status Nondriver 7.3% 27.0% 16.5% 49.2% Driver 18.8% 17.8% 37.1% 26.3% Mobility Impairment Absent 17.6% 18.7% 36.4% 27.4% Present 15.7% 21.5% 16.3% 46.5% Annual Household Income <$15,000 14.4% 13.5% 10.4% 61.7% $15,000 to $24,999 20.2% 19.8% 12.5% 47.4% $25,000 to $34,999 21.9% 17.0% 20.6% 40.5% $35,000 to $49,999 21.1% 18.7% 29.7% 30.5% $50,000 to $74,999 22.2% 22.4% 33.6% 21.8% $75,000 to $99,999 16.8% 19.9% 43.5% 19.8% $100,000+ 12.3% 19.9% 62.1% 5.7% Vehicle Deficit Category of Household Zero-vehicle 15.3% 6.9% 7.8% 70.1% Deficit (hard or soft) 11.2% 28.1% 26.7% 33.9% Nondeficit (sufficient/surplus) 20.0% 16.2% 39.5% 24.3% Note: An order consists of an instance when a household member "order[ed] something online and had it delivered" in the past 30 days. 250 Online Shopping and Travel Behavior Aside from online shopping during stay-at-home orders, such as during the COVID-19 pandemic, the extent to which online shopping replaces in-person shopping trips is an open question. Compared to Georgians who had not placed an online order within the past 30 days, a larger percentage of online shoppers made a trip to buy goods on the travel day (35.5 percent versus 29.7 percent) (table 121). Table 121. Percent of persons ages 16+ who made a trip to buy goods on the travel day. Online shoppers* All persons ages 16+ 35.5% MPO Tier 1. Atlanta MPO 35.0% 2. Medium MPOs 37.2% 3. Small MPOs 37.6% 4. Non-MPO counties 34.3% * Ordered something online and had it delivered within the past 30 days. Others 29.7% 27.8% 31.4% 29.7% 32.2% This may be partially explained by the fact that online shoppers tend to have higher incomes and be more mobile; thus, they tend to make more trips, in general, and to shop more at baseline, in particular. However, the NHTS data have a number of limitations for examining the links between online shopping and travel behavior. First, the NHTS provides a snapshot of a single day of travel behavior; shopping trips might be more accurately measured over a longer time span. Second, it is important to consider shopping trips at the household level. For example, a household may make one trip to the grocery store per week, so whether an individual household member visited the grocery store is dependent on whether or not another household member did so. Third, the type of goods purchased and amount being spent should be considered. To what 251 extent are online shoppers shifting their purchases of staple goods to the internet versus making specialty purchases that they might otherwise have foregone? 252 CHAPTER 5. SOCIAL INCLUSION AND EQUITY CHAPTER 5 SUMMARY Throughout, this report has drawn attention to intergroup differences in Georgians' travel behavior and access to transportation. This chapter will focus specifically on those differences that can result in reduced opportunity or quality of life for Georgians. Many underlying sources of social exclusion and economic inequality are not solely caused by the transportation system. However, a failure to consider the equity implications of transportation investment, policy, and planning decisions can exacerbate the negative impacts of poverty and prejudice. In this chapter, the researchers use key mobility indicators to examine transportation disadvantage based on economic and social exclusion. We devote particular attention to the needs of two groups not covered in depth elsewhere in the report: immigrants and people with mobility impairments. We also examine gendered inequalities stemming from intrahousehold decisions about how to allocate resources and tasks. Overview introduces the topic of social inclusion and discusses key findings from other chapters in this report as they relate to equity. This section also discusses the research team's choice of indicators for measuring equity. Key Equitable Mobility Indicators examines mobility disparities using key indicators such as travel day and chronic immobility, number of trips, and access to vehicles. We consider the travel patterns of immigrant Georgians, finding that they are less mobile and less likely to be drivers than nonimmigrants. They also own fewer vehicles. Ethnic and 253 transportation differences exist between immigrant Georgians of different education levels, but these travel patterns are extant among immigrants of all educational levels. Vehicle Access examines intrahousehold vehicle allocation and travel differences between vehicle-sufficient households and households that are not vehicle-sufficient. We find that captive transit users pay a time penalty for taking transit, not just in comparison to Georgians who use private autos, but also in comparison to choice transit users. Captive nonmotorized trips are also longer than choice nonmotorized trips, although (unlike the case with transit) the difference is based on different purposes for captive and choice trips rather than quality of service. Turning to intra-household vehicle allocation, we find that when there is a shortage of vehicles, women are less likely to be the primary driver of a vehicle. The exception to this is for female caregivers of young children, perhaps because the vehicle is needed for child-serving trips. Women are also less likely to be the recipients of newly purchased vehicles, but somewhat counterintuitively more likely to have the newest vehicle by model year. Among teen drivers, girls are more likely than boys to be given the newest household vehicle, whether because they are considered more trustworthy drivers, or because of greater concern about their safety in case of mechanical difficulties, or both. How Much and What For: The Interrelated Effects of Gender and Age on Mobility and Trip Purpose examines the interrelated effects of gender and age on travel behavior and immobility. This section shows that older women are at an elevated risk of becoming housebound, but because younger women are slightly more mobile than younger men, this risk may not be apparent in averages by gender alone. These results also draw 254 attention to the need to apply an intersectional lens to analysis of transportation needs (considering the interaction between multiple sources of social exclusion). Health and Disability examines equity concerns related to disability status and health. Mobility and physical activity are compared between Georgians with and without mobility impairments, and between different subgroups of Georgians with disabilities. The researchers identify a strong correlation between low income and poor health among both disabled and nondisabled Georgians. Furthermore, this health disparity persists even when comparing Georgians at similar levels of physical activity. We also discuss the ways in which elderly Georgians and those with mobility impairments adapt their travel behavior, and use logistic regression to examine risk factors for immobility among Georgians with mobility impairments. OVERVIEW Unequal mobility and transportation access can diminish Georgians' capability of participating fully in economic, social, and political life (Nussbaum 2003, World Bank 2013). Transportation policymakers and planners can mitigate inequality through social inclusion, or "the process of improving the ability, opportunity, and dignity of people, disadvantaged on the basis of their identity, to take part in society" (World Bank 2013, p. 4). Improved transportation can increase Georgians' employment prospects and educational opportunity, and reduce social isolation (Suzuki, Cervero, and Iuchi 2013; Vasconcellos 2001; World Bank 2013). Transportation disadvantage stems from two interconnected sources: (1) social exclusion, or barriers based on membership in a stigmatized or stereotyped group, and (2) poverty. The cumulative effect of social exclusion is often poverty. For example, 24 percent of Georgians 255 make less than $25,000 per year. However, as shown in figure 28, while just 14.7 percent of white men fall into this income category, 40.1 percent of Black women have a household income of less than $25,000. The effects of mobility impairments are even more striking: more than half of people with mobility impairments live in low-income households,75 and more than 70 percent of Black people with mobility impairments. It should be noted that the NHTS asks specifically about mobility impairments (a "condition or handicap that makes it difficult to travel outside of the home").76 While mobility impairments can also impede the ability to work, the need for a mobility aid such as a wheelchair in and of itself has no bearing on someone's capacity to engage in paid employment. However, a lack of high-quality, accessible, affordable transportation can pose a barrier to employment and education, regardless of the person's capacity to work or study (Bezyak et al. 2019, National Council on Disability 2015). In addition to physical barriers, people with mobility impairments also face social stigma (Papadimitriou 2008). 75 The rate is even higher for mobility-impaired people of working age than for people of retirement age (60 percent versus 46 percent), suggesting that this is not a result of higher rates of disability among the old and infirm. 76 Other disabilities, including intellectual and psychiatric disabilities, can also impact mobility (Bezyak et al. 2019). Based on the NHTS's question wording, it is unclear whether individuals with these conditions would have responded affirmatively to this question. 256 All Adults White male White female Black male Black female Latino male Latina female Other male Other female Adults without mobility impairment White male, nondisabled White female, nondisabled Black male, nondisabled Black female, nondisabled Latino male, nondisabled Latina female, nondisabled Other male, nondisabled Other female, nondisabled Adults with mobility impairment White male, disabled White female, disabled Black male, disabled Black female, disabled Latino male, disabled Latina female, disabled Other male, disabled Other female, disabled 24.0% 14.7% 18.3% 33.2% 40.2% 21.3% 32.1% 22.0% 16.9% 20.8% 12.9% 14.9% 28.8% 36.3% 19.9% 29.0% 20.1% 14.6% 54.1% 40.0% 45.3% 68.7% 69.9% 54.5% 59.6% 49.9% 45.6% Figure 28. Bar graph. Percent of adults with annual household income less than $25,000, by race, gender, and mobility impairment. Figure 28 also illustrates how the effects of multiple forms of oppression overlap. Women of color are doubly disadvantaged by their gender and race. The gender gap in low-income incidence is greatest among Latinos; 21.3 percent of men and 32.1 percent of women live in households earning less than $25,000 per year. Black Georgians are the most likely to be in lowincome households, and the gender gap is narrower than for Latinos. White Georgians have the 257 smallest percentage of low-income households and also the smallest gender gap. Disability is associated with a wider gendered low-income incidence gap among whites compared to the gap among nondisabled whites, but a narrower gap among Blacks and Latinos. To account for such intersectionality, it is important to consider the joint effect of multiple sources of disadvantage. In addition, the effects of social exclusion can persist long after officially sanctioned discrimination has been addressed. For example, while redlining and race-based exclusions from mortgage financing have been banned for 50 years, the assistance provided to white families half a century ago provided them with a "head start" on building wealth. African Americans facing housing and employment discrimination had fewer assets to pass on to their children, contributing to a racial wealth gap. Today, the median net worth of whites is 1,000 percent of that of Blacks (Jan 2017). Mainstreaming Equity One promising template for addressing inequality is what is known as gender mainstreaming, or "the process of assessing the implications for women and men of any planned action.... so that women and men benefit equally and inequality is not perpetuated" (United Nations Economic and Social Council 1997). Gender mainstreaming has two key features. First, gender mainstreaming recognizes that, because of pre-existing inequalities, gender-blind planning (sometimes referred to as gender "neutral" planning) tends to result in unequal outcomes. To achieve equitable outcomes, it is necessary to have a gender-aware process. 258 Second, under a gender mainstreaming approach, women's needs should not be thought of as a "special issue." Rather, transportation professionals should incorporate gender considerations into standard planning and analysis practices, for example by providing gender-disaggregated data and examining policies' differential impacts on men and women. To address a longstanding "gender data gap" (Criado-Perez 2019), studies focused specifically on gender are sometimes needed. However, in addition to such focused analysis, everyday policy and planning processes should also consider whether gender might affect the equity or efficacy of a proposed action. The idea of mainstreaming can be applied to other sources of inequality. As engineers and planners evaluate the effects of proposed projects on congestion mitigation or transit ridership, they should also examine whether the benefits of those projects will mitigate or exacerbate inequality based on gender, race, disability, and other sources of social exclusion. This report has followed an equity mainstreaming approach by incorporating analysis of equity into discussion of general travel and commuting patterns. The researchers will provide a centralized summary of key equity findings from elsewhere in the report before proceeding with more in-depth analyses. Key Equity Findings Race White non-Hispanic Georgians make more trips than Georgians of other races, and are less likely to be immobile on the travel day (see chapter 1, Household and Personal Mobility). Trips made by nonwhite Georgians are shorter in distance, and overall PMT per capita for nonwhite Georgians is lower than PMT per capita for white Georgians. White Georgians' trips are more likely to be by personal occupancy vehicle, and their VMT per capita is higher. There are also 259 racial differences in trip purpose. Whites make more discretionary trips and fewer trips to transport someone else (see chapter 1, Trip Purpose). Focusing on work travel, Black Georgians are more likely to make complex commutes than white Georgians and Georgians of other races (see chapter 2, Overview of Commuters). Black Georgians' commute distances and overall commute PMT are similar to those of whites, but their commute times are longer (see chapter 2, Demographic Differences and Demographic Differences in Commute Duration). Part of this difference in commute duration is due to Black commuters' higher usage of slower modes, such as transit and nonmotorized travel (see chapter 2, Commute Mode by Person). Black Georgians are more likely than Georgians of other races to spend 2 or more hours on their daily commute (see chapter 2, Total Daily Commute Burden). Black Georgians are also more likely to say that transit safety is a concern (see chapter 1, Transit Service Preferences Among Workers). Additionally, white Georgians are more likely to have the flexibility to set their own schedule or work from home than are Black Georgians and Georgians of other races (see chapter 3, Work Flexibility for Whom?). There are also racial differences in the uptake of new services and technologies. White Georgians are more likely to purchase alternate-fuel vehicles than are Georgians of other races, qualifying them for tax credits and other benefits of owning an expensive but efficient vehicle (see chapter 4, Alternative-fuel Vehicles). Similarly, white households are more likely to be online shoppers than nonwhite/mixed race households, and among online shopping households make more purchases on average (see chapter 4, Online Shopping). The users of carsharing and bikesharing, in contrast, are disproportionately Black (see chapter 4, Shared Vehicles: Bikesharing and Carsharing). There were not pronounced differences in the percent of people 260 who used ridehailing by gender or race. However, among ridehailing users, Black riders made more trips than whites and people of other races" (see chapter 4, Ridehailing). Black Georgians walk more than white Georgians and those of other races (e.g., Latino, Asian, Native American), but bike slightly less (see chapter 6, Travel Day Walking and Biking by Georgia Adults). Leisure accounts for 15 percent of Black Georgians' nonmotorized travel trips and legs, less than half the share for whites and Georgians of other races. NMT is more prominent as a mode of transit access/egress travel for Black Georgians. Among children, white children make fewer NMT trips than Blacks and children of other races (see chapter 6, Frequency of Nonmotorized Travel by Children). Among children who did walk or bike, average minutes of NMT is much higher for Black children (38.8) versus other races (25.7) and whites (21.4). In this chapter, the researchers will discuss racial differences in immobility and travel time and compare the mobility of immigrant and nonimmigrant Georgians (see chapter 5, Key Equitable Mobility Indicators). Income and Vehicle Ownership Low-income Georgians make fewer trips and are more likely to be immobile on the travel day (see chapter 1, Household and Personal Mobility). Trips, PMT, and VMT all increase with income (see chapter 1, Household and Personal Mobility). High-income people make more discretionary trips, though discretionary trips also make up a comparatively high share of trips by very low-income people (see chapter 1, Trip Purpose). It is possible that this latter spike is related to unemployment or underemployment: removal of commute and work-related trips from 261 the "pie" will necessarily increase the share (of a smaller total number) of trips pertaining to some other purposes. Nearly one third of the lowest-income households (those making less than $15,000 per year) own no vehicle (see chapter 1, Vehicle Ownership). This share decreases dramatically among the next-lowest income group; 6.9 percent of households making $15,00024,999 own zero vehicles. Just 0.1 percent of the highest-income households own zero vehicles. Low- and moderateincome households are also more likely to have a vehicle deficit (at least one vehicle, but fewer vehicles than potential drivers). Because zero-vehicle households are so uncommon outside of the lowest income bracket, these households are best thought of as carless by necessity rather than "car-free" by choice. Among vehicle-owning households, the lowest and highest income households are more likely to have purchased at least one vehicle in the past 12 months (see chapter 1, Vehicle Ownership). However, while the average age of a vehicle purchased by a household earning at least $100,000 (the top 22 percent) is 5 years, the average vehicle purchased by a household earning less than $15,000 per year (the bottom 17 percent) is more than 11 years old and has more than 130,000 miles on the odometer.77 Vehicles near the end of their useful lives are financially more accessible to low-income households, but higher maintenance costs, high-interest auto loans, and predatory sales practices for used vehicles mean that the poor pay a premium for basic mobility (Karger 2003). It also means that low-income drivers must replace their vehicles more frequently. 77 The difference in median age is even larger. The median age of a vehicle purchased by a high-income household is 2 years, while the median age for low-income households is 12 years. 262 The highest-income households (earning at least $100,000 per year) are substantially more likely to have hybrid or electric vehicles (see chapter 4, Alternative-fuel Vehicles). More than 9 percent of these households have at least one hybrid or electric vehicle, versus 6 percent of households making $75,00099,999 per year and fewer than 1 percent of vehicle-owning households making less than $25,000 per year. The income-based differences in vehicle age and quality are indicative of a de facto tax on poverty, where the poor pay more (Caplovitz 1967, Karger 2003). Trips for the purpose of transporting someone else spike among lower-middle income households ($35,000$44,999 annual income) (see chapter 1, Trip Purpose). These households are more likely to own a vehicle than lower-income households, but more likely to have a vehicle deficit than higher-income households. An adult in one of these lower-middle income households will make 73 more person trips per year than someone in the next income category down, but 22 of those trips, or 30 percent, will be to transport someone else. Vehicle-deficit households are more mobile than carless households, but the disproportionate spike in trips to transport others suggest that the increase in the utility derived from travel is somewhat smaller. Turning to work travel, commute times are longest for zero-vehicle households, and approximately equivalent in vehicle-deficit and nondeficit households (see chapter 2, Commute Times by Mode and Vehicle Ownership). By distance, in contrast, low-income workers have the shortest commutes and upper-middle income workers have the longest (see chapter 2, Demographic Differences). High-income workers are more likely to have the ability to set their own work schedules or work from home and are more likely to take advantage of that opportunity (see chapter 3, Descriptive Statistics on Access to Flexible Schedule and Work Location, and Travel Day Telecommuting). 263 Mode choice has a U-shaped relationship with income, with walking, biking, and transit being most common among low-income commuters, least common among middle-income commuters, and somewhat more common among high-income commuters (see chapter 2, Commute Mode by Person). Low- and high-income commuters' views of what makes transit a "good option" for commutes differ, with higher-income Georgians more likely to prioritize convenience (schedule and location of stops) and lower-income Georgians more likely to prioritize cost and safety (see chapter 1, Transit Service Preferences Among Workers). High-income commuters are more concerned about speed, but there were no significant differences by income in preferences for consistent travel times. However, there is a notable difference in level of service between captive and choice transit commuters: the average transit commute for a captive rider is 20.4 minutes longer than the average choice transit commute, despite the fact that the average distance of captive riders' commutes is 3.1 miles shorter than a choice rider's transit commute (see chapter 2, Commute Times by Mode and Vehicle Ownership). Commuters who, by necessity, use nonmotorized modes likewise walk or bike longer than commuters who have chosen these modes, suggesting that the "captive" pedestrians and cyclists documented in other countries (Pendakur 2011) also exist in Georgia. In chapter 5, Captive Travel will document that this difference between captive and choice travelers also applies to nonwork travel, and that the time penalty for captive transit users persists even after controlling for purpose, distance, and other variables. Turning to new technologies and services, bikeshare and carshare users come disproportionately from vehicle-deficit and zero-vehicle households (see chapter 4, Shared Vehicles: Bikesharing and Carsharing). However, among users, residents of vehicle-deficit and zero-vehicle households 264 use the service less than users from nondeficit households. The pattern for ridehailing usage is the opposite: adoption is higher among high-income people, but among users, low-income people use the service more frequently. Walking and biking are most common among residents of the lowest income (less than $15,000 per year) and highest income ($100,000 or more per year) households, compared to their incidence among other income groups (see chapter 6, Access and Egress Travel). However, lowincome pedestrians and cyclists are more likely to walk/bike to access transit or desired destinations (instrumental travel), whereas higher income Georgians are more likely to walk and bike for leisure. Captive pedestrians' and cyclists' instrumental trips are longer than those of choice pedestrians and cyclists, even after controlling for other factors (see chapter 6, Captive and Choice Nonmotorized Travel). Among children, the percent who walk or bike on the travel day is similar across income brackets. However, the amount of time spent walking or biking varies substantially: an average of 52 minutes for children in the lowest-income households (< $15,000 per year) versus 2226 minutes for children in other income brackets (see chapter 6, Frequency of Nonmotorized Travel by Children). In this chapter, the researchers will discuss the different conditions confronting captive and choice transit users (see chapter 5, Vehicle Access), the relationship between income and vehicle ownership and mobility (see chapter 5, Key Equitable Mobility Indicators), the association between disability and poverty (see chapter 5, Health and Disability), and how vehicle-deficit households allocate family vehicles (see chapter 5, Intra-Household Vehicle Allocation: Who Gets the Family Car?). 265 Gender On average, men and women make a similar number of person trips. However, men make more vehicle trips (see chapter 1, Household and Personal Mobility). The average distance for men's trips is greater, and men generate more PMT and VMT. Additionally, a higher percentage of women are immobile on the travel day. More detailed analysis of the relationship between gender, age, and immobility appears in chapter 5, How Much and What For: The Interrelated Effects of Gender and Age on Mobility and Trip Purpose. Women are more likely to make complex commutes than men, making it critical to accurately measure full work journeys rather than basing estimates on a single leg of the trip (see chapter 2, Overview of Commuters and Summary of Methods Implications for Upcoming Report Sections). Women's commute distances are shorter, but their commute times are not (see chapter 2, Demographic Differences in Commute Duration). Further, when commute distance is disaggregated by education level, it becomes apparent that the gender gap in distance only applies to workers without a college degree (see chapter 2, Demographic Differences). Women's complex commutes are at least in part a reflection of their elevated "second shift" responsibilities in maintaining the household and providing childcare. However, they have less access to work accommodations, such as teleworking and flexible scheduling, that would make it easier to meet these commitments (see chapter 3, Work Flexibility for Whom?). Women are more likely than men to commute by transit (see chapter 2, Commute Mode by Person), and more likely to list safety as an important aspect of transit service (see chapter 1, Transit Service Preferences Among Workers). Fewer women report walking and biking to be their "usual" commute mode, but when observed work journeys on the travel day are examined, 266 there is no meaningful gender difference in the rate of nonmotorized commuting (see chapter 2, Commute Mode by Person). Women are more likely to walk than men, but less likely to bike (see chapter 6, Overview). However, bikeshare users are more evenly divided by gender than cyclists as a whole (see chapter 4, Shared Mobility). Women are more likely than men to be discouraged from walking and biking by safety issues such as lack of sidewalks, heavy traffic, and inadequate night lighting (see chapter 6, Barriers to Walking and Biking More Frequently). Men are more likely to own AFVs than women (see chapter 4, Alternative-fuel Vehicles). Multicaregiver families with children are more likely than single adults to take advantage of online shopping. However, single-parent households, which are predominantly female-headed, are less likely to have placed orders online (see chapter 4, Online Shopping). In this chapter, the researchers will discuss how gender shapes women's mobility differently among different age groups: young women tend to be more mobile than average, while older women are substantially more likely to be immobile than older men and younger Georgians of all ages (see chapter 5, Key Equitable Mobility Indicators). This chapter will also discuss gender differences in vehicle allocation (see chapter 5, Intra-Household Vehicle Allocation: Who Gets the Family Car?) and the interaction between disability and gender (see chapter 5, Risk Factors for Immobility among Adults with Mobility Impairments). Age and Disability Travel day immobility is very high among the elderly and those with a mobility impairment (see chapter 1, Household and Personal Mobility). In addition to making fewer trips, trips by elderly 267 people and those with mobility impairments cover shorter distances. People with mobility impairments report very few work trips (see chapter 1, Trip Purpose). Younger adults--millennials and members of Gen Z--were more likely to use ridehailing apps and made more trips than older users; thus, ridehailing services may be underused as tools for improving seniors' mobility (see chapter 4, Ridehailing). Similarly, elderly Georgians are the least likely to take advantage of online shopping (see chapter 4, Online Shopping). Walking has a higher mode share for Georgians with mobility impairments than for those without; this counterintuitive finding is likely related to the fact that people with mobility impairments disproportionately find themselves in groups that walk more by necessity (lowincome, nondrivers, etc.) (see chapter 6, Travel Day Walking and Biking by Georgia Adults). Otherwise, walking and biking tend to decrease with age, though older adults make more leisure and loop trips than younger adults (see chapter 6, Travel Day Walking and Biking by Georgia Adults, and Model Structure). In this chapter, the researchers provide more in-depth analysis of the travel patterns and immobility of people with mobility impairments (see chapter 5, Health and Disability) and consider how they intersect with gender differences (see chapter 5, How Much and What For: The Interrelated Effects of Gender and Age on Mobility and Trip Purpose). We find that among Georgians with mobility impairments, groups who are already marginalized for other reasons-- the poor, elderly, and women--are more likely to be immobile than Georgians with disabilities from other backgrounds. We further find that existing mobility services for these populations are inadequate. 268 Choice of Equity Indicators The ultimate goal of increasing the equity of transportation is to improve Georgians' quality of life. Therefore, this report focuses on indicators that can best serve as proxies for quality of life in terms of access to destinations and time expended in order to access those destinations. From a utility perspective, the distance of each trip is less important than the existence of the trip and the time the traveler expends on it. As an equity indicator, travel distance (i.e., PMT and VMT) is less directly connected to the utility of that trip. Accordingly, this report's key equity indicators are: 1. Travel day trips per capita, or the total number of trips divided by the total adults in the population or subpopulation. 2. Travel day trips per capita, active travelers only. This measurement, which is based only on adults with at least one trip on the travel day, indicates how many trips are made on days when members of a given group do not stay in the same place all day. Together, indicators 1 and 2 document the total average mobility, and differentiate between mobility differences based on a reduced number of trips on a typical travel day and those stemming from a reduced number of travel days. 3. Travel day immobility (zero trips on the travel day). This indicator captures differences in how often members of different groups stay at home for an entire day. 4. Chronic immobility (zero trips in the past 7 days). This indicator helps differentiate between respondents who happened to stay home on the travel day and those whose travel is severely restricted. 269 The report also considers vehicle sufficiency (having enough vehicles for each potential driver in the household). Chapter 1, Vehicle Ownership has described demographic patterns in vehicle sufficiency at the household level. This chapter provides similar analysis for immigrants, who were not considered as a separate subpopulation in that section. The researchers then consider vehicle access within the household (identified by the primary driver for each vehicle), and vehicle access as a predictor of travel time. We are particularly interested in "captive" transit users, pedestrians, and cyclists. KEY EQUITABLE MOBILITY INDICATORS On an average day, 18.5 percent of Georgians make zero trips. However, over the course of a typical week, 97.6 percent of Georgia adults will leave home at least once. Not having made a single trip in the past 7 days, then, is an indicator of more chronic immobility. Georgians with mobility impairments are more than twice as likely as Georgians without mobility impairments to stay home on the travel day, but they are also six times as likely to be more chronically immobile. While Georgians with mobility impairments are the most likely to have been immobile for the past 7 days, 7-day immobility is also comparatively high among low-income people, residents of carless and vehicle-deficit households, Georgians with low educational attainment, and women. Seven-day immobility is lower among white non-Hispanic and Latino Georgians, and higher among Georgians who identify as Black, Asian, or some other race. In chapter 5, How Much and What For: The Interrelated Effects of Gender and Age on Mobility and Trip Purpose, the researchers use logistic regression to examine the interrelated effects of gender, age, and various demographic characteristics on the risk of chronic immobility. For example, after controlling for 270 other factors, being female is associated with an increased likelihood of being housebound, specifically among older adults. Table 122 shows summary demographic statistics for Georgians, and for two subpopulations of interest: (1) Georgians with mobility impairments, and (2) immigrants. According to the U.S. Census Bureau, Georgia's median household income is $55,679. Since the NHTS income data are categorical, the closest boundary marker available is at $50,000. Based on the NHTS data analyzed in this report, 46 percent of Georgians live in households earning less than $50,000 per year. However, 74 percent of Georgians with disabilities live in households earning less than $50,000. Educational attainment and workforce participation are also lower. As a reminder, participants were asked about mobility impairments and not impairments that directly affect academic or professional functioning. It should therefore not be assumed that these differences stem from impaired scholastic or professional abilities. Immigrants are less likely to reside in households earning less than $50,000 per year than nonimmigrants. As shown in table 123, this pattern is true for all immigrants without a 4-year college degree. Immigrants with a bachelor's degree or higher are somewhat likelier than nonimmigrants of the same education level to live in a lower-income household. There are several potential explanations for this. Some immigrants with professional certifications that are not recognized in the U.S. may find themselves working in a lower-paid industry. Additionally, many international students come to Georgia to study for advanced degrees, during which period they likely have lower incomes than their nonimmigrant counterparts with bachelor's or higher degrees, since many of the latter are already in regular employment whereas F-1 visa holders (the 271 category of visa held by most international students) have restrictions on such employment.78 In view of F-2 visa restrictions preventing the spouses of international students from working, nonimmigrants in graduate school are also more likely than immigrant graduate students to live in households with at least one regularly employed worker. Low-education and high-education immigrants also hail from different parts of the world. The majority of immigrants with a high school education or less are Latino, while a plurality of higher education immigrants identify as Asian or some other race. Table 124 shows differences in mobility by various demographic factors. On a typical day (averaging weekdays and weekends), 18.5 percent of Georgia adults stay home.79 This travel-day immobility is lowest in Atlanta and highest in non-MPO counties. Immobility decreases as education level increases. In terms of household income, immobility is highest among lowincome Georgians and lowest among those living in households earning $75,00099,999 per year. Vehicle-sufficient households have the lowest travel-day immobility. Interestingly, travelday immobility is higher among people in vehicle-deficit households than people in zero-vehicle houses. This may signify a pattern of consolidating trips into a smaller number of days for the former group: on active travel days, Georgians from vehicle-deficit households make more trips than Georgians from zero-vehicle households. On an average day, 18.5 percent of Georgians make zero trips. However, over the course of a typical week, 97.6 percent of Georgia adults will leave home at least once. Not having made a single trip in the past 7 days, then, is an indicator of more chronic immobility. Georgians with 78 https://www.uscis.gov/working-united-states/students-and-exchange-visitors/students-and-employment. 79 For weekdays only, this figure is 16.1 percent. 272 mobility impairments are more than twice as likely as Georgians without mobility impairments to stay home on the travel day, but they are also six times as likely to be more chronically immobile. While Georgians with mobility impairments are the most likely to have been immobile for the past 7 days, 7-day immobility is also comparatively high among low-income people, residents of carless and vehicle-deficit households, Georgians with low educational attainment, and women. Seven-day immobility is lower among white non-Hispanic and Latino Georgians, and higher among Georgians who identify as Black, Asian, or some other race. In chapter 5, How Much and What For: The Interrelated Effects of Gender and Age on Mobility and Trip Purpose, the researchers use logistic regression to examine the interrelated effects of gender, age, and various demographic characteristics on the risk of chronic immobility. For example, after controlling for other factors, being female is associated with an increased likelihood of being housebound, specifically among older adults. 273 Table 122. Demographic characteristics by disability and national origin. Population (thousands)* MPO Tier 1. Atlanta MPO 2. Medium MPOs 3. Small MPOs 4. Non-MPO Race White non-Hispanic only Black and Black multiracial Hispanic (any race) Asian or other Education Level High school or less Some college or associate degree Mobility Impairment National Origin All Adults Absent Present Nonimmigrant 7,704 6,967 732 6,771 Immigrant 927 54.1% 15.8% 10.1% 20.0% 55.1% 15.9% 10.3% 18.7% 44.6% 16.9% 9.2% 29.3% 52.0% 16.3% 10.7% 21.0% 73.5% 13.5% 5.9% 7.1% 55.2% 30.8% 8.5% 5.5% 53.6% 31.5% 9.2% 5.7% 51.4% 37.2% 7.0% 4.5% 57.9% 33.6% 5.8% 2.7% 15.2% 18.1% 36.1% 30.6% 31.8% 34.6% 57.5% 37.3% 31.5% 30.5% 28.7% 25.5% 29.3% 21.0% Bachelor's degree or higher 37.7% 36.7% 17.0% 33.3% 47.4% Annual Household Income <$15,000 14.6% 12.3% 36.8% 15.0% 8.4% $15,000 to $24,999 9.4% 9.3% 16.6% 10.0% 8.4% $25,000 to $34,999 10.1% 10.1% 11.5% 9.7% 13.9% $35,000 to $49,999 12.0% 12.0% 9.1% 11.7% 12.2% $50,000 to $74,999 16.6% 16.6% 12.7% 16.4% 15.4% $75,000 to $99,999 11.8% 12.2% 6.5% 11.4% 15.1% $100,000+ 25.5% 27.6% 6.8% 25.8% 26.7% Household Vehicle Ownership Zero-vehicle 4.9% 4.0% 16.0% 5.1% 3.9% Vehicle-deficit 25.9% 25.3% 38.8% 25.1% 37.0% Vehicle-sufficient 69.2% 70.7% 45.2% 69.8% 59.2% Gender Male 47.9% 49.2% 41.2% 48.9% 45.8% Female 52.1% 50.8% 58.8% 51.1% 54.2% Mobility Impairment Absent 90.5% - - 89.9% 95.2% Present 9.5% - - 10.1% 4.8% National Origin Nonimmigrant 88.0% 87.3% 94.0% - - Immigrant 12.0% 12.7% 6.0% - - * Unweighted sample size: 15,222. Mobility impairment is unknown for 8 respondents (weighted to represent 4,400 Georgians) and national origin is unknown for 7 respondents (weighted to represent 6,000 Georgians). Excluding Black Hispanic. Vehicle-deficit households have at least one vehicle, but fewer vehicles than potential drivers. Vehicle-sufficient households have at least one vehicle for every potential driver. 274 Table 123. Demographics of nonimmigrant and immigrant Georgians by level of education. Category Race White non-Hispanic only Black and Black multiracial Nonimmigrant Low Medium High HS or Less Some College Bachelor's or Associate or Greater Immigrant Low Medium HS or Less Some College or Associate High Bachelor's or Greater 52.6% 39.5% 57.3% 34.7% 69.6% 24.7% 8.3% 15.2% 13.6% 28.0% 19.6% 16.6% Hispanic (any race) 5.5% 5.5% 3.8% 58.4% 35.6% 22.5% Asian or other Annual Household Income <$15,000 2.5% 27.3% 2.5% 12.5% 2.0% 3.6% 18.1% 16.0% 22.8% 5.7% 41.3% 5.0% $15,000 to $24,999 13.6% 11.1% 4.0% 14.2% 8.5% 5.0% $25,000 to $34,999 13.1% 9.8% 6.0% 19.1% 15.4% 7.5% $35,000 to $49,999 11.2% 15.3% 9.4% 14.0% 15.9% 9.7% $50,000 to $74,999 15.2% 19.3% 16.5% 12.9% 19.1% 15.0% $75,000 to $99,999 7.0% 12.2% 15.0% 11.0% 14.6% 18.2% $100,000+ Household Vehicle Ownership Zero-vehicle 12.5% 10.2% 19.8% 3.4% 45.7% 0.8% 12.7% 4.8% 20.6% 1.5% 39.7% 4.1% Vehicle-deficit 36.3% 24.5% 14.6% 48.0% 36.6% 30.2% Vehicle-sufficient MPO Tier 1. Atlanta MPO 53.6% 43.3% 72.1% 49.3% 84.5% 64.0% 47.2% 66.0% 62.0% 73.0% 65.8% 77.9% 2. Medium MPOs 15.8% 17.4% 15.5% 11.5% 10.9% 15.9% 3. Small MPOs 12.0% 12.1% 7.9% 9.9% 6.8% 3.1% 4. Non-MPO Gender Male 28.9% 50.7% 21.3% 47.2% 12.7% 46.8% 12.6% 47.1% 9.3% 40.7% 3.0% 44.8% Female Excluding Black Hispanic. 49.3% 52.8% 53.2% 52.9% 59.3% 55.2% Vehicle-deficit households have at least one vehicle, but fewer vehicles than potential drivers. Vehicle-sufficient households have at least one vehicle for every potential driver. 275 Table 124. Key mobility indicators for Georgia adults by geography, race, education, income, vehicle ownership, gender, and disability. Daily Trips Per Capita Active* Travelers Only All adults 1.86 2.27 MPO Tier 1. Atlanta MPO 3.45 4.17 2. Medium MPOs 3.30 4.08 3. Small MPOs 3.36 4.11 4. Non-MPO 3.10 3.99 Race White non-Hispanic only 3.38 4.10 Black and Black multiracial 3.34 4.17 Hispanic (any race) 3.39 4.14 Asian or other 2.95 3.90 Education Level High school or less 2.84 3.82 Some college or associate 3.30 4.05 Bachelor's degree or higher 3.81 4.38 Annual Household Income <$15,000 2.94 4.04 $15,000 to $24,999 3.04 3.90 $25,000 to $34,999 3.35 4.13 $35,000 to $49,999 3.55 4.27 $50,000 to $74,999 3.37 4.10 $75,000 to $99,999 3.56 4.11 $100,000+ 3.55 4.20 Household Vehicle Ownership Immobility Travel Past Seven Day Days 18.5% 2.4% 17.1% 19.2% 18.0% 21.9% 2.2% 2.5% 2.4% 2.7% 17.4% 19.7% 18.0% 24.1% 1.5% 3.6% 1.6% 6.2% 25.2% 18.6% 12.8% 3.7% 2.4% 1.2% 27.0% 22.1% 18.7% 16.7% 17.8% 13.4% 15.3% 3.9% 3.2% 2.0% 3.1% 2.9% 2.5% 0.6% Zero-vehicle Vehicle-deficit Vehicle-sufficient Gender Male Female Mobility Impairment Absent Present 2.82 3.70 22.9% 5.0% 3.03 4.10 26.1% 4.8% 3.50 4.14 15.4% 1.3% 3.33 3.99 16.5% 1.8% 3.36 4.23 20.3% 2.9% 3.45 4.13 16.3% 1.6% 2.32 3.84 39.5% 10.0% * Reporting at least one trip on the travel day. Excluding Black Hispanic. Vehicle-deficit households have at least one vehicle, but fewer vehicles than potential drivers. Vehicle-sufficient households have at least one vehicle for every potential driver. Table 125 compares mobility by educational attainment between immigrant and nonimmigrant Georgians. Table 126 shows what portion of Georgians are drivers and how many vehicles their 276 household owns, subdivided by national origin, race, and other demographic variables. This report classifies vehicle ownership status based on the ratio between household vehicles and potential drivers (household members ages 16+). A vehicle-deficit household has at least one vehicle, but fewer vehicles than potential drivers. A nondeficit (or vehicle-sufficient) household has at least one vehicle for each potential driver. Table 125. Mobility indicators by national origin and educational attainment among Georgia adults. Taken together, the two tables show that immigrant Georgians are somewhat less mobile than nonimmigrant Georgians at all educational levels. Immigrants of all education levels are less likely to be drivers than nonimmigrants. They are also less likely to live in vehicle-sufficient households and more likely to live in vehicle-deficit households. Because vehicle-deficit households are especially common among immigrants, low-education immigrants are also less likely than low-education nonimmigrants to live in carless households. High-education 277 immigrants, in contrast, are slightly more likely to be carless than their nonimmigrant counterparts. Table 126. Driver status and vehicle ownership by Georgia adults of different demographic characteristics. Vehicle Sufficiency* Driver Status "Does this person drive?" Nondeficit At least one household vehicle for each potential driver Deficit At least one household vehicle, but fewer vehicles than potential drivers ZeroVehicle Carless household All Adults 89.0% 69.2% 25.9% 4.9% Mobility Impairment Absent 95.1% 82.2% 15.1% 2.7% Present 67.2% 59.3% 28.2% 12.5% National Origin Nonimmigrant 92.5% 80.7% 15.5% 3.8% Immigrant 87.8% 69.6% 27.2% 3.2% Nonimmigrants by Education Level High school or less 81.8% 66.4% 24.3% 9.3% Some college or associate degree 95.2% 81.4% 15.9% 2.7% Bachelor's or higher 97.9% 89.7% 9.4% 0.9% Immigrants by Education Level High school or less 73.6% 52.8% 41.3% 5.9% Some college or associate degree 90.8% 76.4% 20.8% 2.8% Bachelor's or higher 92.7% 74.2% 23.5% 2.3% Race White non-Hispanic only 94.9% 85.6% 12.9% 1.5% Black and Black multiracial 84.6% 65.4% 23.9% 10.8% Hispanic (any race) 89.1% 67.2% 28.5% 4.4% Asian or other 88.4% 68.6% 27.1% 4.3% Annual Household Income <$35,000 82.4% 61.1% 26.9% 12.0% $35,000 to $49,999 92.7% 77.1% 22.3% 0.5% $50,000 to $74,999 95.3% 85.5% 13.9% 0.6% $75,000 to $99,999 96.7% 88.1% 11.7% 0.1% $100,000+ 98.3% 93.2% 6.8% 0.1% Sex Male 93.4% 80.5% 16.2% 3.3% Female 91.1% 79.1% 16.7% 4.1% * Like income, vehicle sufficiency is calculated at the household level. The figures here summarize the status of adults within those households. They do not indicate whether an individual is the main driver for any of the household vehicles. This is the question wording used by NHTS. NHTS does not ask about driver's licensing. 278 VEHICLE ACCESS In chapter 5, Key Equitable Mobility Indicators, this report documented demographic differences in vehicle availability. This section examines the effects of vehicle access (or the lack thereof) on the travel time of "captive travelers" who take transit, walk, or bike. The researchers find that captive travelers are doubly penalized. First, transit and walking are, in most cases, slower than driving. Second, captive transit users receive an additional time penalty in comparison to similar trips by choice riders. Finally, we consider how households allocate vehicles among their members. In vehicle-deficit households, we find that women are less likely to be the main driver of a household vehicle than are men. This effect is reversed, however, for female caregivers of young children. Captive Travel Conceptually, transit users, cyclists, or pedestrians are "captive" when they are traveling by that mode out of necessity because a private automobile is not available. A choice traveler, in contrast, has the option of taking a private auto, but opts not to do so. In this analysis, a captive transit or nonmotorized trip meets the following criteria: 1. The traveler is from a carless or vehicle-deficit household. The researchers include vehicle-deficit households in the definition of captive travel because, while a car is available to some household members for some trips, for other trips, household members will have more limited options.80 80 Additionally, in Georgia, vehicle-sufficient households are the norm. The overwhelming majority of Georgia households (74 percent) have a vehicle for every potential driver. Zero-vehicle households are very uncommon (7 percent). In a state or country where zero-vehicle and vehicle-deficit households are more common, it might make sense to subdivide captive travel from travelers in each of those types of households. 279 2. The household earns less than $50,000 per year.81 The income criterion is designed to screen out travel by people who are "car-free" by choice versus those who are financially unable to afford vehicles for every potential driver in the household. Table 127 presents the distribution of trip purposes by mode, distinguishing between choice and captive trips for the transit and NMT modes. Table 127 also presents the unweighted sample sizes for the trips on which all descriptive and statistical analyses in this section are based. As table 127 shows, the purposes of choice transit and NMT trips differ from the purposes of captive trips. A plurality of choice NMT trips are for recreation or fitness, or travel for its own sake. Captive pedestrians and cyclists, in contrast, are usually walking or biking because they are trying to reach a destination. On the transit side, trips to work account for nearly a third of choice trips, compared to just 22 percent of captive trips. Medical and dental trips account for 8 percent of captive transit trips versus just 1 percent of choice transit trips. 81 We chose this value because it is the closest approximation of Georgia's median income ($55,679) achievable with NHTS data. We conducted sensitivity analysis using different income cutoffs, and the models presented here were found to be robust. 280 Table 127. Purpose by mode for captive, choice, and POV trips (weighted percent and unweighted sample sizes). Purpose for Captive*, Choice, and POV Trips (weighted) Purpose NMT (walk or bike) Transit Other Choice Captive Choice Captive POV Mode All Work or school 13.4% 6.4% 31.7% 22.1% 15.9% 20.0% 15.7% Household-serving travel 9.3% 25.7% 24.0% 23.0% 29.0% 12.2% 27.6% Recreation or fitness 40.9% 21.1% 5.7% 1.6% 4.4% 6.6% 6.9% Other discretionary trip 12.1% 13.4% 4.1% 8.4% 14.5% 12.7% 14.2% Medical or dental 0.3% 1.0% 1.0% 7.9% 1.6% 5.8% 1.6% Return home 21.9% 31.0% 29.2% 35.5% 33.9% 36.8% 33.2% Other purpose Unweighted Sample Sizes 2.1% 1.4% 4.2% 1.5% 0.7% 5.8% 0.9% NMT (walk or bike) Transit Choice Captive Choice Captive POV Other Mode Total Work or school 306 55 106 65 6,736 89 7,357 Household-serving travel 270 247 48 85 13,214 53 13,917 Recreation or fitness 1,197 185 12 7 1,950 39 3,390 Other discretionary trip 300 122 17 31 6,591 37 7,098 Medical or dental 11 15 5 24 982 30 1,067 Return home 530 283 79 113 14,845 132 15,982 Other purpose 58 18 11 5 311 19 422 Total 2,672 925 278 330 44,629 399 * Transit and nonmotorized trips are considered captive if the traveler is from a vehicle-deficit or zero-vehicle household with an annual household income <$50,000. Trips by travelers that do not meet these criteria are categorized as choice. Based on sample of trips by Georgia residents that take place entirely within Georgia. Excludes 1,270 POV trips and 131 non-POV trips for which income is unknown, as well as 77 POV and 13 non-POV trips for which duration, purpose, or disability status are unknown. Mean trip duration differs not just by mode, but between captive and choice users of the same mode (table 128). The difference is especially notable among transit users, where the mean travel time for a captive transit trip is 20 minutes longer than that of choice users. The average durations of captive and choice NMT trips are similar, largely because a preponderance of lengthy recreation trips by choice pedestrians and cyclists counterbalances the fact that captive 281 pedestrians' and cyclists' nonrecreational trips are longer than those of choice walkers and bikers. Table 128. Mean duration and distance of captive, choice, and POV trips. Mode Choice NMT (walk or bike) Percent of Trips 5.5% Mean Duration in Minutes All Recreation and Purposes Fitness Trips 18.3 26.4 Nonrec. Only Mean Distance (Miles) Nonrec. Trips Only* 12.7 0.7 Captive NMT (walk or bike) 2.9% 19.2 23.3 18.1 0.8 Choice transit 1.1% 43.0 78.2 40.8 11.2 Captive transit 1.1% 63.4 60.6 63.4 9.7 POV incl. rental car 88.5% 21.9 26.4 21.7 9.1 Other mode 1.0% 31.8 21.7 32.5 12.6 * Recreational trips and fitness trips are excluded from distance because the NHTS's "shortest path" method of calculating distance is expected to be less accurate for trips where the traveler is likely to choose a circuitous route. The distance column also excludes 3 POV trips, 29 NMT trips, 3 transit trips, and 3 other trips for which distance is missing. Transit and nonmotorized trips are considered captive if the traveler is from a vehicle-deficit or zero-vehicle household with an annual household income <$50,000. Trips by travelers that do not meet these criteria are categorized as choice. Because differences in duration between choice and captive travelers are confounded with differences in trip purpose and other factors, we used linear regression to isolate the effect of captivity on trip duration for transit and NMT trips (table 129). We find that captive transit users face a time penalty of 15.7 minutes per trip compared to choice transit users, after controlling for trip purpose, MPO, and demographic factors. Captive pedestrians and cyclists do not face the same captivity time penalty as captive transit users. Rather, the longer average duration of captive NMT trips is better explained by the dominance of purposes that are associated with longer walk or cycle trips. 282 Table 129. Linear regressions: Effect of captive travel on duration of transit and nonmotorized trips. 283 Table 130 shows an alternate estimation of the time penalty on captive transit trips in comparison not just with choice transit trips, but also with trips by POV. The model controls for the same purpose and demographic variables, and additionally controls for trip distance. This model estimates that a trip by public transit will take approximately 10 minutes longer than a comparable trip by POV (or 17.7 minutes in Atlanta). The time penalty for captive transit trips is an additional 20.2 minutes. In other words, the full time penalty for a captive transit user instead of using a POV if it had been available is 38 minutes per trip in Atlanta and 30 minutes per trip elsewhere in the state. A traveler who chooses transit for two trips a day is accepting an average increase of 2035 minutes in travel time, perhaps in exchange for increased convenience, lower cost, or other benefits (see chapter 1, Transit Service Preferences Among Workers). Captive transit users, on the other hand, lose more than an hour of their day, regardless of their mode preferences. 284 Table 130. Linear regression: Duration in minutes of trips by POV and transit. 285 Intra-Household Vehicle Allocation: Who Gets the Family Car? Not every household member has equal access to the family car. As shown in table 133, overall, 90 percent of Georgia drivers (ages 16+) are listed as the main driver for at least one family vehicle (in other words, it is "their") car. In vehicle-deficit households, only 70 percent of drivers have their own car, versus 97.5 percent of drivers in nondeficit households. When vehicles are scarce, to whom do households allocate them? We examine this question by using logistic regression to model the probability that a driver ages 16+ will be the main driver of a household vehicle in three circumstances: 1. Among households with a vehicle deficit (where multiple drivers must compete for a limited number of vehicles), we model the probability that a driver is listed as the main driver for any household vehicle. 2. Among households that acquired a vehicle in the past 12 months, we model the probability that a household member will be listed as the main driver for a newly purchased vehicle (whether that vehicle was new or used at the time of purchase). 3. Among households with at least two vehicles, we model the probability of being the main driver for the vehicle with the most recent model year (i.e., the newest vehicle chronologically). Table 131 and table 132 present the weighted and unweighted sample distributions on which these models are based. While the researchers do not present a model based on drivers in all households with 2+ vehicles (deficit and nondeficit combined), we present descriptive statistics about this population. Accordingly, we also provide a sample distribution for this group. 286 Table 131. Weighted sample characteristics of drivers by household vehicle ownership. Driver Characteristics Household Type 2+ Potential Drivers Vehicle-Deficit 2+ Household Vehicles Newly Purchased Vehicle(s) All drivers ages 16+ Male Female By Caregiver Status 100.0% 51.1% 48.9% 100.0% 47.3% 52.7% 100.0% 52.4% 47.6% 100.0% 50.9% 49.1% Male noncaregiver Female noncaregiver Male caregiver, youngest 04 Female caregiver, youngest 04 Male caregiver, youngest 515 Female caregiver, youngest 515 33.5% 30.5% 9.1% 9.1% 8.5% 9.2% 31.7% 33.3% 8.4% 9.1% 7.2% 10.4% 33.9% 29.6% 9.3% 9.1% 9.2% 8.9% 32.4% 28.7% 9.5% 10.6% 9.0% 9.8% Age Teen (1617) Working-age adult (1864) Senior (6579) Elderly (80+) 2.4% 84.1% 11.9% 1.6% 3.0% 84.3% 10.7% 1.9% 1.9% 85.3% 11.5% 1.3% 3.9% 87.3% 8.0% 0.8% Worker Status Nonworker Worker 34.4% 65.6% 43.5% 56.5% 30.4% 69.6% 33.9% 66.1% Mobility Impairment Absent Present 94.2% 5.8% 90.8% 9.2% 90.8% 9.2% 95.1% 4.9% Vehicle Sufficiency Nondeficit, 2+ potential drivers 73.6% 0.0% 87.8% 78.3% Deficit of 1+ vehicles 26.4% 100.0% 12.2% 21.7% Deficit of exactly 1 vehicle 19.6% 74.3% 9.6% 15.3% Deficit of 2+ vehicles 6.8% 25.7% 2.6% 6.4% Number of Household Vehicles 1 13.8% 52.3% 0.0% 54.6% 2 43.6% 31.4% 54.8% 54.1% 3+ 42.6% 16.3% 41.5% 43.4% Unweighted sample size 11,042 1,954 9,513 3,532 Note: Weighted column percentages shown. For consistency between models and descriptive tables, 348 observations with missing household income have been excluded from this table. Based on drivers in households with two or more members of driving age (16+). Vehicle-deficit households have at least one vehicle, but fewer vehicles than potential drivers. Household has purchased a new or used vehicle within past 12 months. A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. 287 Table 132. Unweighted sample characteristics of drivers by household vehicle ownership. Driver Characteristics 2+ Potential Drivers Household Type Vehicle-Deficit 2+ Household Vehicles Newly-Purchased Vehicle(s) All drivers ages 16+ Male Female By Caregiver Status 11,042 48.3% 51.7% 1,954 45.1% 54.9% 9,513 49.5% 50.5% 3,532 47.9% 52.1% Male noncaregiver Female noncaregiver Male caregiver, youngest 04 Female caregiver, youngest 04 Male caregiver, youngest 515 Female caregiver, youngest 515 Age Teen (1617) Working-age adult (1864) Senior (6579) Elderly (80+) Worker Status Nonworker Worker Mobility Impairment Absent Present Vehicle Sufficiency Nondeficit, 2+ potential drivers Deficit of 1+ vehicles Deficit of exactly 1 vehicle Deficit of 2+ vehicles Number of Household Vehicles 1 2 3+ 36.5% 38.1% 5.8% 6.6% 6.0% 7.1% 1.9% 73.3% 21.7% 3.2% 41.3% 58.7% 93.6% 6.4% 82.3% 17.7% 14.6% 3.1% 0.0% 10.3% 48.4% 41.4% 33.7% 39.4% 5.5% 6.9% 5.8% 8.6% 3.0% 73.6% 18.9% 4.5% 51.9% 48.1% 88.4% 11.6% 0.0% 100.0% 82.6% 17.4% 0.0% 58.0% 29.4% 12.5% 37.1% 36.8% 6.0% 6.7% 6.4% 7.0% 1.5% 74.7% 21.2% 2.6% 37.9% 62.2% 95.1% 4.9% 93.1% 6.9% 5.7% 1.2% 0.0% 0.0% 53.4% 46.6% 35.0% 36.3% 6.5% 7.8% 6.5% 8.0% 3.3% 78.6% 16.3% 1.8% 37.4% 62.6% 94.3% 5.8% 85.6% 14.4% 11.4% 3.0% 0.0% 5.8% 38.8% 55.4% Note: Unweighted column percentages shown. For consistency between models and descriptive tables, 348 observations with missing household income have been excluded from this table. Based on drivers in households with two or more members of driving age (16+). Vehicle-deficit households have at least one vehicle, but fewer vehicles than potential drivers. Household has purchased a new or used vehicle within past 12 months. A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. To understand vehicle access, the research team examined the percentage of drivers who are listed as the "main driver" for a household vehicle (table 133). Because household members may share access to a vehicle, not having a vehicle that is officially "theirs" does not necessarily 288 mean a driver never has vehicle access. However, someone who is the main driver for a vehicle likely has more reliable access than someone who is not. Unsurprisingly, the larger the vehicle deficit is in any given household, the less likely individual drivers are to have their own vehicle. Among households with at least two drivers, 98 percent of drivers in nondeficit households are the main driver of a household vehicle. In contrast, 74 percent of drivers in a household with a one-vehicle deficit are listed as a main vehicle driver, and only 56 percent of those in a household with a deficit of two or more vehicles. On average, women are less likely than men to be the main driver of a household vehicle, though this pattern reverses among caregivers for young children. Nonworkers and people with mobility impairments are less likely to be main drivers than workers and nondisabled household members, respectively. Analysis using logistic regression indicates that these patterns continue to be significant after controlling for other factors (table 134). 289 Table 133. Percent of drivers listed as the "main" driver for a household vehicle. Personal Vehicle Access: Main Driver of a Household Vehicle* All Households with 2+ Potential Drivers Vehicle-Deficit Households Only Nondeficit Households Only All drivers ages 16+ Male Female By Caregiver Status 90.1% 91.4% 88.8% 69.6% 70.5% 68.8% 97.5% 98.1% 96.8% Male noncaregiver Female noncaregiver Male caregiver, youngest 04 Female caregiver, youngest 04 90.6% 87.3% 91.1% 90.3% 69.5% 64.5% 67.8% 72.3% 97.6% 96.6% 98.5% 96.6% Male caregiver, youngest 515 94.6% 77.6% 99.5% Female caregiver, youngest 515 92.2% 79.4% 97.6% Age Teen (1617) 65.8% 18.0% 89.5% Working-age adult (1864) 91.1% 71.6% 98.0% Senior (6579) 89.4% 67.5% 96.2% Elderly (80+) 82.8% 72.3% 87.7% Worker Status Nonworker 81.6% 55.7% 94.6% Worker 94.6% 80.2% 98.8% Mobility Impairment Absent 90.9% 70.8% 97.8% Present 77.5% 57.7% 91.4% Vehicle Sufficiency Nondeficit, 2+ potential drivers 97.5% - 97.5% Deficit of 1+ vehicles 69.6% 69.6% - Deficit of exactly 1 vehicle 74.3% 74.3% - Deficit of 2+ vehicles 56.0% 56.0% - Unweighted sample size 11,042 1,954 9,088 * Participant is listed as the "main driver" for one or more household vehicles. Based on drivers in households with two or more members of driving age (16+). For consistency between descriptive analysis and models, households with missing income values have been excluded from this table. Vehicle-deficit households have at least one vehicle, but fewer vehicles than potential drivers. A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. 290 Table 134. Logistic regression: Vehicle allocation within vehicle-deficit households. Logistic Regression: Probability of Being Main Driver for Any Household Vehicle Odds Robust Ratio SE T P-Value Female 0.633 0.087 Caregiver status by age of youngest child (reference: noncaregiver) -3.32 0.001 *** Caregiver, youngest child 04 years old 0.543 0.135 -2.45 0.014 ** Caregiver, youngest child 515 0.971 0.252 -0.11 0.911 Female x Caregiver, youngest child 04 2.655 0.993 2.61 0.009 *** Female x Caregiver, youngest child 515 1.928 0.702 1.80 0.071 * Age Age2 Annual household income (reference: <$35,000) 1.097 0.0155 0.999 0.000141 6.59 -5.09 0.000 *** 0.000 *** $35,000 to $49,999 0.986 0.109 -0.13 0.896 $50,000 to $74,999 0.912 0.105 -0.80 0.422 $75,000 to $99,999 1.119 0.183 0.68 0.494 $100,000+ 1.056 0.149 0.38 0.702 Worker 3.050 0.393 8.66 0.000 *** Mobility impairment Deficit of 2+ vehicles (versus deficit of exactly 1) 0.687 0.532 0.114 0.057 -2.25 -5.84 0.024 ** 0.000 *** Constant 0.161 0.052 -5.69 0.000 *** Model Indicators Number of cases, N 1,954 Final log likelihood, LL() -1,078 Log likelihood of constants-only model, LL(C) McFadden's Pseudo-R2: 1-LL()/LL(C) -1,198 0.0999 Sample: drivers age 16+ in vehicle-deficit households (households with at least one vehicle, but fewer vehicles than potential drivers). A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. Deficit size is the number of potential drivers in a household minus the number of vehicles. A related question concerns the quality of household vehicles, which we examine by looking at the newest household vehicle. We consider vehicles that were the newest chronologically (most recent model year) and also vehicles that were newly purchased, whether new or used at time of purchase. As shown in table 135, women are less likely to be recipients of a newly purchased 291 vehicle, but, especially for caregivers, more likely to have the newest vehicle by model year. Logistic regression indicates that these patterns hold even after controlling for worker status, mobility impairment, and other factors (table 136). Table 135. Percent of drivers who are listed as main driver for newest household vehicle. By Model Year Newly Purchased All drivers ages 16+ Male Female By Caregiver Status Male noncaregiver Female noncaregiver Male caregiver, youngest 04 Female caregiver, youngest 04 48.1% 41.1% 55.9% 42.2% 52.3% 38.3% 60.9% 48.1% 49.5% 46.6% 49.9% 43.1% 51.8% 51.7% Male caregiver, youngest 515 Female caregiver, youngest 515 Age Teen (1617) Working-age adult (1864) Senior (6579) Elderly (80+) 39.8% 62.6% 23.5% 48.2% 51.1% 55.0% 46.0% 51.6% 43.3% 48.4% 48.9% 34.9% Worker Status Nonworker 48.1% 43.6% Worker 48.1% 50.5% Mobility Impairment Absent 48.2% 48.5% Present 47.3% 40.2% Number of Household Vehicles 1 - 54.6% 2 54.8% 54.1% 3+ 41.5% 43.4% Unweighted sample size 9,513 3,532 * For consistency between descriptive analysis and models, households with missing income values have been excluded from this table. Sample: drivers age 16+ in household with 2+ vehicles and 2+ potential drivers. In case of a tie, both vehicles are coded as newest. Vehicles purchased in the past 12 months, whether purchased new or used. Based on sample of drivers age 16+ in household with newly purchased vehicle(s). A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. 292 Table 136. Logistic regression: Intra-household allocation of new vehicles. Newly Purchased By Model Year Odds Ratio P-Value Odds Ratio P-Value Female 0.692 0.000 *** Caregiver status by age of youngest child (reference: noncaregiver) 1.623 0.000 *** Caregiver, youngest child 04 years old 0.876 0.383 0.838 0.063 * Caregiver, youngest child 515 0.701 0.015 ** 0.813 0.023 ** Female x Caregiver, youngest child 04 1.420 0.155 1.823 0.000 *** Female x Caregiver, youngest child 515 1.908 0.005 *** 1.653 0.002 *** Age Age2 Annual household income (reference: <$35,000) 1.003 1.000 0.750 0.992 1.044 0.9997 0.000 *** 0.000 *** $35,000 to $49,999 0.986 0.856 0.961 0.362 $50,000 to $74,999 1.036 0.609 0.877 0.000 *** $75,000 to $99,999 0.954 0.525 0.843 0.000 *** $100,000+ 0.925 0.233 0.819 0.000 *** Worker 1.242 0.006 *** 1.145 0.006 *** Mobility impairment 0.672 0.010 ** Number of household vehicles (reference: exactly one vehicle) 0.839 0.071 * Exactly two 0.769 0.013 ** - - Three or more 0.554 0.000 *** Number of household vehicles (reference: exactly two vehicles) - - Three or more - - 0.690 0.000 *** Constant 1.431 0.166 0.269 0.000 *** Model Indicators Number of cases, N 3,352 9,513 Final log likelihood, LL() -2411.30 -6366.71 Log likelihood of constants-only model, LL(C) McFadden's Pseudo-R2: 1-LL()/LL(C) -2448.19 0.0151 -6590.44 0.0339 Vehicles purchased in the past 12 months, whether purchased new or used. Sample: drivers age 16+ in household with newly purchased vehicle(s). Sample: drivers age 16+ in household with 2+ vehicles and 2+ potential drivers. In case of a tie, both vehicles are coded as newest. A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. An examination of teen drivers' vehicles specifically indicates that gender affects vehicle access for the newest drivers. As shown in table 137, male teen drivers are slightly more likely to have their own car than female teen drivers. However, teen girls are slightly more likely to have a 293 newly purchased car, and more than three times as likely to have the newest vehicle in the house by model year. Teen girls, it seems, are provided with cars that are in better condition than are teen boys, whether because they are considered more trustworthy drivers, because of greater concern about their safety in case of mechanical difficulties, or both. Table 137. Percent of teen drivers who are the main driver for household vehicles, by sex. Teen Drivers only All Boys Girls Any vehicle 64.9% 65.4% 64.6% Newest by model year 15.0% 6.3% 21.6% Newly-purchased 24.4% 22.4% 25.9% Sample: all drivers ages 16-17 in households with at least one vehicle HOW MUCH AND WHAT FOR: THE INTERRELATED EFFECTS OF GENDER AND AGE ON MOBILITY AND TRIP PURPOSE Gender profoundly influences how and why people travel. However, the effects are not the same across all ages and life stages. How can researchers best understand the interrelated effects of gender and age? Working-age women have more complex tripmaking patterns than men, and are disproportionately responsible for household-serving travel (trips made to maintain the household or its members) (Taylor, Ralph, and Smart 2015). Mothers are more likely than fathers to be responsible for transporting children to school and elsewhere, affecting day-to-day mode choice and long-term career and residential location decisions (Gimenez-Nadal and Molina 2016, Jun and Kwon 2015). Elderly women are likely to give up driving sooner than men, and suffer disproportionately from being immobile or housebound (Loukaitou-Sideris and Wachs 2018). All of these effects can reduce women's quality of life, but many of them remain hidden if researchers do not incorporate age, gender, and the interactions between the two into their analyses. 294 This section examines trends in mobility and trip purpose by gender and age using binary and multivariate probit (MVP) models to analyze travel diary data. The researchers model: (1) the likelihood of being housebound (having made no trips outside the home within the past 7 days) and, (2) among nonhousebound respondents, the likelihood of having made trips of various purposes on the travel day. The MVP structure facilitates efficient modeling of the "simultaneous but separate" choices (Choo and Mokhtarian 2008, p. 147) to make or not make various types of trips in a single day. The section begins by briefly discussing planning for women's needs and the importance of additionally considering the divergent needs of women of different ages. We then discuss the methods used; in particular, we address why an MVP model is useful for analyzing trip purpose. The effects of gender on mobility and trip purpose are found to be substantial, but strongly agedependent. All else equal, young women are not substantially more likely to be chronically housebound than young men. Among older adults, however, being female is associated with an increased likelihood of being housebound, ranging from 2.3 percentage points for seniors ages 7579 to 8 percentage points for ages 85+, even after controlling for disability. This finding suggests that if being female were not associated with an elevated risk of becoming housebound, there would be 26,800 fewer housebound seniors in Georgia. The findings on trip purpose show a consistent pattern. With one exception (trips for dining), among younger adults, being female is associated with an increased likelihood of making every kind of trip. With no exceptions, among older adults, being female is associated with a decreased likelihood of making every kind of trip measured. These findings underscore how the loss of 295 mobility reduces older women's ability to attend social and recreational activities, maintain their household, and even seek medical care. As for younger women and men, we find that equal mobility does not imply equal utility. Working-age women are disproportionately responsible for household-serving travel, and the extra responsibilities for transporting family members fall more heavily on mothers and other female caregivers. This finding holds even after controlling for employment. Further, the age at which men's tripmaking eclipses women's is much younger for trips that benefit the individual traveler (i.e., leisure) than for household-serving trips (i.e., errands, transporting others). The final discussion in the section is the policy implications. It is vital to facilitate older women's mobility. For younger women who are already mobile, it may be more important to focus on providing a high level of service for the household-serving trips for which women are disproportionately responsible, and to facilitate access to recreational and fitness opportunities. This analysis is a reminder that women's travel needs are not constant at every stage of life. Background There is a substantial body of evidence that crafting transportation policy without taking gender into account results in worse outcomes for women (Loukaitou-Sideris 2016, Fainstein and Servon 2005). On average, women's travel behavior and vulnerabilities differ from those of men. In the U.S., the gender gap in household-serving and child-serving travel has persisted, even as gaps in labor force participation and overall mobility have narrowed (Taylor, Ralph, and Smart 2015, Craig and van Tienoven 2019). In part because of this disproportionate responsibility for household labor and childcare, women's commutes tend to be shorter than men's, but their 296 journeys are more complex and involve more trip chaining (McQuaid and Chen 2012, Loukaitou-Sideris 2016). While traditionally, and currently in some developing countries, women have been less mobile than men (Loukaitou-Sideris 2016), many American women today are constrained by an abundance of household- and child-serving trips at the expense of other uses of their time and travel (Loukaitou-Sideris 2016, Craig and van Tienoven 2019); Hanson (2010) refers to this conundrum as hypermobility. Transportation planning centered around travel directly between home and workplace, a pattern more typical of men, disadvantages women who may need to drop off a child on the way to work and run a series of errands on the way home. Researchers and practitioners have argued that women's needs are best served when a gendered lens is applied to policymaking (Fainstein and Servon 2005). While progress has been uneven (Loukaitou-Sideris and Fink 2009), many organizations now incorporate gender into transportation planning and policymaking. However, women's needs are not monolithic. Just as planning for the needs of an "average" traveler often neglects the needs of women, planning for the needs of the "average" woman can disadvantage women who are also poor, disabled, or members of a racial minority. Age strongly affects women's travel behavior and needs. All of the trends described here are true of American women "on average."82 As this report will show, none of these trends are true of elderly women. While there is an increasing amount of literature on older adults, substantially fewer studies explicitly focus on gender. While common to note that women are more affected 82 This includes those studies that discuss "women's needs" but did not include older adults. 297 because more seniors are women, it is significantly less common to examine how being female affects older women's travel behavior in comparison to that of older men. Much transportation research on older adults has focused on driving behavior. Among the oldest seniors, driver licensing rates are lower among women than among men; this is not true of younger senior cohorts (Loukaitou-Sideris and Wachs 2018). For licensed drivers, researchers have focused on safety risks and the self-imposed limitations older drivers may place on their driving. Studies have consistently found that women are more likely to self-limit than men, and do so at younger ages (see Wong et al. 2016 for a review). Advocates, and an increasing number of researchers, have commented on the shortsightedness of a focus on helping seniors know when to stop driving without providing adequate transportation alternatives once they do so (Loukaitou-Sideris and Wachs 2018, Wong et al. 2016). Many U.S. seniors live in areas with anemic public transit coverage, and where transit exists, it may be physically and/or cognitively difficult for elderly riders to navigate. The resulting isolation adversely impacts seniors' mental and physical health (Decker 2006). Gender, then, shapes the travel behavior and quality of life of working-age adults and seniors. Even female children are given less leeway to travel independently than male children (McDonald 2012). To meet women's transportation needs across an entire lifespan, it is critical to understand the separate and interconnected effects of gender and age on travel behavior. Methods Table 138 shows descriptive statistics about the sample. Unweighted data are used for the models of the likelihood of being housebound, and of trip purpose; in general terms, models do 298 not require representative data to identify empirical relationships. Unless otherwise stated, the NHTS's person-weights are applied to descriptive statistics and average marginal effects. Table 138. Sample characteristics. Category Female Mean age Adult 1864 Senior (ages 6579) Elderly (ages 80+) Number of Trips on Travel Day 0 12 35 6+ Out of country (N=57)* Chronically housebound (no trips within past seven days) Full-time worker Part-time worker Nonworker Unknown worker status (N=294) Mobility impairment absent Mobility impairment present Unknown disability status (N=8) Caregiver for child(ren) ages 015 Driver Race: white non-Hispanic Race: Black, Black multiracial, or Black Hispanic Race: other Annual Household Income $0 to $24,999 $25,000 to $34,999 $35,000 to $49,999 $50,000 to $74,999 $75,000 to $99,999 $100,000+ Unknown (N=3,393) Table continues on next page. All Adults (N= 15,222) 55.0% 53.2 70.8% 23.7% 5.5% Unweighted Men (N=6,845) 53.0 71.0% 24.0% 5.1% 16.9% 24.6% 37.6% 20.5% 0.4% 14.7% 27.1% 38.0% 19.9% 0.4% 2.1% 42.3% 10.2% 45.5% 1.9% 89.2% 10.7% 0.1% 21.4% 92.1% 69.9% 1.6% 50.5% 8.5% 39.1% 1.9% 90.8% 9.1% 0.0% 20.7% 93.4% 73.0% 22.2% 7.9% 19.3% 7.7% 19.0% 9.3% 12.2% 17.5% 13.2% 25.4% 3.3% 15.8% 8.7% 12.1% 17.9% 14.3% 28.2% 3.0% Women (N=8,377) 53.3 70.6% 23.5% 5.9% 18.7% 22.6% 37.3% 21.1% 0.4% 2.4% 35.7% 11.6% 50.8% 1.9% 87.9% 12.0% 0.1% 22.0% 91.1% 67.3% 24.6% 8.1% 21.6% 9.8% 12.3% 17.2% 12.3% 23.2% 3.6% 299 Table 138. (Continued). Continued from previous page. Category Weighted (Non-missing Only) All Adults Men Women Female 52.1% Mean age 46.1 45.1 46.9 Adult 1864 83.1% 84.8% 81.6% Senior (ages 6579) 13.7% 13.0% 14.5% Elderly (ages 80+) 3.1% 2.3% 3.9% Number of Trips on Travel Day 0 18.4% 16.4% 20.2% 12 26.2% 29.1% 23.5% 35 36.5% 36.5% 36.5% 6+ Chronically housebound (no trips within past 7 days) 18.9% 2.4% 18.0% 1.8% 19.8% 2.9% Full-time worker 48.7% 58.3% 39.9% Part-time worker 11.9% 9.9% 13.7% Nonworker 39.4% 31.8% 46.4% Mobility impairment absent 90.5% 92.2% Mobility impairment present Caregiver for child(ren) ages 015 Driver 9.5% 31.8% 89.0% 7.8% 30.1% 90.5% 11.1% 33.3% 87.6% Race: white non-Hispanic 55.2% 59.5% 51.3% Race: Black, Black multiracial, or Black Hispanic 31.6% 27.2% 35.7% Race: other 13.1% 13.3% 13.0% Annual Household Income $0 to $24,999 24.0% 20.7% 27.0% $25,000 to $34,999 10.1% 9.5% 10.8% $35,000 to $49,999 12.0% 12.2% 11.8% $50,000 to $74,999 16.6% 16.7% 16.4% $75,000 to $99,999 11.8% 12.9% 10.9% $100,000+ 25.5% 28.1% 23.1% * NHTS automatically records a trip total of zero for participants who were out of the country. For the purposes of this paper, it is more useful to treat these observations as missing data. Unknown here includes any kind of missing data, including "Don't Know," refusal to answer, and left blank. A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. The question asked by the NHTS is "Do you/does this person drive?" Driver rates in the data decrease among both male and female seniors; some seniors categorized as nondrivers likely have driver's licenses. 300 Seventeen percent of adults reported making zero trips on their travel day. Travel patterns vary over the course of the week, so it would be incorrect to assume all those respondents are housebound. We define "chronically housebound" as having made no trips outside the home within the past 7 days; this describes 2.4 percent of all Georgians and 8.7 percent of elderly or disabled residents.83 Table 139 shows annual per-capita rates of tripmaking by purpose, gender, and age. These descriptive statistics show clear gender differences in tripmaking patterns in every age group, but the form of those differences varies by age. Among the elderly population, women are significantly less mobile than men, making only two thirds as many trips. This mobility disadvantage is not present among working-age women. In fact, working-age women travel slightly more than working-age men (an extra 1.2 trips per week). Even though working-age men and women make a similar quantity of trips, there are evident differences in the purposes of those trips. On average, over the course of a week, a working-age man will make 1.4 more work or school trips than a working-age woman, while a working-age woman will make an extra 2.2 household-serving trips. These patterns are obviously entangled with broader demographic trends, such as women's lower participation in the labor market. To clarify the roles of gender and age, a holistic modeling approach is needed. This report uses a two-stage process to model mobility and trip purpose. 83 Excluding residents of institutions such as nursing homes, who were not included in the sample. 301 Table 139. Annual trips per capita by purpose, gender, and age (weighted). Working-Age Senior Elderly Trip Purpose Men Women Men Women 1864 1864 6579 6579 Men 80+ Women 80+ All purposes 1,223 1,285 1,196 1,062 948 636 Mandatory Travel 256 182 83 41 28 4 Work commute 215 149 73 35 26 1 Other work-related travel 18 15 8 4 2 1 Attend school or daycare 23 18 1 2 2 Household-Serving Travel 286 399 405 399 345 242 Transport someone 72 116 61 53 23 28 Shopping or errands 203 260 318 312 284 187 Medical/dental services 11 23 26 33 37 26 Discretionary Travel 262 275 306 259 256 175 Social/recreational 130 136 139 119 133 85 Dining 103 100 117 86 74 45 Community/religious 29 38 49 53 49 45 Other Travel Return home 419 429 403 362 320 215 407 414 389 355 312 209 Other 13 15 14 7 7 7 Based on all adults (N=15,222). Estimates are weighted using NHTS's trip-weights, which are annualized versions of the person-weights. We analyze mobility by using binary probit to model the likelihood that a person is chronically housebound (i.e., zero trips outside the home in the past 7 days). We chose to model chronic immobility rather than number of trips because the NHTS diary covers a single day for each respondent. The models therefore need to distinguish between respondents who completed their diaries on a day they happened to stay home and respondents who reported no trips because their ability to travel is more chronically impaired. We exclude people who were out of the country, as well as people missing either worker or disability status.84 84 Income was the only variable with numerous missing values. To avoid compromising the representativeness of the sample by discarding these observations, a dummy variable was created for missing income. Where applicable, NHTS-imputed values for race, sex, and age are used. 302 The second model is a multivariate discrete choice (probit) model of the likelihood of having made at least one trip for each of five common trip purposes: transporting someone else, shopping/errands, medical/dental, social/recreational/fitness, and dining. Multivariate probit allows for the simultaneous inclusion of multiple dependent variables. It is useful for any situation where there are multiple binary outcomes to be modeled, and where those outcomes are separate but not necessarily independent. The model is based on the same sample as the binary model, but excludes those who are chronically housebound. Although work is a common trip purpose, the research team opted not to include it in this analysis because the overwhelming determinant of whether or not people make work trips is their worker status. Worker status is instead included as a control variable, allowing us to assess whether different rates of household-serving travel persist even between men and women with the same employment status. Trips to return home are also not included because they are essentially the complement of having made any other trip. To capture the separate and interrelated effects of gender and age on travel, the models incorporate multiple interaction terms between gender and age, and between gender and other relevant factors. While this provides more nuanced findings, the large number of interaction terms makes discerning the effects of gender directly from model coefficients more complicated. This report therefore provides visual representations of the average marginal effect of gender by age group and discusses joint significance of interaction terms. Analysis for this project was conducted in Stata 15 software (StataCorp 2017). The MVP model was estimated using the mvProbit program created by Cappellari and Jenkins (2003). 303 Results Table 140 shows the results of the first model: a binary probit model of the likelihood of being housebound. For both men and women, one of the strongest predictors of being housebound is disability. On average, having a disability is associated with an increase of 4.6 percentage points in the probability of being housebound; the average marginal effect (AME)85 of having a disability is even larger for seniors (7.6 percentage points, versus 4.0 points for adults younger than 65). However, the model shows that even after controlling for disability, gender affects the risk of being housebound. When the interactions between gender and age are accounted for, gender per se is not significant. Rather, gender's effects are captured by the interaction terms between gender and age, which are jointly significant (=.05). The combined effect of gender and age is depicted in figure 29. 85 To control for gender-related differences in the explanatory variables, AMEs are calculated here by predicting the probability of being housebound as if the whole sample had a mobility impairment, predicting the probability as if the whole sample did not have a mobility impairment, and subtracting the former from the latter. AMEs by age are the average of respondents from a given age category. 304 Table 140. Binomial probit model of likelihood of being chronically housebound (having made no trips within the past 7 days). Covariate Coefficient P-Value Female 0.251 0.514 Age Age2 -0.0220 0.00019 0.088 * 0.126 Female x Age Female x Age2 -0.0161 0.000213 0.291 0.129 Worker -0.346 <0.001 *** Mobility impairment, adult ages 1864 0.666 <0.001 *** Mobility impairment, adult ages 65+ 0.774 <0.001 *** Vehicle Ownership Category (Reference: Vehicle-Nondeficit Household with 2+ Drivers) Single-driver vehicle-nondeficit -0.174 0.051 * Vehicle-deficit 0.404 <0.001 *** Zero-vehicle 0.141 0.233 Race (reference: white non-Hispanic) Black 0.267 <0.001 *** Other 0.293 0.003 *** Built Environment Type (Reference: Urban/Second City) Rural 0.115 0.164 Small town 0.0135 0.862 Suburban 0.0102 0.908 Annual Household Income (Reference: <$25,000) $25,000 to $34,999 -0.152 0.157 $35,000 to $49,999 -0.00525 0.956 $50,000 to $74,999 0.00719 0.940 $75,000 to $99,999 0.0108 0.923 $100,000+ -0.390 0.001 *** Unknown (missing) -0.0699 0.628 Constant -1.773 0.000 *** Number of cases, N 15,155 Final log likelihood, LL() -1194.2 Initial log likelihood LL(0) Pseudo-R2: 1-LL()/LL(0) -1477.6 0.192 * Denotes significance for = .10 ** Denotes significance for = .05 *** Denotes significance for = .01 The test of joint significance of female x age and female x age 2 returned a P-value of .020, indicating that they are jointly significant for = .05. The P-value for a joint significance test of female, female x age, and female x age 2 was .006. Vehicle-deficit households have at least one car, but fewer cars than potential drivers. Includes Black Hispanic and Black multiracial. 305 Figure 29. Histogram. Average marginal effect of being female on probability of being housebound, by age group (percentage points). All else equal, young women are not substantially more likely to be chronically housebound than young men; the average marginal effect (AME)86 of gender among working-age adults is essentially zero. However, it begins to increase in late middle age. Among seniors ages 7579, being female is associated with an increase of 2.3 percentage points in the probability of being housebound. The AME for seniors in their early 80s is 4.1 percentage points. At age 85, being female is associated with an increase of 8 percentage points. The AME of gender and age is in addition to the effects of age alone; elderly women's total probability of being housebound is 14.5 percent, more than double that of elderly men (figure 30). To put these numbers in perspective, if being an older woman were not associated with an elevated risk of being housebound even after controlling for disability, there would be 26,800 fewer housebound seniors in Georgia. 86 To control for gender-related differences in the explanatory variables, AMEs are calculated here by predicting the probability of being housebound as if the whole sample were male, predicting the probability as if the whole sample were female, and subtracting the former from the latter. AMEs by age are the average of respondents from a given age category. Five-year age groups are used to ensure that each estimate is based on an average over at least 300 respondents. 306 Figure 30. Histogram. Probability of being housebound by gender and age. Trip Purpose Table 141 shows results of the multivariate probit model of trip purpose. Before describing findings, this section briefly assesses the strength of the model as a whole and presents McFadden's pseudo-R2 statistics based on several different benchmarks (Mokhtarian 2016) (table 142). The juxtaposition of a relatively low 2(MS) with higher 2(EL) and 2(NC) indicates that compared to a model with only constant terms, the full model represents a modest improvement in the researchers' ability to predict what trips a person will have made and a larger increase in our power to explain the factors that contribute to the observed trip patterns. The 2(EL) of 0.380 is considered very good, especially for a multivariate model with 64 (=26) possible outcomes. 307 Table 141. Multivariate probit model of likelihood of making trips for various purposes. Covariate Female Community/ Religious -0.040 Transport Other -0.330 Shopping/ Errands 0.235 Medical/ Dental 0.405 Age 0.0156 * 0.0236 *** 0.0401 *** 0.0268 *** Age2 -0.000068 -0.000209 *** -0.000295 *** -0.000087 Female x Age Female x Age2 Mobility-impaired non-senior -0.001277 -0.000004 -0.194 ** 0.0153 -0.000204 ** -0.149 ** -0.000213 -0.000081 -0.0569 -0.00725 -0.000017 0.497 *** Driver 0.064 0.390 *** 0.441 *** -0.0102 Female x Driver 0.246 * 0.245 0.0954 0.230 * Full-time worker -0.305 *** -0.118 *** -0.452 *** -0.456 *** Part-time worker 0.002 0.0859 * -0.234 *** -0.319 *** Caregiver for Child <16 Years Old* (Reference: Noncaregiver) Caregiver - - -0.0322 0.0915 Female x Caregiver - - -0.00338 0.00631 Caregiver Status by Age of Youngest Child in Years (Reference: Noncaregiver) Youngest child 04 0.197 ** 0.723 *** - - Youngest child 515 0.0652 0.796 *** - - Female x Youngest 04 -0.0607 0.196 ** - - Female x Youngest 515 -0.0336 0.122 * - - Vehicle Ownership Category (Reference: Vehicle-Nondeficit Household with 2+ Drivers) Single-driver vehicle-nondeficit 0.0904 * 0.0696 0.291 *** 0.0313 Vehicle-deficit -0.0253 0.0511 0.0324 0.00898 Zero-vehicle 0.142 -0.163 0.413 *** 0.168 * Race (Reference: White Non-Hispanic) Black 0.0827 * 0.0737 * 0.054 * 0.108 ** Other 0.0506 -0.004 0.008 0.053 Table continues on next page. Social/ Recreational -0.189 0.00903 -0.000096 * 0.00228 -0.000061 -0.273 *** 0.118 0.206 ** -0.426 *** -0.225 *** -0.102 ** 0.00592 - - - - 0.292 *** -0.0400 0.213 *** -0.160 *** 0.027 Dining -0.164 0.00347 -0.00003 0.0109 -0.000138 ** -0.0936 0.241 *** -0.0211 -0.0432 0.0447 -0.0354 -0.0889 * - - - - 0.149 *** -0.107 ** -0.0147 -0.140 *** -0.238 *** 308 Table 141. (Continued). 309 Table 142. Summary statistics and indicators for multivariate probit model. Indicator Number of cases, N Final log likelihood, LL() Log likelihood of constants-only (market share) model, LL(MS) Equally likely benchmark, LL(0) Log likelihood without constants, LL(NC) McFadden's Pseudo-R2 Measures Market share benchmark, 2 (MS) : 1-LL()/LL(MS) Equally-likely benchmark, 2 (EL) : 1-LL()/LL(EL) No-constant benchmark, 2 (NC) : 1-LL()/LL(NC) Share of Cases Making Each Type of Trip Community or religious Transport other Shopping/errands Medical/dental Social, recreational, or fitness Dining Correlation Between Error Terms for Outcomes 21 0.1302 31 0.0741 32 0.1637 41 -0.0087 42 0.0825 43 0.1384 51 0.0382 52 0.1348 53 0.1703 54 -0.0282 61 0.1454 62 0.2109 63 0.3011 64 0.1851 65 0.1803 14,854 -38323.0 -40562.7 -61776.0 -38497.8 0.055 0.380 0.377 8.6% 13.0% 42.7% 6.8% 27.9% 25.6% P-value 0.000 *** 0.000 *** 0.000 *** 0.803 0.002 *** 0.000 *** 0.080 * 0.000 *** 0.000 *** 0.236 0.000 *** 0.000 *** 0.000 *** 0.000 *** 0.000 *** 310 As with the model of immobility, the researchers find that gender significantly affects trip purpose, but its effects are dependent on age. The gender/age interactions are jointly significant (=.01) for all trip purposes except community/religious trips.87 There are additional interactions between gender and being a parent or other caregiver. Parenthood is associated with an increase in the likelihood of making a trip to transport someone else, but the increase is larger for female caregivers (e.g., mothers and grandmothers) than it is for fathers and other male caregivers. Female caregiving is also associated with a disproportionate decrease in dining trips. Using the same methods as for the immobility model, figure 31 shows the average marginal effect of being female for each outcome, across the entire sample and incorporating all interactions. The AME of being female on the likelihood of making each type of trip is relatively modest, ranging from an increase of about 4 percentage points in the probability of making a trip for transporting someone or shopping, to an approximately 1 percentage point decrease in the likelihood of making a trip for dining or socialization/recreation. 87 For community/religious trips, gender is significant when all interaction terms are jointly tested (=.01). 311 N=14,854 Figure 31. Bar graph. Average marginal effect of gender on probability of making trips of various purposes on travel day (in percentage points). However, these apparently modest effects are deceptive. Because the size and, more importantly, direction of the effect of being female varies by age, a single average marginal effect understates the degree to which gender affects tripmaking. Figure 32 provides a visual representation of the effect of gender across different ages,88 underscoring the dramatically different effect gender may have on the trip patterns of younger and older adults. For example, while the overall AME of gender on the probability of making trips to shop or run errands is +4.0 percentage points, the gender AME by age ranges from +9.2 percentage points for the youngest adults to -12.3 for the oldest, a range of 21.5 percentage points. The smallest range of effects is 3.5 percentage points (for community/religious trips), and the average is 11.1 percentage points.89 88 Five-year age groups are used rather than individual years to ensure that each estimate is based on an average over at least 300 respondents. 89 The two trip types with the smallest marginal effect, community/religious and medical/dental, were the least common at the outset. The fact that the effects are presented as percentage points limits how much values that are already close to zero can be reduced. 312 Figure 32. Histograms. Average marginal effects of gender by age (in percentage points). There is a consistent pattern: for younger adults, being female is associated with an increased likelihood of making a trip of that type, but at some point the trend reverses; among seniors and 313 the elderly, being female is associated with a decreased likelihood. The sole exception to this is dining, where women are already slightly less likely than men to eat out when young, and become increasingly less likely to do so as they age. Trips to transport others are highest during the years when many people are responsible for the care of a child who is not old enough to drive. However, this caregiver increase is higher for female caregivers than for male caregivers (figure 33). Figure 33. 3D histogram. Predicted probability of making a trip to transport others by gender and age. Another notable variation is the age at which being female switches from being associated with an increased likelihood to being associated with a decreased likelihood. For household-serving travel (i.e., transport others, shopping/errands, medical/dental), this transition occurs somewhere between ages 65 and 75, depending on the specific purpose. For social and recreational travel, 314 this switch occurs much earlier, at age 45.90 The latest transition is for community/religious trips, which also shows the smallest decrease. This suggests that as older women are less able to make trips, they are likely to highly prioritize community and religious excursions. Gender also has significant interactions with driver status. Driving increases the likelihood of making trips for several purposes. For women, this increase is greater, though the interaction terms themselves are not universally significant. Discussion While gender clearly influences both the likelihood of being housebound and the purposes of the trips a person makes, these results suggest that there is no monolithic effect of being female. Rather, the effect is strongly mediated by age, and for some purposes, by parenthood and ability to drive. Other studies have documented how the effects of gender are also influenced by travelers' race, income, and other demographic factors (Loukaitou-Sideris 2016). Failing to attend to the interactions between gender and age can obscure troubling gendered inequalities. The fact that women are, on average, more likely to make shopping trips than men is likely of little comfort to an elderly woman who is waiting to fill a prescription until her adult daughter is able to give her a ride to the pharmacy. That daughter may be making just as many trips as men her age; however, if her travel is disproportionately devoted to chauffeuring family members and doing errands, she may have less time to socialize or go out to eat, and as a result derive fewer personal benefits from her travel. 90 The AME of being female on dining trips is always zero or negative. 315 A simple gendered lens can help craft transportation policy that does not disadvantage the average woman. To attend to the needs of all women, a compound gendered lens that clarifies the layered effects of gender, age, and other sources of inequality is needed. The following subsections discuss some of the key findings by age group. Working-Age Women: Equal Mobility, Unequal Utility Historically in the U.S. and currently in some developing cities, women were or are less mobile than men (Loukaitou-Sideris 2016). This study finds no significant discrepancy in the quantity of trips made by working-age men and women in Georgia. There are, however, gendered differences in the purposes of those trips and, therefore, we would argue, in the personal benefit that men and women derive from their tripmaking. In particular, women are disproportionately likely to engage in household-serving travel, even after controlling for worker status. For Georgians who become parents, the burden of transporting their new family falls more heavily on mothers than on fathers. In contrast, the differences between men and women for trip purposes that directly benefit the traveler (i.e., socialization, recreation, dining) are much smaller. This is consistent with findings about hypermobility, suggesting that despite their high levels of mobility, working-age women may derive less utility from those trips (Craig and van Tienoven 2019). From a policy standpoint, the coexistence of equal mobility and unequal utility is a classic "wicked" problem (Rittel and Webber 1973). Gendered inequality in trip purpose is likely a symptom of other gendered inequalities, such as distributions of household labor, expectations of parents, and, for purposes like dining out, income. Given that transportation professionals have few tools for addressing these structural social forces, how can these planners nevertheless 316 improve transportation for working-age women? One method is to focus on providing a high level of service for the types of trips that women disproportionately make, particularly household-serving travel. Public transit is a good example of a system that can inadvertently provide a lower level of service to women because their trip purposes and patterns differ from men's.91 Radial public transit networks, which are designed to transport workers to and from a city center, are often poorly suited for suburb-to-suburb commuting and the more complex trip patterns required to run errands and drop off or pick up children on the way to or from work. Additionally, if riders are required to pay a new fare for each stop they make, the cost of doing errands can rise quickly. Making sure transit routes adequately serve shopping and residential areas and providing free transfers can make transit more useful for household-serving travel, thereby disproportionately benefiting women. Similarly, household-serving travel, especially transporting children, causes many women to spend long hours driving, leaving limited time for leisure and exercise. Active transport such as bicycling can do "double duty" for completing responsibilities but simultaneously enjoying exercise. Since unsafe traffic conditions are known to particularly deter female would-be cyclists (Emond, Tang, and Handy 2009), improving bicycle infrastructure could result in fewer women stuck behind the steering wheel. 91 As a side note, transit may provide equal mobility but unequal utility in another way: sexual assault during transit trips does not always stop women from riding transit, but it drastically increases the emotional and psychological cost (Kash 2019). 317 Transportation professionals could also improve travel for women who are caregivers by improving travel for students. For example, the need to drop children off at school makes it difficult for working parents (disproportionately, working mothers) to walk or bike for their commute. Improving school transportation may also free up students' parents. Older Women Face a Crisis of Immobility Mobility decreases with age for both men and women, but for women, the effect is much stronger. Elderly men make 77 percent as many trips as working-age men, but elderly women make just half as many trips as working-age women. Even after excluding people who live in a supportive care facility, 13 percent of elderly women (ages 80+) are chronically housebound, more than triple the rate among elderly men (3.6 percent). For every trip purpose measured, there is an age past which women are less likely to make that trip than are men of the same age. However, men's likelihood of making leisure trips eclipses women's in middle age, several decades before they eclipse women's rates of household-serving trips. Given that women are documented to utilize medical services at higher rates than men, it is striking that among older adults, being female is associated with a decreased likelihood of making medical trips. Being a driver increases the likelihood of making medical trips for women, but has no significant effect for men. This suggests that women make fewer medical trips not because they have no need of a doctor, but because they have no means of reaching the doctor's office. The only trip purpose for which there is not a large gender discrepancy among older adults is community and religious activities. This finding may indicate that older women especially value 318 trips to stay connected to community and church, but may also reflect a greater availability of transportation; these trips take place primarily on weekends, when relatives are more likely to be available to provide a ride. Additionally, many churches provide free transportation to congregants in need. The results of this study indicate that the most pressing transportation need for older women is simply more transportation. Rates of driving are lower among senior women than senior men. Often, seniors who stop driving are left with few alternatives to replace their private auto; even where public transit networks are robust, the systems are difficult to navigate for riders who gave up driving due to physical (or cognitive) limitations (Decker 2006). Many social service organizations that serve seniors provide shuttle service to take advantage of programming. Transportation to nonorganizational destinations may be more difficult to come by. Taxi service is not affordable for many seniors. Utilization of ridehailing services such as Uber and Lyft is relatively low among seniors; thus, this category may present an opportunity for improving senior mobility (Shirgaokar 2018). In fact, Fulton County (which includes Atlanta) recently began offering subsidized ridehailing for elderly residents.92 A Change of Pace: Women in Late Middle Age and Younger Seniors This report has discussed hypermobility among young women and immobility among the elderly. Middle-aged women and young seniors are at a point of transitioning between the two states of mobility. Is it possible to help middle-aged women and young seniors maintain their mobility and avoid the challenges facing the cohort before them? As men retire from the 92 For more information, visit the website for Fulton County Senior Services: https://www.fultoncountyga.gov/services/senior-services/transportation. 319 workforce, their overall travel declines, but there is some substitution of leisure and household trips for work trips. Curiously, women's leisure trips decline around the same age many children become more independent or leave home, and decline further after retirement. Are there policies that might encourage women to engage in more recreational travel as their childcare and employment responsibilities diminish? Limitations and Directions for Future Research The principal limitation of this study is that it is cross-sectional rather than longitudinal. The primary advantage of a cross-sectional approach is, of course, that the data are much more readily attainable.93 In a cross-sectional study, it is important to ask whether the observed effects that this study has attributed to age may, in fact, be related to cohort (generation). After all, when today's 18-year-old reaches her 80th birthday, the world will be rather different than it is today. However, gendered differences in trip purpose are evident even among the youngest of respondents, and studies have found that gendered divisions of household labor have diminished only slightly over recent decades (Crane 2007, Loukaitou-Sideris 2016). Though the research team believes the effects we have uncovered are predominantly age effects, they are not inevitable. Transportation professionals, in other words, may be able to prevent the epidemic of immobility currently afflicting older adults from reaching today's young women. However, regardless of whether these same issues will face older women decades from now, there is a clear policy need for better transportation for today's seniors. Future research can provide guidance for evidence-based practice. For example, how could ridehailing services help 93 While the NHTS has been repeated a number of times, it uses a unique pool of respondents each time. Additionally, question structures and sampling methods have changed substantially over the years. 320 more older adults? Is it simply a question of increasing older adults' comfort with technology? Or are there aspects of the service that are unfriendly to elderly adults who may have physical or cognitive limitations? Gender identifiably affects the travel behavior of adults as young as 18, raising the question of when this differentiation begins. Therefore, it would also be valuable to analyze gendered differences in trip purpose and independent mobility among children and adolescents. Finally, just as it is important to pay attention to the intersection of age and gender, more research is needed on other factors that interact with gender. For instance, parenthood affects low-income women and single mothers differently than it does women with working partners. A Compound Gendered Lens While gender strongly influences travel behavior, there is no such thing as a monolithic "women's travel needs." This study has found that the magnitude and direction of the effects of gender vary with age. Examining gender without age therefore produces an incomplete, and in some cases misleading, portrait of the challenges facing women. In particular, hypermobility among working-aged women can obscure the crisis of immobility that leaves many older women isolated in their own homes. To attend to the needs of all women, researchers need a compound gendered lens, where they account for the layered effects of gender, age, and other sources of inequality. Methodologically, these findings underscore the importance of incorporating gender and age into analysis, not only separately, but also jointly. From a policy standpoint, this study's results suggest a need for interventions aimed at improving two distinct transportation problems facing women of different ages. Elderly women, who no longer feel comfortable driving or can no longer navigate public 321 transit, urgently need viable replacements to improve their mobility. Working-age women, in contrast, face equal mobility but unequal utility compared to working-age men. In the face of persistent gender differences in household division of labor, we argue that planners should work to ensure that the level of service for the household-serving trips generally allocated to women is not lower than for the simpler work journeys that have been the traditional focus of planning. Additionally, modes that can do double duty as transportation and exercise (such as cycling) should also be made more friendly to women. The specific policies needed to promote equitable transportation vary by age group, but this study reveals one constant: a gendered lens is critical when planning for the needs of travelers of every age. HEALTH AND DISABILITY This section examines equity concerns related to disability status, health, and physical activity. It first investigates mobility differences by age and health among Georgians with and without a mobility impairment. Even among Georgians with no mobility impairment, the researchers find a strong link between low income and poor health, and indications that captive walking trips among the lowest-income Georgians are not a public health panacea. The report then turns to a more detailed analysis comparing different subgroups of Georgians with disabilities. In the context of this analysis, a mobility impairment is a medical condition that interferes with travel. While it is sometimes used interchangeably with disability, it does not include disabilities that have no direct effect on travel (as defined by the participants). It is also important to note that the NHTS sample excludes institutionalized populations. The disabilities being analyzed here are, therefore, not severe enough to require institutionalization. Given that this most infirm 322 population has already been screened out, the differences in mobility between Georgians with and without mobility impairments are striking. Georgians with mobility impairments face many challenges. Some of these challenges are related directly to the presence of a mobility impairment, but many are not. For example, figure 34 compares the annual household incomes of Georgians with and without mobility impairments. Georgians with mobility impairments are likely to be impoverished. Thirty-seven percent have an annual household income of less than $15,000, putting them near or below the poverty line regardless of household size. Only 12 percent of adults without mobility impairments fall in this category. Georgians with mobility impairments are overrepresented in all income groups under $35,000 per year and underrepresented in all income groups above that point. Regardless of the cause of this discrepancy, low-income people with disabilities have fewer resources to address the mobility challenges that they face. 323 Figure 34. Stacked bar graph. Annual household income of Georgians with and without mobility impairments. Health and Physical Activity among Disabled and Nondisabled Georgians A mobility impairment is not synonymous with being elderly or infirm. As shown in table 143, a plurality of Georgians with mobility impairments are in good health (42 percent) and the majority are younger than age 65 (60 percent).94 However, these shares are far smaller among Georgians with mobility impairments than they are for nondisabled Georgians (93 percent and 86 percent, respectively). 94 Table 143 also contains information about the unweighted and weighted distribution of the sample used for table 144 to table 149. 324 Table 143. Health and age of Georgia adults with and without mobility impairments. All adults Health Good, very good, or excellent Fair health Poor health Age Working age (1864) Senior (6579) Elderly (80+) Population (Weighted) Sample (Unweighted) With Mobility Impairment Without Mobility Impairment With Mobility Impairment Without Mobility Impairment 723,241 6,930,155 1,609 13,487 41.5% 37.3% 21.1% 59.8% 27.8% 12.4% 92.9% 6.4% 0.7% 85.7% 12.2% 2.1% 43.1% 37.1% 19.8% 47.8% 33.4% 18.8% 92.5% 6.7% 0.7% 73.7% 22.4% 3.9% As shown in table 144, 37 percent of Georgians with a mobility impairment live in very lowincome households (those with an annual income of less than $15,000), triple the rate for nondisabled adults. The difference in the share of low-income residents between disabled and nondisabled Georgians is actually largest among the young and those in good health. Among working-age Georgians with a mobility impairment, 44 percent live in a very low-income household, versus just 12 percent of nondisabled adults of the same age. The only subpopulation where disability is not linked to an increased likelihood of living in a low-income household is among Georgians in poor health; more than 40 percent of Georgians in poor health live in the lowest-income households, regardless of the presence or absence of a mobility impairment. 325 Table 144. Driver status and income of Georgians with and without mobility impairments, by health and age. With Mobility Impairment Very Low Income Driver* (<$15,000) Without Mobility Impairment Very Low Income Driver* (<$15,000) All Adults Health 61.6% 36.7% 91.9% 12.1% Good, very good, or excellent Fair health Poor health Age 63.7% 68.4% 36.2% 63.4% 62.9% 55.8% 92.7% 83.7% 62.7% 10.6% 30.8% 42.8% Working age (1864) 63.7% 44.3% 92.1% 12.3% Senior (6579) 68.4% 25.1% 93.5% 10.2% Elderly (80+) 36.2% 26.3% 75.9% 16.6% Note: Within each impairment group (with and without), the numbers represent the share of those possessing the row characteristic who also possess the column characteristic. For example, among those with mobility impairment, 63.7% of those in good health are drivers. * NHTS uses the following question wording: "Do you/does this person drive?" Based on annual household income. Income data are missing for 54 respondents with disabilities (3.4% unweighted) and 421 respondents without disabilities (3.1% unweighted). Statistics are based on nonmissing observations. There is a strong link between income and health among all ages and levels of ability (table 145). On average, 12 percent of Georgians describe themselves as being in fair or poor health.95 However, for low-income Georgians, this figure is 30 percent, compared to fewer than 4 percent of high-income Georgians. The association between poverty and worse health persists for disabled and nondisabled Georgians, both working-age adults and seniors. 95 This report combines fair and poor as measures of being less healthy because just 0.7 percent of nondisabled Georgians described themselves as being in poor health, compared to 21.1 percent of Georgians with a mobility impairment. When focusing specifically on Georgians with mobility impairments, we disaggregate fair and poor health. 326 Table 145. Georgians in poor or fair health by income, age, and disability. All Adults (Ages 18+) Working Age (1864) Older Adults (65+) Annual Household Income All Non- Non- Non- disabled* Disabled disabled Disabled disabled Disabled All Adults 11.9% 7.1% 58.5% 6.3% 54.0% 11.8% 65.1% <$15,000 30.5% 18.6% 68.0% 30.5% 67.6% 28.1% 69.2% $15,000 to $24,999 20.3% 11.7% 60.6% 20.3% 47.3% 14.7% 75.6% $25,000 to $34,999 13.1% 8.1% 53.1% 13.1% 42.8% 10.9% 64.6% $35,000 to $49,999 10.2% 6.6% 58.1% 10.2% 54.9% 11.7% 60.5% $50,000 to $74,999 9.2% 5.7% 56.7% 9.2% 50.0% 9.0% 63.1% $75,000 to $99,999 5.6% 3.5% 44.4% 5.6% 44.8% 8.2% 43.8% $100,000+ 3.7% 2.8% 36.8% 3.7% 26.2% 6.6% 53.7% * No mobility impairment. Note: Health is based on self-assessment with response options poor, fair, good, very good, and excellent. Further, the association between low income and poor health persists even when comparing adults with the same level of physical activity (figure 35). Among Georgians who are physically inactive, 22 percent of adults from the lowest-income households are in fair or poor health, compared to just 15 percent of inactive adults from households earning more than $15,000 per year. Among adults who engage in vigorous physical activity, the percentage of adults in the lowest-income households who consider themselves to be in fair or poor health is 10 times that of adults from other income brackets (11 percent versus 1.1 percent). 327 Figure 35. Bar graph. Health of nondisabled Georgia adults by income and level of physical activity in a typical week. On average, nondisabled adults in poor health walk more than those in better health (table 146), likely due to the strong correlation between poor health and poverty. This fact is a reminder that while promoting walking and cycling is an important public health intervention, other policies are needed to support the health of Georgians who are already walking because they have no choice. A majority of both disabled and nondisabled Georgians report making at least one walk trip in the past 7 days (table 146). On average, nondisabled Georgians are more likely than their mobility-impaired counterparts to report at least one walk trip (73 percent versus 60 percent), and are more likely to be physically active (88 percent versus 65 percent). Among working-age Georgians and those in good health, the difference between disabled and nondisabled Georgians with respect to walking is relatively small. In fact, working-age Georgians with disabilities walk 328 slightly more on average than nondisabled working-age Georgians (6.3 average trips versus 5.8 trips), though a somewhat smaller percentage of them report at least one walking trip. The gap in overall physical activity is somewhat larger, however. The same pattern exists among Georgians in good health. Disability is most strongly associated with decreased physical activity and walking among Georgians who are older or in poor health. Mobility Differences between Georgians with and without Mobility Impairments Georgians with mobility impairments are more likely than nondisabled Georgians to be immobile on the travel day, over the past few days, and over the past week (table 147). On a typical day, 4 in 10 Georgians with mobility impairments will not leave the house, compared to 16 percent of nondisabled Georgians. One out of 10 disabled Georgians reports having made no trips in the past 7 days, compared to 1.6 percent of nondisabled Georgians. Mobility impaired Georgians also report fewer trips per capita, and on active travel days (table 148). Unlike differences in physical activity, these gaps persist among the young and the healthy. This suggests that many obstacles to mobility for Georgians with disabilities are not related to health or physical capability. 329 Table 146. Walking and physical activity among Georgians with and without mobility impairments, by health and age. With Mobility Impairment At Least One Number of Walk Trip Walk Trips (Past 7 Days) Physically Active* Without Mobility Impairment At Least One Number of Walk Trip Walk Trips (Past 7 Days) Physically Active* All adults 60.1% 5.01 64.8% 73.4% 5.81 87.5% Health Good, very good, or excellent 70.5% 5.93 75.6% 74.1% 5.81 89.0% Fair health 55.8% 3.95 65.0% 64.5% 5.46 68.4% Poor health 47.1% 5.07 43.4% 67.2% 8.84 67.5% Age Working age (1864) 70.2% 6.26 69.3% 73.4% 5.78 87.0% Senior (6579) 48.1% 3.67 62.4% 74.8% 6.07 90.6% Elderly (80+) 37.8% 1.97 48.7% 67.7% 5.35 91.3% Note: Within each impairment group (with and without) percentages represent the share of those possessing the row characteristic who also possess the column characteristic. For example, 70.5 percent of people with mobility impairments who are in good health made at least one walk trip in the past 30 days, versus 74.1 percent of people in good health who do not have a mobility impairment. * In a typical week, respondent engages in some light, moderate, or vigorous physical activity. 330 Table 147. Immobility by disability, health, and age. With Mobility Impairment Without Mobility Impairment Travel Day Past Few Days Past Seven Days Travel Day Past Few Days Past Seven Days All adults Health 39.5% 15.8% 10.0% 16.3% 2.7% 1.6% Good, very good, or excellent health Fair health Poor health Age 33.8% 41.7% 46.7% 8.3% 16.8% 28.8% 4.4% 10.3% 20.6% 15.6% 22.8% 41.9% 2.4% 5.6% 16.6% 1.4% 4.2% 6.8% Working age (1864) Senior (6579) Elderly (80+) 34.1% 41.9% 60.1% 13.0% 16.2% 29.0% 8.4% 8.2% 22.2% 14.8% 23.6% 33.8% 2.4% 4.2% 6.1% 1.5% 1.8% 2.7% Note: Immobility refers to zero reported trips during the specified timeframe. Respondents with zero trips on the travel day were asked about the date of their most recent trip. Response options included 1. "The day before," 2. "A few days before," 3. "A week before," 4. "More than a week but within a month," and 5. "More than a month." We consider respondents to be immobile for the past few days if they selected response 3, 4, or 5 and immobile for the past week if they selected response 4 or 5. A respondent who has been immobile for the past week has also been immobile for the past few days and on the travel day. Table 148. Average daily trips by disability, health, and age. With Mobility Impairment Without Mobility Impairment Per Capita All adults 2.33 Health Good, very good, or excellent health 2.53 Fair health 2.21 Poor health 2.14 Age Working age (1864) 2.57 Senior (6579) 2.28 Elderly (80+) 1.28 * Reporting at least one trip on the travel day Active Travelers Only* 3.85 3.82 3.79 4.02 3.89 3.92 3.20 Per Capita 3.46 3.50 3.08 2.34 3.51 3.30 2.55 Active Travelers Only* 4.13 4.14 3.99 4.03 4.12 4.31 3.85 A number of behavioral adaptations take place as mobility becomes more difficult. The NHTS asked about such adaptations of all respondents age 80 and older, and of younger individuals who reported mobility limitations. Seventy percent of Georgians with mobility impairments 331 reported reducing day-to-day travel (table 149). This figure was 77 percent for elderly disabled Georgians, versus just 31 percent for elderly nondisabled Georgians. Table 149. Behavioral adaptations among disabled and older adults. Behavioral Adaptation With Mobility Impairment Working All Adults Age Seniors Elderly 18+ (1864) (6579) (80+) Without Mobility Impairment Elderly (80+) Reduced day-to-day travel 69.6% 67.5% 70.8% 76.8% 30.9% Asked others for rides 39.9% 44.1% 31.0% 38.5% 20.1% Limited driving to daytime 26.7% 24.8% 31.5% 22.9% 34.9% Given up driving altogether 24.9% 18.4% 27.0% 52.0% 15.8% Used the bus or subway less frequently 8.1% 9.8% 5.8% 5.0% 4.0% Used special transportation services such as Dial-A-Ride 9.8% 11.3% 7.5% 7.0% 0.5% Used a reduced fare taxi 1.9% 1.4% 2.1% 3.7% 0.1% Note: Values shown are percentage of people in each age group (column) who reported engaging in the listed behavior (row). Compared to nondisabled elderly Georgians, disabled elderly Georgians are also more likely to ask others for rides and use reduced-fare taxis and special transit services. They are slightly more likely to reduce their usage of regular public transit. More than half of elderly Georgians with a mobility impairment have given up driving, versus just 15.8 percent of those without. Elderly Georgians without a mobility impairment were instead more likely to limit driving to daytime (34.9 percent). Children with Mobility Impairments Among Georgia children ages 517, 2.2 percent have a mobility impairment (table 150). Half of these children were immobile on the travel day, versus 22 percent of children without a mobility impairment. However, the sample contains just 42 children with a mobility impairment, and thus 332 the statistics presented here should be interpreted with caution. Nevertheless, the high frequency of immobility among disabled children is indicative of a mobility disadvantage that might benefit from more targeted data collection and study. Table 150. Comparison of children (ages 517) with and without mobility impairments. Demographics for Children Ages 517 Mobility impairment present Mobility impairment absent Percent (weighted) 2.2% 97.8% Mobility Aids (children with mobility impairments) None 86.4% W heelchair 12.2% Vision aid (white cane) 0.6% Other (not specified) 0.8% Immobility Among Children Ages 517 (Weighted) Mobility impairment present Mobility impairment absent Immobile on Travel Day 50.5% 22.0% Sample Size (unweighted) 42 2,416 32 6 2 2 Immobile for Past Seven Days 7.1% 2.1% Variations among Adults with Mobility Impairments Table 151 shows demographic characteristics of Georgia's 723,000 residents with mobility impairments.96 One out of 20 mobility impairments are short term (having lasted fewer than 6 months).97 Women account for 61 percent of people with mobility impairments. This is not 96 Table 151 also contains sample information for table 152 to table 154. 97 We considered eliminating or separating those with short-term disabilities from the analysis of those with longer-term limitations. However, short-term is not synonymous with temporary: an unknown share of disabilities that are less than 6 months old may, in fact, simply be newly occurring long-term disabilities. In view of this fact, plus the relatively small share of short-term disabilities in the sample (5 percent), we chose to retain all cases of mobility limitation for analysis. Short-term disability is included as a covariate in the regressions. 333 fully accounted for by women's longer life expectancies; disability is also more prevalent among working-age women compared to working-age men. 334 Table 151. Demographic characteristics of Georgians with mobility impairments. Population* Sample* (unweighted) All adults with mobility impairment 723,241 1,609 Mobility Aid Usage None Wheelchair (incl. wheelchair and other) Cane, walker, or other Duration of Mobility Impairment 43.2% 19.1% 37.8% 38.9% 20.4% 40.6% Long-term (more than 6 months) Short-term (6 months or less) 95.0% 5.0% 93.2% 6.8% Age Working age (1864) Senior (6579) Elderly (80+) 59.8% 27.8% 12.4% 47.8% 33.4% 18.8% Men Only Working age (1864) Senior (6579) Elderly (80+) 39.2% 25.1% 11.1% 3.0% 38.3% 18.8% 13.4% 6.2% Women Only Working age (1864) Senior (6579) Elderly (80+) 60.8% 34.7% 16.7% 9.4% 61.7% 29.0% 20.0% 12.7% Driving and Paratransit Driver Nondriver Paratransit user Paratransit nonuser 61.6% 38.4% 10.8% 89.2% 67.2% 32.8% 6.7% 93.3% MPO Tier 1. Atlanta MPO 2. Medium MPOs 3. Small MPOs 4. Non-MPO Annual Household Income 44.0% 16.7% 9.1% 30.2% 23.3% 36.4% 25.1% 15.2% <$15,000 $15,000 to $24,999 $25,000 to $34,999 $35,000 to $49,999 $50,000 to $74,999 $75,000 to $99,999 $100,000+ 36.7% 17.3% 11.9% 8.7% 11.9% 6.7% 6.8% 27.5% 16.5% 12.6% 11.9% 13.0% 8.0% 7.2% Workforce Participation Nonworker 87.7% 89.9% Worker 12.3% 10.1% * Population is based on nonmissing observations and sample includes missing observations. Income data are missing for 54 observations (3.4%) and duration of mobility impairment is missing for one observation. Respondents were asked, "Do you/does this person drive?" Paratransit includes reduced-fare taxis and services like Dial-A-Ride. Includes crutches, white cane, service dog, and other (specify). Other (specify) includes brace, respiratory assistance, and prosthesis. 335 The majority of Georgians with mobility impairments drive, but a substantial minority (38 percent) do not. However, just 11 percent of Georgians with mobility impairments use paratransit services. Given how prevalent immobility is among Georgians with disabilities (see table 151), the difference between the number of Georgians with mobility impairments who do not drive and the number who use paratransit suggests unmet transportation needs. Two out of five people with mobility impairments use no mobility aids. One out of five uses a wheelchair, sometimes in conjunction with other mobility aids. As shown in figure 36, the most common mobility aid reported by respondents was a cane (36.5 percent), followed by a walker (22.9 percent) and a wheelchair (regular, motorized, or motorized scooter) (19.2 percent). Respondents could report multiple aids; 72.1 percent of wheelchair users reported using at least one other mobility aid. For example, someone who uses a wheelchair to travel outside of the house may also use a walker to travel short distances or transfer between rooms in their own home. The much higher prevalence of crutches among Georgians with short-term disabilities (18, versus 2 percent of those with long-term disabilities) suggests that some short-term disabilities are more likely to be acute orthopedic injuries. Figure 36. Bar graph. Mobility aids used by adults with mobility impairments. 336 Figure 34 compared the annual household incomes of adults with mobility impairments and adults without mobility impairments; table 152 further examines demographic differences in annual household income and driving specifically among Georgians with mobility impairments. More than half of people with disabilities have an annual household income of less than $25,000 (figure 34). More than a third fall into the very lowest income category (less than $15,000 per year), including 20 percent of workers (table 152). For comparison, this is more than twice the share of nondisabled workers in very low-income households (8 percent--not shown in the table). Mobility-impaired women of all ages are more likely than men in that same age group to be in a very low-income household, and nearly half of working-age women with disabilities are in very low-income households. Poverty is, perhaps unsurprisingly, less common among Georgians with short-term disabilities, of whom 10 percent live in a very low-income household (compared to 38 percent of those whose disabilities have lasted longer than 6 months). Table 153 shows differences in mobility and immobility among Georgians with mobility impairments. In general, wheelchair users are more likely to be immobile than other Georgians with disabilities, as are those whose disability is short term, the elderly (but markedly more so for elderly women), nonworkers, and nondrivers. These patterns will be explored further in chapter 5, Risk Factors for Immobility among Adults with Mobility Impairments. Table 154 shows walking and physical activity among Georgians with mobility impairments. Wheelchair users are less likely to report walking than other Georgians, but it is notable that 34 percent of wheelchair users report making at least one walking trip. This may reflect part-time wheelchair users, but it is also possible that participants are reporting wheelchair trips as 337 "walking" because the vocabulary used in the NHTS questions does not address pedestrian travel by wheelchair users. 338 Table 152. Driver status and income of Georgia adults with mobility impairments. Driver* Very Low Income (<$15,000) All adults with mobility impairment 61.6% 36.7% Mobility Aid Usage None Wheelchair (incl. wheelchair and other) Cane, walker, or other 61.4% 52.8% 66.2% 36.9% 34.3% 37.8% Duration of Mobility Impairment Long-term (more than 6 months) Short-term (6 months or less) 61.0% 72.2% 38.2% 9.8% Age Working age (1864) Senior (6579) Elderly (80+) 63.7% 68.4% 36.2% 44.3% 25.1% 26.3% Men Only Working age (1864) Senior (6579) Elderly (80+) 61.1% 56.7% 73.2% 52.3% 32.8% 39.8% 21.2% 17.1% Women Only Working age (1864) Senior (6579) Elderly (80+) 61.9% 68.7% 65.2% 31.0% 39.3% 47.6% 27.8% 29.3% Driving and Paratransit Driver Nondriver Paratransit user (e.g., reduced fare taxi or Dial-a-Ride) Paratransit nonuser 100.0% -- 36.2% 64.7% 31.1% 46.0% 70.8% 32.6% MPO Tier 1. Atlanta MPO 2. Medium MPOs 3. Small MPOs 4. Non-MPO Annual Household Income 55.7% 62.0% 59.9% 70.5% 32.2% 37.1% 38.5% 42.6% <$15,000 $15,000 to $24,999 $25,000 to $34,999 $35,000 to $49,999 $50,000 to $74,999 $75,000 to $99,999 $100,000+ 52.5% 65.6% 67.7% 63.8% 68.1% 73.2% 70.9% 100.0% -- -- -- -- -- -- Workforce Participation Nonworker 59.3% 39.0% Worker 77.9% 20.3% * Respondents were asked, "Do you/does this person drive?" Annual household income is missing for 54 observations and duration of mobility impairment is missing for one observation. All figures based on nonmissing observations. Includes crutches, white cane, and service dog, and other (specify). Other (specify) includes brace, respiratory assistance, and prosthesis. 339 Table 153. Immobility and average daily trips of Georgians with mobility impairments. Immobility (Zero Trips in Time Frame) Daily Trips Travel Day Past Few Past Seven Per Days Days Capita Active Travelers Only All adults with mobility impairment 39.5% 15.8% 10.0% 2.33 3.85 Mobility Aid Usage None 33.7% 12.8% 8.9% 2.35 3.54 Wheelchair (incl. wheelchair and other) 50.7% 24.8% 13.2% 1.77 3.58 Cane, walker, or other 40.4% 14.8% 9.7% 2.58 4.34 Duration of Mobility Impairment Long-term (more than 6 months) 39.4% 15.6% 9.6% 2.31 3.82 Short-term (6 months or less) 40.1% 20.9% 19.0% 2.59 4.33 Age Working age (1864) Senior (6579) Elderly (80+) 34.1% 13.0% 8.4% 2.57 3.89 41.9% 16.2% 8.2% 2.28 3.92 60.1% 29.0% 22.2% 1.28 3.20 Men Only Working age (1864) Senior (6579) Elderly (80+) 37.3% 14.8% 11.0% 2.36 3.77 36.2% 15.1% 10.9% 2.39 3.75 39.7% 13.8% 10.6% 2.37 3.92 37.6% 16.2% 12.7% 2.09 3.35 Women Only Working age (1864) Senior (6579) Elderly (80+) Driving and Paratransit 40.9% 16.5% 9.4% 2.31 3.90 32.5% 11.5% 6.5% 2.69 3.99 43.4% 17.7% 6.6% 2.22 3.93 67.2% 33.0% 25.2% 1.02 3.11 Driver Nondriver Paratransit user Paratransit nonuser 31.9% 9.6% 4.5% 2.74 4.02 51.6% 25.9% 18.9% 1.67 3.45 31.8% 10.6% 9.1% 2.39 3.51 40.4% 16.5% 10.1% 2.32 3.89 MPO Tier 1. Atlanta MPO 2. Medium MPOs 3. Small MPOs 4. Non-MPO 39.2% 17.5% 11.7% 2.22 3.66 37.2% 14.1% 9.9% 2.49 3.97 39.9% 17.7% 10.9% 2.57 4.28 40.9% 13.8% 7.3% 2.32 3.92 Annual Household Income <$15,000 $15,000 to $24,999 $25,000 to $34,999 $35,000 to $49,999 $50,000 to $74,999 $75,000 to $99,999 $100,000+ 38.7% 14.1% 8.3% 2.43 3.97 35.4% 15.2% 12.2% 2.41 3.74 48.3% 23.7% 11.9% 1.81 3.50 29.9% 14.8% 11.4% 2.81 4.01 46.1% 19.1% 12.4% 2.14 3.98 34.2% 11.7% 7.1% 2.73 4.15 46.8% 13.7% 7.0% 1.91 3.58 Workforce Participation Nonworker Worker 42.1% 17.2% 10.6% 2.21 3.82 20.9% 6.1% 5.7% 3.16 3.99 Based on 1,609 adults with mobility impairments. Income is missing for 54 observations and duration of impairment for one observation. Reporting at least one trip on the travel day. Includes crutches, white cane, service dog, and other (specify). Other (specify) includes brace, respiratory assistance, and prosthesis. Respondents were asked, "Do you/does this person drive?" Paratransit includes reduced-fare taxis and special services such as Dial-A-Ride. 340 Table 154. Walking and physical activity among Georgians with mobility impairments. Walk Trips (Past 30 Days) Physical Activity Level (Typical Week) Any Walk Trip Number of Walk Trips Active (some light, moderate, or vigorous activity) Inactive (rarely or never) All adults with mobility impairment 60.1% 5.0 64.8% 35.2% Mobility Aid Usage None 72.8% 6.1 Wheelchair (incl. wheelchair and other) 34.4% 2.3 Cane, walker, or other 58.5% 5.1 71.8% 51.8% 63.4% 28.2% 48.2% 36.6% Duration of Mobility Impairment Long-term (more than 6 months) Short-term (6 months or less) 60.4% 5.1 54.6% 2.7 63.9% 81.8% 36.1% 18.2% Age Working age (1864) Senior (6579) Elderly (80+) 70.2% 6.3 48.1% 3.7 37.8% 2.0 69.3% 62.4% 48.7% 30.7% 37.6% 51.3% Men Only Working age (1864) Senior (6579) Elderly (80+) 62.5% 5.4 70.7% 6.4 46.6% 3.6 52.7% 3.2 65.3% 68.7% 56.8% 68.0% 34.7% 31.3% 43.2% 32.0% Women Only Working age (1864) Senior (6579) Elderly (80+) Driving and Paratransit 58.5% 4.8 69.9% 6.2 49.1% 3.7 33.1% 1.6 64.5% 69.7% 66.1% 42.5% 35.5% 30.3% 33.9% 57.5% Driver Nondriver Paratransit user Paratransit nonuser 62.4% 5.6 56.3% 4.1 62.8% 4.5 59.7% 5.1 69.0% 58.1% 63.9% 64.9% 31.0% 41.9% 36.1% 35.1% MPO Tier 1. Atlanta MPO 2. Medium MPOs 3. Small MPOs 4. Non-MPO 57.6% 4.5 58.4% 3.8 65.3% 4.7 63.0% 6.4 65.6% 68.1% 71.9% 59.6% 34.4% 31.9% 28.1% 40.4% Annual Household Income <$15,000 $15,000 to $24,999 $25,000 to $34,999 $35,000 to $49,999 $50,000 to $74,999 $75,000 to $99,999 $100,000+ 69.6% 6.1 50.9% 3.7 60.9% 6.2 55.9% 3.2 52.5% 5.7 62.0% 3.8 43.5% 2.8 69.1% 57.8% 64.8% 70.8% 56.7% 71.6% 60.7% 30.9% 42.2% 35.2% 29.2% 43.3% 28.4% 39.3% Workforce Participation Nonworker Worker 58.0% 4.9 75.1% 6.1 63.4% 75.0% 36.6% 25.0% Based on 1,609 adults with mobility impairments. Income is missing for 54 observations and duration of impairment for one observation. Includes crutches, white cane, service dog, and other (specify). Other (specify) includes brace, respiratory assistance, and prosthesis. Respondents were asked, "Do you/does this person drive?" Paratransit includes reduced-fare taxis and special services such as Dial-A-Ride. 341 Risk Factors for Immobility among Adults with Mobility Impairments The researchers use logistic regression to disentangle the demographic factors associated with immobility among people with mobility impairments (table 155).98 We created three binary logit models looking at the likelihood of being immobile on the travel day, immobile for "the past few days," and immobile for the past week.99 Table 156 isolates the AME of key covariates. Being in poor health is associated with a 12-percentage-point increase in the probability of being immobile on the travel day, compared to a mobility-impaired Georgian in good health. It is also associated with an 11-percentage-point increase in the probability of being immobile for the past few days, and a 7-percentage-point increase in the probability of being immobile for the past week. Being in fair health has a similar effect on the travel day (an 11-percentage-point increase in the probability of travel day immobility), but the effect drops off sharply when examining longer term immobility. In other words, on any given day, disabled Georgians in fair or poor health are equally likely to be housebound, but Georgians in fair health are more likely to be able to leave the house at least a few times per week. Wheelchair use is also associated with an increased likelihood of immobility, and working outside the home and being a driver are associated with a decreased likelihood of immobility. Living alone is associated with a decreased likelihood of immobility, but this may be because people able to live independently are also more able to travel independently. Paratransit use is 98 Income was excluded from the models because it was not significant and was missing for 3.3 percent of observations. 99 Respondents with zero trips on the travel day were asked about the date of their most recent trip. Response options included: (1) "The day before," (2) "A few days before," (3) "A week before," (4) "More than a week but within a month," and (5) "More than a month." This report considers respondents to be immobile for the past few days if they selected response 3, 4, or 5 and immobile for the past week if they selected response 4 or 5. A respondent who has been immobile for the past week has also been immobile for the past few days and on the travel day. 342 associated with a lower probability of immobility, but the effect is not significant. Being female is associated with an increased chance of immobility, but only among elderly residents. Urban neighborhoods appear to confer some mobility benefits; people with mobility impairments living in urban neighborhoods are less likely to be immobile on the travel day and over the past several days; there is no significant difference in weekly immobility. 343 Table 155. Logistic regression: Immobility among Georgia adults with a mobility impairment. 344 Table 156. Selected average marginal effects on probability of being immobile among adults with a mobility impairment, percentage points. 1. Immobile on Travel Day Mean predicted probability x 100 38.58 Average Marginal Effects (percentage points) 2. Immobile, Past Few Days 15.95 3. Immobile, Past Seven Days 10.00 Health (reference: good, very good, or excellent) Fair 10.46 *** 5.17 ** 3.45 ** Poor 11.93 *** 11.39 *** 7.31 *** Wheelchair user 7.74 ** 5.40 ** 2.92 Live alone Worker Status (reference: nonworker) Work outside of home (nonhome-based) -7.75 *** -16.88 *** -2.12 -7.98 0.31 -1.79 Work from home (home-based) 0.65 * -5.58 -4.44 Driver Urban neighborhood -19.18 *** -7.03 ** -14.58 *** -3.84 * -10.26 *** -1.57 * Indicates that model coefficient was significant for = .10, ** for = .05, and *** for = .01. Respondents with zero trips on the travel day were asked about the date of their most recent trip. Response options included 1. "The day before," 2. "A few days before," 3. "A week before," 4. "More than a week but within a month," and 5. "More than a month." This report considers respondents to be immobile for the past few days if they selected response 3, 4, or 5 and immobile for the past week if they selected response 4 or 5. A respondent who has been immobile for the past week has also been immobile for the past few days and on the travel day. Weighted mean predicted probability based on unweighted logit models. This report displays predicted probabilities as a percentage rather than observed means for consistency. (The two values are not identical because survey weights have been applied to results produced by an unweighted model. For all variables, the two values differ by less than one percentage point.) Since the explanatory variables presented are dummy variables indicating the presence of a certain characteristic, the AME for each variable is calculated by predicting the probability of being immobile as if the whole sample did not have the characteristic in question, predicting the probability as if the whole sample did have the characteristic in question, subtracting the former from the latter, and averaging the predicted marginal effect for all observations in the sample. Effects shown are weighted means based on the unweighted logit models. For nonhome-based workers, worker = 1 and work from home = 0. For home-based workers, both worker and work from home = 1. Urban or second-city neighborhood type versus suburban, small-town, or rural. Taken together, the findings presented here indicate that Georgians with disabilities face multiple barriers to mobility. Policy is needed to address accessibility barriers in public and private 345 transportation and on the public right of way, as well as to address related sources of social exclusion (Beyzak et al. 2019, Decker 2006, National Council on Disability 2015). 346 CHAPTER 6. NONMOTORIZED TRANSPORTATION AND ACCESS/EGRESS TRAVEL CHAPTER 6 SUMMARY This chapter examines nonmotorized travel (i.e., walking and biking) and access/egress travel, or travel to reach a primary mode of transportation. Overview provides an overview of how many Georgians walk and bike over a typical week and examines some barriers to walking and biking more frequently. It discusses gender differences in concerns about the safety of NMT and reviews the available data about physical activity. Access and Egress Travel discusses access and egress travel. It reviews how the NHTS measures access/egress legs for transit and nontransit trips and discusses related measurement issues. Access/egress travel is particularly important when studying nonmotorized travel because walking and biking account for a large proportion of access/egress travel; including access/egress legs increases the number of instances of walking and biking per capita by 28 percent versus including separately recorded nonmotorized trips alone. This section especially examines access/egress to public transit, for which NMT is a dominant mode. Travel Day Walking and Biking by Georgia Adults examines travel-day walking and biking by Georgia adults. It incorporates the trips analyzed throughout this report, as well as the legs analyzed in Access and Egress Travel in this chapter. Frequency of walking and biking, purpose of NMT trips/legs, and duration of trips by purpose are examined and demographic differences explored. Time of day is also examined. 347 Captive and Choice Nonmotorized Travel examines captive and choice nonmotorized travel. As with public transit, some pedestrians and cyclists choose to walk and bike, while other use these modes by necessity. Captive pedestrians and cyclists spent more time traveling than their choice counterparts did, after controlling for trip quantity and purpose. Captive nonmotorized travelers also tended to make more nonmotorized trips, compounding the differences in total travel time. While increased walking and biking is broadly considered a public health goal, it is important to remember that some travelers are already walking or riding more than they would like. Children's Nonmotorized and School Travel examines nonmotorized and school travel for children ages 517. After describing the research team's methods for identifying school trips, which are analogous to the methods used to define adults' work commutes in chapter 2, this section reviews children's observed and usual modes of travel to and from school. Walking and biking account for approximately 5 percent of school travel in Georgia. It then describes children's nonmotorized travel for all purposes. Children are more likely to have walked or biked than are adults. The difference is particularly pronounced for cycling; 30.5 percent of children reported cycling in the past 7 days versus 5.5 percent of adults. Cycling is most common among children under the age of 9 and declines steadily as they approach adulthood. 348 OVERVIEW As shown in figure 37, in a typical week, 72.6 percent of Georgia adults will walk, ride a bike, or both.100 Cyclists are, in general, also pedestrians; only 0.4 percent of Georgians reported biking but not walking. Figure 37. Pie chart. Use of nonmotorized modes by Georgia adults (past 7 days). As shown in table 157, walkers average 7.9 walking trips per week, and cyclists average 3.1 bike trips. When walking and biking are combined, nonmotorized travelers make an average of 8.1 nonmotorized trips per week. 100 The precise questions were: "In the past seven days, how many times did you: 1. "...take a walk outside including walks to exercise, go somewhere, or to walk the dog (e.g., walk to a friend's house, walk around the neighborhood, walk to the store, etc.)?" 2. "... ride a bicycle outside including bicycling to exercise, or to go somewhere (e.g., bike to a friend's house, bike around the neighborhood, bike to the store, etc.)? 349 There are some differences between MPO tiers, but the larger differences are by neighborhood type, which can vary substantially within an MPO. NMT is most common in the densest urban neighborhood type (which, in Georgia, is only found in Atlanta). Women are more likely to walk than men, but less likely to bike. Biking is most common among those ages 1852 compared to other age groups, while walking is most common from ages 3764. Pedestrians and cyclists are most common among residents of the lowest-income (less than $15,000 per year) and highest-income ($100,000 or more per year) households, compared to their incidence among other income groups. However, low-income pedestrians and cyclists walk and ride their bikes more frequently than do their wealthier counterparts (9.9 times and 7.3 times a week, respectively). 350 Table 157. Walking and biking among Georgia adults (past 7 days). Percent of Adults who have Used Mode (Past 7 Days) Mean Trips among Mode Users (Past 7 Days) All adults ages 18+ MPO Tier 1. Atlanta MPO 2. Medium MPOs 3. Small MPOs 4. Non-MPO counties Urbanicity (Neighborhood Type) Rural Small town Suburban Second city Urban Sex Male Female Age Cohort Millennial and Gen Z (1836) Gen X (3752) Pre-retirement age Boomer (5364) Retirement Age (65+) Caregiver Status Noncaregiver Caregiver, youngest child ages 015 Race White non-Hispanic only Black, Black multiracial & Black Hisp. Other Table continues on next page. Walk 72.2% Bike 5.5% Any NMT 72.5% Walk 7.9 Bike 3.1 Any NMT 8.1 71.7% 5.1% 72.0% 7.2 2.6 7.3 72.1% 7.6% 72.8% 8.0 4.1 8.3 72.4% 6.1% 73.0% 8.6 3.4 8.9 73.3% 4.4% 73.5% 9.6 3.3 9.8 72.9% 4.0% 73.1% 9.2 2.9 9.3 67.4% 5.6% 68.0% 8.0 2.9 8.2 72.5% 5.5% 72.8% 7.1 3.0 7.3 75.0% 5.3% 75.4% 7.4 3.7 7.7 91.0% 18.5% 91.7% 9.3 3.5 10.0 71.6% 6.8% 72.2% 8.2 3.3 8.4 72.7% 4.3% 72.9% 7.7 2.9 7.9 70.9% 6.2% 71.2% 8.3 3.3 8.5 75.3% 6.3% 75.6% 7.3 3.3 7.6 74.1% 5.3% 74.7% 8.3 3.0 8.4 67.1% 2.8% 67.4% 7.9 2.4 8.0 72.2% 5.1% 72.7% 8.0 3.3 8.2 72.1% 6.3% 72.3% 7.7 2.8 8.0 74.2% 5.9% 74.6% 8.3 3.0 8.5 69.0% 4.8% 69.3% 7.5 3.7 7.7 71.3% 5.2% 71.5% 7.2 2.4 7.4 351 Table 157. (Continued). Barriers to Walking and Biking More Frequently How could Georgians be enticed to walk or bike more? Table 158 shows perceived barriers to walking and biking more frequently. The most common complaints for pedestrians are missing or inadequate sidewalks and insufficient night lighting. For cyclists, heavy traffic is the most common complaint, with missing/inadequate sidewalks and a lack of nearby paths or trails coming in second and third. 352 Table 158. Perceived barriers to walking and biking more frequently (pedestrians and cyclists ages 18+). Reason for not walking/biking more Walk more Bike more No nearby paths or trails 18.1% 20.0% No nearby parks 14.4% 13.3% No sidewalks or sidewalks are in poor condition 22.5% 20.9% Street crossings are unsafe 11.7% 15.6% Heavy traffic with too many cars 17.4% 27.5% Not enough lighting at night 21.4% 19.1% None of the above 38.3% 33.4% Participants were allowed to select multiple response options. Walk questions were asked of people who walked at least once in the past 7 days; bike questions were asked of people who biked at least once in the past 7 days. These questions were only asked of people who had walked or biked at least once in the 7 days previous to the survey. As such, they describe barriers to walking and biking more, but not necessarily barriers to walking and biking at all. The NHTS does not have direct information about the concerns of the 94.5 percent of Georgians who have not biked recently; more information is needed about how to attract new pedestrians and cyclists. As shown in figure 38, there are notable gender differences in perceived barriers to walking and biking. Women selected most barriers more frequently than men (with the sole exception of a lack of nearby parks, which male cyclists selected more than female cyclists). The largest gender gaps among pedestrians focused on safety: night lighting (9.8 percentage point difference between men and women) and sidewalks (7.2 percentage point difference). Among cyclists, the largest gender gaps were around sidewalk condition (9.8 percentage points), heavy traffic (9.5 percentage points) and a lack of nearby paths or trails (9.2 percentage points). A number of studies have documented how unfavorable traffic and safety conditions disproportionately discourage female nonmotorized travelers (e.g., Emond, Tang, and Handy 2009). 353 Figure 38. Bar graphs. Gender differences in perceived barriers to walking and biking more frequently among pedestrians and cyclists ages 18+. 354 Physical Activity In this section, we examine physical activity between different groups of Georgians, and in relationship to walking and biking. As shown in table 159, 85 percent of Georgians are at least somewhat physically active in a typical week; 61.5 percent reported some light or moderate physical activity. Twenty-four percent reported vigorous physical activity; whether these respondents also engaged in light or moderate physical activity is not recorded. Respondents who engaged in light/moderate physical activity reported an average of four sessions of at least 30 minutes per week, while those who reported vigorous activity reported an average of 5.2 sessions. Residents of urban neighborhoods are less likely to be inactive than residents of other types of neighborhoods (6.5 percent versus 1416 percent elsewhere). They are much more likely to engage in vigorous physical activity (41 percent versus 2225 percent elsewhere). However, urban residents engaged in vigorous physical activity report fewer instances per person than their vigorously active counterparts in other types of areas (4 percent versus 4.95.8 percent elsewhere). Men are more active than women. Low-income people, nondrivers, and people with mobility impairments are much more likely to report no physical activity. As discussed in chapter 5, Health and Disability, there is considerable overlap between those groups. 355 All adults ages 18+ MPO Tier 1. Atlanta MPO 2. Medium MPOs 3. Small MPOs 4. Non-MPO counties Urbanicity (Neighborhood Type) Rural Small town Suburban Second city Urban Sex Male Female Age Cohort Millennial and Gen Z (1836) Gen X (3752) Pre-retirement age Boomer (5364) Retirement Age (65+) Driver Status Nondriver Driver Table continues on next page. Table 159. Physical activity among Georgia adults. Physical Activity Level (Row Percentages )* Rarely or Never 14.6% Some Light or Moderate 61.5% Some Vigorous 23.8% Weekly Instances of Physical Activity Light/Moderate (Among Vigorous (Among Light/Moderately Active Respondents with Vigorous Respondents) Physical Activity) 4.03 5.22 14.2% 61.9% 23.9% 3.94 4.84 15.4% 61.1% 23.5% 3.77 5.44 13.6% 63.0% 23.4% 3.98 5.51 15.8% 60.3% 23.9% 4.51 5.96 15.1% 59.5% 25.4% 4.41 5.81 15.7% 61.6% 22.7% 3.88 5.18 13.8% 62.7% 23.5% 3.99 4.92 14.9% 63.4% 21.8% 3.86 5.24 6.5% 52.5% 41.0% 4.29 4.05 13.8% 55.8% 30.4% 4.15 5.48 15.4% 66.9% 17.8% 3.93 4.83 15.5% 56.1% 28.4% 4.08 5.53 13.8% 61.4% 24.7% 3.91 5.04 12.5% 66.0% 21.5% 3.99 5.14 16.8% 67.4% 15.8% 4.17 4.72 27.0% 59.4% 13.6% 4.03 6.64 13.1% 61.8% 25.1% 4.03 5.13 356 Table 159. (Continued). Continued from previous page. Physical Activity Level (Row Percentages )* Weekly Instances of Physical Activity Light/Moderate (Among Vigorous (Among Rarely or Some Light or Some Light/Moderately Active Respondents with Vigorous Never Moderate Vigorous Respondents) Physical Activity) All adults ages 18+ 14.6% 61.5% 23.8% 4.03 5.22 Race White non-Hispanic only 12.7% 60.3% 26.9% 4.23 5.30 Black, Black multiracial & Black Hispanic 17.2% 62.9% 19.9% 3.66 4.85 Other 16.3% 63.6% 20.1% 4.09 5.69 Mobility Impairment: a "condition or handicap that makes it difficult to travel outside of the home." Absent 12.5% 61.8% 25.7% 4.11 5.22 Present 35.4% 59.1% 5.5% 3.22 5.44 Annual Household Income <$15,000 19.7% 64.0% 16.3% 4.13 6.04 $15,000 to $49,999 15.7% 62.6% 21.7% 4.10 5.46 $50,000 to $99,999 13.6% 64.5% 21.8% 4.05 5.32 $100,000+ 11.7% 55.3% 33.0% 3.82 4.73 Vehicle Deficit Category of Household (a vehicle-deficit household owns at least one vehicle, but fewer vehicles than potential drivers.) Zero-vehicle 16.3% 68.1% 15.6% 3.80 5.66 Deficit 20.1% 60.5% 19.3% 3.96 5.60 Nondeficit (sufficient/surplus) 12.5% 61.5% 26.1% 4.07 5.10 Worker Status Nonworker 18.8% 65.4% 15.8% 4.20 5.13 Worker 12.0% 59.1% 28.8% 3.91 5.26 * Question wording: "Which of the following statements best describes how physically active you are in a typical week? (1) I rarely or never do any physical activity; (2) I do some light or moderate physical activity; (3) I do some vigorous physical activities." Respondents who reported light moderate physical activity were asked about light/moderate physical activity. Respondents who reported vigorous activity were only asked about vigorous activity; data on light/moderate physical activity by this group were not collected. N=15,120. Table excludes observations missing number of walk or bike trips, physical activity, health, or disability for later comparison on these variables. 357 How does the physical activity of the 73 percent of Georgians who reported walking and/or biking in the previous week compare to the physical activity of the 27 percent who did not walk or bike? As shown in table 160, pedestrians and cyclists are less likely to be sedentary than Georgians who do not walk or bike. Only 8.8 percent of Georgians who walk or bike reported rarely or never engaging in physical activity, versus 30.2 percent of those who did not walk or bike. Among the physically active, pedestrians and cyclists also report a higher number of activity sessions. Cyclists are the most physically active group. Nearly half of cyclists report vigorous physical activity, versus one quarter of Georgians who walked only and 16 percent of Georgians who neither walked nor biked. 358 Table 160. Usual physical activity of Georgia adults by walking and biking behavior (past 7 days). Physical Activity Level (Row Percentages )* Rarely or Some Light or Never Moderate All adults ages 18+ 14.6% 61.5% Any Nonmotorized Travel (Past Seven Days) None (zero walk and bike trips) 30.2% 54.2% Some (1+ walk or bike trip(s)) 8.8% 64.3% Types of Nonmotorized Travel (Past Seven Days) None 30.2% 54.2% Walk only 9.0% 65.7% Bike only and bike + walk 5.2% 47.2% Number of Nonmotorized Trips (Walk + Bike) (Past 7 Days) 0 30.2% 54.2% 15 (group mean: 3.1 trips) 10.5% 67.0% Some Vigorous 23.8% 15.6% 26.9% 15.6% 25.2% 47.6% 15.6% 22.5% Weekly Instances of Physical Activity Light/Moderate (Among Vigorous (Among Light/Moderately Active Respondents with Vigorous Respondents) Physical Activity) 4.03 5.22 3.41 5.00 4.23 5.27 3.41 5.00 4.23 5.33 4.18 4.87 3.41 5.00 3.44 4.54 6+ (group mean: 13.8 trips) 6.7% 61.2% 32.0% 5.21 5.86 Number of Walk Trips (Past 7 Days) 0 30.1% 54.1% 15.9% 3.41 5.01 15 (group mean: 3.1 trips) 10.3% 66.7% 23.0% 3.44 4.52 6+ (group mean: 13.8 trips) 6.8% 61.7% 31.5% 5.26 5.94 Number of Bike Trips (Past 7 Days) 0 15.2% 62.4% 22.4% 4.02 5.27 12 (group mean: 1.4 trips) 6.2% 49.1% 44.7% 3.99 4.99 3+ (group mean: 5.6 trips) 3.7% 44.6% 51.7% 4.47 4.72 * Question wording: "Which of the following statements best describes how physically active you are in a typical week? (1) I rarely or never do any physical activity; (2) I do some light or moderate physical activity; (3) I do some vigorous physical activities." Respondents who reported light moderate physical activity were asked about light/moderate physical activity. Respondents who reported vigorous activity were only asked about vigorous activity; data on light/moderate physical activity by this group were not collected. Category boundaries were chosen based on the median number of trips of each type. Fifty-three percent of NMT travelers made 15 trips, 55 percent of walkers made 15 walking trips, and 59 percent of cyclists made 12 trips. N=15,120. Table excludes observations missing number of walk or bike trips, physical activity, health, or disability for later comparison on these variables. 359 Not surprisingly, pedestrians and cyclists are more likely to report being active in general. It would be useful to know whether NMT replaces other physical activity, or supplements it. However, the following data limitations make it difficult to document the relationship between number of walking and biking trips and the specific number of sessions of physical activity: Because people who reported engaging in vigorous physical activity were not asked how many times they engage in moderate or light physical activity, there is no measure of total physical activity for each respondent. There is no way to distinguish between someone who went to the gym once and also engaged in light exercise throughout the rest of the week from someone who went to the gym once and engaged in no other exercise. The measures of physical activity are generic (a typical week) while the measures of walking and biking are specific (the past 7 days). Information about how many of the instances of physical activity reported are walking versus biking is not available. The extent of overlap between reported walk/bike trips and recorded physical activity bouts is, therefore, unclear. To facilitate more in-depth analysis of this topic, the NHTS could consider: Asking participants about specific instances of light, moderate, and physical activity in the past 7 days. Recording whether bouts of physical activity are walking/running, biking, or something else. Alternatively, asking participants to report total minutes of light, moderate, and vigorous physical activity over the past week. 360 ACCESS AND EGRESS TRAVEL As shown in table 161, Georgians make more than 950,000,000 nonmotorized trips per year. However, in addition to these trips, which are analyzed throughout this report, Georgians also walk and bike as a way to get to and from other modes of transportation (i.e., access/egress legs). As shown in table 161, Georgians used walking or biking to access/egress another mode more than 260,000,000 times per year.101 Table 161. Total annual trips and access/egress legs by nonmotorized modes. Total Nonmotorized Trips* per Year Walk Bike All NMT All Georgians Ages 5+ 887,174,800 66,663,125 953,837,925 MPO Tier 1. Atlanta MPO 539,657,200 23,250,820 562,908,020 2. Medium MPOs 141,069,100 17,425,210 158,494,310 3. Small MPOs 74,786,200 6,015,215 80,801,415 4. Non-MPO counties 131,662,300 19,971,880 151,634,180 * A trip is a unit of travel from one origin to one destination. Total Nonmotorized Access/Egress Legs per Year Walk Bike All NMT All Georgians Ages 5+ 260,397,290 3,534,818 263,932,108 MPO Tier 1. Atlanta MPO 203,008,300 2,342,461 205,350,761 2. Medium MPOs 32,361,740 1,192,357 33,554,097 3. Small MPOs 12,374,430 - 12,374,430 4. Non-MPO counties 12,652,820 - 12,652,820 An occasion where a respondent walks or bikes to access or egress another mode of transportation (e.g., public transit, commuter train). These legs are usually included as part of the trip by the primary mode rather than as individual trips. Thus, nonmotorized access/egress legs are essentially mutually exclusive with nonmotorized trips. 101 Two additional sources of nonmotorized travel are not reflected in the NHTS: stationary fitness NMT (i.e., a stationary bicycle or treadmill) and walking within a destination (e.g., a shopping mall or amusement park). 361 The following subsections will describe how the NHTS is designed to measure access/egress travel, discuss some deviations between the survey as written and some of the resulting data, and then discuss trends in access/egress travel. Measurement of Access/Egress Travel The NHTS has different procedures for measuring access/egress travel for public transit versus other modes. For any mode besides public transit, each section of a multimodal sequence of trips is recorded as its own trip. So, for example, if someone walks to meet up with a carpool to work, the travel would be recorded as a walk trip for the purpose of changing mode of transportation, followed by a car trip with the purpose of work. The same is true for other mixes of modes (e.g., bike to car, take a taxi to the airport). These sequences would be recorded as two trips, and a longer sequence would be recorded as three or more trips. For public transit, in contrast, access and egress legs are considered part of the main transit trip (and accordingly, throughout this report we will distinguish access and egress "legs" from separately recorded "trips"). So, for instance, a trip where someone walked to the bus stop, took the city bus, and then got off the bus and walked to the office would be recorded as a single trip by public transit (rather than two walking trips and a transit trip); details about the walking are included as additional variables about the transit trip. As a result, these access/egress legs are occasions of walking (or other modes, including bicycling) that are not included in trip counts and mode shares. Incorporating transit access and egress legs into public transit trips (as opposed to treating each portion as a separate trip) has a number of advantages. First, a large quantity of analysis is done at the trip level rather than the more complicated tour level. Including access/egress travel within 362 a transit trip therefore ensures that the full, door-to-door durations of public transit trips are considered. Additionally, it ensures that number of trips is a reasonable proxy for number of destinations accessed, rather than including extra trips that are generated entirely by the need to take public transit. This is important in particular for assessing the mobility of transit-dependent people. However, because walking and biking account for 70 percent of access and egress legs, not including those legs results in an undercount of the total amount of walking and biking. As shown in table 162, walking and biking account for 8.6 percent of all trips made by Georgians ages 5+. However, when all instances of travel, including both trips and access/egress legs by any mode are considered, the apparent mode share of nonmotorized travel increases to 10.6 percent. Table 162. Mode share of walking and biking trips and legs. Trips Only Trips and Access/Egress Legs* Walk Bike All NMT Walk Bike All NMT All Georgians Ages 5+ 8.0% 0.6% 8.6% 10.0% 0.6% 10.6% MPO Tier 1. Atlanta MPO 8.7% 0.4% 9.1% 11.5% 0.4% 11.9% 2. Medium MPOs 8.1% 1.0% 9.1% 9.7% 1.0% 10.7% 3. Small MPOs 6.7% 0.5% 7.2% 7.7% 0.5% 8.2% 4. Non-MPO counties 6.5% 1.0% 7.4% 7.0% 1.0% 8.0% * A leg is an occasion where a respondent walks or bikes to access or egress another mode of transportation (e.g., public transit, commuter train). These legs are usually included as part of the trip by the primary mode rather than as individual trips. As table 163 shows, including access/egress legs also increases the number of instances of travel-day walking and biking per capita by 28 percent, from 100 trips to 127 trips and legs. The effect is even more pronounced in the Atlanta MPO, where including access/egress legs 363 increases the measured instances of walking and biking per capita by 36 percent (from 109 trips to 148 trips and legs). Table 163. Instances of walking and biking per person ages 5+. Trips Only Trips and Access/Egress Legs* Walk Bike All NMT Walk Bike All NMT All Georgians Ages 5+ 92.8 7.0 99.8 120.1 7.3 127.4 MPO Tier 1. Atlanta MPO 104.1 4.5 108.6 143.3 4.9 148.3 2. Medium MPOs 92.1 11.4 103.5 113.3 12.2 125.4 3. Small MPOs 76.9 6.2 83.0 89.6 6.2 95.8 4. Non-MPO counties 70.4 10.7 81.1 77.2 10.7 87.8 * A leg is an occasion where a respondent walks or bikes to access or egress another mode of transportation (e.g., public transit, commuter train). These legs are usually included as part of the trip by the primary mode rather than as individual trips. There is no settled answer as to whether access and egress legs should be counted as separate trips for the purposes of calculating mode share, and indeed the NHTS's approach to the question has not been constant over the years.102 For broad comparability, most of this current analysis has followed the NHTS's practice of considering access and egress legs to be part of a larger trip by the mode accessed by those legs. However, for the purpose of measuring the number of occasions and amount of time Georgians spend walking and biking, including access/egress legs is imperative for getting a more accurate picture. The data contain a second problem: 95 percent of trips by public transit have at least one access and/or egress mode listed, but additionally, 0.4 percent of trips by modes other than public transit 102 In 2009, NHTS respondents were directed not to include trips to change the type of transportation. In 2017, travel to access/egress public transit was intended to be excluded while access/egress legs for modes other than public transit were to be recorded as trips with the purpose of changing mode of transportation. In practice, as NHTS Access/Egress Questions and Instructions will show, this procedure was not always followed. 364 incorrectly have access/egress mode(s) recorded (accounting for 10 percent of the access/egress legs recorded). As table 164 shows, the Georgia subsample of the NHTS includes 1,319 access/egress legs. Of these, 977 were to access/egress public transit, 212 were for transit-like modes (e.g., paratransit, charter or intercity bus, ferry), and 130 for modes that were neither public transit nor transit-like. Legs are divided by the mode used for the leg itself, and the mode of the trip for which the leg served as access or egress (the "parent" mode). Table 164. Access and egress legs by mode and parent mode (unweighted). Leg Mode Mode Being Accessed/Egressed (Parent Mode) Public Transit Transit-like* Other All legs by Georgians ages 5+ 977 212 130 Adults ages 18+ 949 183 122 Walk 670 105 59 Bike 26 0 0 POV 54 7 18 School bus 81 2 0 Public or paratransit 65 8 0 Other bus (e.g. intercity, private, or charter) 0 12 2 Air 0 0 28 Other Walk + other(s) 18 39 10 32 8 3 Multiple, not including walk or bike 3 2 2 Children ages 517 28 29 8 Walk 15 26 6 POV 10 1 1 School bus 3 0 0 Other 0 1 0 Walk + other(s) 0 1 1 Note: Each trip can have both an access and an egress leg. The 1,319 legs here were attached to 666 trips. * Legs to access/egress modes that are similar to public transit. For adults, this includes paratransit (N=39 legs), other bus (139), and ferry (6). For children, this includes other bus (28) and school bus (1). Legs to access/egress a mode that is not transit or transit-like. For adults: air (N=57 legs), POV (33), taxi/ridehail/limo (22), walking (2), and other (10). For children: POV (2) and walking (2). Includes one instance of walk + bike. 365 Access/egress legs for nontransit modes would have been more accurately recorded as trips for the purpose of changing mode of transportation.103 Describing the specific questions asked by the NHTS helps to explain how some (albeit not all) of these mistaken records may have originated. NHTS Access/Egress Questions and Instructions Before reporting about individual trips, respondents were asked: "Did you use a bus, subway, train, or some other type of public transportation during your travel day?" If they answered yes, they were prompted after each reported trip, "Did you take a bus, subway, train, or some other type of public transportation from [PreviousPlace] to [CurrentPlace]?" If they answered in the affirmative, after reporting the trip's mode, they were asked, "How did you get to the [mode]?" and "How did you get from the [mode] to [location]?" The fact that the definition of public transit for these questions includes "some other form of public transportation" likely explains why a number of transit-like modes have associated access/egress legs: some respondents would conclude that an intercity bus or a commuter ferry are public transit. Even if subsequent data cleaning would classify the trip itself as nontransit, the recorded access/egress leg(s) would remain with the data. Some other respondents may have had 103 Interestingly, a disproportionate share of incorrectly reported legs comes from higher income respondents. Respondents with an annual household income of at least $50,000 account for 26 percent of correctly reported legs to access/egress public transit, but 48 percent of incorrectly reported legs to access/egress a nontransit trip (unweighted). Put differently, 38 percent of access/egress legs reported by people earning $50,000 or more are for an inappropriate mode, versus just 21 percent of legs reported by people earning less than $15,000 and 17 percent of legs reported by people making $15,00049,999 (unweighted). 366 expansive views of public transit; air was the most common nontransit and nontransit-like parent mode. The remainder can likely be attributed to user error.104 Access/egress legs for parent modes besides public transit were incorrectly recorded, but they do represent real instances of travel. A majority of these misrecorded legs are nonmotorized, and they are equivalent to close to 5 percent of the total of correctly recorded nonmotorized trips (unweighted). To avoid missing these unintentionally hidden instances of walking, subsequent sections of this chapter recategorize access/egress legs for modes other than transit as trips to change mode of transportation. Access/Egress Travel by Georgia Residents Table 165 shows the weighted mode share of access legs for public transit and other parent modes. 104 The inverse error--reporting transit access/egress travel as separate trips for the purpose of change mode of transportation--also occurred. It was, however, less common. The researchers identified three nonmotorized trips that would have more accurately been considered transit access/egress legs. For the sake of simplicity, those trips were not reclassified. 367 Table 165. Mode share of legs to access/egress public transit and other modes. Nonmotorized modes account for three quarters of transit access/egress legs, and a plurality (46 percent) of legs to access/egress other modes.105 Walking accounts for the overwhelming majority of access/egress legs. The second-most common type of transit access mode is another transit or transit-like mode (17.4 percent). A commuter might, for instance, take a local bus to get to the train station. POVs account for 6 percent of transit access/egress legs. Interestingly, no access/egress legs by taxi or other vehicle-for-hire were recorded. As shown in table 166, the average leg to access/egress public transit is 11.6 minutes. Nonmotorized legs are, on average, shorter than motorized legs. Walk legs are the shortest at 8.6 minutes. 105 The difference between the unweighted and weighted proportion of nonmotorized legs to incorrect access modes is likely due to the fact that high-income people, who account for the majority of these trips, are weighted less heavily in order to account for the greater difficulty in eliciting survey responses from low-income households. 368 Table 166. Mean duration in minutes of access/egress legs by mode. Mode Being Accessed/Egressed (Parent Mode) Leg Mode Public Transit Transit-like* and Other All access/egress legs by Georgians ages 18+ 11.6 15.5 All nonmotorized legs 9.2 7.2 Walk 8.6 7.4 Bike 11.9 - Walk + other(s) 19.3 1.8 All motorized 18.9 21.9 Transit and transit-like 18.2 22.6 POV 20.2 13.7 Air - 74.8 Other or multiple without NMT 21.1 9.8 * Transit-like includes school, intercity, private, or charter bus. Ferry is included as a transit-like parent mode, but was not a listed response option for leg modes. Table 167 shows the mode of public transit access/egress legs disaggregated by various demographic factors. 369 Table 167. Mode share of public transit access/egress legs by demographic characteristics (row percentages). NMT (Walk, Bike, Transit or Other Walk + Other) Transit-like* Motorized* Legs by all adults ages 18+ 74.6% 17.3% 8.0% MPO Tier 1. Atlanta MPO 73.8% 17.7% 8.5% 2. Medium MPOs 80.7% 16.1% 3.2% 3-4. Small MPOs & non-MPO counties 75.0% 12.8% 12.2% Urbanicity (Neighborhood Type) Rural & small town 62.5% 24.9% 12.5% Suburban 77.4% 11.7% 10.9% Second city 67.5% 26.4% 6.1% Urban 87.5% 12.0% 0.6% Sex Male 68.3% 22.6% 9.0% Female 79.4% 13.3% 7.3% Age Cohort Millennial and Gen Z (1836) 77.9% 14.0% 8.0% Gen X (3752) 67.1% 24.1% 8.8% Pre-retirement age Boomer (5364) 75.7% 16.3% 8.0% Retirement age (65+) Caregiver Status 83.4% 11.5% 5.0% Noncaregiver 73.3% 17.8% 8.8% Caregiver, youngest child ages 015 79.4% 15.5% 5.1% Race White non-Hispanic only 78.1% 10.6% 11.3% Black, Black multiracial & Black Hispanic 76.5% 15.5% 8.0% Other 57.7% 40.2% 2.0% Driver Status Nondriver 81.5% 15.1% 3.4% Driver 70.4% 18.7% 10.9% Mobility Impairment: a "condition or handicap that makes it difficult to travel outside of the home." Absent 73.9% 17.7% 8.4% Present 78.7% 15.2% 6.1% Annual Household Income <$15,000 80.7% 14.3% 5.0% $15,000 to $24,999 83.3% 4.5% 12.3% $25,000 to $34,999 88.6% 11.4% 0.0% $35,000 to $49,999 79.6% 17.8% 2.6% $50,000 to $74,999 66.0% 17.6% 16.4% $75,000 to $99,999 52.5% 30.6% 16.9% $100,000+ 59.1% 28.9% 12.0% Table continues on next page. 370 Table 167. (Continued). Continued from previous page. NMT (Walk, Bike, Transit or Legs by all adults ages 18+ Walk + Other) 74.6% Transit-like* Other Motorized 17.3% 8.0% Vehicle Deficit Category of Household Zero-vehicle 81.7% 15.5% 2.9% Deficit 77.0% 12.6% 10.4% Nondeficit (sufficient/surplus) 64.2% 23.7% 12.1% Educational Attainment High school or less 74.8% 19.2% 6.1% Some college or associate degree 82.6% 9.8% 7.6% Bachelor's or higher 67.3% 21.9% 10.8% Worker Status Nonworker 87.9% 9.9% 2.3% Worker 68.0% 21.1% 10.9% Occupational Category (Workers Only) Sales or service 80.2% 15.9% 3.9% Clerical or administrative support Blue collar 65.7% 91.3% 20.5% 4.6% 13.8% 4.0% Professional, managerial, or technical 59.1% 26.7% 14.2% * Transit-like includes paratransit and school, intercity, private, or charter bus. Other motorized includes POV, air, other, and multiple (not including walk or bike). A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. A vehicle-deficit household owns at least one vehicle, but fewer vehicles than potential drivers. Blue collar refers to manufacturing, construction, maintenance, or farming. NMT is the most common transit access/egress mode for every subpopulation examined, with mode shares ranging from 52.5 percent (for travelers with a household income of $75,000$99,999) to 87.9 percent (among nonworkers). There are some differences in mode share by MPO tier, but because the built environment differs strongly within an MPO region, the difference is clearer when looking at neighborhood type. The NMT mode share is highest in urban neighborhoods (87.5 percent) and lowest in small towns and rural areas (62.5 percent). NMT comprises a larger share of transit access/egress legs among women (as compared to men), whites and Blacks (as compared to people of another race), nondrivers versus drivers, and people with mobility impairments versus those without. NMT is a more common access mode among 371 lower-income households compared to those of higher incomes and among nonworkers compared to workers. For travelers from zero-vehicle households, NMT accounts for 81.7 percent of legs, versus 77.0 percent for vehicle-deficit households and 64.2 percent for nondeficit households. Table 168 shows the unweighted sample distribution on which table 167 is based. 372 Table 168. Unweighted sample sizes of public transit access/egress legs. NMT (Walk, Bike, Transit or Other Walk + Other) Transit-like* Motorized* Legs by all adults ages 18+ 728 146 76 MPO Tier 1. Atlanta MPO 420 90 52 2. Medium MPOs 212 41 11 3-4. Small MPOs & non-MPO counties 96 15 13 Urbanicity (Neighborhood Type) Rural & small town 81 21 18 Suburban 282 45 39 Second city 282 71 17 Urban 83 9 2 Sex Male 320 72 32 Female 408 74 44 Age Cohort Millennial and Gen Z (1836) 235 38 21 Gen X (3752) 194 52 22 Pre-retirement age Boomer (5364) 188 43 25 Retirement age (65+) Caregiver Status 111 13 8 Noncaregiver 615 130 67 Caregiver, youngest child ages 015 113 16 9 Race White non-Hispanic only 203 36 31 Black, Black multiracial & Black Hispanic 447 92 43 Other 78 18 2 Driver Status Nondriver 270 52 18 Driver 458 94 58 Mobility Impairment: a "condition or handicap that makes it difficult to travel outside of the home." Absent 611 125 68 Present 117 21 8 Annual Household Income <$15,000 334 64 18 $15,000 to $24,999 79 8 7 $25,000 to $34,999 53 7 0 $35,000 to $49,999 57 10 3 $50,000 to $74,999 41 11 12 $75,000 to $99,999 41 14 11 $100,000+ 108 31 25 Table continues on next page. 373 Table 168. (Continued). Continued from previous page. NMT (Walk, Bike, Transit or Other Walk + Other) Transit-like* Motorized* Legs by all adults ages 18+ 728 146 76 Vehicle Deficit Category of Household Zero-vehicle 363 63 14 Deficit 145 27 18 Nondeficit (sufficient/surplus) 220 56 44 Educational Attainment High school or less 272 60 22 Some college or associate degree 175 22 13 Bachelor's or higher 281 64 41 Worker Status Nonworker 337 45 18 Worker 391 101 58 Occupational Category (Workers Only) Sales or service 113 21 4 Clerical or administrative support Blue collar 59 13 12 41 6 1 Professional, managerial, or technical 176 58 38 * Transit-like includes paratransit and school, intercity, private, or charter bus. Other motorized includes POV, air, other, and multiple (not including walk or bike). A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. A vehicle-deficit household owns at least one vehicle, but fewer vehicles than potential drivers. Blue collar refers to manufacturing, construction, maintenance, or farming. Table 169 shows differences in the mean transit access/egress leg duration by mode and demographic groups. Motorized access/egress legs have an average duration more than twice that of nonmotorized legs (18.9 minutes versus 9.2 minutes). Some of the differences in average access/egress leg duration between different groups shown in table 169 are artifacts of differences in the motorized and nonmotorized mode shares of access/egress legs made by members of those groups (table 167). Groups with a higher mode share of motorized legs (e.g., residents of rural areas) will generally have a higher average leg duration than groups with a higher nonmotorized share (e.g., residents of urban areas). 374 However, there are also intergroup differences in the duration of legs by any given mode. For example, workers are more likely than nonworkers to access transit using a motorized mode. However, because the average duration of a nonmotorized leg is longer for nonworkers than for workers (10.7 minutes versus 8.2 minutes), the difference in the average leg duration for the two groups is smaller than what the mode share might imply. 375 Table 169. Duration in minutes of public transit access/egress legs by demographic characteristics. NMT (Walk, Bike, All Legs Walk + Other) Motorized* Legs by all adults ages 18+ 11.6 9.2 18.9 MPO Tier 1. Atlanta MPO 11.9 9.3 19.2 2. Medium MPOs 10.0 8.7 15.5 3-4. Small MPOs & non-MPO counties 11.1 8.5 19.0 Urbanicity (Neighborhood Type) Rural & small town 17.3 13.7 23.2 Suburban 10.7 7.9 20.1 Second city 12.8 10.4 17.8 Urban 9.4 9.2 11.1 Sex Male 13.4 10.2 20.4 Female 10.2 8.5 17.0 Age Cohort Millennial and Gen Z (1836) 11.3 8.6 21.0 Gen X (3752) 12.6 10.5 16.9 Pre-retirement age Boomer (5364) 11.6 9.2 19.7 Retirement age (65+) 8.9 7.8 14.7 Caregiver Status Noncaregiver 11.9 9.5 18.7 Caregiver, youngest child ages 015 10.4 8.2 19.8 Race White non-Hispanic only 10.8 8.1 20.5 Black, Black multiracial & Black Hispanic 11.9 9.7 19.4 Other 11.4 8.3 15.7 Driver Status Nondriver 10.2 9.1 15.4 Driver 12.5 9.3 20.2 Mobility Impairment: a "condition or handicap that makes it difficult to travel outside of the home." Absent 11.7 9.1 19.1 Present 11.2 9.7 17.0 Annual Household Income <$15,000 11.6 10.2 17.5 $15,000 to $24,999 11.3 9.2 21.5 $25,000 to $34,999 9.6 7.8 23.6 $35,000 to $49,999 9.1 6.4 19.7 $50,000 to $74,999 9.7 6.1 16.7 $75,000 to $99,999 19.9 18.1 21.8 $100,000+ 11.3 6.3 18.5 Table continues on next page. 376 Table 169. (Continued). Continued from previous page. Legs by all adults ages 18+ Vehicle Deficit Category of Household Zero-vehicle Deficit Nondeficit (sufficient/surplus) Educational Attainment High School or less Some college or associate degree Bachelor's or higher Worker Status Nonworker All Legs 11.6 11.5 11.7 11.7 12.4 9.3 12.7 11.4 NMT (Walk, Bike, Walk + Other) 9.2 Motorized* 18.9 10.1 17.5 10.3 16.4 6.5 21.0 10.9 17.1 7.4 18.2 8.8 20.8 10.7 16.7 Worker 11.7 Occupational Category (Workers Only) Sales or service 9.8 8.2 19.3 8.2 16.3 Clerical or administrative support Blue collar 10.9 6.3 19.8 8.0 7.8 10.0 Professional, managerial, or technical 13.7 9.1 20.4 * Motorized includes public transit; transit-like (paratransit; school, intercity, private, or charter bus); POV; air; other; and multiple (not including walk or bike). A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. A vehicle-deficit household owns at least one vehicle, but fewer vehicles than potential drivers. Blue collar refers to manufacturing, construction, maintenance, or farming. TRAVEL DAY WALKING AND BIKING BY GEORGIA ADULTS This section describes travel day trips and transit access/egress legs. As discussed in the previous section of this chapter (Access and Egress Travel), access/egress legs for modes other than public transit have been recoded as trips with a purpose of "change mode of transportation." Georgia adults produce an average of 135 nonmotorized trips and legs per year (table 170). The mean duration for a nonmotorized trip/leg is 16.4 minutes, which averages to 6.0 minutes of walking and biking per person per day. 377 Table 170. Nonmotorized trips/legs per capita and duration of NMT trips/legs. All adults ages 18+ MPO Tier 1. Atlanta MPO 2. Medium MPOs 3. Small MPOs 4. Non-MPO counties Urbanicity (Neighborhood Type) Rural Small town Suburban Second city Urban Sex Male Female Age Cohort Millennial and Gen Z (1836) Gen X (3752) Pre-retirement age Boomer (5364) Retirement age (65+) Caregiver Status Noncaregiver Caregiver, youngest child ages 015 Race White non-Hispanic only Black, Black multiracial & Black Hisp. Other Table continues on next page. Nonmotorized Trips and Legs per Adult Annually 134.9 Mean Duration of NMT Trips/Legs (Minutes) 16.4 Average Daily Minutes of Walking/Biking* 6.0 158.8 15.3 6.7 133.8 17.3 6.3 96.2 19.1 5.0 90.4 19.0 4.7 72.5 16.9 3.4 79.4 22.8 5.0 175.4 15.7 7.6 170.6 14.3 6.7 587.4 12.9 20.8 140.5 16.9 6.5 129.7 15.8 5.6 154.5 15.6 6.6 134.8 15.9 5.9 133.2 16.0 5.9 97.1 20.5 5.5 146.9 17.4 7.0 108.9 13.5 4.0 122.5 16.2 5.4 154.5 15.4 6.5 139.2 19.8 7.6 378 Table 170. (Continued). Continued from previous page. Nonmotorized Trips Mean Duration of Average Daily and Legs per Adult NMT Trips/Legs Minutes of Annually (Minutes) Walking/Biking* All adults ages 18+ 134.9 16.4 6.0 Driver Status Nondriver 309.8 16.1 13.7 Driver 113.2 16.4 5.1 Mobility Impairment: a "condition or handicap that makes it difficult to travel outside of the home." Absent 132.1 16.5 6.0 Present 161.0 15.5 6.9 Annual Household Income <$15,000 279.0 17.6 13.4 $15,000 to $49,999 109.5 15.8 4.7 $50,000 to $99,999 88.1 16.3 3.9 $100,000+ 141.6 15.5 6.0 Vehicle Deficit Category of Household Zero-vehicle 688.9 16.4 31.0 Deficit 134.8 16.5 6.1 Nondeficit (sufficient/surplus) 95.8 16.3 4.3 Worker Status Nonworker 152.0 19.2 8.0 Worker 124.1 14.2 4.8 * Average duration x annual per capita total 365. A caregiver is defined as any adult ages 18+ in a household with a child of less than 5 years old, and any adult ages 22+ in a household with a child of 515 years old. A vehicle-deficit household owns at least one vehicle, but fewer vehicles than potential drivers. Residents of urban areas make 587 NMT trips/legs per year; this is more than three times as many NMT trips/legs as second-city and suburban residents, and more than seven times as many trips/legs as residents of small towns and rural areas. The average urban resident spends more than 20 minutes per day walking and/or biking, compared to less than 8 minutes in all other neighborhood types. 379 Residents of zero-vehicle households make 689 NMT trips/legs per year. Low income, nonworker and nondriver status, and younger age are all associated with increased NMT trips/legs. Mobility-impaired adults generate more NMT trips/legs per capita than other adults, though the duration of each trip/leg is somewhat shorter. This counterintuitive finding is likely related to the fact that people with mobility impairments disproportionately find themselves in groups that walk more by necessity (low-income, nondrivers, etc.). See chapter 5, Health and Disability for more discussion of the challenges facing people with mobility impairments. Table 171 shows the mode and purpose of NMT trips/legs. Walking accounts for 94.8 percent of nonmotorized travel while biking accounts for 5.2 percent. Biking is less prominent in the Atlanta MPO overall, but more prevalent in the urban neighborhoods within Atlanta, where it accounts for 9.6 percent of nonmotorized travel. Cycling also accounts for a comparatively high proportion of the nonmotorized travel by people in medium MPOs and small towns, men, members of Gen X, low-income people, people with mobility impairments, and residents of zero-vehicle and vehicle-deficit households. 380 Table 171. Mode and purpose of NMT trips/legs by demographic factors. All adults ages 18+ MPO Tier 1. Atlanta MPO 2. Medium MPOs 3. Small MPOs 4. Non-MPO counties Urbanicity (Neighborhood Type) Rural Small town Suburban Second city Urban Sex Male Female Age Cohort Millennial and Gen Z (1836) Gen X (3752) Pre-retirement age Boomer (5364) Retirement age (65+) Caregiver Status Noncaregiver Caregiver, youngest child ages 015 Race White non-Hispanic only Black, Black multiracial & Black Hisp. Other Table continues on next page. Annual NMT Trips and Legs per Adult 134.9 158.8 133.8 96.2 90.4 72.5 79.4 175.4 170.6 587.4 140.5 129.7 154.5 134.8 133.2 97.1 146.9 108.9 122.5 154.5 139.2 Mode* Purpose Bike 5.2% 2.8% 10.5% 6.1% 10.1% 3.4% 8.7% 4.3% 3.3% 9.6% 6.8% 3.7% 5.0% 7.2% 4.9% 1.9% 5.0% 5.8% 6.7% 3.0% 5.8% Walk 94.8% Leisure 26.2% Instrumental 73.8% 97.2% 89.5% 93.9% 89.9% 25.2% 22.2% 27.4% 34.6% 74.8% 77.8% 72.6% 65.4% 96.6% 91.3% 95.7% 96.7% 90.4% 34.6% 37.1% 22.8% 23.8% 17.3% 65.4% 62.9% 77.2% 76.2% 82.7% 93.2% 96.3% 27.8% 24.5% 72.2% 75.5% 95.0% 92.8% 95.1% 98.2% 21.0% 25.7% 29.2% 39.0% 79.0% 74.3% 70.8% 61.0% 95.0% 94.2% 93.3% 97.0% 94.2% 27.8% 21.5% 32.8% 15.1% 31.4% 72.2% 78.5% 67.2% 84.9% 68.6% 381 Continued from previous page. Table 171. (Continued). Annual Mode* Purpose NMT Trips and Legs per Adult Bike Walk Leisure Instrumental All adults ages 18+ 134.9 5.2% 94.8% 26.2% 73.8% Driver Status Nondriver 309.8 6.8% 93.2% 12.6% 87.4% Driver 113.2 4.7% 95.3% 30.8% 69.2% Mobility Impairment: a "condition or handicap that makes it difficult to travel outside of the home." Absent 132.1 4.9% 95.1% 27.1% 72.9% Present 161.0 7.6% 92.4% 19.1% 80.9% Annual Household Income <$15,000 279.0 8.7% 91.3% 15.8% 84.2% $15,000 to $49,999 109.5 3.4% 96.6% 24.4% 75.6% $50,000 to $99,999 88.1 2.6% 97.4% 35.6% 64.4% $100,000+ 141.6 Vehicle Deficit Category of Household 5.1% 94.9% 32.7% 67.3% Zero-vehicle 688.9 8.0% 92.0% 8.3% 91.7% Deficit 134.8 7.2% 92.8% 23.7% 76.3% Nondeficit (sufficient/surplus) 95.8 2.8% 97.2% 36.5% 63.5% Worker Status Nonworker 152.0 4.7% 95.3% 28.1% 71.9% Worker 124.1 5.6% 94.4% 24.7% 75.3% Row percentages shown. * Includes multimodal access/egress legs. One walk+bike egress leg is included in both the walk and bike columns. Leisure includes nonloop and loop trips with the purposes of fitness and leisure. Instrumental includes all other trip/leg purposes. A caregiver is defined as any adult ages 18+ in a household with a child of less than 5 years old, and any adult ages 22+ in a household with a child of 515 years old. A vehicle-deficit household owns at least one vehicle, but fewer vehicles than potential drivers. The purpose of NMT varies for different groups of people, as well. Overall, 26.2 percent of NMT trips/legs are for leisure: recreation and fitness.106 The remaining 73.8 percent of trips/legs 106 As discussed elsewhere, recreation and fitness trips are a mix of trips where the purpose of the trip itself is leisure and trips to access a leisure destination (e.g., a gym or movie theater). 382 are instrumental trips taken for the purpose of accessing a destination (work, school, shopping, public transit, etc.). Many of the same groups that have elevated numbers of NMT trips/legs per capita also devote a higher percentage of those trips to instrumental travel. More than 90 percent of NMT trips/legs by people from zero-vehicle households are for instrumental purposes, versus just 63.5 percent of the NMT by people from nondeficit households. Nondrivers, low-income people, millennials and Gen Zers, people with mobility impairments, and residents of urban neighborhoods likewise report more NMT, with a higher share of it devoted to instrumental travel. In contrast, men and childfree adults all report higher-than-average rates of NMT, but with a higher than usual share of it devoted to leisure. Table 172 compares the duration of leisure and instrumental trips/legs by walking and biking. Bike trips/legs are longer than walk trips/legs on average, though the difference is larger for leisure travel than for instrumental travel. Table 172. Mean duration in minutes of leisure and instrumental NMT trips/legs by Georgia adults. All purposes Leisure (recreation and fitness) Instrumental (all other purposes) All NMT 16.4 25.5 13.1 Walk 16.1 25.1 12.7 Bike 21.8 37.8 19.1 Table 173 provides a more detailed breakdown of the purposes of NMT trips and legs. Among instrumental NMT trips/legs, the most common purposes are to access/egress public transit (20.6 percent), return home (19.0 percent) and household-serving travel (11.1 percent). Compared to walking trips, a higher percentage of bike trips are for instrumental travel 383 (85.6 percent versus 73.2 percent). The higher percentage of return-home trips by bicycle suggests a number of possible explanations; perhaps cyclists are not trip chaining or stopping along the way as often as pedestrians, or pedestrians are more likely to start a walking circuit from a location besides their home. The more detailed purposes of bicycle trips should be treated with caution, as some of the unweighted cell sizes are quite small (see table 174). Table 173. Purpose and duration of NMT trips/legs. All purposes Leisure Recreation nonloop Fitness nonloop Loop recreation or fitness Instrumental Access/egress public transit (legs) Work School Medical Household-serving Change mode of transportation Percent of Trips/Legs (Column Percent) All NMT Walk* Bike* 100% 100% 100% 26.2% 26.8% 14.4% 3.0% 2.9% 3.9% 7.2% 7.3% 6.8% 16.0% 16.7% 3.8% 73.8% 73.2% 85.6% 20.6% 21.4% 6.5% 6.3% 6.2% 7.2% 2.2% 2.3% 1.3% 0.4% 0.4% 0.1% 11.1% 10.7% 18.5% 4.1% 4.3% 0.5% Mean Duration (Minutes) All NMT 16.4 25.5 14.2 14.6 32.5 13.1 9.2 12.4 20.9 23.8 13.3 9.3 Discretionary: dining, visit friends/relatives 8.9% 8.7% 13.4% 12.1 Community, volunteer and religious activities 1.0% 1.0% 0.2% 19.7 Return home 19.0% 18.0% 37.8% 17.4 Other 0.1% 0.1% 0.0% 23.7 * Includes walk + other and bike + other legs. One egress leg with the modes walk + bike is included in both the walk and bike columns (but only counted once when combining all NMT trips and legs). Loop trips with a purpose of home or work-from home have been recoded as recreation/fitness. However, since it is not possible to distinguish between recreation and fitness loops, loop recreation/fitness is listed as its own subcategory. Loop trips with an instrumental purpose are included with that purpose. See chapter 7 for more details on loop trip classification. Includes buy goods and services, general errands, transport others, and accompany others. Trips to change mode of transportation and access/egress legs for modes besides public transit (e.g., airplane, ferry, long- distance train, etc.). 384 Table 174 contains the unweighted data on which table 173 is based. Table 174. Unweighted trips/legs by mode and purpose. Time of Day Figure 39 shows the start times of NMT trips/legs on weekdays and weekends. The figure on the left depicts the annual total in millions, and the figure on the right shows trips as a percentage of weekday or weekend trips. Table 175 and table 176 show the weighted and unweighted data on 385 which these figures are based. While there are gentle peaks around 9:00 a.m. and 5:00 p.m., trips/legs are distributed throughout the day. Figure 39. Bar graphs. Weekday and weekend NMT by time of day. 386 Table 175. Weighted NMT trips/legs by time of day 387 Table 176. Unweighted NMT trips/legs by time of day. 388 Table 177 shows the time of day by day of week and MPO tier, collapsed into 3-hour categories (and a 6-hour overnight period). Cyclists and pedestrians get a slightly later start on the weekends.107 Table 177. Time of day of NMT trips/legs by day and MPO tier. All Day of Week All NMT Trips/Legs Weekday Weekend Weighted Column Percentages 12:005:59 am 2.0% 1.8% 6:008:59 am 15.4% 16.7% 9:0011:59 am 20.4% 19.0% 12:002:59 pm 19.7% 20.3% 3:005:59 pm 22.0% 22.2% 6:008:59 pm 15.3% 15.3% 9:0011:59 pm 5.3% 4.7% Unweighted Sample Sizes All times of day 4,830 4,127 12:005:59 am 74 60 6:008:59 am 787 692 9:0011:59 am 998 833 12:002:59 pm 938 817 3:005:59 pm 1,079 924 6:008:59 pm 729 627 9:0011:59 pm 225 174 Unweighted Column Percentages 12:005:59 am 1.5% 1.5% 6:008:59 am 16.3% 16.8% 9:0011:59 am 20.7% 20.2% 12:002:59 pm 19.4% 19.8% 3:005:59 pm 22.3% 22.4% 6:008:59 pm 15.1% 15.2% 9:0011:59 pm 4.7% 4.2% Includes all trips and legs by walking, biking, or walk + other. 2.6% 11.1% 24.9% 17.7% 21.5% 15.2% 7.1% 703 14 95 165 121 155 102 51 2.0% 13.5% 23.5% 17.2% 22.0% 14.5% 7.3% Tier 1 Atlanta 2.3% 16.3% 19.6% 19.5% 21.1% 15.9% 5.2% 1,870 39 324 360 373 385 290 99 2.1% 17.3% 19.3% 19.9% 20.6% 15.5% 5.3% MPO Tier Tier 2 Medium MPOs Tier 3 Small MPOs 1.8% 13.3% 24.6% 19.3% 22.0% 13.6% 5.2% 1,742 24 267 387 331 394 261 78 1.4% 15.3% 22.2% 19.0% 22.6% 15.0% 4.5% 0.6% 14.3% 19.8% 22.3% 23.1% 16.4% 3.5% 840 8 131 180 156 199 136 30 1.0% 15.6% 21.4% 18.6% 23.7% 16.2% 3.6% Tier 4 NonMPO Counties 1.4% 14.0% 19.5% 19.6% 25.7% 13.4% 6.3% 378 3 65 71 78 101 42 18 0.8% 17.2% 18.8% 20.6% 26.7% 11.1% 4.8% 107 The weekend trips/legs in the overnight period are primarily in the 12:0012:30 a.m. hour, while those on weekdays are more evenly divided between late night and early morning. 389 CAPTIVE AND CHOICE NONMOTORIZED TRAVEL As with public transit, some pedestrians and cyclists choose to walk and bike, and others use these modes by necessity. This section compares the nonmotorized travel of captive and choice travelers. We use linear regression to examine the time penalties incurred by captive pedestrians and cyclists as compared to choice pedestrians/cyclists. We model the total minutes of walking/biking on the travel day for leisure (fitness and recreation) and instrumental purposes (all other purposes) as a function of being a captive traveler and a number of control variables. Unless otherwise stated, descriptive statistics are weighted using values provided by NHTS, and models are based on unweighted data. Defining Captive Travel We base our definition of captive travel on mode, household vehicle ownership, and income. This allows us to differentiate the needs of travelers who are carless by necessity from those who are car-free by choice (Brown 2017). We also include transit and nonmotorized trips by residents of low- and moderate-income vehicle-deficit households, or households that own at least one automobile but not enough for each potential driver in the household (Blumenberg, Brown, and Schouten 2018). In this study, a potential driver is any household resident ages 16+.108 In Georgia, these households substantially outnumber carless households, and because 108 The researchers chose to use all individuals ages 16+ because the NHTS defines drivers by asking "does this person drive?" Whether or not a person drives may be in part determined by the availability of a vehicle. 390 the family car is not available for all trips, some household members will obligatorily use transit or nonmotorized modes. Table 178 shows vehicle sufficiency among Georgia households earning less than $50,000 per year and those earning $50,000 or more per year. We chose this threshold because it is the closest approximation of Georgia's median income (estimated by the American Community Survey to be $55,679 for the years 20142018) achievable with NHTS categorical income data. We conducted sensitivity analysis using different income cutoffs, and the models we present were found to be robust. Georgians who have sufficient income largely choose to have a vehicle for each driver. Among households earning at least $50,000 per year, vehicle ownership is nearly universal; 99.5 percent of households own at least one vehicle. Vehicle-deficit households are also somewhat uncommon (12.7 percent), and just 0.5 percent have chosen to be car-free. While the majority of households in the lower half of the income distribution also own cars, 38.1 percent of these households are vehicle-deficit or carless. Additionally, higher-income vehicle-deficit households are more likely to have a vehicle deficit due to the presence of a driver under the age of 18 (10.9 percent of vehicle-deficit households earning at least $50,000 versus 3.3 percent of those earning less than $50,000). 391 Table 178. Vehicle sufficiency of Georgia households and adults. In this analysis, a captive nonmotorized trip is a walking or biking trip by a traveler who lives in a captive household. A captive household meets the following criteria: 1. The household is carless or vehicle-deficit. We include vehicle-deficit households in the definition of captive travel because, while a car is available to some household members for some trips, for other trips, household members will have more limited options. 2. The household earns less than $50,000 per year. The income criterion is designed to screen out travel by people who are "car-free" by choice versus those who are financially unable to afford vehicles for every potential driver in the household. Model Structure We model the total amount of NMT on the travel day based on the sample of people who had at least one nonmotorized trip or leg. We estimate one model for leisure travel (fitness and 392 recreation) and one model for instrumental travel (all other purposes). The dependent variable for each model is the total time in minutes spent on nonmotorized travel of that type. The leisure model includes fitness and recreation trips. The instrumental model includes trips for all other purposes, as well as time spent walking/biking to and from transit stations.109 Per-person totals are modeled rather than individual trips because, unlike transit trips, nonmotorized trips are easily subdivided by travelers in ways that produce idiosyncrasies in the data. Consider, for example, two respondents who walk to work past a coffee shop. The choice pedestrian, with more disposable income, may choose to stop, while a lower-income captive pedestrian will walk past, regardless of her desire for coffee. To avoid the kind of "apples to oranges" comparisons that can result from such differences in whether respondents stop along the way, the research team chose to model the daily total duration of nonmotorized travel while controlling for purpose and number of trips. We separate leisure travel from instrumental travel because we believe different processes guide people's choices to walk for leisure or instrumental purposes. While in most cases longer trips are considered a disutility, for some recreational and fitness trips the longer duration is part of the purpose. Separating the two forms of travel allows the models to accommodate the general tendency to try and minimize instrumental travel time but maximize leisure time (including leisure walking and biking). It also allows us to incorporate the amount of time spent walking for nonleisure purposes as an independent variable into the model, explaining the amount of time people spend walking for leisure. 109 As discussed in Travel Day Walking and Biking by Georgia Adults of this chapter, the NHTS does not report access/egress nonmotorized travel as separate "trips," so this report describes them as "legs." 393 Multiple specifications were explored for each model; the effects of captive travel described here were consistent in both magnitude and significance across alternate specifications. Results As shown in table 179, captive pedestrians and cyclists make an average of 3.4 nonmotorized trips and legs per day, versus 2.3 for choice pedestrians and cyclists. Choice pedestrians and cyclists are more likely to use NMT for leisure (58 percent, versus 30 percent of captive travelers), and somewhat less likely to use NMT for instrumental purposes. Among instrumental purposes, work and school are more common destinations for choice pedestrians and cyclists than for captive travelers, who more commonly use nonmotorized travel for household-serving travel and accessing public transit. 394 Table 179. Duration, quantity, and purpose of nonmotorized travel for captive and choice travelers. All Nonmotorized Travelers Captive* Nonmotorized Travelers Choice Nonmotorized Travelers At least one walk trip or leg 97.1% 95.6% 97.8% At least one bike trip or leg Percent with any instrumental NMT 4.4% 73.2% 7.2% 84.4% 3.2% 68.2% Percent with any leisure (fitness or recreation) NMT 49.6% 30.6% 58.1% Mean NMT trips and legs, all purposes 2.64 3.42 2.29 Mean walk trips and legs, all purposes 2.50 3.17 2.21 Mean bike trips and legs, all purposes 0.14 0.25 0.08 Total NMT minutes 39.91 51.33 34.80 Leisure NMT minutes 15.37 9.97 17.78 Instrumental NMT minutes 24.55 41.36 17.02 Percent of Travelers using NMT for Each Purpose Recreation nonloop 7.2% 3.7% 8.7% Fitness nonloop 15.2% 9.3% 17.9% Loop recreation or fitness 29.5% 19.6% 33.9% Access/egress public transit (legs) 16.1% 29.5% 10.1% Work 13.8% 10.4% 15.3% School 3.7% 1.6% 4.6% Medical Household-serving Change mode of transportation 0.8% 20.3% 5.1% 1.6% 32.8% 3.7% 0.5% 14.8% 5.7% Discretionary: dining, visit friends/relatives 20.7% 22.0% 20.1% Community, volunteer, and religious activities 2.5% 2.9% 2.3% Return home 49.3% 49.3% 35.4% * From household with fewer vehicles than potential drivers (incl. zero vehicles) and annual household income <$50,000. Instrumental NMT includes trips for any purpose besides fitness or leisure, along with public transit access/egress legs. Includes buy goods and services, general errands, transport others, and accompany others. Trips to change mode of transportation and access/egress legs for modes besides public transit (e.g., airplane, ferry, long-distance train, etc.). Table 180 shows the effect of being a captive traveler on the total amount of time spent walking and/or biking for instrumental purposes over the course of the day. Being a captive traveler does not, by itself, have a significant effect on the amount of time spent walking or biking. However, 395 it increases the expected duration of each expected walk trip by just over 3 minutes. While there was no significant interaction between captivity and the duration of bike trips or transit access/egress legs, the fact that captive users tend to make more NMT trips and legs for instrumental purposes results in a longer overall expected duration of instrumental NMT. Additionally, using a private auto at any point on the travel day is associated with a 6-minute reduction in the duration of nonmotorized instrumental travel. While this reduction applies to captive and choice travelers alike, choice travelers are more likely to have used a POV at some point on the travel day. In terms of trip purpose, school travel, medical travel, and travel to return home have the longest durations. The coefficient on medical trips is not significant, but as this trip was the least common (reported by just 23 respondents), it may be worth revisiting the issue of nonmotorized medical travel with a larger or more targeted dataset. Table 181 models the total duration of leisure (recreation and fitness) NMT. The amount of instrumental (nonleisure) NMT is negatively associated with the amount of leisure travel. This may reflect the fact that travel that does "double duty" as both instrumental and leisure will be recorded as instrumental. For example, a decision to walk to a nearby coffee shop rather than drive will be recorded by the purpose at the destination (in this case, dining). However, it likely also reflects a reduced availability of leisure time. In particular, the duration of a recreational trip to wander around a local park will be constrained by the traveler's amount of free time to, say, choose a more circuitous route or pause to admire wildlife. 396 As per the descriptive statistics in table 179, the average minutes of leisure NMT is lower for captive users than for choice users.110 However, after controlling for time spent on instrumental nonmotorized travel, captivity does not have a significant effect. It seems a key reason captive nonmotorized travelers spend less time on leisure NMT is because they spend more on instrumental NMT. 110 Interactions between captivity and number, purpose, and mode of trips were also found to be insignificant and were therefore excluded from the final model. 397 Table 180. Linear regression: Daily total minutes of instrumental walking and biking. Captive traveler Used POV on travel day Used public transit on travel day Instrumental walk trip (excluding public transit access legs) Captive* walk trip Instrumental bike trip (excluding public transit access legs) Captive* bike trip NMT leg to access/egress public transit Captive* NMT transit access/egress leg Purpose(s) of nonmotorized trips (1 = yes) Work (commute or work-related business) School Medical Household-serving (transport other, buy goods/services, general errands) Change mode of transportation, including access/egress non-transit mode Discretionary: dining, visit friends/relatives Community, volunteer and religious activities Return home Female Race (reference: white non-Hispanic only) Black and Black multiracial Other Annual household income (reference: $100,000+) <$15,000 $15,000 to $24,999 $25,000 to $34,999 $35,000 to $49,999 $50,000 to $74,999 $75,000 to $99,999 Mobility impairment Age Caregiver for child ages 04 Neighborhood type (reference: rural) Small town Suburban Second city Urban Weekend Constant Coefficient 1.12 -6.14 -5.79 6.59 3.35 11.32 -1.78 8.62 0.26 5.96 18.09 13.94 4.47 2.91 5.45 9.72 11.16 -1.72 1.34 2.11 4.88 -1.70 -5.42 -3.74 -2.24 0.88 1.30 -0.07 -5.85 3.55 3.61 4.16 0.46 2.69 3.32 P-value 0.737 0.002 *** 0.090 * 0.000 *** 0.005 *** 0.000 *** 0.587 0.000 *** 0.849 0.011 ** 0.003 *** 0.130 0.035 ** 0.312 0.016 ** 0.004 *** 0.000 *** 0.227 0.477 0.450 0.135 0.613 0.023 ** 0.080 * 0.264 0.654 0.719 0.165 0.001 *** 0.108 0.120 0.064 * 0.873 0.198 0.467 Model indicators: N = 1,288 R 2 = 0.471 Adjusted R 2 = 0.457 From household with fewer vehicles than potential drivers (incl. zero vehicles) and annual household income <$50,000. 398 Table 181. Linear regression: Daily total minutes of leisure walking and biking. Leisure Travel (New, with Loop Separated) Captive traveler Any instrumental NMT Minutes of instrumental NMT Minutes2 of instrumental NMT Used POV on travel day Used public transit on travel day Walk recreation legs (nonloop) Bike recreation legs (nonloop) Walk fitness legs (nonloop) Bike fitness legs (nonloop) Walk loop leisure travel (fitness or recreation) Bike loop leisure travel (fitness or recreation) Female Race (reference: white non-Hispanic only) Black and Black multiracial Other Annual household income (reference: $100,000+) <$15,000 $15,000 to $24,999 $25,000 to $34,999 $35,000 to $49,999 $50,000 to $74,999 $75,000 to $99,999 Mobility impairment Age Caregiver for child ages 04 Neighborhood type (reference: rural) Small town Suburban Second city Urban Weekend Constant Coefficient -2.72 -15.59 0.50 -2.72E-03 -3.64 -5.52 -0.16 1.44 1.78 3.88 12.48 36.04 3.32 0.24 1.26 6.43 2.16 0.33 -2.35 0.41 4.32 -9.99 0.10 -1.14 -2.16 -1.63 -2.99 -0.69 4.14 21.06 P-value 0.394 0.000 *** 0.000 *** 0.001 *** 0.121 0.238 0.943 0.830 0.374 0.298 0.000 *** 0.001 *** 0.048 ** 0.937 0.667 0.123 0.549 0.903 0.418 0.874 0.105 0.002 *** 0.050 * 0.723 0.448 0.585 0.305 0.865 0.063 * 0.000 *** Model indicators: N = 1,059 R 2 = 0.228 Adjusted R 2 = 0.206 From household with fewer vehicles than potential drivers (incl. zero vehicles) and annual household income <$50,000. Nonmotorized travel for a purpose other than fitness or leisure. Since promoting walking and cycling are often pursued as public health interventions, it is also important to note that the association between health and nonmotorized travel is not the same for 399 captive and choice travelers (results not tabulated). Among choice pedestrians and cyclists who walked or biked on the travel day, 96.1 percent consider themselves to be in good, very good, or excellent health. This is slightly higher than the rate among residents of similar households who did not walk or bike (92.3 percent). However, only 76.7 percent of captive pedestrians and cyclists consider themselves to be in those categories of good health. Not only is this lower than the rate among choice pedestrians, it is also lower than the rate among residents of captive households who did not walk or bike (79.2 percent). Captive pedestrians and cyclists spent more time traveling than their choice counterparts did, after controlling for trip quantity and purpose. Captive nonmotorized travelers also tended to make more nonmotorized trips, compounding the differences in total travel time. While increased walking and biking is broadly considered a public health goal, it is important to remember that some travelers are already walking or riding more than they would like. However, the health benefits of increasing walking and biking may not accrue equally to captive nonmotorized travelers. Not only were captive pedestrians and cyclists less likely to be in good health than choice pedestrians and cyclists, they were also less likely to be in good health than travelers from captive households who did not walk and bike. While this finding is descriptive, it suggests the need for follow-up study to examine differences in the relationship between nonmotorized travel and health among captive pedestrians and cyclists. As with transit, there may be equity issues related to access to walking and biking infrastructure and other streetscape features that are known to affect wellbeing, as well as exposure to air pollution (Maurer Braun, Read, and Ricklin 2016). 400 CHILDREN'S NONMOTORIZED AND SCHOOL TRAVEL This section discusses children's nonmotorized travel, as well as their travel to and from school. Walking and biking historically accounted for a large portion of school travel, though this is less true today. The section begins with a discussion of how children's school travel has been identified in this study. Identifying School Trips In addition to children's reported "usual" modes of travel to and from school, we examine school travel on the travel day. A school trip is a trip to school at the start of the school day and a trip home from school, typically after classes and extracurricular activities have concluded. The methods used to identify school travel are closely related to the techniques for identifying work journeys described in chapter 2, in that we consider trips with the purpose of school and trips to and from the school location, some of which may not have a listed purpose of school. Most school journeys (82 percent) are simple: directly from home to school or school to home. Figure 40 provides a sample itinerary to illustrate how we categorize more complex travel patterns. 401 Figure 40. Diagram. Example child's travel itinerary with trips to and from school. This child travels from home to school. Her class then embarks on a field trip to a local museum. At the end of the field trip, the child returns to the school to change her mode of transportation. She goes to a friend's house for a few hours, and then returns home. Following the model of constructing work journeys precisely, trip 1 would be a simple school journey and trips 4 and 5 would be a complex school journey. Trips 2 and 3, from the school to a nonhome location and back, would not be part of a school journey. Because the goal is not to quantify the duration of school travel, the researchers' needs here are somewhat simpler than they are in chapter 2. For simplicity, then, in complex school journeys, the research team identifies them by the leg that is connected to the school location (the last leg of a journey to school or the first leg of a journey from school). In this example, trips 1 and 4 would be considered school travel. 402 It is important to note that the definition of school travel used in this section differs from trips with the purpose of school reported elsewhere. Trip purpose is defined by purpose at destination; as a result, all trips from school to home are classified as having a purpose besides school. Ninety percent of school trips identified here have a purpose of either school or home, with the remaining 10 percent split between purposes such as running errands, visiting friends or relatives, leisure, etc. The following steps provide a few more technical details. Some readers may wish to proceed to the next section. Step 1: In identifying travel to and from school for children ages 517, the researchers initially included all trips that started or ended at the respondent's school location, all trips with the listed purpose of school, and all trips where the listed origin-purpose (the primary activity at the origin location) was school. Trips to school: A total of 1,412 trips with a destination and/or purpose of school were identified. Of those, 54 trips were screened out because they were loops or because the origin and destination purposes were both school; 1,290 trips with a purpose of school that did not originate from the school location were identified as school trips; and 68 trips to the school location with a different listed purpose were flagged for manual review. Trips from school: A total of 1,440 trips with an origin location or origin purpose of school were identified. Of those, 56 were disqualified because they were loops or the origin and destination purposes were both school. A total of 1,311 trips were classified as trips from school based on having an origin-purpose of school and nonschool destinations 403 and destination-purposes. Seventy-three trips from the school location with a nonschool origin-purpose were flagged for manual review. Step 2: Flagged trips from step 1 were reviewed manually based on other trips in each child's travel itinerary; see chapter 2 for an example of the kinds of factors considered. The researchers identified one additional trip to school and six trips from school. The remainder were disqualified as school trips. Step 3: To further screen out field trips and similar travel from school, the total number of school trips for each child was calculated. Seventeen children were listed as having three or more school trips (in the single travel day measured by the diary). These children's trips were manually reviewed. As a result, six trips to school and nine trips from school were disqualified, leaving 1,285 trips to school, 1,308 trips from school, and 3,967 nonschool trips. As in the rest of this chapter, this analysis of school travel additionally screened out 8 children who did not answer questions about frequency of nonmotorized travel. Table 182 shows the mode of trips to and from school. 404 Table 182. Mode of travel-day trips to and from school by children ages 517 (unweighted). Additionally, respondents (or their parents) were asked about how they "usually get to school" (table 183). These two data sources are used for the analysis. Table 183. Usual mode of travel to and from school by children ages 517 (unweighted). Usual Mode.... Walk Bike POV School bus Public or paratransit Other bus (e.g., intercity, private, or charter) Vehicle for hire (taxi, ridehail, limo) Air Other Total (all modes) To School 87 (3.8%) 9 (0.4%) 1,190 (52.5%) 958 (42.3%) 8 (0.4%) 2 (0.1%) (0.0%) 1 (0.0%) 11 (0.5%) 2,266 From School 100 (4.4%) 9 (0.4%) 995 (43.9%) 1,122 (49.5%) 6 (0.3%) 16 (0.7%) (0.0%) (0.0%) 18 (0.8%) 2,266 405 Due to the small sample size of school travel by bicycle, the remaining analysis combines walking and biking into the more general category of nonmotorized travel. Children's School Travel Table 184 shows children's observed (travel day) and usual modes of travel to and from school. Table 184. School travel: Mode split of trips and usual mode(s). Travel Day School Trips N=2,593 Trips Usual School Mode N=2,273 Children To and/or To From Total To From From* Nonmotorized (walk or bike) 4.2% 4.9% 4.6% 3.9% 4.5% 5.4% POV 49.5% 47.0% 48.2% 44.8% 37.6% 47.6% School bus 45.0% 46.7% 45.8% 49.7% 55.2% 58.8% Other 1.3% 1.4% 1.4% 1.6% 2.7% 3.0% Weighted column percentages shown. * Mode is reported as usual mode of transportation either to school, from school, or both. Because 14.9 percent of children have a different usual mode to and from school, this column sums to more than 100 percent. School bus is the most commonly reported "usual" mode of transportation to and from school; 58.8 percent of caregivers reported that their child usually takes the bus to school, from school, or both. The second most common usual school travel mode is POV, which is the usual mode to and/or from school for 47.6 percent of children. In contrast, for observed (travel day) school travel, the private auto eclipses school bus, accounting for 48.2 percent of trips (versus 45.8 percent by school bus). Students' day-to-day travel diverges, to some degree, from their caregivers' description of their "usual" travel behavior. For both observed and "usual" mode of school travel, nonmotorized travel comes in a distant third. Depending on which measure is used, nonmotorized travel accounts for slightly more or 406 slightly less than 5 percent of school travel. The mode share of nonmotorized travel is higher in the afternoon (trips home from school) than in the morning; some children arrive to school by auto or bus and then walk home in the afternoon. As shown in table 184, 4.6 percent of observed school trips are by walking and biking. 5.4 percent of children are listed as usually using nonmotorized travel in at least one direction (0.9 percent only to school, 1.5 percent only from school, and 3.0 percent in both directions). Table 185 shows demographic differences in the prevalence of nonmotorized school travel. 407 Table 185. Walking and biking as percentage of school travel by geography, age, driver status, sex, race, income, and household vehicles. Travel Day School Usual School Mode Trips (To and From) (Either Direction) N=2,593 Trips N=2,273 Children All children 517 4.6% 5.4% MPO Tier 1. Atlanta MPO 4.9% 5.5% 2. Medium MPOs 7.2% 8.1% 3. Small MPOs 3.1% 4.4% 4. Non-MPO counties 2.6% 3.5% Urbanicity (Neighborhood Type) Rural 1.2% 0.6% Small town 2.1% 3.5% Suburban 7.4% 7.2% Second city 6.6% 9.8% Urban 18.6% 22.2% Age and Driver Status 59 1013 1415 1617 Nondriver, ages 1617 Driver, ages 1617 5.7% 5.8% 5.7% 5.6% 1.2% 6.1% 2.8% 3.5% 3.9% 5.8% 1.3% 1.2% Sex Male 5.2% 5.1% Female 3.9% 5.8% Race White non-Hispanic only 2.3% 3.3% Black, Black multiracial & Black Hispanic 7.0% 7.1% Other 4.6% 7.6% Annual Household Income <$15,000 4.2% 7.5% $15,000 to $49,999 6.6% 6.6% $50,000 to $99,999 2.5% 3.6% $100,000+ Vehicle Deficit Category of Household 4.1% 5.4% Zero-vehicle 8.4% 8.9% Deficit 8.0% 8.3% Nondeficit (sufficient/surplus) 3.0% 4.0% Note: Trips denotes the percentage of all school trips (in either direction) that was made by walking or biking by children in the row category. Usual school mode depicts the percentage of children in each row category who "usually" walk or bike to school, from school, or both. A vehicle-deficit household owns at least one vehicle, but fewer vehicles than potential drivers. 408 Walking and biking to/from school is most common in medium-MPO counties, where they account for 7.2 percent of observed school trips and are listed as the usual mode of school travel in at least one direction for 8.1 percent of children. It is the least common in non-MPO counties, where just 3.5 percent of children usually walk or bike to or from school and NMT accounts for just 2.6 percent of observed school trips. While nonmotorized school travel is not quite as common in the Atlanta MPO compared to smaller MPOs, it is much more prevalent in the dense urban neighborhoods only found in Atlanta; 18.6 percent of school trips in these neighborhoods are by walking or biking. Walking and biking are less common among children who are old enough to drive (ages 1617). However, this effect mainly applies to children who are themselves drivers. Nondrivers ages 1617 are comparatively more likely to walk/bike to or from school. White children are also less likely to walk to school than children of other races. Children in zero-vehicle households and vehicle-deficit households are more likely to walk or bike to school than children in nondeficit households. There is a dip in walking/biking among children in upper-middle income households (earning $50,000$99,999 per year) compared to both poorer and wealthier households. This may be a sign that walkability is an amenity for some high-income families. Children's Nonmotorized Travel for All Purposes We turn now from school travel to nonmotorized travel for any purpose. Table 186 shows the purposes of nonmotorized trips made by children, as well as access/egress legs.111 Because the 111 Access/egress legs were added to the data following the methods that were described with regard to adults in Access and Egress Travel in this chapter. 409 most common purposes of children's NMT differ from those of adults, rather than duplicate the adults' categories, the research team has created categories more tailored to children's observed travel. We divide children's NMT into four types of purposes. Travel for "fun" purposes like leisure and socialization (i.e., exercise, recreation, visiting friends and relatives, and dining out) account for a plurality of children's nonmotorized travel (35.4 percent). Trips with the purpose of school account for 7.0 percent of children's NMT.112 Other instrumental purposes (such as changing mode of transportation, household-serving travel, work, community, volunteer, and religious activities) account for 27.6 percent, and the remaining 30.0 percent of trips and legs have the purpose of returning home. 112 As opposed to the definition of school travel used in Children's School Travel in this chapter, here, we are returning to the standard method of defining the purpose of a trip by its destination. School trips (both to and from school) account for 14.5 percent of children's nonmotorized trips and legs. 410 Table 186. Purpose of children's nonmotorized trips (weighted and unweighted). Table 187 shows variations in the purpose of nonmotorized trips between different subpopulations of children. Unweighted sample sizes of nonmotorized trips by each group are included in the right-most column to allow the reader to cautiously interpret findings about groups with especially small sample sizes (e.g., children in urban neighborhoods, who made a total of 23 reported nonmotorized trips on the travel day). Leisure and socialization account for more than half of children's nonmotorized travel in rural and small-town neighborhoods (51.9 percent and 51.5 percent, respectively), as well as in nonMPO counties more generally (51.5 percent). Elsewhere in the state, this travel for "fun" purposes is outnumbered by other instrumental travel. 411 "Fun" purposes account for a higher share of nonmotorized travel by children 1015 years old as compared to older and younger children. However, while 1617-year-old nondrivers devote just 19.8 percent of their nonmotorized trips to leisure and socialization, 60.5 percent of nonmotorized trips by their peers who drive are devoted to leisure and socialization. This is largely because children who can drive begin making fewer nonmotorized trips overall. 412 Table 187. Purpose of children's nonmotorized trips by geography, age, driver status, sex, race, income, and household vehicle ownership. Purposes of Nonmotorized Trips (Weighted Row Percentages) Leisure and Attend School Other Return Socialization or Daycare Instrumental Home All children ages 517 35.4% 7.0% 27.6% 30.0% MPO Tier 1. Atlanta MPO 33.6% 7.3% 28.8% 30.4% 2. Medium MPOs 27.2% 10.3% 28.0% 34.5% 3. Small MPOs 36.6% 1.8% 26.4% 35.2% 4. Non-MPO counties 51.5% 5.9% 23.2% 19.3% Urbanicity (Neighborhood Type) Rural 51.9% 1.0% 20.8% 26.3% Small town 51.5% 6.1% 13.8% 28.7% Suburban 30.2% 9.3% 32.1% 28.5% Second city 25.3% 6.2% 36.1% 32.4% Urban 15.3% 15.2% 23.1% 46.5% Age 59 27.5% 10.6% 30.3% 31.5% 1013 40.3% 8.6% 19.7% 31.3% 1415 43.1% 2.3% 18.4% 36.2% 1617 33.1% 2.3% 42.9% 21.7% Driver Status Underage, ages 515 36.0% 8.3% 23.5% 32.3% Nondriver, ages 1617 19.8% 2.5% 57.8% 19.9% Driver, ages 1617 60.5% 1.9% 12.3% 25.3% Sex Male 38.0% 6.3% 23.1% 32.6% Female 31.5% 8.0% 34.2% 26.3% Race White non-Hispanic only 42.8% 3.3% 21.7% 32.1% Black & Black multiracial 30.8% 10.2% 29.8% 29.1% Other 30.4% 7.0% 34.9% 27.7% Annual Household Income <$15,000 31.5% 3.2% 38.0% 27.4% $15,000 to $49,999 26.0% 10.6% 32.0% 31.5% $50,000 to $99,999 43.4% 6.0% 19.0% 31.6% $100,000+ 40.6% Vehicle Deficit Category of Household 6.7% 22.7% 30.0% Zero-vehicle 10.0% 5.3% 58.7% 26.0% Deficit 40.2% 6.9% 25.7% 27.2% Nondeficit (sufficient/surplus) 38.4% 7.6% 20.5% 33.5% A vehicle-deficit household owns at least one vehicle, but fewer vehicles than potential drivers. Unweighted Sample Size 602 216 231 115 40 87 145 160 187 23 202 200 84 116 486 70 46 353 249 307 200 95 93 164 135 203 66 172 364 413 Leisure and socialization also account for a higher proportion of travel among wealthier children, and a greatly reduced portion of trips (just 10.0 percent) made by children from zero-vehicle households. Frequency of Nonmotorized Travel by Children Table 188 shows the mode(s) of nonmotorized transportation used by children in the past 7 days. Children are more likely to report walking and biking than adults; 78.8 percent walked or biked in the past 7 days compared to 72.5 percent of adults. Children are slightly more likely to walk than adults (75.5 percent versus 72.2 percent of adults), and much more likely to bike (30.5 percent versus 5.5 percent). Additionally, some children exhibited a mobility pattern that was virtually absent from adults. Just 0.3 percent of adults reported biking but not walking. As shown in table 188, 3.3 percent of children reported just biking. 414 Table 188. Nonmotorized travel of children ages 517 (past 7 days). 415 Table 189 consolidates the data from table 188 by presenting the total percentage of children who walked (regardless of whether or not they biked) and the percentage who biked (regardless of whether or not they walked). It also shows the average number of trips in the past 7 days by users of each mode. Statewide, 78.8 percent of children ages 517 have walked and/or biked within the past 7 days. This share is higher than average in non-MPO counties, small towns, and rural areas and lower than average in second-city and urban neighborhoods. NMT usage is more common among younger children than older ones. However, while the percentage of children who walk remains relatively constant between ages 515, biking declines earlier; 44.4 percent of children ages 59 reported riding a bike, versus 34.9 percent of 1013 year olds, 15.6 percent of 1415 year olds, and just 7.4 percent of teenagers age 1617. There is not an analogous age-related decrease in the number of nonmotorized trips among those children who do walk or bike. 416 Table 189. Percentage of children who have walked or biked in the past 7 days and average number of trips among mode users. NMT Usage Past Seven Days Avg. Trips among Mode Users Any NMT Total Walk* Bike* (Walk and/or Bike) Walk Bike NMT All children ages 517 75.5% 30.5% 78.8% 7.5 4.2 8.9 MPO Tier 1. Atlanta MPO 74.3% 24.4% 77.0% 6.9 3.5 7.8 2. Medium MPOs 75.0% 37.7% 78.2% 8.9 4.8 10.9 3. Small MPOs 70.8% 40.9% 75.7% 8.1 4.4 10.0 4. Non-MPO counties 82.6% 36.6% 86.9% 7.7 5.1 9.4 Urbanicity (Neighborhood Type) Rural 80.1% 36.3% 84.9% 8.1 5.0 9.8 Small town 75.9% 34.3% 79.2% 7.2 4.4 8.8 Suburban 74.6% 24.0% 77.6% 7.1 3.5 7.9 Second city 72.2% 29.5% 74.6% 8.2 3.9 9.5 Urban 63.5% 13.3% 63.5% 7.9 2.2 8.4 Age 59 78.1% 44.4% 83.1% 7.2 4.3 9.1 1013 76.8% 34.9% 80.6% 6.8 3.9 8.2 1415 76.4% 15.6% 77.4% 9.0 5.2 9.9 1617 66.6% 7.4% 67.4% 8.4 4.1 8.7 Driver Status Underage, ages 515 77.3% 35.1% 81.0% 7.4 4.3 8.9 Nondriver, ages 1617 67.0% 6.7% 67.0% 10.7 5.7 11.3 Driver, ages 1617 66.2% 8.1% 67.8% 5.8 2.6 6.0 Sex Male 78.2% 33.1% 81.0% 7.7 4.6 9.3 Female 72.8% 27.8% 76.6% 7.3 3.8 8.4 Race White non-Hispanic only 78.9% 32.0% 82.3% 7.4 4.3 8.7 Black & Black multiracial 72.2% 27.4% 74.8% 7.7 4.3 9.0 Other 73.9% 34.1% 78.5% 7.5 4.2 8.9 Annual Household Income <$15,000 78.6% 38.3% 80.1% 8.8 4.2 10.7 $15,000 to $49,999 74.8% 29.3% 78.2% 9.0 5.4 10.6 $50,000 to $99,999 77.4% 29.1% 79.9% 6.6 3.7 7.7 $100,000+ 75.8% Vehicle Deficit Category of Household 30.3% 80.6% 5.8 3.4 6.8 Zero-vehicle 81.1% 39.1% 81.1% 6.5 3.4 8.2 Deficit 74.3% 22.7% 76.0% 8.1 5.0 9.4 Nondeficit (sufficient/surplus) 75.6% 33.2% 79.8% 7.4 4.1 8.7 * Percentage of row group who have walked or biked, regardless of their usage of the other mode (i.e., walk includes walk only and both walk and bike; bike includes bike only and both walk and bike). A vehicle-deficit household owns at least one vehicle, but fewer vehicles than potential drivers. 417 We turn now to trips observed on the travel day itself. As shown in table 190, 9.7 percent of trips by children ages 517 were by walking and 1.0 percent were by biking, for a total of 10.6 percent of all children's trips. The percentage of children who made at least one nonmotorized trip on the travel day (12.9 percent) is slightly higher than the percentage of the trips themselves that are nonmotorized. Table 190. Nonmotorized travel on the travel day: Percent of trips/legs and percent of children. Travel Day Trips/Legs by Children Ages 517 Number of Percent of Trips Trips (Weighted) (Unweighted) Unweighted Percent Walk* 9.7% 540 8.2% Bike 1.0% 62 0.9% All NMT (walk + bike) 10.6% 602 9.2% All other modes 89.4% 5956 90.8% Total 6558 Modes Used on Travel Day by Each Child Ages 517 Percent of Number of Children Children Unweighted (Weighted) (Unweighted) Percent 1+ walk trips/legs 11.8% 276 11.3% 1+ bike trips/legs 1.4% 33 1.3% 1+ nonmotorized trip/leg (walk and/or bike) 12.9% 306 12.5% No nonmotorized travel 87.1% 2,145 87.5% Total 2,451 * Includes two legs of walk + other. Three children made both walk and bike trips on the travel day. As shown in table 191, a higher percentage of children walk and/or bike in large and medium MPOs than in small MPOs and non-MPO counties. In urban neighborhoods of Atlanta, more than a quarter of children walked or biked on the travel day compared to less than 10 percent in rural and small-town neighborhoods. 418 Table 191. Percent of children who walked/biked on the travel day and minutes of nonmotorized travel. Percent of Children who Total Minutes of Nonmotorized Walked and/or Biked on the Travel on Travel Day (among Travel Day children with 1+ minutes of NMT) All children ages 515 12.9% 29.3 MPO Tier 1. Atlanta MPO 13.7% 31.3 2. Medium MPOs 13.8% 22.0 3. Small MPOs 10.9% 23.4 4. Non-MPO counties 11.1% 33.7 Urbanicity (Neighborhood Type) Rural 7.9% 31.4 Small town 9.9% 22.0 Suburban 15.3% 39.5 Second city 18.7% 21.7 Urban 26.1% 27.6 Age 59 11.8% 27.2 1013 14.4% 27.7 1415 10.6% 28.4 1617 14.6% 36.6 Driver Status Underage, ages 515 12.6% 27.6 Nondriver, age 1617 18.4% 39.2 Driver, age 1617 10.5% 31.8 Sex Male 14.9% 28.1 Female 10.9% 31.0 Race White non-Hispanic only 11.9% 21.4 Black & Black multiracial 13.8% 38.8 Other 13.8% 25.7 Annual Household Income <$15,000 16.4% 51.8 $15,000 to $49,999 14.9% 26.4 $50,000 to $99,999 8.6% 21.7 $100,000+ 12.8% 26.3 Vehicle Deficit Category of Household Zero-vehicle 23.4% 36.7 Deficit 17.2% 32.4 Nondeficit (sufficient/surplus) 10.2% 25.6 A vehicle-deficit household owns at least one vehicle, but fewer vehicles than potential drivers. There is not a clear trend in the percentage of children who walk/bike by age. However, among teenagers old enough to drive, nondrivers are more likely to walk/bike than drivers. Additionally, 419 nondrivers who walk or bike do so for an average of 39.2 minutes compared to 31.8 minutes among pedestrians/cyclists who are able to drive. Boys are more likely to report walking/biking than girls, but female pedestrians/cyclists spend slightly longer on nonmotorized travel than do male ones. Compared to white non-Hispanic children, children of all other races were more likely to walk/bike on the travel day. However, Black children who walk/bike do so for an average of 38.8 minutes, while children of other races average 25.7 minutes (white children average 21.4 minutes). The lowest-income children (<$15,000 per year) are somewhat more likely to walk/bike than their wealthier peers. However, the average number of minutes spent on NMT by pedestrians/cyclists from these households (51.8 minutes), is nearly double the amount of time spent by more affluent pedestrians and cyclists. As with adults, there is evidence that some children are walking and biking by necessity. 420 CHAPTER 7. TRAVEL FOR ITS OWN SAKE CHAPTER 7 SUMMARY This chapter explores the positive utility of travel by focusing on travel undertaken for its own sake. Overview and Methods defines and reviews examples of travel for its own sake (TFIOS). We discuss the methodological challenges of measuring TFIOS in general and with regard to NHTS data. We focus on loop trips (trips with the same start and end location) as an easily identifiable form of TFIOS, and discuss the NHTS's methods of soliciting information about loop trips. The purpose of a loop trip generally differs from the activities conducted at the origin and destination. This section discusses how the researchers have reclassified the purposes of loop trips (this classification has also been used throughout chapter 1 to chapter 6). Loop Trips in the 2017 Georgia NHTS Subsample reviews empirical findings on loop trips in the state of Georgia and compares them to nonloop trips. While nonloop travel is dominated by the private auto, the overwhelming majority of loop trips are on foot. Likewise, while recreation and fitness account for 85 percent of loop trips, only 6 percent of nonloop trips are for the purpose of recreation or fitness. OVERVIEW AND METHODS Studies of transportation typically frame travel as a disutility: a cost that must be paid to access a desired destination. However, a number of studies have pointed to the intrinsic value of travel, 421 beyond its utilitarian purpose of getting people from A to B (Mokhtarian 2019, Mokhtarian et al. 2015). Travel can provide positive utility via several avenues. First, utility can come from enjoyment of the trip itself, independent of destination (for example, a walk through one's own neighborhood to experience fresh air and exercise113). Travel can also provide positive utility as an enjoyable activity on the way to a desired destination (e.g., an enjoyable bike ride to meet a friend at a coffee shop).114 Finally, travel can provide positive utility through the opportunity it affords to conduct additional productive or pleasant activities during the trip; this has been the focus of some research on mode choice between public transit and private autos and the emerging literature on autonomous vehicles (Frei, Hyland, and Mahmassani 2017; Frei, Mahmassani, and Frei 2015; Malokin, Circella, and Mokhtarian 2019). For this study's purposes, TFIOS comprises trips that afford positive utility in their own right, instead of or in addition to the utility of the destination accessed. TFIOS can include: A trip with the same origin and destination (loop trip), made for fitness, exercise, leisure, to enjoy riding a motorcycle (or train, boat, or beloved car), to see the sights, or simply to "get out of the house." Travel where the nominal destination is the excuse for the trip rather than the other way around. In other words, the destination was generated by the trip and would not have been visited otherwise. Travel that was intended to be a loop trip but included an impulsive stop along the way. 113 As will be discussed in the section on Loop Trips in the 2017 Georgia NHTS Subsample in this chapter, such trips are often categorized (or, as these researchers argue, miscategorized) as trips to return home. 114 The difference between these two cases is that in the second case, the positive utility is a purpose in addition to the recorded purpose at the destination while in the first, the purpose at destination (e.g., to go home) is not an accurate reflection of the trip itself. 422 A sightseeing tour with stop(s) along the way. A trip where the main purpose is the destination accessed, but the traveler chooses a longer route or slower mode for pleasure. This would include choosing a circuitous, deliberately longer route on a bike trip or enjoying a drive along one of Georgia's Scenic Byways115 when a more direct route is available. Challenges of Identifying and Measuring TFIOS Many studies address the positive utility of travel in a general sense, with survey questions about a liking for travel in general, by various modes, and for various purposes (e.g., Ory and Mokhtarian 2005). Research that investigates a particular trip often focuses on directly measuring travelers' subjective well-being, direct satisfaction with the travel, and perceptions about whether the travel time itself was well spent or wasted (Ettema et al. 2011, Friman et al. 2012). Other studies focus on qualitative aspects of the trip themselves, including pleasant or unpleasant experiences during travel and activities conducted during travel (Abou-Zeid et al. 2012, Gripsrud and Hjorthol 2012, Susilo et al. 2012, Lin 2012).116 Some studies use the thought experiment of teleportation ("If you could instantly teleport yourself to the destination, would you do so?") to assess the extent to which the trip itself is desirable (Russell and Mokhtarian 2015). 115 See http://www.dot.ga.gov/DS/Travel/Scenic for a list of the 15 corridors (as of this writing) that have been designated as Georgia Scenic Byways by the Georgia DOT. 116 However, there is not perfect overlap between studies of the positive utility of travel and TFIOS; travel can have incidental benefits even if it is undertaken primarily for a separate purpose. 423 NHTS does not contain data about travelers' internal experiences, well-being, satisfaction, or motives for making a trip.117 Without these direct data, identifying and measuring TFIOS poses a number of challenges: No category of purpose at destination can unambiguously indicate that a trip ought to be considered TFIOS. A trip with the purpose of "leisure," for example, would include going to a local park for a stroll (TFIOS) or going to a movie theater to sit and watch a film (not TFIOS). For loop trips, the purpose at destination will never capture the true purpose of the trip. For example, as discussed in Loop Trips in the 2017 Georgia NHTS Subsample later in this chapter, 72 percent of Georgia's loop trips nominally have the destination purpose of "regular home activities." For nonloop trips, it is difficult to distinguish between (a) a nominal destination generated by a trip and (b) a standard destination that generated a trip. Further, the intent of such a trip may vary between members of the same party. A parent may coax their child to go for a walk in the park by promising them ice cream. For the parent, the purpose of the trip might be to get out of the house and enjoy the greenery (TFIOS). For the child, the point is definitely the ice cream (not TFIOS). Using loop trips as a proxy for TFIOS will result in many false negatives (i.e., failing to identify nonloop trips that would more accurately be considered TFIOS). Classifying all loop trips as TFIOS results in a small but nonzero number of false positives, such as a drive to charge up a car battery or walking the family dog in inclement weather. 117 Trip purpose in NHTS is defined by the primary activity conducted at the destination, rather than reasons for conducting that activity or for making the trip at all. 424 Additionally, some loop trips may actually be a series of nonloop trips for which the respondent neglected to report interim stops. Unless the reason for mode choice is asked--which it is not for the NHTS--no given mode can be assumed to be TFIOS. Even a stereotypically pleasurable mode such as the bicycle may be only reluctantly "chosen" by the traveler who owns a car and would prefer to drive but whose car is in the shop. Conversely, for the traveler who just bought a new car, even a trip to the grocery store may be TFIOS, invented as an excuse to drive. With walking for its own sake, it is unclear if and how walking within a destination such as a botanical garden or zoo is reflected in survey data and may vary based on how detailoriented the traveler filling out the survey is. While using a service such as a Google API is likely more accurate than self-reported distance for many trips, it will not detect trips where a circuitous route is chosen and will underestimate the distance of such trips. Because trip duration is self-reported, the duration will still be accurate. However, an unexpectedly long duration, particularly for a trip by private auto, is more likely to be a result of congestion than of TFIOS. NHTS does not include travel outside of the country, which will miss trips taken while on vacation abroad. With the caveats listed above, in the NHTS data set, loop trips are the most reliable proxy for TFIOS. While using loop trips as a proxy will neglect to include nonloop TFIOS, a high proportion of loop trips can safely be considered TFIOS. 425 Measurement of Loop Trips and TFIOS in the 2017 and 2009 NHTS Figure 41 shows an excerpt from the 2009 NHTS instructions for completing the travel diary.118 A trip was defined as "whenever you travel from one address to another." While this instruction would seem to exclude loop trips, respondents were given the clarification "if you started and ended in the same place, list the farthest point you reached and record a return trip." Figure 41. Instructions. Excerpt from 2009 NHTS travel diary instructions. 118 Page 13 of the instructions sent to respondents, reproduced as p. 124 of the 2009 "Questionnaire and Field Documents" file available for download on the NHTS site: https://nhts.ornl.gov/documentation.shtml. 426 The 2017 NHTS, in contrast, prompted respondents to include loop trips without adding a destination in the middle (figure 42). Additionally, compared to 2009, loop trips figured more prominently in the instructions; both of the instructional graphics shown to respondents included a clearly marked loop. Figure 42. Instructions. Excerpt from 2017 NHTS travel diary instructions. When 2017 NHTS respondents reported a loop trip, they were asked to report the total distance traveled (as opposed to nonloop trips, where the Google API was used to calculate the shortest path distance). Misra (2017), studying data from the Cycle Atlanta mobile app, which recorded location traces of users' bicycle trips, found that the majority of trips deviated from the shortest distance path, with a mean deviation of 20 percent and a median of 2 percent (p. 161). The probability and size of the deviation was influenced by built environment characteristics (e.g., 427 traffic volume and speed, availability of bike facilities, slope) and rider demographics; women and older adults were more likely to choose a longer route (p. 168). Self-reported distances are therefore potentially more accurate, depending on the precision of respondents' estimates of their own travel distance. Trip Purpose for Loop Trips With nonloop trips, the primary activity at the destination (destination purpose) is a workable proxy for the trip purpose; a nonloop trip to a location where the primary activity is working for pay can reasonably be considered a trip with the purpose of work, and a nonloop trip where the primary activity at the destination is regular home activities can reasonably be considered a trip to return home. However, this logic breaks down with loop trips. Table 192 shows destination purposes for the 1,294 loop trips reported by 928 Georgians ages 5 and up; 4.6 percent of Georgians (weighted) made at least one loop trip on the travel day. 428 Table 192. Destination purpose of loop trips made by Georgians ages 5+ (unweighted). Destination Purpose* Number of Loop Unweighted Trips Percent All purposes Home 1294 956 100.0% 73.9% Regular home activities (chores, sleep) 933 72.1% Work from home (paid) Recreation and Fitness 23 1.8% 181 14.0% Recreational activities (visit parks, movies, bars, museums) 24 1.9% Exercise (go for a jog, walk, walk the dog, go to the gym) 157 12.1% Work, School, and Daycare 91 7.0% Work 65 5.0% Work-related meeting / trip Attend school as a student 6 0.5% 19 1.5% Attend adult care 1 0.1% Other 66 5.1% Volunteer activities (not paid) 2 0.2% Drop off / pick up someone 8 0.6% Change type of transportation 5 0.4% Buy goods (groceries, clothes, appliances, gas) 8 0.6% Buy services (dry cleaners, banking, service a car, pet care) 2 0.2% Buy meals (go out for a meal, snack, carry-out) 13 1.0% Other general errands (post office, library) 3 0.2% Visit friends or relatives 19 1.5% Health care visit (medical, dental, therapy) 1 0.1% Religious or other community activities 3 0.2% Unknown 2 0.2% * Primary activity at the end location of the loop trip. This table shows original trip destination purposes as defined by NHTS. For the rest of this analysis, home loop trips ("regular home activities" and "work from home") are reclassified as recreation/fitness trips. We also reclassify some trips with the destination purpose of "attend school as a student." Five trips with the purpose of "attend school" and a location of the respondent's school were reclassified as recreation/fitness; school trips at nonschool locations were left with their original classification (14). In subsequent analysis, the researchers combine recreational activities and exercise into a single recreation/fitness category because, absent further data, it is difficult to distinguish recreational purposes (e.g., visiting a park) from exercise purposes (e.g., walking the dog), or ambiguous cases such as walking a dog at the park. 429 Seventy-four percent of these trips (unweighted) have a destination purpose of home (regular home activities or work from home). In fact, the majority of loop trips (58.8 percent unweighted) are walking trips where the primary activity at both the origin and destination is home (not tabulated). In other words, the respondent is engaging in home activities, perhaps reading the news or doing housework. She leaves the house to take a walk, returns home, and resumes household activities (perhaps making dinner or watching television). In this sequence of events, it is not logical to conclude that the purpose of the trip was home activities. The loop trip, rather, was a break from those home activities. The true purpose of these trips could be characterized more accurately as exercise ("go for a jog, walk, walk the dog, go to the gym") or recreation ("visit parks, movies, bars, museums"). Absent further data,119 it is difficult to distinguish between the two, and it is easy to think of trips for which the classification would be ambiguous (e.g., a walk through a local park to exercise while walking the dog). Therefore, throughout this report, the research team has reclassified loop trips with a destination purpose of home and placed them in a combined category of fitness/recreation, along with loop and nonloop trips with the destination purposes of exercise or recreation. We have likewise reclassified five loops with the purpose of attending school as fitness/recreation.120 Some of the remaining loop trips are also likely misclassified, either because the destination purpose does not reflect the purpose of the trip itself or because the respondent neglected to 119 NHTS technical documentation indicates that respondents who reported making a loop trip were prompted as to whether it was for fitness or "something else." However, the variable this question would have generated, "TPURP_LOOP," was not included with either the public or private versions of the dataset. The reasons for this omission are, at the time of writing, unknown. 120 The researchers did not reclassify school trips at nonschool locations on the grounds that the travel might be part of an educational activity. We likewise did not classify work-based loops on the grounds that they might be work errands. 430 report interim stops. Loop trips with the destination purposes of buying goods/services/meals or running errands readily support the interpretation that some of these loops are simply trip chains with missing interim stops. However, these trips are relatively few in number. To be conservative, the research team has not reclassified the purposes of these other types of loops. We do, however, include them in this chapter's analysis of travel for its own sake, presented in the next section. LOOP TRIPS IN THE 2017 GEORGIA NHTS SUBSAMPLE In 2017, Georgians ages 5 and older made 233,173,400 loop trips, or 2.1 percent of all trips. This figure varies considerably by mode: loops account for 22.5 percent of all walk trips but only 5.4 percent of bike trips and just 0.3 percent of POV trips (table 193). There are likewise variations by purpose. Loop trips account for 24.5 percent of recreation and fitness trips versus 0.3 percent of trips for all other purposes (not tabulated). 431 Table 193. Loop trips as a percent of total trips by mode and purpose (Georgians ages 5+). All modes and purposes Mode Nonmotorized (walk, wheelchair, bike) Walk or wheelchair Bike Motorized POV (including rental car) Other ground or water Purpose Recreation and fitness* Work, school, and daycare Other Return home* Note: Weighted row percentages shown. Loop Trips 2.1% Nonloop Trips 97.9% 21.3% 22.5% 5.4% 0.3% 0.3% 0.5% 78.7% 77.5% 94.6% 99.7% 99.7% 99.5% 24.5% 1.2% 0.3% 0.0% 75.5% 98.8% 99.7% 100.0% As shown in table 194, Georgians made 24.4 loop trips per capita in 2017, of which 21.2 were nonmotorized. The number of loop trips per capita is highest in Atlanta and non-MPO counties, and lower in small and medium MPOs. Similarly, the number of loop trips per capita is highest at the lowest and highest ends of the income spectrum. Older adults make more loop trips than younger adults, and those with disabilities make more loop trips than those without. One plausible explanation for this is availability of free time and/or limited access to destinations that are farther afield. Working Georgians make fewer loop trips than nonworkers, but drivers make more loop trips than nondrivers. Black Georgians make fewer loop trips than white Georgians and those of other races. Young children make the fewest loop trips of any group, likely due to their limited autonomy compared to adults. 432 Table 194. Loop trips per capita in 2017 (ages 5+). Nonmotorized* (Walk and Bike) All Other Modes All Georgians ages 5+ 21.2 3.2 MPO Tier 1. Atlanta MPO 23.8 2.9 2. Medium MPOs 14.8 3.1 3. Small MPOs 15.2 1.5 4. Non-MPO counties 22.6 4.9 Sex Male 23.5 3.1 Female 19.2 3.2 Age Cohort Minor ages 517 8.3 2.8 Millennial and Gen Z (1836) 20.9 3.2 Gen X (3752) 20.4 3.8 Pre-retirement age Boomer (5364) 29.3 3.7 Retirement age (65+) 32.4 1.7 Race White non-Hispanic only 24.6 2.9 Black, Black multiracial & Black Hispanic 13.4 4.0 Other 27.1 2.0 Mobility Impairment: a "condition or handicap that makes it difficult to travel outside of the home." Absent 20.9 3.2 Present 24.8 2.3 Annual Household Income <$15,000 26.2 4.5 $15,000 to $49,999 16.4 3.1 $50,000 to $99,999 19.3 2.0 $100,000+ 26.7 3.6 Vehicle Deficit Category of Household Zero-vehicle 26.7 5.9 Deficit 23.1 2.1 Nondeficit (sufficient/surplus) 20.1 3.4 Driver Status (Ages 16+ Only) Nondriver 21.9 2.0 Driver 24.1 3.4 Worker Status (Ages 16+ Only) Nonworker 30.4 2.5 Worker 19.4 3.8 * Walk, bike, and wheelchair. A vehicle-deficit household owns at least one vehicle, but fewer vehicles than potential drivers. Total 24.4 26.7 18.0 16.7 27.4 26.6 22.3 11.1 24.1 24.2 33.0 34.1 27.5 17.4 29.1 24.2 27.1 30.7 19.5 21.2 30.3 32.5 25.2 23.5 23.9 27.5 32.9 23.1 Table 195 shows the mode and purpose of Georgians' loop trips. Of loop trips, 87.1 percent were nonmotorized, primarily by walk/wheelchair (85.5 percent), with biking a distant second 433 (1.6 percent). The remaining 12.9 percent were by other motorized modes: POV (11.7 percent) and other ground or water (1.2 percent). No loop trips by air were reported. Table 195. Mode and purpose of loop trips by Georgians ages 5+. Weighted Unweighted Column Number of Column Percent Trips Percent All loop trips 1,294 Mode* Nonmotorized 87.1% 1,147 88.6% Walk or wheelchair 85.5% 1,122 86.7% Bike 1.6% 25 1.9% Motorized 12.9% 147 11.4% POV (including rental car) Other ground or water 11.7% 1.2% 132 10.2% 15 1.2% Purpose: Loop Trips by All Modes (N=1,294) * Recreation and fitness 84.7% Work, school, and daycare 8.9% 1,142 86 88.3% 6.6% Other 6.3% Purpose: Nonmotorized Loop Trips (N=1,147) Recreation and fitness 88.6% Work, school, and daycare 6.8% 66 1,060 50 5.1% 92.4% 4.4% Other 4.5% Purpose: Motorized Loop Trips (N=147) Recreation and fitness 58.6% Work, school, and daycare 23.0% 37 3.2% 82 55.8% 36 24.5% Other 18.4% 29 19.7% * Column percentages shown are based on total sample of 1,294 loop trips. Includes school bus (3), public transit (3), other bus (1), taxi/ridehail/limo (1), golf cart/Segway (3), boat/ferry/water taxi (2), and unspecified (1). Includes original destination purposes of recreation, exercise, home, and school trips at the school location. Includes work, work-related meeting/trip, attend school as a student, and attend daycare or adult daycare. Column percentages shown are based on subtotal of loop trips by each mode category (nonmotorized or motorized). 434 Recreation/fitness was the dominant purpose of loop trips. These purposes accounted for 84.7 percent of all loop trips and 88.6 percent of nonmotorized trips (after the reclassification described previously in Trip Purpose for Loop Trips). Recreation/fitness was still the most common purpose of motorized loops, but by a narrower margin (58.6 percent). Table 196 compares the characteristics of loop and nonloop trips. Walking accounted for 85.5 percent of loop trips versus just 6.3 percent of nonloop trips. Private autos accounted for 87.1 percent of nonloop trips versus 11.7 percent of loop trips. While recreation and fitness are the dominant purpose of loop trips, these trips make up just 5.6 percent of nonloop trips. 435 Table 196. Mode, purpose, distance, and duration of loop and nonloop trips. Mode (Column Percent) Walk or wheelchair Bike POV (including rental car) Other ground or water Air Purpose (Column Percent) Recreation and fitness* Work, school, and daycare Other Return home* Mean Distance (Miles) All modes Walk or wheelchair Bike POV (including rental car) Other ground or water Mean Duration (Minutes) All modes Walk or wheelchair Bike POV (including rental car) Other ground or water Loop Trips 85.5% 1.6% 11.7% 1.2% 0.0% Loop Trips 84.7% 8.9% 6.3% n/a Loop Trips 4.2 1.0 4.7 26.6 0.9 Loop Trips 46.5 31.3 53.6 160.4 93.0 Nonloop Trips 6.3% 0.6% 87.1% 5.8% 0.2% Nonloop Trips 5.6% 16.4% 43.7% 34.3% Nonloop Trips 12.0 0.7 2.2 10.9 9.0 Nonloop Trips 23.6 13.9 18.1 23.0 38.5 Note: All statistics are weighted. * For loop trips, home trips (regular home activities, work from home) are reclassified as recreation/fitness. School loop trips at the school location are also reclassified as recreation/fitness. Nonloop trips with these purposes are unchanged. Includes work, work-related meeting/trip, attend school as a student, and attend daycare or adult daycare. Because so many loop trips are nonmotorized, their average distance is much shorter than that of nonloop trips (4.2 miles versus 12.0 miles). However, when comparing individual modes, loop trips by walking/wheelchair, biking, and POV are all longer, on average, than nonloop trips by these same modes. In terms of duration, loop trips are longer than nonloop trips in the aggregate and across every mode. 436 Table 197 shows these same statistics but narrowed to the sample of only fitness and recreation trips. The differences between loop and nonloop fitness/recreation trips are similar to those between loop and nonloop trips in general. This may suggest that differences in trip purpose between loop and nonloop trips do not explain differences in distance and duration, but it is also likely that a much higher fraction of nonloop trips in the fitness and recreation category represent trips to a sedentary recreation activity (such as a movie) than for loop trips. In other words, it is likely that the nature of this composite category (created for reasons explained in Trip Purpose for Loop Trips in this chapter) varies substantially between loop and nonloop trips. 437 Table 197. Distance and duration of loop and nonloop trips for fitness and recreation. Percentage of Trips that are for Fitness or Recreation (by Mode) * Loop Trips Nonloop Trips All modes 84.7% 5.6% Walk or wheelchair 88.5% 16.7% Bike 96.5% 14.0% POV (including rental car) Other ground or water 56.3% 4.8% 81.3% 4.9% Mean Distance of Fitness/Recreation Trips (Miles) Loop Trips Nonloop Trips All modes 3.7 10.8 Walk or wheelchair 1.1 0.7 Bike 4.9 2.4 POV (including rental car) 33.2 13.0 Other ground or water 0.9 18.5 Mean Duration of Fitness/Recreation Trips (Minutes) Loop Trips Nonloop Trips All modes 42.9 25.5 Walk or wheelchair 32.2 13.9 Bike 55.3 25.7 POV (including rental car) 162.5 26.7 Other ground or water 103.1 50.2 Note: All statistics are weighted. * For loop trips, home trips (regular home activities, work from home) are reclassified as recreation/fitness. School loop trips at the school location are also reclassified as recreation/fitness. Nonloop trips with these purposes are unchanged. 438 APPENDIX: SAMPLE SIZE TABLES 439 This appendix contains detailed sample breakdowns of the different populations and subpopulations analyzed for this report. As shown in table 198, NHTS provides data on (A) households, (B) vehicles owned by each household, (C) the people that live in those households, and (D) trips made by the people living in each household. NHTS is a national dataset. Unless otherwise stated, all analysis in this report is based specifically on the samples of Georgia households, residents, vehicles, and trips. For our analysis, we used information in the trip files to additionally derive samples of (E) work journeys (see chapter 2), (F) legs of travel to access/egress another mode of transportation (chapter 6), and (G) total nonmotorized travel (pooling nonmotorized trips and nonmotorized access/egress legs). 440 Table 198. Overview of Georgia NHTS population types and samples used in this report. All analyses in this report are based on one of these seven samples, or a subsample thereof (for example, workers are a subset of persons). The remainder of this appendix provides more detailed information on each sample and key subsamples.121 Each section begins with a list of which tables in the body of the report use the relevant samples and subsamples. The tables also note populations for which sample tables can be found in the main body of the report. 121 Tables in chapter 1 through chapter 7 may exclude additional observations from their sample based on missing data. Unless otherwise stated, missing data are omitted from relevant rows/columns rather than the entire table. Where missing data occur, it represents a small fraction of observations. 441 HOUSEHOLD AND VEHICLE SAMPLE TABLES Table 199. Overview of household and vehicle subsamples. Population Subpopulation Households n/a (all households) Households Online shoppers (1+ orders in past month) Sample Size* Tables Using (Unweighted) Subpopulation 8,611 1.37, 1.16, 1.25, 1.26, 4.14, 4.15, 6.22 5,714 4.14, 4.15 Detailed Sample Table(s) A3 n/a Vehicles n/a (all vehicles) 16,947 1.36, 1.2729 A4 Vehicles Seldom-used 832 1.30, 1.31 n/a Vehicles Newly purchased (past 2,369 1.26 n/a 12 months) Vehicles Recent (model year 20042017) 11,588 4.3, 4.5, 4.6 n/a Vehicles Alternative-fuel vehicles 313 4.16 4.2 Vehicles Conventional-fuel 16,606 4.2, 4.4 4.2 vehicles * Additional exclusion criteria may be listed in individual tables. Unless otherwise stated, missing data are omitted from relevant rows/columns rather than the entire table. 442 Table 200. Household sample table. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N Percent (All Observations) All Georgia households - 8,611 - MPO Tier 1. Atlanta MPO counties 53.6% 2,532 29.4% 2. Medium MPO counties 16.3% 3,144 36.5% 3. Small MPO counties 10.2% 2,004 23.3% 4. Non-MPO counties 19.9% 931 Addresses are classified by whether the county is part of an MPO. Medium MPOs have an MPO population of 10.8% 200,0011,000,000 people; small MPOs have a population of 200,000 or less. See chapter 1, Geographic Divisions for Analysis for more details. MPO Albany 1.2% 229 2.7% Athens 2.5% 502 5.8% Atlanta 53.6% 2,532 29.4% Augusta 3.8% 746 8.7% Brunswick 1.1% 245 2.9% Cartersville 0.8% 167 1.9% Chattanooga 1.1% 70 0.8% Columbus 2.7% 503 5.8% Dalton 1.1% 201 2.3% Gainesville 2.3% 512 6.0% Hinesville 0.7% 115 1.3% Macon 1.9% 350 4.1% Rome 0.8% 167 1.9% Savannah 3.8% 811 9.4% Valdosta 1.1% 202 2.4% Warner Robins 1.6% 328 3.8% Non-MPO 19.9% 931 10.8% Neighborhood Type (Urbanicity) Rural 22.3% 2,050 23.8% Small town 26.9% 2,725 31.7% Suburban 27.8% 1,751 20.3% Second city 20.5% 1,945 22.6% Urban 2.7% 140 1.6% Urbanicity is a measure of built environment type. See chapter 1, Geographic Divisions for Analysis for detailed definitions. Table continues on next page. 443 Continued from previous page: Household sample table. Weighted Percent (Nonmissing Observations) All Georgia households - Annual Household Income <$15,000 16.8% $15,000 to $24,999 10.4% $25,000 to $34,999 11.0% $35,000 to $49,999 12.1% $50,000 to $74,999 16.4% $75,000 to $99,999 10.9% $100,000+ 22.4% Missing Household Composition One adult, no children 18.5% 2+ adults, no children 20.8% One adult, youngest child 05 2.1% 2+ adults, youngest child 05 11.0% One adult, youngest child 615 4.0% 2+ adults, youngest child 615 12.3% One adult, youngest child 1621 1.4% 2+ adults, youngest child 1621 5.2% One adult, retired, no children 8.4% 2+ adults, retired, no children 16.3% Number of Workers 0 25.5% 1 41.2% 2+ 33.3% Table continues on next page. Unweighted Sample Size, N 8,611 Percent (All Observations) - 1,119 818 865 1,072 1,487 1,048 1,904 298 13.0% 9.5% 10.1% 12.5% 17.3% 12.2% 22.1% 3.5% 1,635 1,793 93 657 225 710 87 290 1,129 1,992 19.0% 20.8% 1.1% 7.6% 2.6% 8.3% 1.0% 3.4% 13.1% 23.1% 2,972 3,262 2,377 34.5% 37.9% 27.6% 444 Continued from previous page: Household sample table. Weighted Percent (Nonmissing Observations) All Georgia households - Unweighted Sample Size, N Percent (All Observations) 8,611 - Race of Household Member(s) White non-Hispanic only 54.0% 5,680 66.0% Black and Black multiracial only (incl. Black Hisp.) 29.5% 1,949 22.6% Other race only 7.1% 382 4.4% Multiracial household (multiple races) 9.5% 600 7.0% This definition differs from NHTS' measure of household race, which is based only on the race of the primary respondent. Female-headed Household Not female-headed 71.6% 6,206 72.1% Female-headed 28.4% 2,405 27.9% A female-headed household is one with no adult males, or with adult males between the ages of 1821 who are the child or other dependent of an older adult woman. Mobility impairment Absent 82.4% 7,111 82.6% Present 17.6% 1,496 17.4% Missing 4 0.1% A mobility impairment is defined as a "condition or handicap that makes it difficult to travel outside of the home." Vehicle Deficit Category of Household Zero-vehicle 6.9% 444 5.2% Deficit (hard or soft) 18.9% 1,055 12.3% Nondeficit (sufficient/surplus) 74.1% 7,112 82.6% A vehicle-deficit household owns at least one vehicle, but fewer vehicles than household members ages 16+ (i.e., potential drivers). See chapter 1, Vehicle Ownership for more details. Vehicle Deficit Category by Household Size Nondeficit, single potential driver 26.4% 2,685 31.2% Nondeficit, 2+ potential drivers 47.7% 4,427 51.4% Deficit 18.9% 1,055 12.3% Zero-vehicle 6.9% 444 5.2% Table continues on next page. 445 Continued from previous page: Household sample table. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N Percent (All Observations) All Georgia households - 8,611 - Household Vehicle Count 0 6.9% 444 5.2% 1 33.6% 2,737 31.8% 2 34.6% 3,281 38.1% 3+ 24.9% 2,149 25.0% Household Purchased 1+ vehicles in past 12 months No 73.9% 6,532 75.9% Yes 26.1% 2,079 24.1% Newly purchased vehicles can be new or used so long as the household acquired them within the past 12 months. Transit Funding Status of County of Residence No transit funding 10.5% 948 11.0% Rural (on-demand) 29.4% 2,528 29.4% Urban (whole county) 46.0% 3,031 35.2% Urban & rural 12.7% 2,039 23.7% City only 1.4% 65 0.8% County-level transit funding information provided by Garrow et al. (2019). For more information, see chapter 1, Transit Availability and Use. County-level Transit Access None 4.5% 228 2.7% Partial* 35.4% 3,248 37.7% Full 60.1% 5,135 59.6% * Partial transit access includes counties with no fixed-route transit service that have: (a) on-demand rural transit service, (b) access to fixed-route transit in a different county in the MPO, or (c) both. See chapter 1, Transit Availability and Use for more information. Transit Funding Category of County of Residence No transit funding 10.5% 948 11.0% Rural (on-demand) only 29.4% 2,528 29.4% Transit funding 60.1% 5,135 59.6% 446 Table 201. Sample table: Vehicles owned by Georgia households. All vehicles MPO Tier 1. Atlanta MPO counties 2. Medium MPO counties 3. Small MPO counties 4. Non-MPO counties Weighted Percent (Nonmissing Observations) - Unweighted Sample Size, N 16,947 Percent (All Observations) - 52.8% 16.0% 10.1% 21.1% 4,866 6,197 3,878 2,006 28.7% 36.6% 22.9% 11.8% Addresses are classified by whether the county is part of an MPO. Medium MPOs have an MPO population of 200,0011,000,000 people; small MPOs have a population of 200,000 or less. See chapter 1, Geographic Divisions for Analysis for more details. MPO by County Albany 1.0% 392 2.3% Athens 2.5% 1,006 5.9% Atlanta 52.8% 4,866 28.7% Augusta 3.4% 1,398 8.3% Brunswick 1.1% 472 2.8% Cartersville 1.0% 352 2.1% Chattanooga 1.2% 149 0.9% Columbus 2.5% 934 5.5% Dalton 1.1% 415 2.5% Gainesville 2.7% 1,162 6.9% Hinesville 0.7% 232 1.4% Macon 1.8% 650 3.8% Rome 0.8% 338 2.0% Savannah 3.7% 1,548 9.1% Valdosta 1.1% 381 2.3% Warner Robins 1.6% 646 3.8% Non-MPO 21.1% 2,006 11.8% Vehicle Type Car/wagon 49.5% 7,988 47.1% Van 5.4% 840 5.0% SUV 23.6% 3,978 23.5% Pickup 17.5% 3,369 19.9% Other truck 0.6% 97 0.6% RV 0.7% 113 0.7% Motorcycle 2.4% 477 2.8% Something else 0.3% 54 0.3% Missing 31 0.2% Table continues on next page. 447 Continued from previous page: Sample of vehicles owned by Georgia households. All vehicles Weighted Percent (Nonmissing Observations) - Unweighted Sample Size, N 16,947 Percent (All Observations) - Newly Purchased (past 12 months) Not newly purchased 84.2% 14,500 85.6% Newly purchased 15.8% 2,369 14.0% Missing 78 0.5% Newly purchased vehicles can be new or used so long as the household acquired them within the past 12 months. Newly Purchased (past 12 months) Not newly purchased 84.2% 14,500 85.6% Newly purchased 15.8% 2,369 14.0% Missing 78 0.5% Newly purchased vehicles can be new or used so long as the household acquired them within the past 12 months. Vehicle Age Cohort Pre-LEV (pre-1993) 4.6% 882 5.2% LEV1 (19932003) 26.7% 4,346 25.6% LEV2 (20042014) 56.3% 9,534 56.3% New vehicles (20152017) 12.3% 2,054 12.1% Missing 131 0.8% Vehicle Mileage 049,999 mi 28.6% 4,240 25.0% 50,00099,999 mi 24.1% 3,598 21.2% 100149,999 mi 20.6% 2,763 16.3% 150199,999 mi 14.4% 1,863 11.0% 200,000+ mi 12.3% 1,481 8.7% Missing 3,002 17.7% Table continues on next page. 448 Continued from previous page: Sample of vehicles owned by Georgia households. Weighted Unweighted Percent Sample Size, N Percent (Nonmissing (All Observations) Observations) All vehicles - 16,947 - Fuel Type Gas 96.0% 16,229 95.8% Diesel 2.1% 367 2.2% Hybrid, electric, or other alternative fuel 1.8% 307 1.8% Some other fuel 0.1% 18 0.1% Missing 26 0.2% Seldom Used No 95.9% 15,823 93.4% Yes 4.1% 832 4.9% Missing 292 1.7% A seldom-used vehicle is defined as a vehicle with fewer than 1,050 annual miles that is driven less than half as many miles as would be expected given the household annual miles driven and number of vehicles. See chapter 1, Vehicle Fleet Characteristics. Annual Household Income <$15,000 8.4% 1,154 6.8% $15,000 to $24,999 7.6% 1,162 6.9% $25,000 to $34,999 9.3% 1,488 8.8% $35,000 to $49,999 12.0% 2,073 12.2% $50,000 to $74,999 18.4% 3,187 18.8% $75,000 to $99,999 13.7% 2,478 14.6% $100,000+ 30.6% 4,852 28.6% Missing 553 3.3% Vehicle Deficit Category of Household Deficit (hard or soft) 14.4% 1,452 8.6% Nondeficit (sufficient/surplus) 85.6% 15,495 91.4% A vehicle-deficit household owns at least one vehicle, but fewer vehicles than household members ages 16+ (i.e., potential drivers). See chapter 1, Vehicle Ownership for more details. 449 PERSON SAMPLE TABLES Table 202. Overview of person subsamples. Unweighted Detailed Sample Tables Using Sample Population Subpopulation Size* Subpopulation Table(s) Persons Ages 5+ All persons ages 5+ n/a (all) 17,681 1.36, 1.8, 1.17, A6 Drivers ages 5+ n/a (all) 1.18, 6.7, 7.3 14,292 1.3, 1.8 n/a Adults Ages 18+ All adults 18+ n/a (all) 15,222 2.1, 2.4, 3.2, A7, 5.17 5.1.5, 5.1721, 6.1, 6.3, 6.4, 6.14, 6.22 Workers ages 18+ n/a (all) 8,293 2.1, 2.2, 2.4, 2.7, A8-9 2.20, 2.22, 2.24, 2.31, 3.1, 3.2, 3.46, 3.1113, 3.23, 3.2629, 3.31 Workers ages 18+ College-educated 4,202 2.3, 3.7, 3.9 n/a Workers ages 18+ Noncollege-educated 4,081 2.3, 3.8, 3.10 n/a Workers ages 18+ Telecommute-eligible 1,079 3.1722, 3.24, 3.25 n/a Adults 18+ Active commuters (on travel day) 5,039 2.5, 2.6, 2.8, 2.9, A10 2.27, 2.28, 2.36, Adults 18+ Active workers (on travel day) 5,720 3.2, 3.3 n/a Adults 18+ Active telecommuters (on travel day) 713 3.2 n/a Adults 18+ With mobility impairment 1,632 5.2235 5.30 Adults 18+ Without mobility impairment 13,582 5.225.27 5.30 Adults 18+ Recent pedestrian or cyclist (past 7 days) 11,111 6.2, 6.2325 n/a Adults 18+ Travel-day pedestrian or cyclist 2,019 6.236.25 n/a Persons Ages 16+ All persons ages 16+ n/a (all) 15,605 4.8, 4.10, 4.11, 4.14, A11 4.1719 Persons ages 16+ Ridehailing users 1,176 4.7, 4.1113 4.7, 4.12 Persons ages 16+ Carsharing users 104 4.74.9 4.7, 4.8 Persons ages 16+ Bikesharing users 51 4.74.9 4.7, 4.8 Persons ages 16+ Other recent cyclists (past 7 days) 842 4.8 4.8 Persons ages 16+ Workers 8,363 1.3, 1.4, 1.3436 A12 Persons ages 16+ Drivers 14,236 1.3, 1.4, 5.1015 5.11 Other Age Ages 1617 Teen drivers in vehicle-owning households 212 5.16 n/a Children 517 n/a (all) 2,459 5.29, 6.3235 5.29 Children 517 Enrolled in public or private school 2,275 6.276.29 6.27 Older adults ages 80+ Without mobility impairment 534 5.28 n/a * Additional exclusion criteria may be listed in individual tables. Unless otherwise stated, missing data are omitted from relevant rows/columns rather than the entire table. 450 Table 203. Person sample table: All persons ages 5+. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, Percent N (All Observations) All persons ages 5+ MPO Tier 1. Atlanta MPO counties 2. Medium MPO counties 3. Small MPO counties 4. Non-MPO counties - 54.2% 16.0% 10.2% 19.6% 17,681 5,284 6,459 4,035 1,903 - 29.9% 36.5% 22.8% 10.8% Addresses are classified by whether the county is part of an MPO. Medium MPOs have an MPO population of 200,0011,000,000 people; small MPOs have a population of 200,000 or less. See chapter 1, Geographic Divisions for Analysis for more details. MPO Albany Athens Atlanta Augusta Brunswick Cartersville Chattanooga Columbus Dalton Gainesville Hinesville Macon Rome Savannah Valdosta Warner Robins Non-MPO Age Child 515 Teen 1617 Adult 1864 Senior 6579 Elderly 80+ NHTS imputed age for 50 people. Age Cohort (Adults Only) Millennial and Gen Z (1836) Gen X (3752) Pre-retirement age Boomer (5364) Retirement age (65+) {Children & teens} Table continues on next page. 1.1% 2.4% 54.2% 3.8% 1.0% 0.8% 0.9% 2.7% 1.2% 2.4% 0.8% 1.9% 0.7% 3.8% 1.1% 1.6% 19.6% 16.2% 3.2% 67.0% 11.1% 2.5% 34.2% 28.7% 20.3% 16.9% 436 987 5,284 1,567 473 326 129 1,035 421 1,089 257 702 338 1,652 400 682 1,903 2,076 383 10,771 3,609 842 3,244 3,571 3,956 4,451 2,459 2.5% 5.6% 29.9% 8.9% 2.7% 1.8% 0.7% 5.9% 2.4% 6.2% 1.5% 4.0% 1.9% 9.3% 2.3% 3.9% 10.8% 11.7% 2.2% 60.9% 20.4% 4.8% 18.4% 20.2% 22.4% 25.2% 13.9% 451 Continued from previous page: Sample of persons ages 5+. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, Percent N (All Observations) All persons ages 5+ Driver Status by Age Nondriver ages 16+ Driver ages 16+ - 10.5% 73.3% 17,681 1,368 14,236 - 7.7% 80.5% Underage driver (Excluded from analysis unless otherwise stated) 0.4% 56 0.3% Underage nondriver 15.8% 2020 11.4% Missing 1 0.0% NHTS does not ask about drivers' licensing; rather, they ask "do you/does this person drive?" Some children under 16 were reported as drivers, perhaps due to learner's permits. Unless otherwise stated, drivers in the report refers only to drivers ages 16+. Worker Status by Age Nonworker ages 18+ 31.1% 6,929 39.2% Worker ages 18+ 49.6% 8,293 46.9% Nonworker ages 1617 2.7% 313 1.8% Worker ages 1617 0.4% 70 0.4% Child under 16 2076 11.7% NHTS defines a worker as someone who worked for pay or profit, or was temporarily absent from paid employment, in the week before completing the travel survey ("last week"). Unless otherwise specified, all references to "workers" in this report refer to NHTS-defined workers ages 18+. Sex Male Female NHTS imputed sex for 16 people. Race (Detailed) White non-Hispanic only Black non-Hispanic only Latino (white Hispanic) only Asian/Pacific Islander only Native American only Other (single race) Black multiracial (incl. Black Hispanic) 48.5% 51.5% 53.4% 30.8% 7.6% 3.9% 0.2% 0.2% 2.3% 8,142 9,539 12,057 3,847 625 521 49 15 256 46.1% 54.0% 68.2% 21.8% 3.5% 3.0% 0.3% 0.1% 1.5% Other multiracial (not including Black or Black Hispanic) 1.6% 311 1.8% NHTS imputed race and/or Hispanic status for 74 people. See chapter 1, Assorted Definitions and Notes for more details on how race is categorized in this report. Race (Categories) White non-Hispanic only Black and Black multiracial (incl. Black Hispanic) Other Table continues on next page. 53.4% 33.0% 13.6% 12,057 4,103 1,521 68.2% 23.2% 8.6% 452 Continued from previous page: Sample of persons ages 5+. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, Percent N (All Observations) All persons ages 5+ Race/Ethnicity (used in Chapter 5: Equity) - 17,681 - White non-Hispanic only 53.4% 12,057 68.2% Black and Black multiracial (excl. Black Hispanic) 31.9% 4,027 22.8% Hispanic (any race) 9.0% 724 4.1% Asian or other Mobility Impairment Absent 5.6% 91.9% 873 15,998 4.9% 90.5% Present 8.1% 1,674 9.5% Missing 9 0.1% A mobility impairment is defined as a "condition or handicap that makes it difficult to travel outside of the home." Annual Household Income <$15,000 14.3% 1,846 10.4% $15,000 to $24,999 9.8% 1,475 8.3% $25,000 to $34,999 10.2% 1,622 9.2% $35,000 to $49,999 11.8% 2,143 12.1% $50,000 to $74,999 16.2% 3,034 17.2% $75,000 to $99,999 11.8% 2,346 13.3% $100,000+ 25.9% 4,673 26.4% Missing Education Level 542 3.1% 45 mi 5.5% 339 4.1% Missing 1,381 16.7% Road network distance, in miles, between respondent's home location and work location, as calculated by NHTS. Table continues on next page. 469 Continued from previous page: Sample of workers ages 18+. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, Percent N (All Observations) All workers ages 18+ Occupational Category Sales or service 26.9% Clerical or administrative support 9.3% Blue collar* 17.7% Professional, managerial, or technical 46.0% Other 0.1% Missing * Blue collar refers to manufacturing, construction, maintenance, or farming. Worker Type Full-time (35+ hours per week) 19.6% Part-time 80.4% Missing Work Flexibility Type Schedule and location 22.3% Location only 4.9% Schedule only 18.7% Neither schedule nor location 54.2% Missing 8,293 1,930 799 1,143 4,112 10 299 1,553 6,446 294 1,839 317 1,690 4,142 305 23.3% 9.6% 13.8% 49.6% 0.1% 3.6% 18.7% 77.7% 3.6% 22.2% 3.8% 20.4% 50.0% 3.7% Schedule refers to flextime. Location refers to telecommute-eligible workers and home-based workers. See Flextime and Telework Eligibility below for definitions, and chapter 3, Overview for more details. Flextime Ineligible 59.1% Eligible 40.9% Missing "At your (primary) job, do you have the ability to set or change your own start time?" Telework Eligibility Telecommute-ineligible worker 73.8% Telecommute-eligible worker 13.0% Home-based worker 13.2% 4,459 3,529 305 6,129 1,079 1,085 53.8% 42.6% 3.7% 73.9% 13.0% 13.1% A home-based worker "usually work(s) from home." A telecommute-eligible worker does not usually work from home, "have the option of working from home or an alternate location instead of going into your/their primary workplace." See chapter 3, Worker Telework Eligibility Categories for details. Travel Day Teleworking Did not work Telework (exclusive) Telework (mixed) Conventional commute 37.2% 5.5% 2.3% 54.9% 2,573 501 212 5,007 31.0% 6.0% 2.6% 60.4% A respondent teleworked on the travel day if they reported working from home for pay. They engaged in a conventional commute if they reported a trip with the purpose of working for pay (not at home). Mixed teleworkers reported both teleworking and a conventional commute. See chapter 3, Travel Day Work and Telework Categories for more details. 470 Table 207. Person sample table: Active commuters ages 18+. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, Percent N (All Observations) All active commuters ages 18+ 5,113 - An active commuter is a person who made one or more work journeys on the travel day (see chapter 2). NHTS-defined Worker Status Nonworker ages 18+ 1.3% 74 1.4% Worker ages 18+ 98.7% 5,039 98.6% NHTS defines a worker as someone who worked for pay or profit, or was temporarily absent from paid employment, in the week before completing the travel survey ("last week"). A small number of people who were not NHTS-designated workers nevertheless reported work travel on their travel day, perhaps reflecting irregular employment situations. MPO Tier 1. Atlanta MPO counties 57.2% 1,650 32.3% 2. Medium MPO counties 14.9% 1,858 36.3% 3. Small MPO counties 9.8% 1,131 22.1% 4. Non-MPO counties 18.2% 474 9.3% Addresses are classified by whether the county is part of an MPO. Medium MPOs have an MPO population of 200,0011,000,000 people; small MPOs have a population of 200,000 or less. See chapter 1, Geographic Divisions for Analysis for more details. Age Cohort (Adults Only) Millennial and Gen Z (1836) 40.2% 1,517 29.7% Gen X (3752) 35.9% 1,752 34.3% Pre-retirement age Boomer (5364) 19.7% 1,456 28.5% Retirement age (65+) 4.2% 388 7.6% Sex Male 56.2% 2,650 51.8% Female 43.8% 2,463 48.2% Annual Household Income <$35,000 27.9% 1,048 20.5% $35,000 to $49,999 12.8% 646 12.6% $50,000 to $74,999 18.3% 951 18.6% $75,000 to $99,999 14.3% 806 15.8% $100,000+ 26.7% 1,556 30.4% Missing 106 2.1% Table continues on next page. 471 Continued from previous page: Sample of active commuters ages 18+. Weighted Unweighted Percent Sample Size, Percent (Nonmissing N (All Observations) Observations) All active commuters ages 18+ 5,113 - An active commuter is a person who made one or more work journeys on the travel day (see chapter 2). Race (Categories) White non-Hispanic only 53.4% 3,493 68.3% Black and Black multiracial (incl. Black Hispanic) 32.7% 1,182 23.1% Other 13.9% 438 8.6% NHTS imputed race and/or Hispanic status for 74 people. See chapter 1, Assorted Definitions and Notes for more details on how race is categorized in this report. Occupational Category Sales or service 27.2% 1,159 22.7% Clerical or administrative support 9.1% 499 9.8% Blue collar* 19.9% 795 15.5% Professional, managerial, or technical 43.6% 2,553 49.9% Other 0.1% 5 0.1% Not an NHTS-defined worker 74 1.4% Missing 28 0.5% * Blue collar refers to manufacturing, construction, maintenance, or farming. Worker Type Part-time (<35 hours per week) 15.4% 708 13.8% Full-time (35+ hours per week) 84.6% 4,305 84.2% Not an NHTS-defined worker Unknown College-educated No 4-year degree Bachelor's or higher Missing 59.4% 40.6% 78 22 2,654 2,453 6 1.5% 0.4% 51.9% 48.0% 0.1% 472 Table 208. Person sample table: Persons ages 16+. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, Percent N (All Observations) All persons ages 16+ 17,618 - MPO Tier 1. Atlanta MPO counties 54.4% 4,627 26.3% 2. Medium MPO counties 15.8% 5,683 32.3% 3. Small MPO counties 10.1% 3,592 20.4% 4. Non-MPO counties 19.8% 1,703 9.7% Addresses are classified by whether the county is part of an MPO. Medium MPOs have an MPO population of 200,0011,000,000 people; small MPOs have a population of 200,000 or less. See chapter 1, Geographic Divisions for Analysis for more details. Age Cohort Teenager (1617) 3.8% 383 2.2% Millennial and Gen Z (1836) 32.9% 3,244 18.4% Gen X (3752) 27.6% 3,571 20.3% Pre-retirement age Boomer (5364) 19.5% 3,956 22.5% Seniors (6579) 13.2% 3,609 20.5% Elderly (80+) 3.0% 842 4.8% Sex Male 47.9% 7,038 39.9% Female 52.1% 8,567 48.6% Annual Household Income <$15,000 14.5% 1,650 9.4% $15,000 to $24,999 9.5% 1,303 7.4% $25,000 to $34,999 10.1% 1,447 8.2% $35,000 to $49,999 11.9% 1,904 10.8% $50,000 to $74,999 16.6% 2,731 15.5% $75,000 to $99,999 11.7% 2,061 11.7% $100,000+ 25.7% 3,995 22.7% Missing 514 2.9% Table continues on next page. 473 Continued from previous page: Sample of persons ages 16+. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, Percent N (All Observations) All persons ages 16+ 17,618 - Race (Categories) White non-Hispanic only 54.8% 10,852 61.6% Black and Black multiracial (incl. Black Hispanic) 32.1% 3,507 19.9% Other 13.1% 1,246 7.1% NHTS imputed race and/or Hispanic status for 74 people. See chapter 1, Assorted Definitions and Notes for more details on how race is categorized in this report. Mobility Impairment Absent 90.7% 13,957 79.2% Present 9.3% 1,640 9.3% Missing 8 0.0% A mobility impairment is defined as a "condition or handicap that makes it difficult to travel outside of the home." Driver Status Nondriver ages 16+ 12.6% 1,368 7.8% Driver ages 16+ 87.4% 14,236 80.8% Missing 1 0.0% Vehicle Deficit Category of Household Zero-vehicle 5.0% 589 3.3% Deficit (hard or soft) 27.0% 2,697 15.3% Nondeficit (sufficient/surplus) 68.0% 12,319 69.9% A vehicle-deficit household owns at least one vehicle, but fewer vehicles than household members ages 16+ (i.e., potential drivers). See chapter 1, Vehicle Ownership for more details. Transit Use, Past 30 Days No 88.9% 14,344 81.4% Yes 11.1% 1,246 7.1% Missing 15 0.1% Walking, Past 30 Days No 28.0% 4,257 24.2% Yes 72.0% 11,285 64.1% Missing 63 0.4% 474 Table 209. Person sample table: Workers ages 16+. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, Percent N (All Observations) All workers ages 16+ 8,363 - MPO Tier 1. Atlanta MPO counties 57.7% 2,808 33.6% 2. Medium MPO counties 15.7% 3,033 36.3% 3. Small MPO counties 10.0% 1,794 21.5% 4. Non-MPO counties 16.5% 728 8.7% Addresses are classified by whether the county is part of an MPO. Medium MPOs have an MPO population of 200,0011,000,000 people; small MPOs have a population of 200,000 or less. See chapter 1, Geographic Divisions for Analysis for more details. Age Cohort Teenager (1617) 0.9% 70 0.8% Millennial and Gen Z (1836) 37.8% 2,361 28.2% Gen X (3752) 36.5% 2,843 34.0% Pre-retirement age Boomer (5364) 19.9% 2,352 28.1% Seniors (6579) 4.7% 701 8.4% Elderly (80+) 0.2% 36 0.4% Sex Male 53.7% 4,203 50.3% Female 46.3% 4,160 49.7% Annual Household Income <$15,000 8.2% 481 5.8% $15,000 to $24,999 7.7% 478 5.7% $25,000 to $34,999 10.1% 721 8.6% $35,000 to $49,999 12.2% 986 11.8% $50,000 to $74,999 17.7% 1,500 17.9% $75,000 to $99,999 13.2% 1,260 15.1% $100,000+ 30.9% 2,743 32.8% Missing Table continues on next page. 194 2.3% 475 Continued from previous page: Sample of workers ages 16+. All workers ages 16+ Weighted Percent (Nonmissing Observations) Unweighted Sample Size, Percent N (All Observations) 8,363 - Race (Categories) White non-Hispanic only 56.1% 5,789 69.2% Black and Black multiracial (incl. Black Hispanic) 30.7% 1,862 22.3% Other 13.2% 712 8.5% NHTS imputed race and/or Hispanic status for 74 people. See chapter 1, Assorted Definitions and Notes for more details on how race is categorized in this report. Mobility Impairment Absent 98.1% 8,197 98.0% Present 1.9% 163 1.9% Missing 3 0.0% A mobility impairment is defined as a "condition or handicap that makes it difficult to travel outside of the home." Driver Status Nondriver ages 16+ 5.2% 257 3.1% Driver ages 16+ 94.8% 8,105 96.9% Missing 1 0.0% Vehicle Deficit Category of Household Zero-vehicle 2.5% 143 1.7% Deficit (hard or soft) 21.1% 1,086 13.0% Nondeficit (sufficient/surplus) 76.4% 7,134 85.3% A vehicle-deficit household owns at least one vehicle, but fewer vehicles than household members ages 16+ (i.e., potential drivers). See Section 1.6.1 for more details. Transit Use, Past 30 Days No 88.3% 7,628 91.2% Yes 11.7% 729 8.7% Missing 6 0.1% Walking, Past 30 Days No 27.7% 2,179 26.1% Yes 72.3% 6,164 73.7% Missing 20 0.2% 476 TRIP, WORK JOURNEY, AND ACCESS/EGRESS TRAVEL SAMPLE TABLES Samples of trips and trip-derived variables (WJs, access/egress legs) are broken down by who is doing the traveling and what kind of travel is being discussed (person trips, vehicle trips, work journeys, access/egress legs).122 Additional subsamples are based on these primary divisions (e.g., nonmotorized person trips made by adults ages 18+). To clarify these relationships, table 210 summarizes the sample sizes of traveler populations and instances of travel by members of those populations. Table 211 provides a more detailed overview of specific subsamples, where these subsamples are used in the report, and the location of more detailed subpopulation sample tables. In addition to tables focusing on trips or other instances of travel, travel behavior is frequently included in tables in a normalized fashion (e.g., trips per capita). In recognition of this reality, we have included sample tables for person trips and vehicle trips by adults ages 18+. These samples were not the independent basis of tables in the report, but were frequently used to calculate figures found in other tables. 122 See Key Terms in chapter 1 for more discussion of the difference between person trips and vehicle trips. Discussions of the methods for deriving work journeys and access/egress legs can be found in chapter 2 and chapter 6. 477 Table 210. Summary of traveler populations and instances of travel. Traveler Group Group A. Number of Persons in Group Trips B. Total Trips Made by Group C. VehicleTrips D. Nonmotorized Trips Work E. Work Journeys Access/Egress Legs NMT F. Non- H. Non- motorized G. Motorized motorized Legs Legs Travel (D+F) All Georgians ages 5+ 17,681 59,706 40,635 4,480 10,490 952 367 5,432 Ages 16+ 15,605 54,271 40,606 4,023 10,490 923 355 4,946 Adults ages 18+ 15,222 53,203 40,196 3,927 10,490 903 351 4,830 Children ages 517 2,459 6,503 439 553 - 49 16 602 Workers ages 16+ 8,363 31,623 25,484 2,041 10,340 466 218 2,507 Workers ages 18+ 8,293 31,356 25,333 2,017 10,340 466 218 2,483 Drivers, all ages 14,292 51,597 40,606 3,394 10,199 586 244 3,980 Drivers ages 16+ 14,236 51,597 40,606 3,394 10,199 586 244 3,980 Drivers ages 18+ 14,022 50,924 40,196 3,351 10,199 583 243 3,934 * Derived from trips as described in chapters 2 and 6. Also described as person-trips. Vehicle trips are private auto or rental car trips where the respondent is driving the vehicle. Total motorized trips (all modes except walking and biking) can be calculated by subtracting column D from column B. 478 Table 211. Overview of trip, work journey, and access/egress travel subsamples. Trip or Instance Category Traveler Type Trip or Instance Type Unweighted Sample Size* Tables Using Population Detailed Sample Table(s) Person-trips (all trips by all modes) Person-trip Persons ages 5+ All 59,706 1.5, 1.6, 1.912, 1.14, 1.2024, 7.2, A1518 7.5, 7.6 Person-trip Persons ages 5+ Loop 1,294 7.1, 7.4 7.1, 7.4 Person-trip Adults ages 18+ All 53,203 n/a A2225 Person-trip Adults ages 18+ Trips entirely within Georgia 50,270 5.65.9 5.6 Person-trip Children ages 517 School trips 2,593 6.26, 6.2829 6.26 Person-trip Persons ages 16+ Vehicle-for-hire 205 4.12, 4.13 4.12 Vehicle-trips (motorized trips where respondent is driver) Vehicle-trip Persons ages 5+ All 40,635 1.5, 1.6, 1.9, 1.15 A1921 Vehicle-trip Adults ages 18+ All 40,196 n/a A2629 Access/Egress Legs (legs to access or egress another mode of transportation, included in data as part of the trip by the parent mode. See chapter 6.) Access/egress legs Persons ages 5+ All 1,319 6.8, 6.9 6.8 Access/egress legs Adults ages 18+ Legs to access/egress transit 949 6.116.13 6.8 Access/egress legs Adults ages 18+ All 1,254 6.10 6.8 Nonmotorized Trips and Legs (total of nonmotorized trips + nonmotorized access/egress legs) Person-trips + legs Persons ages 5+ Nonmotorized 5,432 6.5, 6.6 n/a Person-trips + legs Adults ages 18+ Nonmotorized 4,830 6.1521 6.18, 6.20 Person-trips + legs Children ages 517 Nonmotorized 602 6.30, 6.31, 6.34 6.30, 6.31, 6.34 Work Journey/Commute (note: WJs and commutes differ in distance and duration, but are synonymous for sample size.) WJ/commute Adults ages 18+ All (incl. supercommutes >100 mi) 10,463 2.1013, 2.15, 2.20, 2.21, 2.23, 2.30 A3032 WJ/commute Adults ages 18+ WJ excl. supercommutes >100 mi 10,378 2.1619, 2.25, 2.26, 2.29, 2.3235 n/a WJ/commute Adults ages 18+ Complex WJs 6,218 2.16 2.16 WJ/commute Adults ages 18+ Matched Complex WJs 1,585 2.17 2.16 * Sample size given is for instances of travel (trips, legs, or WJs) rather than the number of individuals in the traveler population (e.g., adults, children). Additional exclusion criteria may be listed in individual tables. Unless otherwise stated, missing data are omitted from relevant rows/columns rather than the entire table. In addition to tables listed here, trip data are often the denominator in tables of normalized data (e.g., trips per capita, WJ per worker, etc.). 479 Table 212. Trip sample table: All trips by all persons ages 5+ by location. All trips by all persons ages 5+ Weighted Percent (Nonmissing Observations) - Unweighted Sample Size, N Percent (All Observations) 59,706 - MPO Tier (Residence of Traveler) 1. Atlanta MPO counties 55.7% 17,958 30.1% 2. Medium MPO counties 15.8% 22,160 37.1% 3. Small MPO counties 10.1% 13,563 22.7% 4. Non-MPO counties 18.4% 6,025 10.1% Addresses are classified by whether the county is part of an MPO. Medium MPOs have an MPO population of 200,0011,000,000 people; small MPOs have a population of 200,000 or less. See chapter 1, Geographic Divisions for Analysis for more details. MPO (Residence of Traveler) Albany 1.0% 1,396 2.3% Athens 2.4% 3,622 6.1% Atlanta 55.7% 17,958 30.1% Augusta 3.9% 5,446 9.1% Brunswick 1.3% 1,756 2.9% Cartersville 0.8% 1,146 1.9% Chattanooga 0.7% 404 0.7% Columbus 2.6% 3,432 5.7% Dalton 1.1% 1,419 2.4% Gainesville 2.4% 3,741 6.3% Hinesville 0.7% 770 1.3% Macon 1.8% 2,302 3.9% Rome 0.7% 1,271 2.1% Savannah 3.8% 5,515 9.2% Valdosta 1.1% 1,334 2.2% Warner Robins 1.5% 2,169 3.6% Non-MPO Table continues on next page. 18.4% 6,025 10.1% 480 Continued from previous page: Sample of trips by all persons ages 5+. Weighted Percent (Nonmissing Observations) All trips by all persons ages 5+ - Trip Location (Based on Origin and Destination) In Georgia 95.8% Partly in Georgia 1.5% Not in Georgia 2.7% MPO Tier (Trip Origin) 1. Atlanta MPO 55.9% 2. Medium MPOs 15.8% 3. Small MPOs 10.5% 4. Non-MPO 17.8% Out of state MPO (Trip Origin) Albany 1.1% Athens 2.6% Atlanta 54.0% Augusta 3.7% Brunswick 1.4% Cartersville 0.8% Chattanooga 0.5% Columbus 2.5% Dalton 1.0% Gainesville 2.3% Hinesville Macon 0.7% 1.8% Rome 0.7% Savannah 3.8% Valdosta 1.1% Warner Robins 1.5% Non-MPO 17.2% Out of state 3.5% Unweighted Sample Size, N Percent (All Observations) 59,706 - 57,007 1,000 1,699 95.5% 1.7% 2.8% 18,085 20,525 12,830 6,077 2,189 30.3% 34.4% 21.5% 10.2% 3.7% 1,337 3,477 18,085 5,110 1,644 1,000 297 3,246 1,285 3,134 725 2,316 1,175 5,261 1,309 2,039 6,077 2,189 2.2% 5.8% 30.3% 8.6% 2.8% 1.7% 0.5% 5.4% 2.2% 5.2% 1.2% 3.9% 2.0% 8.8% 2.2% 3.4% 10.2% 3.7% 481 Table 213. Trip sample table: All trips by all persons ages 5+ by purpose. All trips by all persons ages 5+ Purpose Home Work commute Other work-related travel Attend school or daycare Transport someone Shopping or errands Medical/dental services Social/recreational or fitness Dining (restaurant or carryout) Community, religious, and volunteer Other Missing Purpose Type Mandatory Household-serving Discretionary Return home Other Missing Weighted Percent (Nonmissing Observations) - Unweighted Sample Size, Percent N (All Observations) 59,706 - 33.6% 11.1% 1.0% 4.1% 6.7% 18.7% 1.4% 11.1% 7.9% 3.0% 1.3% 19,520 6,770 698 1,888 3,643 12,092 1,187 6,552 4,954 1,658 722 22 32.7% 11.3% 1.2% 3.2% 6.1% 20.3% 2.0% 11.0% 8.3% 2.8% 1.2% 0.0% 16.3% 26.8% 22.1% 33.6% 1.3% 9,356 16,922 13,164 19,520 722 22 15.7% 28.3% 22.0% 32.7% 1.2% 0.0% 482 Table 214. Trip sample table: All trips by all persons ages 5+ by mode. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, Percent N (All Observations) All trips by all persons ages 5+ - 59,706 Mode Pedestrian (walk/wheelchair) 8.0% 4,201 Bike 0.6% 287 POV, including rental car 85.5% 52,675 School bus 2.9% 1,213 Public transit 1.4% 515 Paratransit 0.1% 56 Other bus 0.2% 141 Taxi/ridehail/limo 0.6% 222 Air 0.2% 105 Other 0.4% 289 Missing 2 Mode Category POV, including rental car 85.5% 52,675 Nonmotorized (walk, bike, wheelchair) 8.6% 4,490 Public transit or other bus/train 4.6% 1,869 Other ground or water 1.2% 565 Air 0.2% 105 Missing 2 Vehicle or Person Trip Person trip only 37.9% 19,071 Vehicle trip and person trip 62.1% 40,635 See chapter 1, Key Terms for more explanation of the relationship between person trips and vehicle trips. - 7.0% 0.5% 88.2% 2.0% 0.9% 0.1% 0.2% 0.4% 0.2% 0.5% 0.0% 88.2% 7.5% 3.1% 0.9% 0.2% 0.0% 31.9% 68.1% 483 Table 215. Trip sample table: All trips by all persons ages 5+ by demographic factors. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N Percent (All Observations) All trips by all persons ages 5+ - 59,706 - Sex Male 48.4% 27,472 46.0% Female 51.6% 32,234 54.0% NHTS imputed sex for 16 people. Age Child 515 12.5% 5,435 9.1% Teen 1617 2.5% 1,068 1.8% Adult 1864 72.6% 38,862 65.1% Senior 6579 10.7% 12,333 20.7% Elderly 80+ 1.6% 2,008 3.4% Age by Sex Male: child 515 6.7% 2,966 5.0% Male: teen 1617 1.2% 524 0.9% Male: adult 1864 34.6% 17,067 28.6% Male: senior 6579 5.2% 5,958 10.0% Male: elderly 80+ 0.7% 957 1.6% Female: child 515 5.8% 2,469 4.1% Female: teen 1617 1.3% 544 0.9% Female: adult 1864 38.0% 21,795 36.5% Female: senior 6579 5.6% 6,375 10.7% Female: elderly 80+ 0.9% 1,051 1.8% Age Cohort (Adults Only) Millennial and Gen Z (1836) 32.8% 11,175 18.7% Gen X (3752) 31.3% 13,382 22.4% Pre-retirement age Boomer (5364) 21.3% 14,305 24.0% Retirement age (65+) 14.6% 14,341 24.0% {Children & teens} 6,503 10.9% Driver Status by Age Driver ages 18+ 79.1% 50,924 85.3% Nondriver ages 18+ 5.9% 2,279 3.8% Driver ages 1617 1.4% 673 1.1% Nondriver ages 1617 1.2% 395 0.7% Child ages 515 12.5% 5,435 9.1% NHTS does not ask about drivers' licensing; rather, they ask "do you/does this person drive?" Some children under 16 were reported as drivers, perhaps due to learner's permits. Unless otherwise stated, drivers in the report refers only to drivers ages 16+. Table continues on next page. 484 Continued from previous page: Sample of trips by all persons ages 5+. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N Percent (All Observations) All trips by all persons ages 5+ - 59,706 - Worker Status by Age Nonworker ages 18+ 28.0% 21,847 36.6% Worker ages 18+ 57.0% 31,356 52.5% Nonworker ages 1617 2.0% 801 1.3% Worker ages 1617 0.5% 267 0.4% Child under 16 12.5% 5,435 9.1% Race (Categories) White non-Hispanic only 54.7% 41,603 69.7% Black and Black multiracial (incl. Black Hisp.) 32.2% 13,330 22.3% Other 13.0% 4,773 8.0% NHTS imputed race and/or Hispanic status for 74 people. See chapter 1, Assorted Definitions and Notes for more details on how race is categorized in this report. Race/Ethnicity (used in chapter 5, Equity) White non-Hispanic only 54.7% 41,603 69.7% Black and Black multiracial (excl. Black Hisp.) 31.1% 13,101 21.9% Hispanic (any race) 9.1% 2,296 3.8% Asian or other 5.1% 2,706 4.5% Mobility Impairment Absent 94.2% 55,664 93.2% Present 5.8% 4,019 6.7% Missing 23 0.0% Annual Household Income <$15,000 12.8% 5,587 9.4% $15,000 to $24,999 8.7% 4,428 7.4% $25,000 to $34,999 10.0% 5,182 8.7% $35,000 to $49,999 12.3% 7,281 12.2% $50,000 to $74,999 16.3% 10,536 17.6% $75,000 to $99,999 12.2% 8,236 13.8% $100,000+ 27.7% 16,907 28.3% Missing 1,549 2.6% Education Level High school or less 30.9% 14,333 24.0% Some college or associate degree 28.5% 15,289 25.6% Bachelor's degree 21.9% 12,999 21.8% Graduate or professional degree 18.7% 12,568 21.0% Missing 4,517 7.6% Age <14 0.0% Table continues on next page. 485 Continued from previous page: Sample of trips by all persons ages 5+. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N Percent (All Observations) All trips by all persons ages 5+ - 59,706 - College-educated No 4-year degree 59.4% 29,622 49.6% Bachelor's or higher 40.6% 25,567 42.8% Missing 4,517 7.6% Age <14 4,450 7.5% Immigrant Nonimmigrant (born in US) 89.8% 55,487 92.9% Immigrant (born outside of US) 10.2% 4,206 7.0% Missing 13 0.0% Immigrant by Education Level US HS or less 27.9% 13,487 22.6% US some college/assoc. 26.5% 14,540 24.4% US bachelor's+ 34.7% 23,225 38.9% Imm. HS or less 2.9% 846 1.4% Imm. some college/assoc. 2.3% 876 1.5% Imm. bachelor's+ 5.7% 2,342 3.9% N/A (age <14) or missing 4,390 7.4% Caregiver Status Noncaregiver 65.4% 40,602 68.0% Caregiver, youngest child ages 015 34.6% 12,601 21.1% N/A (child under 18) 6,503 10.9% A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. Caregiver Status by Gender Male noncaregiver 32.3% 18,754 31.4% Female noncaregiver 15.3% 5,228 8.8% Male caregiver 33.1% 21,848 36.6% Female caregiver 19.3% 7,373 12.3% Missing 6,503 10.9% Caregiver Status by Age of Youngest Child Noncaregiver 70.6% 47,105 78.9% Youngest ages 04 14.1% 5,687 9.5% Youngest ages 515 Table continues on next page. 15.3% 6,914 11.6% 486 Continued from previous page: Sample of trips by all persons ages 5+ Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N Percent (All Observations) All trips by all persons ages 5+ - 59,706 - Vehicle Deficit Category of Household Zero-vehicle 4.3% 1,804 3.0% Deficit (hard or soft) 23.9% 8,730 14.6% Nondeficit (sufficient/surplus) 71.8% 49,172 82.4% A vehicle-deficit household owns at least one vehicle, but fewer vehicles than household members ages 16+ (i.e., potential drivers). See chapter 1, Vehicle Ownership for more details. Vehicle Deficit Category by Household Size Nondeficit, single potential driver 15.5% 11,675 19.6% Nondeficit, 2+ potential drivers 56.4% 37,497 62.8% Deficit 23.9% 8,730 14.6% Zero-vehicle 4.3% 1,804 3.0% Transit Use, Past 30 Days No 86.1% 53,866 90.2% Yes 13.9% 5,796 9.7% Missing 44 0.1% Walking, Past 30 Days No 22.9% 13,826 23.2% Yes 77.1% 45,688 76.5% Missing 192 0.3% Biking, Past 30 Days No 89.7% 53,934 90.3% Yes 10.3% 5,751 9.6% Missing 21 0.0% 487 Table 216. Trip sample table: All vehicle trips by all persons ages 5+ by location. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N Percent (All Observations) All vehicle trips by all persons ages 5+ 40,635 - MPO Tier (Residence of Traveler) 1. Atlanta MPO counties 55.9% 11,937 29.4% 2. Medium MPO counties 15.3% 15,095 37.1% 3. Small MPO counties 10.3% 9,460 23.3% 4. Non-MPO counties 18.6% 4,143 10.2% Addresses are classified by whether the county is part of an MPO. Medium MPOs have an MPO population of 200,0011,000,000 people; small MPOs have a population of 200,000 or less. See chapter 1, Geographic Divisions for Analysis for more details. MPO (Residence of Traveler) Albany 1.1% 1,001 2.5% Athens 2.2% 2,336 5.7% Atlanta 55.9% 11,937 29.4% Augusta 3.7% 3,720 9.2% Brunswick 1.4% 1,219 3.0% Cartersville 0.9% 845 2.1% Chattanooga 0.8% 319 0.8% Columbus 2.6% 2,443 6.0% Dalton 1.0% 972 2.4% Gainesville 2.4% 2,577 6.3% Hinesville 0.8% 567 1.4% Macon 1.7% 1,579 3.9% Rome 0.7% 855 2.1% Savannah 3.5% 3,700 9.1% Valdosta 1.2% 920 2.3% Warner Robins 1.5% 1,502 3.7% Non-MPO Table continues on next page. 18.6% 4,143 10.2% 488 Continued from previous page: Sample of vehicle trips by all persons ages 5+. Weighted Unweighted Percent (Nonmissing Sample Size, N Percent (All Observations) Observations) All vehicle trips by all persons ages 5+ 40,635 - Trip Location (Based on Origin and Destination) In Georgia 97.2% 39,348 96.8% Partly in Georgia 1.2% 641 1.6% Not in Georgia 1.6% 646 1.6% MPO Tier (Trip Origin) 1. Atlanta MPO 56.2% 12,287 30.2% 2. Medium MPOs 15.3% 14,151 34.8% 3. Small MPOs 10.8% 9,094 22.4% 4. Non-MPO 17.6% 4,143 10.2% Out of state 960 2.4% MPO (Trip Origin) Albany 1.2% 970 2.4% Athens 2.4% 2,310 5.7% Atlanta 55.0% 12,287 30.2% Augusta 3.6% 3,537 8.7% Brunswick 1.5% 1,173 2.9% Cartersville 0.8% 723 1.8% Chattanooga 0.6% 233 0.6% Columbus 2.5% 2,324 5.7% Dalton 0.9% 889 2.2% Gainesville 2.4% 2,187 5.4% Hinesville 0.8% 537 1.3% Macon 1.8% 1,626 4.0% Rome 0.7% 801 2.0% Savannah 3.5% 3,560 8.8% Valdosta 1.3% 930 2.3% Warner Robins 1.6% 1,445 3.6% Non-MPO 17.2% 4,143 10.2% Out of state 2.2% 960 2.4% 489 Table 217. Trip sample table: All vehicle trips by all persons ages 5+ by purpose. All vehicle trips by all persons ages 5+ Purpose Home Work commute Other work-related travel Attend school or daycare Transport someone Shopping or errands Medical/dental services Social/recreational or fitness Dining (restaurant or carryout) Community, religious, and volunteer Other Missing Purpose Type Mandatory Household-serving Discretionary Return home Other Missing Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N Percent (All Observations) 40,635 - 33.8% 14.9% 1.2% 1.3% 8.4% 20.5% 1.4% 7.8% 7.5% 2.7% 0.6% 13,463 5,900 530 401 2,987 9,008 801 3,123 3,113 1,060 235 14 33.1% 14.5% 1.3% 1.0% 7.4% 22.2% 2.0% 7.7% 7.7% 2.6% 0.6% 0.0% 17.4% 30.3% 17.9% 33.8% 0.6% 6,831 12,796 7,296 13,463 235 14 16.8% 31.5% 18.0% 33.1% 0.6% 0.0% 490 Table 218. Trip sample table: All vehicle trips by all persons ages 5+ by demographic factors. All vehicle trips by all persons ages 5+ Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N Percent (All Observations) 40,635 - Sex Male 49.7% 19,343 47.6% Female NHTS imputed sex for 16 people. 50.3% 21,292 52.4% Age Child 515 0.1% 29 0.1% Teen 1617 1.2% 410 1.0% Adult 1864 84.6% 29,696 73.1% Senior 6579 12.6% 9,214 22.7% Elderly 80+ 1.5% 1,286 3.2% Age by Sex Male: child 515 0.1% 13 0.0% Male: teen 1617 0.5% 182 0.4% Male: adult 1864 41.6% 13,553 33.4% Male: senior 6579 6.6% 4,902 12.1% Male: elderly 80+ 0.9% 693 1.7% Female: child 515 0.0% 16 0.0% Female: teen 1617 0.7% 228 0.6% Female: adult 1864 43.0% 16,143 39.7% Female: senior 6579 6.0% 4,312 10.6% Female: elderly 80+ 0.6% 593 1.5% Age Cohort (Adults Only) Millennial and Gen Z (1836) 30.8% 8,110 20.0% Gen X (3752) 33.4% 10,576 26.0% Pre-retirement age Boomer (5364) 21.5% 11,010 27.1% Retirement age (65+) 14.3% 10,500 25.8% {Children & teens} 439 1.1% Driver Status by Age Driver ages 18+ 98.7% 40,196 98.9% Nondriver ages 18+ 0.0% 0.0% Driver ages 1617 1.2% 410 1.0% Nondriver ages 1617 0.0% 0.0% Child ages 515 0.1% 29 0.1% NHTS does not ask about drivers' licensing; rather, they ask "do you/does this person drive?" Some children under 16 were reported as drivers, perhaps due to learner's permits. Unless otherwise stated, drivers in the report refers only to drivers ages 16+. Table continues on next page. 491 Continued from previous page: Sample of vehicle trips by all persons ages 5+. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N Percent (All Observations) All vehicle trips by all persons ages 5+ 40,635 - College-educated No 4-year degree 55.9% 20,785 51.2% Bachelor's or higher 44.1% 19,825 48.8% Missing 25 0.1% Age <14 0.0% Immigrant Nonimmigrant (born in US) 88.7% 37,682 92.7% Immigrant (born outside of US) 11.3% 2,944 7.2% Missing 9 0.0% Immigrant by Education Level US HS or less 22.1% 8,292 20.4% US some college/assoc. 28.6% 11,305 27.8% US bachelor's+ 38.0% 18,076 44.5% Imm. HS or less 2.3% 510 1.3% Imm. some college/assoc. 2.8% 678 1.7% Imm. bachelor's+ 6.1% 1,749 4.3% N/A (age <14) or missing 25 0.1% Caregiver Status Noncaregiver 62.5% 30,005 73.8% Caregiver, youngest child ages 015 37.5% 10,191 25.1% N/A (child under 18) 439 1.1% A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. Caregiver Status by Gender Male noncaregiver 32.5% 14,767 36.3% Female noncaregiver 17.3% 4,381 10.8% Male caregiver 30.0% 15,238 37.5% Female caregiver 20.2% 5,810 14.3% Missing 439 1.1% Caregiver Status by Age of Youngest Child Noncaregiver 63.0% 30,444 74.9% Youngest ages 04 17.8% 4,583 11.3% Youngest ages 515 Table continues on next page. 19.2% 5,608 13.8% 492 Continued from previous page: Sample of vehicle trips by all persons ages 5+. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N Percent (All Observations) All vehicle trips by all persons ages 5+ 40,635 - Worker Status by Age Nonworker ages 18+ 27.6% 14,863 36.6% Worker ages 18+ 71.1% 25,333 62.3% Nonworker ages 1617 0.8% 259 0.6% Worker ages 1617 0.4% 151 0.4% Child under 16 0.1% 29 0.1% Race (Categories) White non-Hispanic only 56.8% 29,034 71.5% Black and Black multiracial (incl. Black Hisp.) 31.4% 8,781 21.6% Other 11.8% 2,820 6.9% NHTS imputed race and/or Hispanic status for 74 people. See chapter 1, Assorted Definitions and Notes for more details on how race is categorized in this report. Race/Ethnicity (used in Chapter 5: Equity) White non-Hispanic only 56.8% 29,034 71.5% Black and Black multiracial (excl. Black Hisp.) 30.3% 8,650 21.3% Hispanic (any race) 8.2% 1,322 3.3% Asian or other 4.7% 1,629 4.0% Mobility Impairment Absent 95.5% 38,412 94.5% Present 4.5% 2,205 5.4% Missing 18 0.0% Annual Household Income <$15,000 9.1% 2,944 7.2% $15,000 to $24,999 8.2% 2,970 7.3% $25,000 to $34,999 10.5% 3,636 8.9% $35,000 to $49,999 13.1% 5,222 12.9% $50,000 to $74,999 18.1% 7,585 18.7% $75,000 to $99,999 13.0% 5,797 14.3% $100,000+ 28.0% 11,350 27.9% Missing 1,131 2.8% Education Level High school or less 24.4% 8,802 21.7% Some college or associate degree 31.5% 11,983 29.5% Bachelor's degree 23.8% 10,192 25.1% Graduate or professional degree 20.3% 9,633 23.7% Missing 25 0.1% Age <14 Table continues on next page. 0 0.0% 493 Continued from previous page: Sample of vehicle trips by all persons ages 5+. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N Percent (All Observations) All vehicle trips by all persons ages 5+ 40,635 - College-educated No 4-year degree 55.9% 20,785 51.2% Bachelor's or higher 44.1% 19,825 48.8% Missing 25 0.1% Age <14 0.0% Immigrant Nonimmigrant (born in US) 88.7% 37,682 92.7% Immigrant (born outside of US) 11.3% 2,944 7.2% Missing 9 0.0% Immigrant by Education Level US HS or less 22.1% 8,292 20.4% US some college/assoc. 28.6% 11,305 27.8% US bachelor's+ 38.0% 18,076 44.5% Imm. HS or less 2.3% 510 1.3% Imm. some college/assoc. 2.8% 678 1.7% Imm. bachelor's+ 6.1% 1,749 4.3% N/A (age <14) or missing 25 0.1% Caregiver Status Noncaregiver 62.5% 30,005 73.8% Caregiver, youngest child ages 015 37.5% 10,191 25.1% N/A (child under 18) 439 1.1% A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. Caregiver Status by Gender Male noncaregiver 32.5% 14,767 36.3% Female noncaregiver 17.3% 4,381 10.8% Male caregiver 30.0% 15,238 37.5% Female caregiver 20.2% 5,810 14.3% Missing 439 1.1% Caregiver Status by Age of Youngest Child Noncaregiver 63.0% 30,444 74.9% Youngest ages 04 17.8% 4,583 11.3% Youngest ages 515 Table continues on next page. 19.2% 5,608 13.8% 494 Continued from previous page: Sample of vehicle trips by all persons ages 5+. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N Percent (All Observations) All vehicle trips by all persons ages 5+ 40,635 - Vehicle Deficit Category of Household Zero-vehicle 0.4% 115 0.3% Deficit (hard or soft) 20.1% 4,888 12.0% Nondeficit (sufficient/surplus) 79.5% 35,632 87.7% A vehicle-deficit household owns at least one vehicle, but fewer vehicles than household members ages 16+ (i.e., potential drivers). See chapter 1, Vehicle Ownership for more details. Vehicle Deficit Category by Household Size Nondeficit, single potential driver 18.7% 9,386 23.1% Nondeficit, 2+ potential drivers 60.8% 26,246 64.6% Deficit 20.1% 4,888 12.0% Zero-vehicle 0.4% 115 0.3% Transit Use, Past 30 Days No 91.7% 38,348 94.4% Yes 8.3% 2,263 5.6% Missing 24 0.1% Walking, Past 30 Days No 25.6% 10,058 24.8% Yes 74.4% 30,458 75.0% Missing 119 0.3% Biking, Past 30 Days No 94.2% 38,158 93.9% Yes 5.8% 2,471 6.1% Missing . 6 0.0% 495 Table 219. Trip sample table: All trips by all persons ages 18+ by location. All trips by adults ages 18+ Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N 53,203 Percent (All Observations) - MPO Tier (Residence of Traveler) 1. Atlanta MPO counties 55.8% 15,850 29.8% 2. Medium MPO counties 15.6% 19,661 37.0% 3. Small MPO counties 10.1% 12,227 23.0% 4. Non-MPO counties 18.5% 5,465 10.3% Addresses are classified by whether the county is part of an MPO. Medium MPOs have an MPO population of 200,0011,000,000 people; small MPOs have a population of 200,000 or less. See chapter 1, Geography Divisions for Analysis for more details. MPO (Residence of Traveler) Albany 1.1% 1,285 2.4% Athens 2.4% 3,292 6.2% Atlanta 55.8% 15,850 29.8% Augusta 3.7% 4,747 8.9% Brunswick 1.3% 1,606 3.0% Cartersville 0.9% 1,059 2.0% Chattanooga 0.8% 375 0.7% Columbus 2.5% 3,031 5.7% Dalton 1.0% 1,250 2.3% Gainesville 2.4% 3,291 6.2% Hinesville 0.7% 679 1.3% Macon 1.7% 2,070 3.9% Rome 0.8% 1,135 2.1% Savannah 3.8% 4,925 9.3% Valdosta 1.1% 1,202 2.3% Warner Robins 1.5% 1,941 3.6% Non-MPO Table continues on next page. 18.5% 5,465 10.3% 496 Continued from previous page: Sample of trips by all adults ages 18+. Weighted Percent (Nonmissing Observations) All trips by adults ages 18+ Trip Location (Based on Origin and Destination) In Georgia 95.7% Partly in Georgia 1.6% Not in Georgia 2.8% MPO Tier (Trip Origin) 1. Atlanta MPO 55.9% 2. Medium MPOs 15.8% 3. Small MPOs 10.5% 4. Non-MPO 17.8% Out of state MPO (Trip Origin) Albany 2.7% Athens 53.9% Atlanta 3.6% Augusta 1.3% Brunswick 0.8% Cartersville 0.6% Chattanooga 2.4% Columbus 1.0% Dalton 2.2% Gainesville 0.6% Hinesville 1.8% Macon 0.7% Rome 3.7% Savannah 1.2% Valdosta 1.5% Warner Robins 17.2% Non-MPO 3.6% Out of state Unweighted Sample Size, N 53,203 Percent (All Observations) - 50,724 930 1,549 95.3% 1.7% 2.9% 15,992 18,171 11,532 5,507 2,001 30.1% 34.2% 21.7% 10.4% 3.8% 1,229 3,175 15,992 4,432 1,493 920 280 2,862 1,121 2,739 633 2,086 1,044 4,683 1,188 1,818 5,507 2,001 2.3% 6.0% 30.1% 8.3% 2.8% 1.7% 0.5% 5.4% 2.1% 5.1% 1.2% 3.9% 2.0% 8.8% 2.2% 3.4% 10.4% 3.8% 497 Table 220. Trip sample table: All trips by all persons ages 18+ by purpose. All trips by adults ages 18+ Purpose Home Work commute Other work-related travel Attend school or daycare Transport someone Shopping or errands Medical/dental services Social/recreational or fitness Dining (restaurant or carryout) Community, religious, and volunteer Other Missing Purpose Type Mandatory Household-serving Discretionary Return home Other Missing Weighted Percent (Nonmissing Observations) 32.8% 13.0% 1.2% 1.4% 7.1% 19.9% 1.6% 10.8% 8.2% 3.0% 1.1% 15.5% 28.6% 22.0% 32.8% 1.1% Unweighted Sample Size, N 53,203 Percent (All Observations) - 17,050 6,715 693 455 3,371 11,465 1,130 5,723 4,541 1,480 558 22 32.0% 12.6% 1.3% 0.9% 6.3% 21.5% 2.1% 10.8% 8.5% 2.8% 1.0% 0.0% 7,863 15,966 11,744 17,050 558 22 14.8% 30.0% 22.1% 32.0% 1.0% 0.0% 498 Table 221. Trip sample table: All trips by all persons ages 18+ by mode. Weighted Unweighted Percent (Nonmissing Sample Size, N Percent (All Observations) Observations) All trips by adults ages 18+ 53,203 - Mode Pedestrian (walk/wheelchair) 7.9% 3,710 7.0% Bike 0.6% 225 0.4% POV, including rental car 88.0% 47,929 90.1% School bus 0.4% 111 0.2% Public transit 1.6% 495 0.9% Paratransit 0.1% 53 0.1% Other bus 0.2% 121 0.2% Taxi/ridehail/limo 0.6% 203 0.4% Air 0.2% 105 0.2% Other 0.4% 249 0.5% Missing 2 0.0% Mode Category POV, including rental car 88.0% 47,929 90.1% Nonmotorized (walk, bike, wheelchair) 8.4% 3,937 7.4% Public transit or other bus/train 2.2% 727 1.4% Other ground or water 1.2% 503 0.9% Air 0.2% 105 0.2% Missing 2 0.0% Vehicle or Person Trip Person trip only 27.8% 13,007 24.4% Vehicle trip and person trip 72.2% 40,196 75.6% See chapter 1, Key Terms for more explanation of the relationship between person trips and vehicle trips. 499 Table 222. Trip sample table: All trips by all persons ages 18+ by demographic factors. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N Percent (All Observations) All trips by adults ages 18+ 53,203 - Sex Male 47.6% 23,982 45.1% Female 52.4% 29,221 54.9% NHTS imputed sex for 16 people. Age Child 515 Teen 1617 Adult 1864 85.4% 38,862 73.0% Senior 6579 12.6% 12,333 23.2% Elderly 80+ 1.9% 2,008 3.8% Age by Sex Male: adult 1864 40.7% 17,067 32.1% Male: senior 6579 6.1% 5,958 11.2% Male: elderly 80+ 0.8% 957 1.8% Female: adult 1864 44.7% 21,795 41.0% Female: senior 6579 6.5% 6,375 12.0% Female: elderly 80+ 1.1% 1,051 2.0% Age Cohort (Adults Only) Millennial and Gen Z (1836) 32.8% 11,175 21.0% Gen X (3752) 31.3% 13,382 25.2% Pre-retirement age Boomer (5364) 21.3% 14,305 26.9% Retirement age (65+) 14.6% 14,341 27.0% Driver Status by Age Driver ages 18+ 93.1% 50,924 95.7% Nondriver ages 18+ 6.9% 2,279 4.3% NHTS does not ask about drivers' licensing; rather, they ask "do you/does this person drive?" Some children under 16 were reported as drivers, perhaps due to learner's permits. Unless otherwise stated, drivers in the report refers only to drivers ages 16+. Table continues on next page. 500 Continued from previous page: Sample of trips by all adults ages 18+. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N Percent (All Observations) All trips by adults ages 18+ 53,203 - Worker Status by Age Nonworker ages 18+ 32.9% 21,847 41.1% Worker ages 18+ 67.1% 31,356 58.9% Race (Categories) White non-Hispanic only 55.8% 37,628 70.7% Black and Black multiracial (incl. Black Hisp.) 31.7% 11,625 21.9% Other 12.5% 3,950 7.4% NHTS imputed race and/or Hispanic status for 74 people. See chapter 1, Assorted Definitions and Notes for more details on how race is categorized in this report. Race/Ethnicity (used in chapter 5, Equity) White non-Hispanic only 55.8% 37,628 70.7% Black and Black multiracial (excl. Black Hisp.) 30.8% 11,474 21.6% Hispanic (any race) 8.6% 1,822 3.4% Asian or other 4.9% 2,279 4.3% Mobility Impairment Absent 93.4% 49,257 92.6% Present 6.6% 3,923 7.4% Missing 23 Annual Household Income <$15,000 12.7% 4,988 9.4% $15,000 to $24,999 8.5% 3,990 7.5% $25,000 to $34,999 10.1% 4,739 8.9% $35,000 to $49,999 12.7% 6,597 12.4% $50,000 to $74,999 16.6% 9,595 18.0% $75,000 to $99,999 12.5% 7,340 13.8% $100,000+ 26.9% 14,467 27.2% Missing 1,487 2.8% Education Level High school or less 27.0% 12,295 23.1% Some college or associate degree 30.1% 15,283 28.7% Bachelor's degree 23.2% 12,999 24.4% Graduate or professional degree 19.7% 12,566 23.6% Missing Table continues on next page. 60 0.1% 501 Continued from previous page: Sample of trips by all adults ages 18+. Weighted Percent (Nonmissing Observations) All trips by adults ages 18+ College-educated No 4-year degree 57.1% Unweighted Sample Size, N 53,203 Percent (All Observations) - 27,578 51.8% Bachelor's or higher Immigrant Nonimmigrant (born in US) 42.9% 88.6% 25,565 49,199 48.1% 92.5% Immigrant (born outside of US) Missing Immigrant by Education Level US HS or less US some college/assoc. US bachelor's+ 11.4% 24.2% 27.6% 36.8% 3,993 11 11,535 14,407 23,223 7.5% 0.0% 21.7% 27.1% 43.6% Imm. HS or less Imm. some college/assoc. Imm. bachelor's+ 2.8% 2.5% 6.1% 760 876 2,342 1.4% 1.6% 4.4% N/A (age <14) or missing Caregiver Status Noncaregiver 65.4% 60 40,602 0.1% 76.3% Caregiver, youngest child ages 015 34.6% 12,601 23.7% A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. Caregiver Status by Gender Male noncaregiver Female noncaregiver 32.3% 15.3% 18,754 5,228 35.2% 9.8% Male caregiver Female caregiver Missing 33.1% 19.3% 21,848 7,373 41.1% 13.9% Caregiver Status by Age of Youngest Child Noncaregiver Youngest ages 04 65.4% 16.6% 40,602 5,687 76.3% 10.7% Youngest ages 515 Table continues on next page. 18.0% 6,914 13.0% 502 Continued from previous page: Sample of trips by all adults ages 18+. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N Percent (All Observations) All trips by adults ages 18+ 53,203 - Vehicle Deficit Category of Household Zero-vehicle 4.1% 1,603 3.0% Deficit (hard or soft) 23.4% 7,517 14.1% Nondeficit (sufficient/surplus) 72.5% 44,083 82.9% A vehicle-deficit household owns at least one vehicle, but fewer vehicles than household members ages 16+ (i.e., potential drivers). See chapter 1, Vehicle Ownership for more details. Vehicle Deficit Category by Household Size Nondeficit, single potential driver 16.2% 10,870 20.4% Nondeficit, 2+ potential drivers 56.3% 33,213 62.4% Deficit 23.4% 7,517 14.1% Zero-vehicle 4.1% 1,603 3.0% Transit Use, Past 30 Days No 87.5% 48,610 91.4% Yes 12.5% 4,549 8.6% Missing 44 0.1% Walking, Past 30 Days No 23.4% 12,378 23.3% Yes 76.6% 40,645 76.4% Missing 180 0.3% Biking, Past 30 Days No 93.3% 49,577 93.2% Yes 6.7% 3,605 6.8% Missing 21 0.0% 503 Table 223. Trip sample table: All vehicle trips by all persons ages 18+ by location. All vehicle trips by adults ages 18+ Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N 40,196 Percent (All Observations) MPO Tier (Residence of Traveler) 1. Atlanta MPO counties 55.8% 11,790 29.3% 2. Medium MPO counties 15.3% 14,926 37.1% 3. Small MPO counties 10.3% 9,358 23.3% 4. Non-MPO counties 18.7% 4,122 10.3% Addresses are classified by whether the county is part of an MPO. Medium MPOs have an MPO population of 200,0011,000,000 people; small MPOs have a population of 200,000 or less. See chapter 1, Geographic Divisions for Analysis for more details. MPO (Residence of Traveler) Albany 1.1% 989 2.5% Athens 2.2% 2,290 5.7% Atlanta 55.8% 11,790 29.3% Augusta 3.8% 3,692 9.2% Brunswick 1.3% 1,204 3.0% Cartersville 0.9% 841 2.1% Chattanooga 0.8% 311 0.8% Columbus 2.6% 2,430 6.0% Dalton 1.0% 959 2.4% Gainesville 2.4% 2528 6.3% Hinesville 0.8% 567 1.4% Macon 1.7% 1,554 3.9% Rome 0.7% 846 2.1% Savannah 3.5% 3,675 9.1% Valdosta 1.2% 907 2.3% Warner Robins 1.5% 1,491 3.7% Non-MPO 18.7% 4,122 10.3% Table continues on next page. 504 Continued from previous page: Sample of vehicle trips by all adults ages 18+ All vehicle trips by adults ages 18+ Trip Location (Based on Origin and Destination) In Georgia 97.2% Partly in Georgia 1.2% Not in Georgia 1.6% MPO Tier (Trip Origin) 1. Atlanta MPO 56.1% 2. Medium MPOs 15.4% 3. Small MPOs 10.8% 4. Non-MPO 17.7% Out of state MPO (Trip Origin) Albany 1.2% Athens 2.4% Atlanta 54.9% Augusta 3.6% Brunswick 1.4% Cartersville 0.8% Chattanooga 0.6% Columbus 2.5% Dalton 0.9% Gainesville 2.3% Hinesville 0.8% Macon 1.8% Rome 0.7% Savannah 3.5% Valdosta 1.3% Warner Robins 1.7% Non-MPO 17.3% Out of state 2.2% 40,196 38,920 634 642 12,140 13,988 8,994 4,120 954 958 2,264 12,140 3,508 1,158 720 228 2,311 877 2,142 537 1,600 791 3,535 917 1,436 4,120 954 - 96.8% 1.6% 1.6% 30.2% 34.8% 22.4% 10.2% 2.4% 2.4% 5.6% 30.2% 8.7% 2.9% 1.8% 0.6% 5.7% 2.2% 5.3% 1.3% 4.0% 2.0% 8.8% 2.3% 3.6% 10.2% 2.4% 505 Table 224. Trip sample table: All vehicle trips by all persons ages 18+ by purpose. All vehicle trips by adults ages 18+ Purpose Home Work commute Other work-related travel Attend school or daycare Transport someone Shopping or errands Medical/dental services Social/recreational or fitness Dining (restaurant or carryout) Community, religious, and volunteer Other Missing Purpose Type Mandatory Household-serving Discretionary Return home Other Missing Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N 40,196 Percent (All Observations) - 33.7% 15.0% 1.3% 1.1% 8.4% 20.6% 1.5% 7.7% 7.4% 2.7% 0.6% 13,291 5,868 528 311 2,967 8,969 801 3,081 3,082 1,052 232 14 33.1% 14.6% 1.3% 0.8% 7.4% 22.3% 2.0% 7.7% 7.7% 2.6% 0.6% 0.0% 17.4% 30.5% 17.9% 33.7% 0.6% 6,707 12,737 7,215 13,291 232 14 16.7% 31.7% 17.9% 33.1% 0.6% 0.0% 506 Table 225. Trip sample table: All vehicle trips by all persons ages 18+ by demographic factors. Weighted Percent (Nonmissing Observations) Unweighted Sample Size, N Percent (All Observations) All vehicle trips by adults ages 18+ 40,196 - Sex Male 49.8% 19,148 47.6% Female NHTS imputed sex for 16 people. 50.2% 21,048 52.4% Age Child 515 0.0% Teen 1617 0.0% Adult 1864 85.7% 29,696 73.9% Senior 6579 12.8% 9,214 22.9% Elderly 80+ 1.5% 1,286 3.2% Age by Sex Male: adult 1864 42.2% 13,553 33.7% Male: senior 6579 6.7% 4,902 12.2% Male: elderly 80+ 0.9% 693 1.7% Female: adult 1864 43.5% 16,143 40.2% Female: senior 6579 6.1% 4,312 10.7% Female: elderly 80+ 0.6% 593 1.5% Age Cohort (Adults Only) Millennial and Gen Z (1836) 30.8% 8,110 20.2% Gen X (3752) 33.4% 10,576 26.3% Pre-retirement age Boomer (5364) 21.5% 11,010 27.4% Retirement age (65+) 14.3% 10,500 26.1% Driver Status by Age Driver ages 18+ Nondriver ages 18+ NHTS does not ask about drivers' licensing; rather, they ask "do you/does this person drive?" Some children under 16 were reported as drivers, perhaps due to learner's permits. Unless otherwise stated, drivers in the report refers only to drivers ages 16+. Table continues on next page. 507 Continued from previous page: Sample of weekday trips by all adults ages 5+. Weighted Unweighted Percent Sample Size, Percent (Nonmissing N (All Observations) Observations) All vehicle trips by adults ages 18+ 40,196 - Worker Status by Age Nonworker ages 18+ 28.0% 14,863 37.0% Worker ages 18+ 72.0% 25,333 63.0% Race (Categories) White non-Hispanic only 56.5% 28,685 71.4% Black and Black multiracial (incl. Black Hisp.) 31.6% 8,716 21.7% Other 11.9% 2,795 7.0% NHTS imputed race and/or Hispanic status for 74 people. See chapter 1, Assorted Definitions and Notes for more details on how race is categorized in this report. Race/Ethnicity (used in chapter 5, Equity) White non-Hispanic only 56.5% 28,685 71.4% Black and Black multiracial (excl. Black Hisp.) 30.5% 8,591 21.4% Hispanic (any race) 8.2% 1,301 3.2% Asian or other 4.7% 1,619 4.0% Mobility Impairment Absent 95.5% 37,973 94.5% Present 4.5% 2,205 5.5% Missing 18 0.0% Annual Household Income <$15,000 9.1% 2,919 7.3% $15,000 to $24,999 8.3% 2,968 7.4% $25,000 to $34,999 10.5% 3,614 9.0% $35,000 to $49,999 13.2% 5,183 12.9% $50,000 to $74,999 18.2% 7,514 18.7% $75,000 to $99,999 13.1% 5,717 14.2% $100,000+ 27.7% 11,158 27.8% Missing 1,123 2.8% Education Level High school or less 23.5% 8,370 20.8% Some college or associate degree 31.9% 11,979 29.8% Bachelor's degree 24.1% 10,192 25.4% Graduate or professional degree 20.6% 9,631 24.0% Missing 24 0.1% Table continues on next page. 508 Continued from previous page: Sample of weekday trips by all adults ages 5+. Weighted Unweighted Percent Sample Size, Percent (Nonmissing N (All Observations) Observations) All vehicle trips by adults ages 18+ 40,196 - College-educated No 4-year degree 55.4% 20,349 50.6% Bachelor's or higher Immigrant Nonimmigrant (born in US) 44.6% 88.6% 19,823 37,275 49.3% 92.7% Immigrant (born outside of US) Missing Immigrant by Education Level US HS or less US some college/assoc. US bachelor's+ Imm. HS or less Imm. some college/assoc. Imm. bachelor's+ N/A (age <14) or missing Caregiver Status Noncaregiver 11.4% 21.2% 29.0% 38.5% 2.3% 2.9% 6.1% 62.5% 2,912 9 7,892 11,301 18,074 478 678 1,749 24 30,005 7.2% 0.0% 19.6% 28.1% 45.0% 1.2% 1.7% 4.4% 0.1% 74.6% Caregiver, youngest child ages 015 37.5% 10,191 25.4% A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. Caregiver Status by Gender Male noncaregiver Female noncaregiver Male caregiver Female caregiver Missing Caregiver Status by Age of Youngest Child Noncaregiver Youngest ages 04 Youngest ages 515 Table continues on next page. 32.5% 17.3% 30.0% 20.2% 62.5% 18.0% 19.5% 14,767 4,381 15,238 5,810 36.7% 10.9% 37.9% 14.5% 30,005 4,583 5,608 74.6% 11.4% 14.0% 509 Continued from previous page: Sample of weekday trips by all adults ages 5+. Weighted Unweighted Percent Sample Size, Percent (Nonmissing N (All Observations) Observations) All vehicle trips by adults ages 18+ 40,196 - Vehicle Deficit Category of Household Zero-vehicle 0.4% 112 0.3% Deficit (hard or soft) 20.1% 4,817 12.0% Nondeficit (sufficient/surplus) 79.5% 35,267 87.7% A vehicle-deficit household owns at least one vehicle, but fewer vehicles than household members ages 16+ (i.e., potential drivers). See chapter 1, Vehicle Ownership for more details. Vehicle Deficit Category by Household Size Nondeficit, single potential driver 18.9% 9,383 23.3% Nondeficit, 2+ potential drivers 60.5% 25,884 64.4% Deficit 20.1% 4,817 12.0% Zero-vehicle 0.4% 112 0.3% Transit Use, Past 30 Days No 91.7% 37,940 94.4% Yes 8.3% 2,232 5.6% Missing 24 0.1% Walking, Past 30 Days No 25.5% 9,894 24.6% Yes 74.5% 30,183 75.1% Missing 119 0.3% Biking, Past 30 Days No 94.3% 37,744 93.9% Yes 5.7% 2,446 6.1% Missing 6 0.0% 510 Table 226. Work journey/commute sample table: Location, distance, and mode. Weighted Unweighted Percent (Nonmissing Sample Size, N Percent (All All work journeys by adults ages 18+ Observations) Observations) 10,490 MPO Tier (Residence of Traveler) 1. Atlanta MPO counties 56.9% 3,334 31.8% 2. Medium MPO counties 15.2% 3,848 36.7% 3. Small MPO counties 9.9% 2,330 22.2% 4. Non-MPO counties 18.0% 978 9.3% MPO (Residence of Traveler) Albany 1.2% 274 2.6% Athens 2.1% 605 5.8% Atlanta 56.9% 3,334 31.8% Augusta 3.9% 990 9.4% Brunswick 1.1% 242 2.3% Cartersville 0.8% 195 1.9% Chattanooga 0.9% 88 0.8% Columbus 2.4% 620 5.9% Dalton 1.1% 243 2.3% Gainesville 1.7% 523 5.0% Hinesville 1.0% 186 1.8% Macon 1.9% 429 4.1% Rome 0.5% 137 1.3% Savannah 4.1% 1,022 9.7% Valdosta 1.0% 244 2.3% Warner Robins 1.2% 380 3.6% Non-MPO 18.0% 978 9.3% WJ Type Simple (no stops) 75.8% 7,872 75.0% Complex (one or more stops) 24.2% 2,618 25.0% Supercommute (WJ > 100 mi) Not a supercommute (WJ PMT <100 mi) 99.2% 10,378 98.9% Supercommute (WJ >100 mi) 0.8% 85 0.8% Missing 27 0.3% Mode POV 92.8% 9,994 95.3% Nonmotorized (walk, bike wheelchair) 2.3% 171 1.6% Public transit or other bus/train 2.6% 149 1.4% Other ground or water 0.9% 50 0.5% Air or air multimodal 0.1% 6 0.1% Multimodal with POV 0.8% 98 0.9% Multimodal, no POV 0.5% 20 0.2% Missing 2 0.0% Note: Work journeys and commutes are the same in terms of sample size; the two differ in terms of distance and duration. See chapter 2 for more details. 511 Table 227. Work journey/commute sample table: Stops, driving, and time of day. Weighted Unweighted All work journeys by adults ages 18+ Percent (Nonmissing Observations) Sample Size, N 10,490 Percent (All Observations) WJ Type Simple (no stops) 75.8% 7,872 75.0% Complex (one or more stops) 24.2% 2,618 25.0% Number and Duration of WJ Stops (Complex WJ Only) 0.0% Direct (no stops) 75.8% 7,872 75.0% Single short stop (<30 min) 13.3% 1,345 12.8% Single long stop (30+ min) 5.0% 616 5.9% Multiple short 2.5% 265 2.5% Short + long or multiple long 3.5% 392 3.7% Purpose of Stop(s) (Complex WJ Only) 0.0% Shopping 10.7% 1,162 11.1% Transport others 6.9% 673 6.4% Dining 5.0% 566 5.4% Social, recreational or fitness 2.6% 324 3.1% Other work 0.7% 110 1.0% Other nonwork 3.3% 329 3.1% Note: Categories are not mutually exclusive. WJ may contain stops for more than one purpose. Drive Alone for Entire POV No 31.5% 2,723 26.0% Yes 68.5% 7,767 74.0% Driver Status for WJ Drive alone all POV legs 69.1% 7,833 74.7% Family sharing: drive with household passenger for 1+ legs 6.7% 648 6.2% Carpool driver: drive with nonhousehold passenger for 1+ legs 9.1% 963 9.2% Both carpool and family sharing 1.7% 157 1.5% Drive alone + ride as pax 0.2% 33 0.3% All POV legs as passenger 6.7% 458 4.4% No POV legs 6.4% 398 3.8% Peak Category AM peak (6am9:59am) 35.2% 3,873 36.9% Midday (10am2:59 pm) 15.8% 1,589 15.1% PM peak (3pm6:59 pm) 35.1% 3,846 36.7% Overnight (7pm6:59am) 13.9% 1,182 11.3% Note: Work journeys and commutes are the same in terms of sample size; the two differ in terms of distance and duration. See chapter 2 for more details. 512 Table 228. Work journey/commute sample table: Traveler characteristics. Weighted Unweighted All work journeys by adults ages 18+ Percent (Nonmissing Observations) Sample Size, Percent N (All Observations) 10,490 Worker Status of Commuter Not an NHTS-defined worker 1.2% 150 1.4% NHTS-defined worker 98.8% 10,340 98.6% NHTS defines a worker as someone who worked for pay or profit, or was temporarily absent from paid employment, in the week before completing the travel survey ("last week"). A small number of people who were not NHTS-designated workers nevertheless reported work travel on their travel day, perhaps reflecting irregular employment situations. Occupational Category Sales or service 27.2% 2,365 22.5% Clerical or administrative support 9.0% 1,045 10.0% Blue collar* 19.8% 1,603 15.3% Professional, managerial, or technical 43.9% 5,260 50.1% Other 0.2% 11 0.1% Missing 206 2.0% * Blue collar refers to manufacturing, construction, maintenance, or farming. Worker Type Full-time 16.4% 1,608 15.3% Part-time 83.6% 8,830 84.2% Missing 52 0.5% Sex Male 56.3% 5,437 51.8% Female 43.7% 5,053 48.2% Age Cohort (Adults Only) Millennial and Gen Z (1836) 40.5% 3,145 30.0% Gen X (3752) 35.3% 3,543 33.8% Pre-retirement age Boomer (5364) 19.9% 2,992 28.5% Retirement age (65+) 4.3% 810 7.7% Driver Status Nondriver 5.6% 291 2.8% Driver 94.4% 10,199 97.2% Race (categories) White non-Hispanic only 53.4% 7,199 68.6% Black and black multiracial (incl. black Hispanic) 32.3% 2,394 22.8% Other 14.3% 897 8.6% See chapter 1, Assorted Definitions and Notes for more details on how race is categorized in this report. Table continues on next page. 513 Continued from previous page: Sample of work journeys by adults ages 18+. Weighted Unweighted Percent Sample Size, Percent (Nonmissing N (All Observations) Observations) All work journeys by adults ages 18+ 10,490 Mobility impairment Absent 98.7% 10,336 98.5% Present 1.3% 150 1.4% Missing 4 0.0% A mobility impairment is defined as a "condition or handicap that makes it difficult to travel outside of the home." Annual Household Income <$35,000 28.0% 2,148 20.5% $35,000 to $49,999 13.2% 1,345 12.8% $50,000 to $74,999 18.0% 1,965 18.7% $75,000 to $99,999 14.4% 1,657 15.8% $100,000+ 26.3% 3,161 30.1% Missing 214 2.0% Education Level High school or less 27.6% 2,226 21.2% Some college or associate degree 31.4% 3,152 30.0% Bachelor's degree 23.3% 2,641 25.2% Graduate or professional degree 17.7% 2,460 23.5% Missing 11 0.1% College-educated No 4-year degree 59.0% 5,378 51.3% Bachelor's or higher 41.0% 5,101 48.6% Missing 11 0.1% Caregiver Status Noncaregiver 63.8% 7,520 71.7% Caregiver, youngest child ages 015 36.2% 2,970 28.3% A caregiver is defined as any adult age 18+ in a household with a child of less than 5 years old, and any adult age 22+ in a household with a child of 515 years old. Caregiver Status by Gender Male noncaregiver 36.3% 3,850 36.7% Female noncaregiver 20.0% 1,587 15.1% Male caregiver 27.5% 3,670 35.0% Female caregiver 16.2% 1,383 13.2% Caregiver Status by Age of Youngest Child Noncaregiver 63.8% 7,520 71.7% Youngest ages 04 16.9% 1,308 12.5% Youngest ages 515 19.3% 1,662 15.8% Vehicle Deficit Category by Household Size Nondeficit, single potential driver 14.7% 1,905 18.2% Nondeficit, 2+ potential drivers 60.5% 7,098 67.7% Deficit 22.3% 1,314 12.5% Zero-vehicle 2.6% 173 1.6% Note: Work journeys and commutes are the same in terms of sample size; the two differ in terms of distance and duration. See chapter 2 for more details. A vehicle-deficit household owns at least one vehicle, but fewer vehicles than household members ages 16+ (i.e., potential drivers). See chapter 1, Vehicle Ownership for more details. 514 ACKNOWLEDGMENTS The authors thank Stacey Bricka, Ali Etezady, Anthony Fucci, Laurie Garrow, Habte Kassa, Sung Hoo Kim, Yongsung Lee, Nancy McGuckin, Atiyya Shaw, and Xinyi Wang. 515 REFERENCES Abou-Zeid, M., Witter, R., Bierlaire, M., Kaufmann, V., and Ben-Akiva, M. (2012). "Happiness and Travel Mode Switching: Findings from a Swiss Public Transportation Experiment." Transport Policy, 19(1), pp. 93104. Available online: https://doi.org/10.1016/j.tranpol.2011.09.009. Bezyak, J.L., Sabella, S., Hammel, J., McDonald, K., Jones, R.A., and Barton, D. (2019). "Community Participation and Public Transportation Barriers Experienced by People with Disabilities." Disability and Rehabilitation, 42(23), pp. 19. Available online: https://doi.org/10.1080/09638288.2019.1590469. Blumenberg, E. (2016). "Why Low-Income Women in the US Still Need Automobiles." Town Planning Review, 87(5), pp. 525545. Available online: https://doi.org/10.3828/tpr.2016.34. Blumenberg, E., Brown, A., and Schouten, A. (2018). Auto-Deficit Households: Determinants, Travel Behavior, and the Gender Division of Household Car Use. Final Report. UCConnect, The University of California Center for Economic Competetiveness in Transportation, Los Angeles, CA. Available online: http://ucconnect.berkeley.edu/auto-deficit-households-determinantstravel-behavior-and-gender-division-household-car-use. Blumenberg, E. and Pierce, G. (2014). "A Driving Factor in Mobility? Transportation's Role in Connecting Subsidized Housing and Employment Outcomes in the Moving to Opportunity (MTO) Program." Journal of the American Planning Association, 80(1), pp. 5266. Available online: https://doi.org/10.1080/01944363.2014.935267. Brown, A.E. (2017). "Car-Less or Car-Free? Socioeconomic and Mobility Differences Among Zero-Car Households." Transport Policy, 60(November), pp. 152159. Available online: https://doi.org/10.1016/j.tranpol.2017.09.016. Caplovitz, D. (1967). The Poor Pay More: Consumer Practices of Low-Income Families. Free Press; Collier-Macmillan, New York; London. Available online: http://books.google.com/books?id=zr9AAAAAIAAJ. Cappellari, L. and Jenkins, S.P. (2003). "Multivariate Probit Regression Using Simulated Maximum Likelihood." The Stata Journal, 3(3), pp. 278294. Choo, S. and Mokhtarian, P.L. (2008). "How Do People Respond to Congestion Mitigation Policies? A Multivariate Probit Model of the Individual Consideration of Three Travel-Related Strategy Bundles." Transportation, 35(2), pp. 145163. Available online: https://doi.org/10.1007/s11116-007-9142-8. Claritas. (2018). Assessing the Role of Urbanicity. Claritas, LLC. Available online: https://nhts.ornl.gov/assets/Assessing_the_Role_of_Urbanicity.pdf. 516 Clark, S.K., Jeglic, E.L., Calkins, C., and Tatar, J.R. (2016). "More Than a Nuisance: The Prevalence and Consequences of Frotteurism and Exhibitionism." Sexual Abuse: A Journal of Research and Treatment, 28(1), pp. 319. Available online: https://doi.org/10.1177/1079063214525643. Concas, S. and Winters, P.L. (2007). "Impact of Carpooling on Trip-Chaining Behavior and Emission Reductions." Transportation Research Record: Journal of the Transportation Research Board, 2010(1), pp. 8391. Available online: https://doi.org/10.3141/2010-10. Conway, M., Salon, D., and King, D. (2018). "Trends in Taxi Use and the Advent of Ridehailing, 19952017: Evidence from the US National Household Travel Survey." Urban Science, 2(3), p. 79. Available online: https://doi.org/10.3390/urbansci2030079. Craig, L. and van Tienoven, T.P. (2019). "Gender, Mobility and Parental Shares of Daily Travel with and for Children: A Cross-National Time Use Comparison." Journal of Transport Geography, 76(April), pp. 93102. Available online: https://doi.org/10.1016/j.jtrangeo.2019.03.006. Crane, R. (2007). "Is There a Quiet Revolution in Women's Travel? Revisiting the Gender Gap in Commuting." Journal of the American Planning Association, 73(3), pp. 298316. Available online: https://doi.org/10.1080/01944360708977979. Criado-Perez, C. (2019). Invisible Women: Data Bias in a World Designed for Men. Abrams Press, New York. Decker, A. (2006). "Stuck at Home: When Driving Isn't a Choice." Access Magazine, 29(Fall), pp. 2833. Research at the University of California Transportation Center. Available online: https://www.accessmagazine.org/fall-2006/stuck-home-driving-isnt-choice/. Emond, C.R., Tang, W., and Handy, S.L. (2009). "Explaining Gender Difference in Bicycling Behavior." Transportation Research Record: Journal of the Transportation Research Board, 2125(1), pp. 1625. Available online: https://doi.org/10.3141/2125-03. Ettema, D., Grling, T., Ericksson, L., Friman, M., Olsson, L.E., and Fujii, S. (2011). "Satisfaction with Travel and Subjective Well-Being: Development and Test of a Measurement Tool." Transportation Research Part F: Traffic Psychology and Behaviour, 14(3), pp. 16775. Available online: https://doi.org/10.1016/j.trf.2010.11.002. Fainstein, S.S. and Servon, L.J. (2005). "Introduction: The Intersection of Gender and Planning." In Gender and Planning: A Reader, eds. Fainstein, S.S. and Servon, L.J., pp. 112, Rutgers University Press, New Brunswick, NJ. Frei, C., Hyland, M., and Mahmassani, H.S. (2017). "Flexing Service Schedules: Assessing the Potential for Demand-Adaptive Hybrid Transit via a Stated Preference Approach." Transportation Research Part C: Emerging Technologies, 76(March), pp. 7189. Available online: https://doi.org/10.1016/j.trc.2016.12.017. 517 Frei, C., Mahmassani, H.S., and Frei, A. (2015). "Making Time Count: Traveler Activity Engagement on Urban Transit." Transportation Research Part A: Policy and Practice, 76(June), pp. 5870. Available online: https://doi.org/10.1016/j.tra.2014.12.007. Friman, M., Fujii, S., Ettema, D.F., Grling, T., and Olsson, L.E. (2012). "Psychometric Analysis of the Satisfaction with Travel Scale." Transportation Research. Part A, Policy and Practice, 48, pp. 132145. Available online: https://www.researchgate.net/profile/DickEttema/publication/257640475_Rules_for_aggregated_satisfaction_with_work_commutes/links/ 55641afe08ae6f4dcc98c0e9/Rules-for-aggregated-satisfaction-with-work-commutes.pdf. Garrow, L., Douthat, T., Yang, W., Nord, A., Rao, P., and Douglass, S. (2018). Rural and Small Urban Transit Systems in Georgia. Technical Report, Georgia Institute of Technology, School of Civil and Environmental Engineering, Atlanta, GA. Available online: http://garrowlab.ce.gatech.edu/sites/default/files/201812%20Rural%20and%20Small%20Urban %20Transit%20in%20GA.pdf. Gimenez-Nadal, J.I. and Molina, J.A. (2016). "Commuting Time and Household Responsibilities: Evidence Using Propensity Score Matching." Journal of Regional Science, 56(2), pp. 332359. Available online: https://doi.org/10.1111/jors.12243. Gripsrud, M. and Hjorthol, R. (2012). "Working on the Train: From `Dead Time' to Productive and Vital Time." Transportation, 39(5), pp. 941956. Available online: https://doi.org/10.1007/s11116-012-9396-7. Hanson, S. (2010). "Gender and Mobility: New Approaches for Informing Sustainability." Gender, Place & Culture, 17(1), pp. 523. Available online: https://doi.org/10.1080/09663690903498225. Jan, T. (2017). "White Families Have Nearly 10 Times the Net Worth of Black Families. And the Gap Is Growing." The Washington Post, September 28. Available online: https://www.washingtonpost.com/news/wonk/wp/2017/09/28/black-and-hispanic-families-aremaking-more-money-but-they-still-lag-far-behind-whites/. Jun, M.-J. and Kwon, K.-H. (2015). "Why Dual-Earner Households in Seoul Live Closer to the Wife's Workplace than the Husband's?" Urban Policy and Research, 33(2), pp. 217232. Available online: https://doi.org/10.1080/08111146.2014.994739. Karger, H.J. (2003). "No Deals on Wheels: How and Why the Poor Pay More for Basic Transportation." Journal of Poverty, 7(4), pp. 93112. Available online: https://doi.org/10.1300/J134v07n04_06. Kash, G. (2019). "Always on the Defensive: The Effects of Transit Sexual Assault on Travel Behavior and Experience in Colombia and Bolivia." Journal of Transport & Health, 13(June), pp. 23446. Available online: https://doi.org/10.1016/j.jth.2019.04.004. Kass, A. (2016). "Fulton County Approves Uber Rides for Seniors." Atlanta Journal Constitution, Aug 16. Available online: https://www.ajc.com/news/local-govt--politics/fultoncounty-approves-uber-rides-for-seniors/ZQM6L7ycqmKCtmWsEfwOjO/. 518 Keane, C. (1998). "Evaluating the Influence of Fear of Crime as an Environmental Mobility Restrictor on Women's Routine Activities." Environment and Behavior, 30(1), pp. 6074. Available online: https://doi.org/10.1177/0013916598301003. Kim, S.H., Mokhtarian, P.L., and Circella, G. (2019). The Impact of Emerging Technologies and Trends on Travel Demand in Georgia. Report number FHWA-GA-19-1631, Georgia Department of Transportation, Atlanta, GA. Available online: http://g92018.eos-intl.net/G92018/OPAC/Index.aspx. Lin, W. (2012). "Wasting Time? The Differentiation of Travel Time in Urban Transport." Environment and Planning A: Economy and Space, 44(10), pp. 24772492. Available online: https://doi.org/10.1068/a4525. Loukaitou-Sideris, A. (2014). "Fear and Safety in Transit Environments from the Women's Perspective." Security Journal, 27(2, SI), pp. 242256. Available online: https://doi.org/10.1057/sj.2014.9. ------. (2016). "A Gendered View of Mobility and Transport: Next Steps and Future Directions." Town Planning Review, 87(5), pp. 547565. Available online: https://doi.org/10.3828/tpr.2016.38. Loukaitou-Sideris and Fink. (2009). "Addressing Women's Fear of Victimization in Transportation Settings A Survey of US Transit Agencies." URBAN AFFAIRS REVIEW 44 (4): 55487. https://doi.org/10.1177/1078087408322874. Loukaitou-Sideris, A. and Wachs, M. (2018). Transportation for an Aging Population: Promoting Mobility and Equity for Low-Income Seniors. Report No. 18-10, Project 1704, Mineta Transportation Institute, San Jos, CA. Available online: https://transweb.sjsu.edu/sites/default/files/1704-Sideris-Transportation-Aging-PopulationEquity-Mobility.pdf. Lu, W., Scott, D.M., and Dalumpines, R. (2018). "Understanding Bike Share Cyclist Route Choice Using GPS Data: Comparing Dominant Routes and Shortest Paths." Journal of Transport Geography, 71(July), pp. 172181. Available online: https://doi.org/10.1016/j.jtrangeo.2018.07.012. Malokin, A., Circella, G., and Mokhtarian, P.L. (2019). "How Do Activities Conducted While Commuting Influence Mode Choice? Using Revealed Preference Models to Inform Public Transportation Advantage and Autonomous Vehicle Scenarios." Transportation Research Part A: Policy and Practice, 124(June), pp. 82114. Available online: https://doi.org/10.1016/j.tra.2018.12.015. Maurer Braun, L., Read, A., and Ricklin, A. (2016). "A29 The Benefits of Street-Scale Features for Walking and Biking." Journal of Transport & Health, 3(2), pp. S22S22. Available online: https://doi.org/10.1016/j.jth.2016.05.060. McDonald, N.C. (2012). "Is There a Gender Gap in School Travel? An Examination of US Children and Adolescents." Journal of Transport Geography, 20(1), pp. 8086. Available online: https://doi.org/10.1016/j.jtrangeo.2011.07.005. 519 McFarland, J., Hussar, B., Zhang, J., Wang, X., Hein, S., Diliberti, M., Forrest Cataldi, E., Bullock Mann, F., and Barmer, A. (2019). The Condition of Education 2019. NCES 2019-144, U.S. Department of Education, National Center for Education Statistics, Washington, DC. Available online: https://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2019144. McGuckin, N. and Fucci, A. (2018). Summary of Travel Trends: 2017 National Household Travel Survey. Report No. FHWA-PL-18-019 , Federal Highway Administration, Washington, DC. Available online: https://nhts.ornl.gov/assets/2017_nhts_summary_travel_trends.pdf. McGuckin, N., Zmud, J., and Nakamoto, Y. (2005). "Trip-Chaining Trends in the United States: Understanding Travel Behavior for Policy Making." Transportation Research Record: Journal of the Transportation Research Board, 1217, pp. 199204. Available online: https://doi.org/10.3141/1917-22. McQuaid, R.W. and Chen, T. (2012). "Commuting Times The Role of Gender, Children and Part-Time Work." Research in Transportation Economics, 34(1), pp. 6673. Available online: https://doi.org/10.1016/j.retrec.2011.12.001. Misra, A. (2017). Mapping Bicyclist Route Choice Using Smartphone Based Crowdsourced Data. Ph.D. dissertation, Georgia Institute of Technology, Atlanta, GA. Available online: https://smartech.gatech.edu/bitstream/handle/1853/58619/MISRA-DISSERTATION2016.pdf?sequence=1&isAllowed=y. Misra, A. and Watkins, K. (2017). "Modeling Cyclists' Willingness to Deviate from Shortest Path Using Revealed Preference Data." 96th Annual Meeting Compendium of Papers, Transportation Research Board, Washington D.C. ------. 2018. "Modeling Cyclist Route Choice Using Revealed Preference Data: An Age and Gender Perspective." Transportation Research Record: Journal of the Transportation Research Board, 2672(3), pp. 145154. Available online: https://doi.org/10.1177/0361198118798968. Mokhtarian, P.L. (2016). "Discrete Choice Models' 2: A Reintroduction to an Old Friend." Journal of Choice Modelling, 21(December), pp. 6065. Available online: https://doi.org/10.1016/j.jocm.2016.02.001. ------. (2019). "Subjective Well-Being and Travel: Retrospect and Prospect." Transportation, 46(2), pp. 493513. Available online: https://doi.org/10.1007/s11116-018-9935-y. Mokhtarian, P.L., Papon, F., Goulard, M., and Diana, M. (2015). "What Makes Travel Pleasant and/or Tiring? An Investigation Based on the French National Travel Survey." Transportation, 42(6), pp. 11031128. Available online: https://doi.org/10.1007/s11116-014-9557-y. National Council on Disability. (2015). Transportation Update: Where We've Gone and What We've Learned. Report, National Council on Disability, Washington DC. Available online: https://ncd.gov/publications/2015/05042015. 520 National Household Travel Survey (NHTS). (2018). 2017 NHTS Version 1.1 Release Notes. NHTS, Federal Highway Administration, Washington, DC. Available online: https://nhts.ornl.gov/documentation. Nussbaum, M. (2003). "Capabilities as Fundamental Entitlements: Sen and Social Justice." Feminist Economics, 9(23), pp. 3359. Available online: https://doi.org/10.1080/1354570022000077926. Oldenburg, R. (1989). The Great Good Place: Cafs, Coffee Shops, Community Centers, Beauty Parlors, General Stores, Bars, Hangouts, and How They Get You through the Day. 1st ed., Paragon House, New York. O'Regan, K.M. and Quigley, J.M. (1998). "Cars for the Poor." Access, 12(Spring), pp. 2025. Available online: https://www.accessmagazine.org/spring-1998/cars-poor/. Ory, D.T. and Mokhtarian, P.L. (2005). "When Is Getting There Half the Fun? Modeling the Liking for Travel." Transportation Research Part A: Policy and Practice, 39(23), pp. 97123. Available online: https://doi.org/10.1016/j.tra.2004.09.006. Paleti, R., Bhat, C.R., and Pendyala, R.M. (2013). "Integrated Model of Residential Location, Work Location, Vehicle Ownership, and Commute Tour Characteristics." Transportation Research Record: Journal of the Transportation Research Board, 2382(1), pp. 162172. Available online: https://doi.org/10.3141/2382-18. Papadimitriou, C. (2008). "Becoming En-wheeled: The Situated Accomplishment of Re- embodiment as a Wheelchair User after Spinal Cord Injury." Disability & Society, 23(7), pp. 691704. Available online: https://doi.org/10.1080/09687590802469420. Pendakur, V.S. (2011). "Non-Motorized Urban Transport as Neglected Modes." In Urban Transport in the Developing World: A Handbook of Policy and Practice, eds. Dimitriou, H.T. and Gakenheimer, R.A., Edward Elgar Publishing, Northampton, MA. Rittel, H.W.J. and Webber, M.M. (1973). "Dilemmas in a General Theory of Planning." Policy Sciences, 4, pp. 155169. Available online: https://link.springer.com/article/10.1007/BF01405730. Roth, S.B., Dai, Y., and DeMatteis, J. (2017). 2017 NHTS Weighting Report. Task P, prepared for U.S. Federal Highway Administration by Westat, Rockville, MD. Available online: https://nhts.ornl.gov/documentation. Russell, M. and Mokhtarian, P. (2015). "How Real Is a Reported Desire to Travel for Its Own Sake? Exploring the `Teleportation' Concept in Travel Behaviour Research." Transportation, 42(2), pp. 333345. Available online: https://doi.org/10.1007/s11116-014-9546-1. Shen, Q. and Sanchez, T.W. (2005). "Residential Location, Transportation, and Welfare-to-Work in the United States: A Case Study of Milwaukee." Housing Policy Debate, 16(34), pp. 393 431. Available online: https://doi.org/10.1080/10511482.2005.9521551. 521 Shirgaokar, M. (2018). "Expanding Seniors' Mobility through Phone Apps: Potential Responses from the Private and Public Sectors." Journal of Planning Education and Research, 40(4), pp. 405415. Available online: https://doi.org/10.1177/0739456X18769133. Smart, M.J. and Klein, N.J. (2018). "Disentangling the Role of Cars and Transit in Employment and Labor Earnings." Transportation, 47, pp. 12751309. Available online: https://doi.org/10.1007/s11116-018-9959-3. StataCorp. (2017). Stata Statistical Software: Release 15. StataCorp LLC, College Station, TX. Susilo, Y.O, Lyons, G., Jain, J., and Atkins, S. (2012). "Rail Passengers' Time Use and Utility Assessment: 2010 Findings from Great Britain with Multivariate Analysis: Travel Behavior 2012." Transportation Research Record: Journal of the Transportation Research Board, 2323(1), pp. 99109. Available online: https://doi.org/10.3141%2F2323-12. Suzuki, H., Cervero, R., and Iuchi, K. (2013). Transforming Cities with Transit: Transit and Land-Use Integration for Sustainable Urban Development. Urban Development Series, World Bank, Washington, DC. Available online: https://doi.org/10.1596/978-0-8213-9745-9. Taylor, B.D., Ralph, K., and Smart, M. (2015). "What Explains the Gender Gap in Schlepping? Testing Various Explanations for Gender Differences in Household-Serving Travel." Social Science Quarterly, 96(5), pp. 14931510. Available online: https://doi.org/10.1111/ssqu.12203. United Nations Economic and Social Council (ECOSOC). (1997). UN Economic and Social Council Resolution 1997/2: Agreed Conclusions. 18 July, 1997/2. Available at: https://www.refworld.org/docid/4652c9fc2.html. ------. 2011. "Equity Evaluation of Urban Transport." In Urban Transport in the Developing World: A Handbook of Policy and Practice, eds. Dimitriou, H.T. and Gakenheimer, R.A. Edward Elgar Publishing, Northampton, MA. Vasconcellos. (2001). Venolia, G., Erickson, T., Tang, J., Mazza, B., and Herring, S. (2014). "Lifestyle Teleworkers Speak Out!" In CSCW Companion '14: Proceedings of the Companion Publication of the 17th ACM Conference on Computer Supported Cooperative Work & Social Computing, pp. 117120. ACM Press, Baltimore, MD. Available online: https://doi.org/10.1145/2556420.2556848. Wong, I.Y., Smith, S.S., Sullivan, K.A., and Allan, A.C. (2016). "Toward the Multilevel Older Person's Transportation and Road Safety Model: A New Perspective on the Role of Demographic, Functional, and Psychosocial Factors." The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 71(1), pp. 7186. Available online: https://doi.org/10.1093/geronb/gbu099. World Bank. (2013). Inclusion Matters: The Foundation for Shared Prosperity. New Frontiers of Social Policy series, World Bank, Washington, DC. Available online: https://doi.org/10.1596/978-1-4648-0010-8. 522 Zhu, G., Corcoran, J., Shyy, P., Pileggi, S.F., and Hunter, J. (2018). "Analysing Journey-to-Work Data Using Complex Networks." Journal of Transport Geography, 66(January), pp. 6579. Available online: https://doi.org/10.1016/j.jtrangeo.2017.11.006. 523