Measuring Goal Attainment for GeorgiaGain and Mf~rit System Reform Data Analysis Report for the Georgia Merit System Submitted to Marjorie H. Young, Commissioner Georgia Merit System October 10, 2000 Mr. Steven T. Elmore, Principal Investigator Dr. Grady L. Cornish, Project Director Vinson Institute of Government The University of Georgia Data Analysis Report Table of Contents II. Background Overview GeorgiaGainHistorical Background Meri! System Reform Historical Background Overlapping Tirnelines 9 !) !) 10 II III. General Methodology Approaches to Measuring Goal Attainment Identifying Goals Measurable through Data Analysis Analysis of Existing Data 13 1:3 1:3 15 IV. Data Selection and Summarization.. ~_. Overview of Data Available for Analysis Detail Review of Employee Extract Data Detail Reviev of Performance Evaluation Data Detail Review of Personnel Transaction History Data Factors Complicating Data Analysis .._....... . . 17 17 18 l!) 20 20 V. GeorgicJGain Goals Measurable with GMS Data Assure Fair Performance Ratings.... __ __ Approach Selection of Records for Analysis Examination of Performance Evaluations by Overall Score or Rating Examination of Performance Evaluations by Ethnic Group Examination of Performance Evaluations by Gender Examination of Performance Evaluations by Age Examination of Performance Evaluations by Tenure Examination of Performance Evaluations by Pay Grade Examination of Performance Evaluations by Salary Range Examination of Performance Evaluations by -Iob Title Examination of Performance Evaluations by Agency Examination of Terms & Conditions Ratings Observations Reward Best Performers _.~ ._ _ _ ~_ __ App roach Selection of Records for Analysis Examination of Relationship between Performance and Increases Observations 23 25 25 2~) 2H :32 :35 :>7 :>9 .41 .4;') ,48 52 5fi 57 59 5H GO ()j g:3 October 10, 2000 Page i Data Analysis Report Help Manager-s Document Poor Performance,"" Approach Selection of Records for Analysis Examination of Low Performance Evaluation Scores and Ratings Examination of l';mployeeProfiles of Low Score and Ratings Observations ".~" ~."" 65 (if) (if) ()(:i (;7 (;H VI. Merit System Reform Goals Measurable with GMS Data 71 Advance Employees on the Basis of Abilities, Knowledge and Skills "._._.... 73 Approach 7;l Selection of Records for Analysis 74 Examination of Promotions for High Performers 75 Examination of Promotions for Average Performers 77 Observations 78 Provide Equttable Compensation based on Merit and Performance 79 App roach 7~) Selection of Records for Analysis 7H Examination of Salary Lines by Performance Group 81 Examination of Salary Lines by Compensation Group 8:3 Examination of Salary Lines by Ethnic Group 8;j Examination of Salary Lines by Gender 87 Examination of Salary Lines by Hire Date 8!) Observations H1 Retain Employees on Basis of their Performance Approach Selection of Records for Analysis Examination of Separations by Performance Group Examination of Performance Evaluations and Separations by Year Examination of Tenure of Active and Separated Employees Observations 93 !):3 !):3 H5 H7 !H) 100 Take Action to Address Inadequate Performance Approach Selection of Records for Analysis Examination of Corrective Actions Taken Observations .,"."_. .._ ~.~__ 103 10il 10:3 104 108 VII. Summary of Observations ..... .. 109 VIII. Conclusion and Recommendations Conclusion Recommendations Begin Data Analysis Prior to Implementing the Project.. Ensure the HRMS Can Capture Necessarv Data Acquire Data Analysis Tools 113 1 I:l I 1;~ 114 114 115 Page ii October 10, 2000 Data Analysis Report IX. Appendices Appendix A Personnel Transaction Codes Used in (}EMS Appendix B Summary of Employees by Personal Characteristics Appendix C PAl Scores fix Classes with 100+ Evaluations Appendix I) PIVIF Ratings for -Iobs with 100+ Evaluations Appendix E Reprint of Complex Charts with Data Values 117 1W 12:~ 121, ]:O I:l7 October ]O. 2000 Page iii Data. Analysis Report Pageiv October 10,2000 Data Analysis Report List of Tables Table I Goals of GeorgiaGain Table 2 Goals of tvlerit System Reiorin Tahlc S Goals Partially Measurable through Data Analysis Table 4 Summary of Employee Data in IIRMS Table f> Summary of Performance Appraisal Data in HHMS Table G Distribution ofI\\1 Performance Evaluations by Period Table 7 Distribution of PMF Evaluations by Period Table 8 Description of Performance Appra.isal Instrument Scores Tab le D Distr-ibution of PAl Scores by 1/10 Point... Table 10 Distribution of PAIP;valuations by Score Table 11 Distribution of I).MF Ratings Table 12 Ideal Distribution of Performance Evaluation Ratings Table 1:{ Distribution of PAl Scores by Ethnic Group Table 11 Distribution of PNIF Ratings by Ethnic Group Table 15 Distribution of PAl Scores by Gender.. Table] G Distribution of PM F Ratings by Gender Table 17 Dist.ribution of PAl Scores by Age Table 18 [Jist.ribu tinn of PMF Ratings by Age Table H) I)istribution of PAl Scores by Tenure Table 20 Distribution of PMF Ratings by Tenure Table 21 Dist.ribution of PAl Scores by Pay Grade Table 22 Distribution of PM F Ratings by Pay Grade Table 2;~ Distribution of PAl Scores by Salary Range Table 24 Distribution of PMF Ratings by Salary Range Table 25 Distribution of PAl Scores for 25 Most Frequently Evaluated Class Titles Table 26 Distribution of PMFRatingsf(lr 25 Most Frequently Evaluated -Iob Titles Table 27 Distribution of PAl Scores by Agency Table 28 Distribution of PMF Ratings within Agencies Table 29 Distribution of Terms & Conditions Rating Table :~o Performance Groups Table ;3) Distribution of Selected Increases by Year., Table ;32 Distribution of Performance Evaluations by Year Table :{:J Performance Evaluations versus Increases under PAl System 'fable :J4 Performance Evaluations versus Increases under PMF System Table ;~5 Percentage Distrihution of Increases under PAl System Tahle of Percentage Distribution of Increases for PMF System Table S? Low Performance Evaluations by Evaluations System Table :{8 Detail of Low Scores under PAl System Table :3!J Promotional Authority Codes Used in GEMS Table 40 Number of Promotions per Year Table 11 Performance Evaluation Data Available for All Types of Promotions Table ,12 Number of Regular Promotions by Performance Group Table 4;{ Percentage Distribution of Regular Promotions by Performance Group Table 44 Distribution of Performance Evaluations by Performance Group P.:!~c: 10 10 14. 18 l!) 2(; 27 2~j 2~) :W :w ;H ;l2 ;J;{ ;{5 :w :)7 :l8 ;m .40 .4] .4:J Af> 4H 48 50 52 51 5(; 5!J GO ()() (i2 ()2 f;2 (;;{ fig fi(; 7:~ 74 74 75 7fi 7fi October 10, 2000 Page v [lata Analysis Report Table 45 Salary Increases with Matching Performance Data 80 Tablcdf Budgeted Salary Increases 80 Table 47 Salary Increase Rates by Performance Groups 81 Table 48 Salary Increase Rates by Compensation Group 8:J Table 4H Salary Increase Rates by Ethnic Group 85 Table fiO Salary Increase Rates by (lender. 87 Table 51 Salary Increase Rates by Hire Date 8!) Table 52 Summary of Separation Data H4 Table 5::1 Separations by Performance Group !)5 Table 54 Approximate Percentage of Performance Groups Separating Annually !t7 Table 55 Tenure in Months of Separated Employees m) 'I'able 5(; Number of Low Performance Evaluations per year. lOi~ Table 57 Low Performers as It Percentage of Employees and Evaluations 104 Table (')8 One-'(ear Follow-up on Low Performers 105 Table 5!) Pre- and Post-lv!erit S:.",5/(I11 Reiorm Activity for LowPerformers 10;) Table no Separation Types for Low Performers 10(; 'I'able (il Separation Types for All Employees 10(; Table 62 Personnel Transaction Types for Low Performers 107 Table 6:) Personnel Transaction Hates for All Employees 107 Page vi October 10.2000 Data Analysis Report List of Figures Page Figure I ~ Gcorgi;JGain and Merit. Systetn Reiortn T'imeline 11 Figure 2 PAl Performance Evaluations by Month 2(; Figure ;J - prvlF Evaluations by Month 28 Figure 4 PAl Scores by Year ao Figure 5 PMF Ratings by Year :~1 Figure G PAl Scores by Ethnic Group :)2 Figure 7 PMF Ratings by Ethnic Group :3:J Figure 8 PAl SCOI'es by Gender ;);'") Figure D PMF Ratings by Gender ;W Figure 10 PAl Scores by Age :)7 Figure 11 PMF Ratings by Age :)8 Figure 12 PAl Scores by Tenure :JH Figure 1;) PMF Ratings by Tenure 40 Figure 14 PAl Scores by Pay Grade 42 Figure 15 PMFH.atings by Pay Grade .41 Figure l() PAl Scores by Salary Range ,45 Figure 17 PIVlF Ratings by Salary Range .4fl Figure 18 PAl Scores for 25 Most Frequently Evaluated Class Titles .4!l Figure 19 - Pl'vlF Scores for 25 Mostly Frequently Evaluated -Iobs Titles ;")1 Figure 20 PAl Scores for Larger Agencies 5;) Figure 21 PMFRatings for Larger Agencies 55 Figure 22 Terms & Conditions by Responsibility Rating 5(; Figure 2:) Comparison of LO\v Scores and Ratings by Ethnic Group G7 Figure 24 Comparison of Low Scores and Ratings by Tenure (,8 Figure 25 Comparison of Low Scores and Ratings by Salary Range ('8 Figure 2H Comparison of Evaluations and Promotions for High Performers 77 Figure 27 _. Comparison of Evaluations and Promotions for Average Performers 78 Figure 28 Cumulative Salary Lines by Performance Group 82 Figure:m Cumulative Salary Lines by Compensation Group 84 Figure ;10 Cumulative Salary Lines by Ethnic Group 8(; Figure :)1 Cumulative Salary Lines by Gender 88 Figure ;32 Cumulative Salary Lines by Hire Date ~}O Figure a:) Percentage of Separations by Performance Group ~)5 Figure:H Percentage of Performance (},'oups Separating (approx.) ~)8 Figure :)5 Trend Lines for Service at Separat.ions 100 October 10, 2000 Page vii Data Analysis Report Page viii October 10. 2()()() Data Analysis Report Preface GeorgiaGain a performance-based compensation system, was implemented in I !)!}(; as the response to then-Governor Zell Miller's call for a way to better motivate, reward, and retain high quality employees within Georgia state government. This new approach was predicated on the notion that managers and employees are jointly accountable for optimum job performance. Although it began as a project to develop a new performance-based compensation system, Geor;giaGain was expanded to revitalize and re-engineer many of the State's personnel processes. Convinced of the need to put Georgia's government on a more business-like foundation and thereby help it achieve the type of efficiencies derna nded by the public, Governor Miller initiated legislation in lH!)(i to reform Georgia's civil service system. Act 8 Hi, now commonly referred to as Merit System Reiarm, marked the most significant change in Georgia's personnel system in fifty years. This report sets forth the results of an analysis of data available to the Georgia Merit System through its human resources management system. Available data permitted comprehensive measurement of three of the eight goals of Gl'or;giaGain and four of the seven goals of Merit System Reiortn. The principal researcher, Mr. Steven T.l'~lmore, has examined both initiatives mindful of the need to clearly describe the findings and to present the results in terms that have practical implications for the State's human resources administrators and policy makers. Mr. Elmore is perhaps the most qualified person to conduct this analysis. In addition to a degree in mathematics and information systems, he is thoroughly familiar with the data in the State's human resources management system. Since H)()S and up to his retirement in e:lrly-2000, Mr. Elmore held a variety of positions with the Georgia Merit System and played a significant leadership role in the design, development, and implementation of the State's human resources management system. Among a long list of other accomplishments, Mr. Elmore: Prepared the first comprehensive report of employee of turnover in state government; Designed computer programs to move 40,nOO employees and positions to new agencies as part of Governor Carter's reorganization of state government: Co-managed the development and implementation of the State's first comprehensive human resource management system, which included employee classification. position control. applica nt testing and selection, personnel transactions, and payroll; Co-managed the development of software to implement a comprehensive employee reclassification project known as the Hay Study; and Co-managed the system and data conversion activities necessary to implement Geor;giaGain and Meri! System Reform in the State's human resources management system. October 10, 2000 Pageix DataAnalysis Report Thanks to Lynn Seymour, Ph.D. and Assistant ProfcsaorcDepartment of Statistics, The University of Georgia, for her critique of this report. \Ve also wish to express appreciation for the valuable input and aasiata nce provided by the Commissioner and her staff in the preparation and execution of this study. Grady L. Cornish Project Director Page x October 10.2000 Data Analysis Report Author's Note: This paper documents a practical review of the results of two radical changes to the State's personnel program. The statistical measures of significance and confidence that one might expect to see in such a study are not provided in this paper. Such measures help assure the reader that. the values. averages and trends calculated from samples are representative of the larger universe of data. Throughout this effort, however. a unique opportunity presented itself the ability to analyze every transaction and evaluation from Hm;5 to HHH). In the preparation of this paper. no sampling was performed. The entire universe of transactions was analyzed in every instance. During the period covered by this study, changes occurred in several naming conventions associated with the data used. These changes OCCUlTed in H)HG and are somewhat related to the outcomes and de live rables of the Georgietlein and Merit System Reiorm projects. The state agency official referred to as the State Merit System of Personnel Administration changed its commonly used name from State Merit System to Georgia Merit System. This name change was part of the internal transition effort after lv/edt System Reioni: that moved the agency from a regulatory role to a new service/consultative role. In this paper, the agency is consistently referred to as the Georgia Merit System although the source documents before 199G use the State Merit System designation. The nomenclature of employee performance evaluations changed completely as part of the GeorgiaGtlin implementation. Prior to GeorgiaGain. employee performance evaluations were referred to as performance appraisals. The employee received a numerice! performance appraisal score on a form called a periormence appraisal instrument (PAL). The new performance evaluation system developed as part of GeorgiaGain changed this nomenclature. Employee performance evaluations are referred to as performance evaluations, and the employee receives a verbal performance eva luationrgjjjjg on a form called a pertormnnce management Iorm (PMF). In this paper, performance evaluation is used to reference the general process of evaluating an employee's performance or to identity any formal evaluation document. The system in place prior to GeOlgiaGain is referred to as the PAl system; and for symmetry. the newer system is referred to as thePMF system. This may cause concern among human resources purist - the older system had no formal name and the new system is official known as the Pertortuence Menegetnent Process. As part of GeorgiaGain implementation, the term class, meaning a group of positions or employees with similar duties. was changed to the 19.12. In this report. class or class title will be used to reference a specific type of work before Ceorgieciein implementation and job or job title will be used after GeorgiaGain implementation although the concepts of class and job are nearly identical. October 10. 2000 Page xi Data Analysis Report Under the PAl-based system, performance evaluations contained only one overall Score. However, under the PMF-based system performance evaluations contained two ratings one as a measure of performance on job, individual, and statewide responsibilities and another as a measure of compliance with agency established terms and conditions of employment. Several of the reports required comparative analysis of performance evaluations with low, average, and high scores or ratings. For most of these reports the responsibility rating was the only rating considered. In the seventh report, which focused on tracking activity fill' low performers, performance evaluations with low ratings in the terms and conditions area were also considered in the group of low performers. Page xii October] 0, 20C)O Data Analysis Report l'~xecutive Summary 1. Executive Summary Two projects implemented in W!)(;, CeorgiaGain and Merit Svstem Reiorm. radically changed the personnel program of the State of Georgia. These projects had related objectives of improving employee performance and linking compensation to performance. Ceorgietlein focused on the need to restructure employee compensation and rewards through development of a pay-for-performance system. Merit System Reiotin addressed concerns that the existing civil service system protected low-performing employees and limited the ability of managers to operate their organizations in the same manner as private businesses. Although not the first attempts at changing the State's personnel program, these changes were the most fully implemented and most directly driven by executive forces, There is considerable overlap in the goals for Ceotgietlein and Merit System Reiorm. The concepts of measuring performance, training employees to perform, rewarding good performance, addressing poor performance, and assuring basic fairness can be found in both. Because of the broad nature of these goals, several approaches must be undertaken in a coordinated manner to measure the degree to which CeorgiaGain and Merit System Reiotm have met their goals. Surveys. interviews, and analysis of hard data must be used together to properly assess goal attainment and measure project success. This report only addresses those goals that can he measured through analysis of data available in the computer systems of the (}eorgia Merit System. The following are the GeorgiaGain and Merit System Reiorm goals that are measurable at least in part through the data available to the Georgia Merit System: GeorgiaGaill GOltlli Provide mechanisms to assure fair performance ratings Reward best performers through variable increases to base pay -- Help managers document poor performance fl:'tfl.LiL~S'yst('rnBefgml Cklll] Recruit, select, and advance employees based on ability, knowledge, and skills -- Provide equitable and adequate compensation based on merit and performance Retain employees on the basis of the adequacy of their performance Correct inadequate performance where possible, and separate employees whose performance is inadequate The preceding list is less than half of the stated goals for these projects. Measuring the level of attainment of other goals, and Some dimensions of these goals, requires data not available to the Georgia Merit System through its data systems. Other methods should be used to measure success in those areas. Several types of data are available to the Georgia Merit System for data analysis. Periodic snapshots of the entire employee population, or 'employee extracts'. have proven very useful in responding to interrogatories in federal and state court actions where accurate historical October 10, 2000 Page 1 Executive Summary Da ta Ana lysis Report analysis has been required. In the normal process of granting annual and performancebased increases, the final rating or score on 'performance evaluations' were captured. 'Transaction histories' exist for nearly every personnel transaction, particularly those transactions that affected compensation or employment status. Because the implementations of Merit System Reionn and GeorgiaCain occurred in close proximity in l!H)(j (July and October respectively), goal attainment for both was measured with a common set of data. Although not necessary for the measurement and comparison of trends, equal timeframes both before and after these changes were selected. The most recent data available is from October HHW, or three years after October H)!)(i. Therefore, an early cutoff point for data of three years before -Iuly H)!)(j (eluly W!);n was chosen. Three factors added to the complexity of the task and diminished the accuracy of the results. The Georgia Employment Management Systems (GEMS) database contains nearly ;')00,000 performance evaluations, however, this is not the complete set of performance evaluations created during the period. Not all performance evaluations associated with disciplinary actions may be present in the database. During GeorgiaGain implementation, employees were moved individually to the new jobs that best fit their responsibilities and assignments. This individual approach to movement makes it is impossible to cleanly compare trends by job or pay grade across this point. Finally, several attempts to simplify data entry procedures resulted in changes in computer coding structures and elimination of many distinctions in personnel activity recorded in the system. Any attempts to extend the following reports by incorporating future transactions will be very difficult. GEMS was replaced as the State's human resources management system (HRMS) in OctoberH)~m with a purchased software package. Few modifications were made in the new software to make it compatible with the historical data. For each measurable goal, a report was prepared showing summary statistics and a appropriate charts and graphs along with some general observations. Reports 1 through address GeorgiaGain goals; Reports 4 through 7 address Merit System Reform goals. Report 1 (GeorgiaGiILn> -- Assure Fair Performance Ratin.,g This report examined the change in influence of several non-performance related factors on performance evaluations between the older performance appraisal instrument (PAl) system and new performance management form (PMF) system developed as part of Gcorgietleu. The personal and employment characteristics examined include ethnic group, gender, age, tenure, job title, pay grade, compensation level. and employing agency. During the period from April Hma to October W!H), two performance evaluation systems were used. The pre-GeorgiaGain system used the performance appraisal instrument (PAl) and for the period April 19!):3 to December HH);'), a total of 18:>,0(jO performance evaluations were available for analysis. In ,January 1996, the State implemented a new Performance Management Process that used the performance management form (PMF). For the period -Ianuary 1996 to October I!)!)!), a total of :J0(j,(;40 performance eva luations were available for analysis. Several very similar 'ideal' rating distributions for the new PMF system have been proposed since the beginning of the Georgtetlein project. The percentage of Page 2 October 10. 2000 Data Analysis Report Executive Summary performance evaluations in the Meets and Exceeds categories has approached these 'ideal' distributiousfor several years. The percentage of Does Not Meet and Far Exceeds continues to be lower than anticipated. Despite attempts at universal evaluation for all employees, the number of employees granted a performance-based increase without a full performance evaluation exceeds the number receiving ratings of Does Not Meet and Far Exceeds combined. Ratings under the newer PIVfF system are more equitably distributed as evidenced by a compar-ison of performance evaluations by gender and ethnic group. The influence of age and tenure on performance evaluation ratings bas been reduced. With nota ble exceptions. the high degree of variability seen by agency in the previous system has been greatly reduced. However, there continues to be strong influence of compensation-related factors in the distribution of ratings and there is strong va riabrlity in ratings distribution by job title. The universal evaluation ofevery employee against the terms and conditions of employment standards may be unproductive. Over m)% of performance evaluations indicate the highest rating possible in this area and only ;~28 salary increases may have been withheld solely on the basis of the terms and conditions evaluation when otherwise warranted based on the responsibility evaluation. R~~.Il9rt 2 (GeOfJjiaGain) ~.. Reward. Best Perfi:mners This report examined the number of salary increases (other than promotions) given and then identified the performance evaluation that immediately preceded the salary increase. Three types of salary increases and adjustments were examined performance-based increases, salary adjustments, and criterion-based adjustments. These increases have some measure of agency discretion associated with them. In order to facilitate comparison, delineation between high, average, and low performers was constructed and this delineation is used in subsequent reports. Low performers fall below the standards for receiving annual or performance-based increases and high performers qualify for higher than average performance-based Increases. The number of discretionary salary adjustments processed by agencies has increased over 1000% following GeorgiilGain implementation. Positive differences exist between the proportion of discretionary increases granted to high pcrforrncra versus low performers: however, average performers received proportionately more discretionary increases than high performers Average performers also received more criterion-based increases than did high performers. The actual number of salary adjustments given to reward high performers cannot be discerned because H2'% of all salary adjustments were given to average and low performers. October 10, 2000 Page ;~ Executive Summary Data Analysis Report a Report (Gf!Q[gJ.1:LGain) m Help ManHger~J)oeumentPoor Per!i)rman, This report examined the level of confidence managers demonstrate in assigning low performance evaluations. Additionally, the internal make up of the low performing group was analyzed by ethnic group, tenure, and salary. Performance evaluations under two evaluation systems, PAl and PMF, were reviewed. Low scores and ratings represent a very small portion of performance evaluations under both systems consistently falling below the ideal distribution. More low scores proportionately were given on PAl performance evaluations than low ratings on Pl\11 F performance evaluations. The distribution of low performance evaluations under thePMF system is more equitable. The proportion of black employees in the group of Low Performers under the P1\1F system decreased more than 10'% from that seen under the PAl system. The proportion of low performance evaluations assigned to employees with 20 to 2H years of service increased from ~)%. to 14'%. Although managers assign more low performance evaluations to employees with higher salaries under the PMF-based system, this shift is probably due more to overall increases in salaries than to changes in manager behavior. l{eport.::Li!5'e{orm) - AdvanceEmnJ~>"ys~es on the Basis of Abiliti~~s_ Knowledge, and Skills This report examined promotional rates for employees. Performance evaluations were used as a proxy for "ability, knowledge, and skills". Higher scores and ratings on performance evaluations were considered as indicative of higher "ability, knowledge, and skills". A general reversal of previous trends is seen in the years H)!)() through I!)!)!) following Merit System Reform. The percentage of promotions given to those employees with high performance evaluations is now higher than their representation in the general body of evaluations and the percentage of promotions given to those employees with average performance evaluations is nov',' lower than their representation in the general body of evaluations. 'I'he changes in previous trends may be in part due to the continuing impact of Georgietsein rather than changes in management practices. The Georg/aGain job structures eliminated many purely promotional jobs thus making a promotion a true movement to a new job rather than a time-in-grade salary increase with a new title. This trend to fewer promotions may be a permanent impact of Ceorgistlein. Report f) (Reform)-, Provide Equitable (;ompeIlsation based on Merit and Performance This report examined whether changes in average salary lines for employees are attributable to performance or to other factors, such as ethnic group, gender, current compensation, or hire date. A salary line was constructed for each employee based on career salary progression including normal and special increases, promotions, and general adjustments. Page 4 October 10. 2000 Data Analysis Report Executive Summary Using the data from the annual budget reports, the average salary line fill' an employee who received all 'normal' increases hut no promotions or adjustments should show a 2(U,%, increase from HH):J to 1!)!)!). The actual average sala ry line for all employees was up :$8.82'~;J. Very 44.330 95712 77177 43110 9570E 77 363 41914 96.11C 77584 40462 2451.92 2012554 1 501.293 Unclassifed Full-time Part-time Employees Employees Employees 9.995 99?O 10741 11 OR,:! 11 151 11?M 11s14 11950 12.129 12.298 12.656 12~ 13. 14 ??fi 1!'>fi14 17.1'\61 19362 20B70 22170 24985 26.648 284'\1 30434 32.382 34.067 35429 37.122 511.261 67823 67761 6Rfi10 70019 70 7<;2 71 1?7 719R7 73137 73654 73803 73.984 74.519 7C, 1fi1 7S ?C,7 74582 74265 73BP 73.429 74123 74.079 74.316 745Q9 74958 75541 76259 7fi4RR 76667 1980952 869 854 1195 1 ??Il 1.221 1 197 1 19~ 1224 127C 1.263 1.314 1.291 11~ 110~ 1257 1.284 1247 125 1.234 1 187 1187 1 160 1 151 1 171 918 87~ 917 31 60~ Temporary 14.642 14128 13.967 14424 14 ?OQ 14fill 14405 14409 1,.217 15761 15.088 1~.332 is ?77 1f; sss 16112 16813 17.354 16.433 17.255 17.398 18.039 ts.szs 18~{O9 17998 18.542 18.345 18526 439369 Page 18 October 10, 2000 Data Analysis Report Data Selection and Summarization On September :m, IHHG, prior to implementation of Georgietlein. no extract was taken of the employee database. An extract had been taken on August :n, U)!)(; and that data was used in lieu of the missing September :JO, lH~)() data. On September ao, IHH7, performance- based increases for October 1, lHH7 were applied to the database before the extract was created resulting in salaries being misstated. The August aI, I~)~)7 extract "vas unusable; therefore, the -Iuly :n. WH7 extract was used in lieu of the September :JO, lHH7 data. The number of employees reported here is the number of employees on the payroll and does not purport to represent the number of full-time equivalent (FTE) employees. In fact, many employees reported as 'temporary employees' worked on an as needed basis. These employees did not. receive salary increases or formal performance evaluations. petail Review of Performance Evaluation Data From April wm~ to Oetober WH9, a total of 492, 18~) performance evaluations were recorded. Not all evaluations were recorded accurately and some could not be matched fully against employee data. The total number of performance evaluations available for analysis is 489,700 or over 9~U)%) of the all evaluations. Table 5 displays the number of performance evaluations from each evaluation system and the number and type of errors. Table 5 Summary of Performance Appraisal Data in JIRMS Rating unknown unknown unknown Rating Year 1993 1995 1997 Total Ratings inlolRMS 1 2 4 Invalid Salarv 1 Invalid Race Code I Inva id data in fRMS Invalid Wrong Sex Code System Invalid Ratina Total Unusable Ratlnas Total Usable 1 1 0 2 0 Hlil niiiMiMiiiH iiiII1II I 0 PAl PAl PAl PAJ kiM 1993 1994 1995 1996 iH1: 44 930 68255 70135 287 "ill"!l 157 9 7 61 3 2 38 3 2 17 I; ::';15 Hi'!':111 iiH' 159 44771 62 68193 39 70096 287 0 ~8'7 i./i 1 , 1 ; 1 1 1 ' - PMF 1994 PMF 1995 PMF 1996 PMF 1997 PMF 1998 PMF 1999 1 115 98208 69198 7054'1 70508 1264 230 191 128 492189 2089 1 115 48 47 1 7 6 4 3 iM 75 67 403 1 115 1265 231 19.4 ~ 7 24 0 0 96943 68967 70351 70379 489700 October 10,2000 Page H) Data Selection and Summarization Data Analysis Report During the period covered by this study, two performance evaluation systems were in use. The older system using the performance appraisal instrument (PAl) existed until December W95. For the period AprilW9:3 to December Im)5, a total of 18:3,0(;0 evaluations are available for analysis. In -Ianuary ImHi, the State implemented the new Performance Management Process that used the performance management form (PMF). For the period Jun a Jul s: C Au 0 ~ Sep Oct Nov Dec Total Evaluations Evaluations Evaluations Total in 1993 in 1994 in 1995 Evaluations 0 0 0 1,613 5,288 13,618 4,762 4,086 3,571 3,341 3,943 4,090 6,445 6,370 6,831 13,572 4,470 4,857 4,180 4,022 4,180 5,233 68,193 4,378 4,562 6,429 6,222 6,694 12,691 5,407 5,236 4,520 4,581 4,240 5,136 70,096 8,321 8,652 12,874 14,205 18,813 39,881 14,639 14,179 12,271 11,944 1 Performance evaluations under the older PAl-based system were primarily associated with delivery of an annual increase, sometimes referred to as a 'within-grade-increase', or WIGL Before the advent of GeorgiaGain, salary increases were delivered throughout the year on employees' anniversary dates and the distribution of performance evaluations by date of evaluation bears this out (see Table () and Figure 2), 15,000 10,000 El1993 81994 81995 5,000 o Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 2 ~ PAl Performance Evaluations by Month PAl performance evaluations may have been prepared at other times, such as upon delivery of a special increase or a delayed annual increase, as part of a disciplinary action (dismissal, etc.), and at end of working test when not coincident with the employee's Page 2G October 10,2000 Data Analysis Report Assessment of C;eorgiaGLjin Goals anniversary date. The GEMS database did not differentiate the reason a performance evaluation was prepared. however, the only specialized data entry screens fin' collection of performance evaluation data were those that were part of the annual increase data entry process. Table 7 l)itrilmti()Jl of PM Fl<~val!Jjl t iimiJ2y'j?!~.Iiml Jan Feb Mar c 0 ~ m:J Apr May w> Jun '0 Jul .c C Aug 0 :2 Sep Oct Nov Dec Total Evaluations Evaluations Evaluations Evaluations Total in 1996 in 1997 in 1998 in 1999 Evaluations 3,989 18 14 248 4,269 4,540 5,548 6,060 4,671 878 10,053 19,762 16,381 25,018 20 23 96,943 19 27 269 65 161 16,591 20,983 5,944 24,873 13 4 68,967 8 17 3 4 273 19,132 17,734 8,167 24,979 11 9 70,351 11 14 4 5 298 13,398 14,452 6,593 35,356 70,379 4,578 5,606 6,336 4,745 1,610 59,174 72,931 37,085 110,226 44 36 306,640 Performance evaluations prepared under the new PMF-based system were primarily associated with annual delivery of performance-based increases (PHI). After GeorljiaGain implementation, salary increases were delivered only on October I of each year following preparation of performance evaluations for every eligible employee. These performance evaluations were prepared between the end of the evaluation period, usually -Iune :m. and the increase date of October 1. This is reflected in the distribution ofperformance evaluations by effective date (see Table 7 and Figure :3). PMF performance evaluations may have been prepared at other times, such as upon delivery of a special increase, as part of a disciplinary action (dismissal. etc.), and at end of working test when not coincident with the annual evaluation. The GEIVIS database did not differentiate the reason a performance evaluation was prepared, however, the only specialized data entry screens for collection of performance evaluation data were those that were part. of the performance-based increase data entry process. DuringWH5 and W9(j, a transition was underway from the PAl-based system to the PMFbased system. As performance evaluations were completed in 1995 under the PAl system, PMF planning documents were prepared for use in the W!)() evaluation. Those employees whose anniversary dates in 19!Hi fell prior to ,July W!)(; present an interesting subset of employees for further investigation they were evaluated using the PMFbased process but they were still classified and compensated on the older pre- G('orgiaGain job titles and pay ranges. These same employees were also re-evaluated along with all other employees for another salary increase in October 199(;. This accounts for the larger than expected number of performance evaluations in ImHi. Performance evaluations completed in October 10, 2000 Page 27 Assessment of Georgiecleiti Goals Data Analysis Report -Ianuary through May ofHH)(j were at a rate consistent with the older anniversary date PAl-based system. 40,000 30,000 81996 81997 111998 -1999 20,000 10,000 Average number of performance evaluations per month under PAl System o JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Figure 3 PMF Evaluations by Month The following detail examinations of PAl and PMF performance evaluation scores and ratings were conducted: Examination of Performance Evaluations by Overall Score or Rating Examination of Performance Evaluations by Ethnic Group Examination of Performance Evaluations by Gender Examination of Performance Evaluations by Age Examination of Performance Evaluations by Tenure I<:xamination of Performance Evaluations by Pay Grade Examination of Performance Evaluations by Salary Range Examination of Performance Evaluations by -lob Title Examination of Performance Evaluations by Agency Page 28 October 10, 2000 Data Analysis Report Assessment of GeorgiaGain Goa ls The original design of the PAl-based employee performance evaluation system called for assigning a single-digit score as a measure of employee performance. The scores and their meanings are shown in Table 8. Table 8 DJ,:'~J'iption of Perf()nnance ApPI:!~i111ntDUm~!ltJi~;52.res ~:i<;Q!:(~ 5 4 a 2 Ex})Innation 01:' Des~in!ion Performance consistently exceeds requirements Performance frequently exceeds requirements Performance meets requirements Performance frequently falls below requirements Performance consistently falls below requirements Prior to I n!);{, the system was modified to permit supervisors to assign scores of one-tenth point resulting in an expansion of the scoring system from 5 to 41 values. Table H Distribution of PAl Seores by 1/JO Poil!! Score 10 1.1 1.2 1.3 1.4 1. 1.6 1.7 1.8 1.9 Count Score 6 2.0 2 21 4 22 7 23 4 2.4 2.5 2.6 2.7 2.8 29 Count 50 11 702 T'ahle B displays the distribution of PAl scores during the period April l!m;{ to December l!)!)[) using the 41 possible scores. The most frequent score assigned was 4.0, and this score is described as 'frequently exceeds' job requirements. For purposes of simplifying further analysis, the ,II scores were grouped into nine ranges as follows: 1.0 to 1.4 2.0 to 2.4 :U) to ;3.4 4.0 to 4.4 5.0 1.;"> to I.!) 2.5 to 2.9 ;1.5 to :U) 4.5 to 4.!) Table 10 displays the PAl scores by calendar year using these reduced ranges. Consistent with the detail distribution of scores, the range from 4.0 to 4.4 contained the largest number of performance evaluations. October 10, 2000 Page 2H Assessment of GeorgiaC.;ain Goals Data Analysis Report 11993 11994 11995 Total 1.0 to 1.4 3 10 10 23 'fable 10 Distribution of PAl Evaluations 12Y...!~c()re 1.5 to 1.9 25 31 21 77 2.0 to 2.4 111 195 172 478 5;,,.,,,rAilit 2.5 3.0 to to 2.9 3.4 316 8 131 531 14201 477 14987 1324 37319 3.5 to 3.9 10370 15978 16712 43060 4.0 to 4.4 14126 21084 21572 56.782 4.5 to 4.9 8342 12031 11 922 32295 5.0 3347 4132 4223 11702 Total 44771 68193 70096 183060 When viewed graphically. a subtle shift is apparent during the period as scores begin a slight but consistent shift. toward more moderate ratings. During this same period, supervisors and managers received intensive training in using the new Performance Management Process and were exposed to the concept that most ratings should be at the Meets level. ,A..ldlhotHll 110 40% 01993 .1994 30% .1995 20% 10% 0% L _ _-...... c : l _........- l l i 1.0 to 15 to 2.0 to 2.5 to 3.0 to 3.5 to 4.0 to 4.5 to 5.0 1.4 1.9 2.4 2.9 3.4 3.9 4.4 4.9 Figure 4 .~ PAl Scores by Year The distribution of performance evaluations under the new PMFhased system is shown in Table 11. 'l'ahle 11 Distribution of PMI<' Ratings 11111f\ 11997 11998 11999 Total Does Not C Meet 1 37!) 561 1614 543 2345 503 ? !)fifl 7902 2169 Meets 86763 58838 57439 259465 Exceeds 7407 7455 9452 10.020 34334 Far Exceeds 837 517 612 804 2770 Total 96943 68967 70351 70379 306640 Page 80 October 10.2000 Data Analysis Report Assessment of GoorgiaC;tlin Goals Various 'ideal' rating distributions for the new PMF-based system have been voiced since the beginning of the G('o(giaGain project. A.II fit in the follow ranges: Table 12 Ideal Distribution ofJ'erformance Evaluatllin Ratings N"-'''' -- Ratings 'Ideal' Distribution Does Not. Meet .',)100/ :~ (Xl Meets 80%, ~ 8:3%, f - - _E..._x-c-.eeds Far Exceeds 10'% - 15%, 5(% .m As can be seen in Figure 5, the percentage of PMF-based performance evaluations containing Meets and Exceeds ratings approached the 'ideal' distribution for several years. The percentage of Does Not Meet and Far Exceeds ratings continued to be lower than anticipated. 100% 80% 60% m1996 m1997 111998 .1999 Oldeal 40% 20% 0% c DNM Meets Exceeds Figure.') PMF Ratings by Year Far Exceeds For administrative convenience, a new rating of'C' was devised to indicate that a full performance evaluation was not. prepared, but that. the supervisor had 'considered' the employee's performance to-date and no issues were present to warrant withholding the salary increase. 'I'his code was originally designed to address concerns of supervisors when evaluating the performance of new hires and recent transfers, The number of employees granted a performance-based increase without a full performance evaluation exceeded the number receiving Does Not Meet and Far Exceeds ratings combined. In the balance of these reports. performance evaluation ratings of 'C' will be reported separately unless ratings are grouped, in which case the 'C' ratings will be grouped with average performers (those receiving a Meets rating). October 10. 2000 Page :31 Assessment of Geotgistieir: Goals [lata Analysis Report Examination of P~!:formanceEvaluations by Ethnic CBmp This examination of performance evaluations focused on a personal characteristic. However, the distinction between performance evaluation systems was retained. (}EMS recorded race/ethnic group information using standard categories required by state and federal legislation - white, black. Asian, Hispanic, American Indian, and multi-racial. For purposes of this examination, the Asian, Hispanic. American Indian, and multi-racial categories were combined into a group labeled 'other'. Table 1a shows the distribution of PAl scores for the modified ethnic groups. Table ia Distribution of PAl Scores by Ethnic C1roup 1.0 to 1.4 13 23 1.5 to 1.9 36 39 77 2.0 to 2.4 215 256 7 478 Total 116011 64877 2172 183060 These ratings from the 199:3 to 1995 period were representative of the distribution of the overall employee population (see Appendix B on page 12:3). The study shows that ;3;'').4%) of performance evaluations were for black employees in a period when black employees comprise :~5.:3% of the workforce. Other employees comprise] .2% of the study data and 1.;~(X) of the period workforce. 30% 20% CWhite -Black Other 10% 0% 1.0 to 1.5 to 2.0 10 2.5 to 3.0 to 3.5 to 40 to 4.5 to 5.0 1.4 19 2.4 2.9 3.4 3.9 4.4 4.9 Figure () PAl Scores by Ethnic Group Page :>2 October 10. :2000 Data Ana lysis Report Assessment of GeorgiaGain Goals For all three ethnic groups, the 4.0 to 4.4 range was the modal range. The distribution for black employees appears more evenly distributed with 2i)f%1 to ;)()I% of performance evaluations falling in the :3,0 to :3.4, :15 to :3.H, and 4.0 to ,1.4 ranges. White employees received proportionately twice the number of 5.0 ratings as did black employees (7Jjl!;(1 and 4.1'% respectively). Although the numbers are small, black employee received proportionately more than twice the number of low ratings (ratings below :Ul) as did white employees (1.5W% and 0.7:V% respectively). The distribution of PMFbased performance evaluation ratings by ethic group is shown in Table 14. Table 14 Distribution of PMFRatings bV~Jhqif: (~roup iWhite IBlack IOther Total Does Not C Meet ?, 778 1088 3915 1043 ')()Q 38 7902 2169 Far Meets Exceeds Exceeds 151230 104151 4084 25323 8521 490 2232 491 47 259465 34334 2770 Total 183651 118121 48681 3066401 These ratings from theW!)(j to 1!)H!) period are representative of the distribution of the overall employee population (see Appendix B on page 12:3) though slightly skewed to include more performance evaluationsfor black employees. The study shows that :38.;">% of performance evaluations were for black employees in a period when black employees comprise :31 .7%. of the workforce. Other employees comprise 1.5'% of the study data and l.H% of the period workforce. 100% 80% 60% CWhite -Black -Other 40% 20% 0% C DNM Meets Exceeds Far Exceeds Figure 7 - PMF Ratings by Ethnic Group For all three ethnic groups, the Meets rating was the most common rating received and for all three groups the number of Meets ratings was near the 'ideal' distribution. Proportionately, white employees received nearly twice the number of Exceeds ratings as October 10. 2000 Page :t3 Assessment of (;eorgiaC;;:.lin Goals Data Analysis Report black employees (1:3.78% and 7.21% respectively) and nearly three times thenumber of Far Exceeds ratings (1.21 %. and 0.42%. respectively). Page :34 October 10. 2000 Data Analysis Report Assessment of GeorgiaGaitl Goals ExaminatiolIQJ Performance Evalt,t.ottions by Gender This examination of performance evaluations focused on a personal characteristic. However, the distinction between performance evaluation systems was retained. Table 15 shows the distribution of PAl scores during the period for both male and female employees. Table 15 Distribution of Pt\J Scores by G(~nder Male Total 1.0 to 1.4 20 3 23 1.5 to 1.9 49 28 77 2.0 to 2.4 256 222 478 S~nrA 2.5 3.0 to to 2.9 3.4 710 17 541 614 1 -- 1 324 37319 3.5 to 3.9 21957 1 103 43060 4.0 to 4.4 33437 23.345 56782 4.5 to 4.9 22388 9907 32295 5.0 8824 2878 11 702 Total 105182 77 878 183060 These ratings from the !HHi3 to !!)95 period were representative ofthe distribution of the overall employee population (see Appendix 13 on page 12:n, though slightly skewed to include more performance evaluations for female employees. The study showed that 57.5%. of performance evaluations were for female employees in a period when female employees comprised 5G.0% of the workforce. 40% 30% IIFemale -Male 20% 10% 0% 10 to 1.5 to 20 to 2.5 to 3.0 to 35 to 40 to 4.5 to 50 1.4 1.9 24 29 3.4 3.9 4.4 4.9 Figure 8 -PAl Scores by Gender For both groups, the -1.0 to 4.4 range was the modal range. The distribution for male employees appears more evenly distributed with 25% to :30(% of performance evaluations falling in the a.o to :lA, :3.5 to :U), and 4.0 to 4.4 ranges. Female employees received October 10, 2000 Page :35 Assessment of Georgietsein Goals Data Analysis Report proportionately more than twice the number of 5.0 ratings as did male employees (8.:W'X> and ;~.70(% respectively). Male and female employees received proportionately the same number of low ratings (ratings below a.O). The distribution PMFbased performance evaluations by gender is shown in Table Hi. Table Hi Distribl!!ion ofPMFH,f!Jings by (}endeJ: ..IFemale Total Does Not C Meet 4899 ~ on~ 1 207 7902 21691 Far Meets Exceeds Exceeds 149827 20740 1~ !'i~4 1617 1 1531 259465 34 334 2770 Total 178290 128.350 306640 These ratings from the I HHf) to 1999 period were representative of the distribution of the overall employee population. The study shows that 58.1 % of performance evaluations were for female employees in a period when female employees comprise 58.2% of the workforce. 80% -Female -Male 60% 40% 20% 0% L __. . . ._L~~ c DNM ~_ Meets Exceeds Figure D PMF Ratings by Gender Far Exceeds As can be seen in Figure !}, there were only slight differences of approximately 1'% in the distribution of ratings for male and female employees. Page :j(:; October 10,2000 Data Analysis Report. Assessment of C;eOtIjiaC;ain Goals Examina.!1Qfl of Perf~~rntanceEvallUltions by Ag.~ This examination of pe rforrna nee evaluations focused on a personal characteristic. However. the distinction between performance evaluation systems was retained. The employee's age at the time of the performance evaluation was calculated and employees were grouped into decade cohorts. These groupings were not static since an employee might move to the next grouping during the study. Table 17 DititrilHltion of PAl Scores bY.I:\gg 1.0 to 1.4 110'0:; [20's 4 ~ 10 6 3 Total" 23 1.5 to 1.9 15 19 19 21 3 77 2.0 to 2.4 84 147 142 79 25 1 478 Score - 2.5 3.0 to to 2.9 3.4 1 25 223 8281 435 11975 409 10289 213 5340 41 1 317 2 92 1 324 37319 3.5 to 3.9 14 6971 13719 13868 6847 1560 43060 4.0 to 4.4 16 6478 16578 20812 10337 2412 14~ 56782 4.5 to 4.9 1 2365 8349 13439 6489 1538 114 32295 Total Evaluation 5.0 0:; 2~= 57 25031 53941 4796 63780 2745 32074 7663 514 11 183060 Disallowing the very small number of evaluations for the oldest and youngest groups (]()'s and 70'8+), the internal distribution of ratings for employee 40 and older was remarkably similar with only slight differences. For those employees in the 20's and ao's, the number of performance evaluations in the higher ranges (4.5 and above) was less and the number of rating below 4.0 was considerably higher. 100% 80% 60% 40% 20% 10's 20's 30's 40's 50's 60's Figure 10 PAl Scores by Age 70's+ 114.5to 49 04.0 to 4.4 03.5 to 39 t'J30 to 3.4 02.5 to 2.9 02.010 24 015 to 1.9 01.0 to 1.4 October IO, 2000 Page ;l7 Assessment of GeorgiaGain Goals Data Analysis Report The distribution of PMF-based performance evaluations by age is shown in Table IS. Table ]S I2jstribution of PME.Jii.!Jings by Age 10... 20's 30's 40... SO's .60's Does Not C Meet 77 5 12 241 2358 572 1849 732 780 509 115 97 11 13 7902 2169 Far Meets Exceeds Exceeds 280 37600 71075 R6384 52644 10770 712 2 3164 9349 13283 7420 1066 50 197 807 1 08~ 58~ 8f ~ 259465 34334 277C Total 364 43914 84161 103337 61 938 12.137 789 306640 A change can be seen from the earlier PAl distribution in that employees in the modal age category (40's) have the largest percentage of Exceeds and Far Exceeds ratings; and this percentage grows smaller in other groups the further removed the group is from this modal group. 100% 60% 40% 20% 0% 10'5 20's 30'5 40's 50's 60's Figure 11 PMF Ratings by Age 70'5+ Page as October 10, 2000 Data Analysis Report Assessment of GeoQjhlGain Goals .E.~.mJnationof Performance Evaluations!ll'.J~!lure This examination of performance evaluations focused on an employment characteristic. However, the distinction between performance evaluation systems was retained. The employee's tenure with the State at the time of the performance evaluation was calculated and employees with similar tenure were grouped together. These groupings were not static since an employee might move through one or more groupings during the period covered by the study. Table W J2i.trj!?ution of PAl Scores bv Tenure <2 2-4 5-9 1019 20-29 30-39 ~ 1.0 to 1.4 4 2 8 5 4 23 1.5 to 1.9 8 20 24 17 8 77 2.0 to 2.4 101 110 128 91 46 2 478 Score 2.5 3.0 to to 2.9 3.4 293 11 198 322 8606 328 8184 253 6423 126 2806 1 100 1 2 1 324 37319 3.5 4.0 to to 3.9 4.4 758R ;;; ..,."" 9615 9846 11092 14788 995 15654 460t 9228 20E : 548 ') 9 43060 56782 4.5 to 4.9 .... 382 4415 7660 10220 7115 497 6 32295 - Total 5.0 690 28979 1426 34:i62 2728 44940 3946 46560 2657 26596 240 1 594 !l 29 11702 183060 In all groupings, as tenure increased the percentage of lower scores decreased and the percentage of higher scores increased. 100% -80% 60% 40% 20% 0% <2 24 5-9 10-19 20-29 30-39 Figure 12 PAl Scores by Tenure -45to 49 1:]4.0 to 44 I:] 3.5 to 3.9 I:] 3.0 to 34 025to 2.9 020to 24 01.5to 19 010 to 1.4 40+ October 10. 2000 Page :39 Assessment of (;eor:qiaGain Goals Data Analysis Report 'I'he distribution of PMF-based performance evaluations by tenure is shown in Table 20. Consiatent with expectations, a large portion ()!r%) of employees rated with the 'C' code were employees with less than two years of tenure.Howf~ver,iU'% of such perforrna nee evaluations were received by more senior employees. Table 20 Ui~Jrit>~!cltigl1.9fJ~.MJ" }{~ltings by T~:I1~!r(~ <2 24 5.9 10.19 ?O.?q 30.39 40+ Total Does Not C Meet 5453 349 873 448 650 449 644 572 270 315 11 36 1 7902 2169 Me.~ExC"'. Far Exceeds 476 3132 277 51 724 5969 455 55321 7449 532 64412 10 fiflQ 890 37185 6514 538 3185 579 78 29 2 259465 3 2770 Total 56820 59469 64401 77 207 44R22 3889 32 306640 Although not as marked as under the PAl-based system, with increased tenure comes a higher percentage of Exceads and Far Exceeds ratings. For example, employees with 80 or more years of service were 4 times as likely as other employees to have received a Far Exceeds (2.01% versus 0.49%1) and nearly three times as likely to have received an Exceeds (l4.8H% versus 5.5]%). 100% 80% 60% II Far Exceeds II Exceeds mMeets DDNM lie 40% 20% 0% <2 2-4 5-9 10-19 20-29 30-39 40+ Figure 1:3 - PMF Ratings by Tenure Page 40 October 10,2000 Data Analysis Report Assessment of Gf'Ol~fjhIGainGoals Examination of PerformaJ!~!~w.Evaluationsby Pay Grade This examination of performance evaluations focused on an employment characteristic. However, the distinction between performance evaluation systems was retained. 'I'he employee's pay grade at the time of the performance evaluation was determined through a review of employee extracts. Table 21 Distribution of Pl,~JJi9Qr~2';:\ by Pay Grade 1.0 Pay to Grade 14 OA 2 OB 1 n 14 15 16 17 ~ 2 4 18 19 1 20 21 1 22 23 24 1 25 26 27 1 ~8 1 9 e; 10 1 1 32 2 n 39 40 41 42 43 44 45 46 47 48 51 52 53 "" 23 1.5 to 1.Q 3 5 ~ 2 7 5 4 3 ? 2 1fl 1 4 11 4 1 1 :l 1 1 1 1 77 2.0 to ?A 75 B 19 21 12 1q 14 26 2 24 25 11 27 15 9 16 24 11 16 l' 21 7 ." 1 1 478 2.5 to ?Q 69 1R ?7 80 12 7Q 21 74 1H 8 :;3 0 107 'iO 121 R? I' R1 1 7 1 1 1 7 11 1 7 1 1 3.0 to 1.A :l41 159, 1l1' 1718 5~7 1 Qat; 459 Q7 1 Po'q "f;9 1 148 1 '09 77R 4814 1 100 6471 1417 100 15M 1122 1.096 517 778 11 2lfi 2 ~2 144 7~ 143 t; 85 24 ?4 6 1 7 27 1 1324 373 3.5 to 1.0 2 oss 1299 ">AA 1 iss 5h.i 1924 "'id 11 j <1,1 7:;Q 1707 187" 1277 2 1111 1910 55RS ? Q1A 1.?7 ? 'iR1 2 1117 ? 7QR 1 In? 1 >41 7r::.. 17 7 ~8 0 3 ~ 16 21 " 16 4 2<; 7 43060 4.0 to AA 4743 27M RO~ R7Cl 7112 ?Oq 91 i1 1f;Q 7"'7 211 3.25 1979 244R 21M 15?9 40fl1 4~ 1.s: ~ III 7 1." 1 r. 2.8 f'. 127 138 105 477 i'P;, 11Q 417 17" 100 1A 74 4 A 57 ?c1 4Q :'>'1 56782 4.5 to <10 1427 2 :l20 191 21'" 242 771 4~1 44 1 2 10S1 1 ~99 1 OR? 1M ? 14 ?~i ?~O 1 "R 174 1 10 1~ 47 ~q~ 1 27 47 4 R7 1 4 16 17F 34 89 1 6 11; ?7 41 2 '1'7 32295 5.0 Total Evaluations 9~~ 12720 1 1'11; 9 'iM 1n 2 ?'ill ~ 4202 2152 '(':017 18 . 263:l 1 37 1';7 10'I'>',Q 'i 24)4 1"'liM -;:;:528 4821 1SO 5483 4R7 12097 111 f;74R H;~ 14250 554 11677 1.7 1 21'>1 74 11230 ~1 '1i''''70 J 2~f il74R 4590 11157 1 R7" ~ '<11 531 4067 nf 2495 1Rr 1.347 ?1f ?1t;.F o: 283 153 1 '<21 1 50 55" 9F '" 24~ c 14 , 100 1 70 152 2 M 1170 183060 October 10, 2000 Page 41 Assessment of Georgietsein (Joals Data Analysis Report Over 12%1 of performance evaluations were for employees on non-standard pay plans that did not have pay grade identifiers or which had non-standard pay grades. This includes unclassified employees and teachers. Two artificial pay grades were created for this examination to help categorize these employees, Employees earning below S:W,OOO at the time of the performance evaluation were assigned to pay grade OA and those earning $:30,000 and above were assigned to pay grade OB. The number of performance evaluations on each pay grade approximated the number of employees on each pay grade including the very low number of employees on pay grades 18 and 28. Entry-level professional job titles were frequently assigned to pay grades 2f) and 27 and entry-level Caseworker was assigned to pay grade 24. Correctional Officer 2 was assigned to pay grade 2(;. A small number of performance evaluations were prepared for employees on the very high pay grades of 4() and above. These pay grades were generally used for medical professionals. The distribution of scores within each pay grade up to 4;') is shown in Figure 14. Immediately evident is the larger proportion of 4.5 and higher scores on pay grades :Ja and above and the very small proportion of 4.5 and higher scores on pay grades Ld through 16 used primarily for service, maintenance, and entry-level clerical class titles. 100% 80% 60% 40% 20% 0% -4.5 to 49 EJ4.0 to 4.4 EJ3.5 10 3.9 EJ3.0 to 3.4 02.5 to 2.9 02.0 to 2.4 01.5 to 1.9 01.0 to 1.4 Figure 14 .... PAl Scores by Pay Grade Another point of interest is the very high proportion of a.o to ~t4 scores on pay grade 2(i and, to a lesser degree, pay grade 24. This proportion was higher that than seen on any other pay grade. The distribution PMF -based performance evaluations by pay grade is shown in Table 20. Because PMF-based performance evaluations in 1996 were prepared for both old and new pay grades, only data from lH97 through 19~m were included. Page 42 October 10,2000 Data Analysis Report Assessment of GeorgiaGain Goals Since over l()');(. of evaluations were for employees on non-standard pay plans which did not have pay grade identifiers or which had non-standard pay grades, two artificial pay grades were created for this examination to help categorize these employees. Employees earning below $;30,000 at the time of the evaluation were assigned to pay grade OA and those earning S:m,ooo and above were assigned to pay grade OB. Table 22 Distribution of PM f2Jtatings by Pay Grade Pay Grade OA os 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 25 26 Total Does Not C Meet 456 66 292 68 27 14 91 25 335 93 423 145 653 190 96 30 2339 319 451 170 640 248 283 108 186 50 96 30 58 21 39 19 30 8 15 2 10 1 7 1 6527 1608 Far Meets Exceeds Exceeds 8 181 804 68 8810 2410 422 1 708 61 2 3165 116 6 11633 381 13 14097 1 196 48 21058 2973 162 4762 831 67 41019 4750 189 11 848 2075 139 19558 3492 163 10812 2371 118 6382 1864 141 3925 1326 120 2318 801 63 1 784 704 77 998 490 89 251 114 16 300 136 19 93 27 10 2 2 1 1 172 702 26927 1 Total 9575 12002 1812 3403 12455 15909 25036 5786 48616 14683 24101 13692 8623 5497 3261 2623 1615 398 466 138 2 3 1 209697 The number of ratings on each pay grade approximates the number of employees on each pay grade. After GeocgiaGain implementation, entry-level professional job titles were are frequently assigned to pay grades 12 and 1a. Entry-level Case Manager was assigned to pay grade 11. as was Correctional Officer. An extremely small number of performance evaluations were prepared for employees on the very high pay grades of 2:3 and above. Medical professionals were assigned to another pay plan and are reported under pay grade DB. The distribution of performance eva luation ratings within each pay grade is shown in Figure 14. Immediately evident is the very small proportion of Exceeds and Far Exceeds ratings on pay grades 05 through 07, which were primarily used for service and maintenance jobs. Also, the proportion of Exceeds and Far Exceeds ratings increases beginning on pay grade 12 and continuing through pay grade 19. This continuing increase in the number of Exceeds and Far Exceeds ratings reached then surpassed the 'ideal' distribution by pay grade 15. October 10. 2000 Page 4:3 Assessment of GcorgiaGaill Goals Data Analysis Report Pay grade 11, used for Correctional Officer and Case Manager, continued the pattern observed previously of much smaller than anticipated numbers of high ratings. 100% 80% 60% 40% 20% 0% Figure 15 PMF Ratings by Pay Grade October 10, 2000 Data Analysis Report Assessment of GeorgiaGain Goals This examination of performance evaluationsfocuaed on an employment characteristic. However, the distinction between performance evaluation systems was retained. The employee's annual salary at the time of the performance evaluation was determined through a review of employee extracts. ~ !l;fiOK $70K+ Total Table 2:J Distribution of PAl Scores by Salary Range 1.0 to 1.4 12 8 3 2~ 1.5 to 1.9 31 33 10 1 1 1 77 2.0 to 2.4 248 168 43 17 2 478 ScnrA - 2.5 3.0 to to 2.9 3.4 664 21 153 467 12304 149 2622 36 748 5 279 1 81 2 n? 1 324 37319 3.5 to 3.9 17R12 18288 5312 1 174 320 73 81 43060 4.0 to 4.4 18570 23726 10101 2963 871 248 303 56782 4.5 to 4.9 7 2~2 13115 7367 2925 927 279 450 32295 5.0 241R 439t; 3030 1087 442 138 192 11 702 Total 68140 72504 28637 8951 2847 820 1 161 183060 During the H)!):J to 1995 period represented by the performance evaluations shown in Table 2:3, the average annual salary for state employees was approximately $25.000. The evaluation data appears to be consistent with the average salary for employees. 40% --O--<$20K --o-$20K -Q--$30K 30% -6.-$40K -D-$50K - -0- $60K --<>-$70K+ 20% 10% 0% 10 to 1.5 to 2.0 to 2.5 10 3.0 to 3.5 to 4.0 to 4.5 to 5.0 1.4 1.9 24 2.9 3.4 3.9 4.4 4.9 Figure 16 PAl Scores by Salary Range Figure 16 shows a considerable difference in the distribution pattern of performance evaluations scores based on salary range. For example, the modal score received by October 10, 2000 Page 4;") Assessment of Geor:giaGain Goals Data Analysis Report employees making less than $20,000 (about :37'7'0 of the population) was in the :U) to 8.4 range while the modal score received by all those making over $50,000 (2.;)17'0 ofthe population) was ,I.;') to 4.n. In the middle incomes of $20,000 to $4H,OOO (60.5'% of the population), the modal score received was 4.0 to 4.4. No salary range group had :3.5 to :U) as a modal range of ratings. Increasing strength in modality is seen as income rises. For the under $20,000 group, the modal score was received by :31 %1 of the group; at $:JO,OOO, the modal score was received by :35% of the group; and at over $70,000. the modal scores was received by nearly :3n% of the group. The distribution of PMFbasod performance evaluations by salary group is shown in Table 24. During the period H)9fl to 19!)9, the average salary of employees rose to nearly S2n,OOO and the evaluations show a similar increase. Table 24 I)istribution of PMF Ratings by Salary Rfmge d20K $20K $30K !.40K $50K 560K 570K+ Total Does Not Far C Meet Meets Exceeds Exceeds 4047 519 68918 3707 214 2569 1005 117949 14 116 81 756 425 44087 8508 706 272 142 17 392 4458 447 99 45 6.211 1918 262 57 11 2238 81::\ 1 102 ~ 2670 814 1 7902 2 259465 34 334 277 Total 77 405 136452 54482 22711 8535 3.256 3799 306640 100% 80% 60% --<>--<:$20K --O-$20K --{r-$30K -6-$40K -o-$50K - {J- $60K --+--$70K+ 40% 20% 0% c DNM Meels Exceeds Far Exceeds Figure 17 -PMF Ratings by Salary Range Page 4t:> October 10.2000 Data Analysis Report Assessment of GeorgiaC'ain Goals Unlike the distribution under the PAl-based system, under the PMF-based system the modal rating for all salary groups was the same Meets (see Figure ]7). All salary ranges followed the same distribution model and tracked to some degree the 'ideal' distribution. There was some difference, however, in how close each group came to the 'ideal' distribution. For example, beginning at under 520,000 the Meets ratings accounted for 8~)'~i. of performance evaluations hut that percentage dropped steadily as income increased until at $70,000 Meets ratings accounted for only 6WX. of evaluations. Similarly, the percentage of Exceeds ratings increased as income increased from a low of 4.8'J(. for employees earning under $20,000 to nearly 25(~1i {(II' those earning $(;0,000. Asincome increased, the distribution of performance evaluation racings changed from one with too many Meets ratings for employees earning under $20.000. to one that was rich in Exceeds and Fa r Exceeds ratings for those employee earning above $tiO.OOO. I n the S:JO,OOO salary range, the distribution was closest to the 'ideal' where 81(% of performance evaluations contained Meets ratings and 15% contained Exceeds ratings. October 10. 2000 Page 47 Assessment of Georgieilein Goals Data Analysis Report Examination of Performance Evaluations b'y Job Title This examination of performance evaluations focused on an employment characteristic. However, the distinction between performance evaluation systems was retained. The employee's job classification at the time of the performance evaluation was determined through a review of employee extracts. Table 25 l}is~[ilmJjonof PAl Scoresf{)t 2!5J'vlost Frequently Evalm\fed C:lass 'fiUes 1,0 to Class Title 1.4 Accountina Tech 1 Caseworker Prin 3 Caseworker Sr 1 Clerk Admin k Prin k Sr 4 correcuonat Ofe 1 1 Correctional Ofc 2 Enuioment Operator 2 Health Svcs Tech Health Sves Tech Sr Health Svcs Tech Lead Housekeeoer Houseoarent 2 Human Svcs Tech <::., Human Svcs Provider Instructor LPN Sr Nurse Sr IComm/Cln\ Probation/Parole Ofc 2 Secretarv. Prin Secretarv Sr Secretarv/Tvnist Social SVGS Soec 1 Technical lnsuuctor Other Job Titles 1 H1 Total Evaluations 23 1.5 2.0 2.5 3.0 3.5 4.0 4.5 to to to to to to to 1.9 2.4 2.9 3.4 3.9 4.4 4.9 5.0 Total 6 9 201 395 647 411 137 1806 2 10 15 143 306 94 .187 311 2920 10 26 511 1094 1 290 1 7 966 204 5414 7 9 300 462 729 581 221 2309 10 3" 543 672 807 429 122 2618 2 13 20 402 480 775 414 155 2265 7 59 3010 483 23 3 3586 1 5 97 6212 4761 660 50 2 11 788 7 4 677 185 1005 43 28 1949 4 10 t;t; 749 474 334 61 19 1706 15 53 1349 1 516 1757 629 156 5475 3 1 12 15 18 240 607 343 406 ~ 632 446 102 23 1669 1670 1 1 4 3 5 24 18 8 542 340 233 402 487 372 307 650 621 125 415 350 1i 1427 2003 1665 1 6 13 19 43q 402 436 464 609 760 263 451 : 69 187 ~ 3 190 360 67R 524 213 1971 1 3 23 257 867 1214 251 , 2622 3 6 104 250 525 419 20 1 514 4 13 35 554 1004 1 971 1763 93 6277 2 17 27 743 1 135 1759 992 5046 5 18 38 496 737 1 168 590 3144 = ! 815 147 476 237 4 40 ~ ~, 77 1324 37319143060 56782 32295111 2 101 105996 183060 The twenty-five class titles shown in Table 25 bad the largest. number of performance evaluations and comprise over 42%. of all performance evaluations. This list is consistent with the list of the most populous class titles. Two of the class titles (marked with an asterisk "") were used exclusively for unclassified employees. The remaining class titles (over 2,(00) were grouped into Other -Iob Titles. Within the grouping of selected class titles at least five class series or families are represented ~ Clerk, Secretary. Correctional Officer. Health Services Technician, and Caseworker. Page 48 October 10. 2000 Data Analysis Report Assessment of GeorgiaGiJj1J noah Figure 18 displays the percentage distribution of PAl scores for each of the selected class titles. Immediately evident is the very high proportion (over 8:3% of Correctional Officer 2 performance evaluations with scores of :3.0 to :~.'1 followed closely by a similar high proportion (over 5211.,) of Correctional Officer 1 performance evaluations with this same rating. Certain low skill class trtles (Housekeeper, Houseparent and Health Services Technician) also showed large proportion of scores below :>.5 and smaller than average proportion of scores above 4.5. Strong differences existed in the distribution patterns. For example, compare the Correctional Officer performance evaluation pattern with that of Secretaries where approximately 75% of performance evaluations were for 4.0 or higher. Accounting Tech 1 Caseworker, Pnn Caseworker. Sr CleCrkle,rkA,dPmrin Clerk, Sr IE:;SE~::i~EE::=:;~I:;St~ ::::::~:z: Correctional Officer 1 CorrectIOnal OffICer 2 Equipment Operator 2 Health svcs Tech Health svcsTech Sr ' Health Svcs Tech, Lead Housekeeper Houseparent Human Services Tech, Instruclor LPN, Sr Nurse. Sr(Commlelnj Probation/Parole Ofc 2 Secretary, Prin Secretary, Sr SecrelarylTypist Soctal Service Spec 1 Technicallnstruclor Other JobTTiotletasl B===:z=:c====::::r======~~ 0% 20% 40% 80% 01.0 101.4 015101.9 02.0102.4 2,5102.9 103,0 10 3 4 03,5103,9 104 ,0 10 4 ,4 -4,5104,9 -5,0 100% Figure 18 PAl Scores for 25 Most Frequently Evaluated Class Titles The distribution of PMF-based performance evaluations by job title is shown in Table 2fi. Because PMF performance evaluations in 199(i were prepared for both new jobs titles and old class titles, only l!}!}7 through 1999 data were used. During GeorgiaGajn implementation, employees were moved individually to new job titles by agencies following a review of job duties. Part. of the design of the new classification structure involved combining of some class titles that had been part of traditional promotional hierarchies into fewer new job titles. This occurred predominantly in class series or families with large numbers of employees. For example. the number of titles was reduced in the Correctional Officer, Probation/Parole Officer, Clerk, and Secretary series. At the same time. hundreds of new job titles were created in other areas to permit clearer definition of compensation criteria and performance measures. For example, the number of job titles created in the social services area grew and the Caseworker series was expanded to include many different job titles based on specialty and service areas. October 10,2000 Page ,m Assessment of GeorgiaGain Goals Data Analysis Report Agencies were also encouraged to develop new job titles within the standard classification plan to accommodate the thousands of unstructured non-merit class titles identified only by a title and code. Several agencies used the GeorgiaGain project to develop new jobs or move employees into common jobs within the state classification plan. Table 2() Uitrih!!tjg.ll 0 f PMF R!LtingJ~lr~~M(l!J:~rS~Sl ue n t Iv EVl!!!!!lt~g....!!9h.Ti!Je--,'S Job Title Acet Paraorotessronat Clerk 1 General Clerk 2 General Correctional Officer DOL Services Snec EauiD Ooerator 2 EQuiD Onerator 3 Fam Ind Case Mar 1 Health Svcs Tech Housekeener Houseparent Instructor .IIIV Corr Officer 1 l P N !Innatient Svcs) Probation Officer 1/2 Proaram Assistant Public Health Nurse Secretarv 1 Secretarv 2 rseroaent (GDC\ Social Svcs Case Mar Social Svcs Prov 1 Social Svcs Tech Social Svcs Tech Sr Troooer 1st Class lttes Total Does Not Far Total C Meet Meets Exceeds Exceeds Evaluations 59 97 91 I 1252 31 4 18 2044 30 2032 12 2184 65 19558 42 1462 8 20P 514 159 321 1385 95 148 22 2657 11 19 ~ ') 22262 " 1634 1 2.173 1 3 1.225 186 4 1419 213 86 4233 398 3 4933 174 25 4610 80 4889 19 13 1415 34 2 1483 89 32 2175 53 7 2356 21 12 1720 87 1840 496 5 2501 56 6 3066 21 8 1 526 40 1595 64 17 1 741 221 7 2050 159 34 4639 810 12 5654 36 5 1609 109 3 1762 117 23 2164 288 12 2604 93 25 2761 870 70 3819 42 R 1 341 206 1 1598 122 76 3435 473 1 4107 66 50 1675 211 17 2019 38 18 1410 108 6 1580 42 43 1755 153 15 2008 1 II 6527 13 9:17 1608 1 164 100 172702 332 26927 28 1538 5: 1?t; RQt; 209697 The twenty-five job titles shown in Table 2() accounted for 40% of all PMF-base performance evaluations in 1HH7 to 1~)9H. Some job titles included here were not part of the twentyfive most frequently evaluated class titles under the older system. Department of Labor (DOL) Services Specialist, and Trooper 1"1 Class ,vere new' entries while others. such as -Iuvenile Correctional Officer, were new job titles previously reported under other class titles, such as Correctional Officer. Performance evaluations for employee in Housekeeper, Houseparent, and Health Service Technician jobs showed lower than average numbers of Exceeds and Far Exceeds ratings. Secretary 2 and Trooper 1st Class showed higher than average number of Exceeds ratings and less than 'ideal' number of Meets ratings. In the job title -Iuven ilo Correctional Officer, more than IBYu of performance evaluations were not true performance evaluations, but were 'considerations' of performance. Page 50 October 10, 2000 Data Analysis Report Assessment of Geotgietlntr; Goals Cler 1,General Clerk 2,General Correctional Officer DOL Services Spec E<1uip Oyerator 2 EqUifl Operator 3 Fam Ind Case Mgr 1 Health SIICS Tech Housekeeper Houseparent Instructor JUII Corr Officer 1 LPN (Inpabent Svcs) 1121Iii~ili~~gll~~~II; Program Assistant PPrOubbOC DOE DOL OCR DOT OPS TA- GFC P&P OtherTotal LO Sch Figure 21 PMF Ratings fill' Larger Agencies October 10, 2000 Page 55 Assessment of Georgietlein Goals Data Analysis Report Examination of Terms & C~mditions Ratings This examination of PMF ratings for the period HmG to W99fix:used on an internal aspect of PMF-IHIS(~d performance evaluations. During the development of the Performance Management Process, concerns were raised about the possibility of employees performing adequately in the areas related to job, individual. and statewide responsibilities but failing to meet basic terms and conditions of employment. A separate 'Terms & Conditions' section was added to the performance management form and during each evaluation period, supervisors evaluated each employee in two areas- Responsibilities and Terms & Conditions. Both values were captured in the GEMS database. The earlier examinations have been limited to the rating assigned to the responsibility area. >- ;e~ C) til r;; c o ; ~ 0.0:: til 4) 0:: C Did Not Meet Meets Exceeds Far Exceeds Total - Table 29 Distribution of Terms & Conditions Rating - TArm!'1. Ratinn Needs C Did Not Meet Improvement Meets 7902 7902 622 794 1251 328 1312 257.825 14 3432c 1 2769 950 1 621 29616' Total 7902 2169 259465 34334 2770 306640 During the four-year period covered by this study that the PMF-based system was used, as few as :328 salary increases were withheld based on Terms & Conditions rating when otherwise warranted based on the rating of job and individual responsibilities. 100% 80% 60% 40% 20% 0% lOt-Meets ,lIIt-NI -t-DNM .Ot-e 0.00% 0.00% 100,00% 13.55% 28,68% 0.00% 99,37% 0.13% 0.00% 0.00% 0.00% 0.00% Figure 22 - Terms & Conditions by Responsibility Rating Figure 22 examines Terms & Conditions rating mapped to the corresponding responsibility rating. Nearly fHU)% of employee receiving ratings of Meets or higher on job and individual Page SCi October 10, 2000 Data Analysis Report Assessment of GeorgiaGain Goals responsibilities, also received the highest Terms & Conditions rating of Meets. Two-thirds of those receiving- a Does Not Meets rating in the Terms & Conditions section also received a Does Not Meets rating on reeponsibi.litiee. Observations 1. Changes in performance evaluation patterns occurred following GeorgiaGain implementation. Before GeorgiaGain 48.7%1 of evaluations indicated that the employee 'frequently' exceeded requirements of the job and (j,4%1 'consistently' exceeded job requirements. LInder the Performance Management Process developed as part of Ceotgietlein, only 11.2%, 'Exceed' expectations and O.!)(X, 'Far Exceed' expectations. 2. Ratings under the Performance Management Process are more equitably distributed as evidenced by a comparison of performance evaluations by gender and ethnic group. Total equity has not been attained; white and female employees still receive proportionately more of the higher ratings. a. Implementation of the Performance Management Process has reduced the influence of age and tenure on performance evaluations. Increased age and tenure still impact positively on P]\lF ratings through age 50; but age does not positively influence ratings after that point. 4. There continues to be strong influence of compensation-related factors in the distribution of performance evaluation ratings. Consistent patterns of higher ratings are seen for employees on higher pay grades and with higher annual salaries. More than 2;)% of employees on pay grades 15 and higher received ratings of Exceeds or Far Exceeds as compared to less than 5% of those on pay grades 7 and below. 5. There is strong variability in the distribution of performance evaluation ratings by job title. Performance evaluations for employees in jobs such as Housekeeper, Houseparent, and Health Services Technician contain less than :VYi. of higher ratings compared with 25% for employees in other jobs such as Secretary and Trooper lq Class. G. With notable exceptions. the high degree of variability seen in the previous system by agency bas been greatly reduced. 7. The universal evaluation of every employee against the Terms & Conditions standards may be unproductive. Over D~)'Y.. of performance evaluations indicate the highest rating possible in this area. Only :~28 increases may have been withheld solely on the basis of Term & Conditions when otherwise warranted based on responsibility rating. October 10.2000 Page 57 Assessment of Ceorgietlein Goals Data Analysis Report Page 58 October 10, 2000 Data Analysis Repott Assessment of GeorgiaGaiTl Goals Report 2 (GeorgiaGain Goal 4) Rewa rei Best Performers The performance evaluation data was linked to personnel transactions for salary increases and adjustments recorded in the Georgia Employment Management Systems (GEMS). For purposes of establisbing linkage between a performance evaluation and a personnel transaction, a five-month window beginning on the first of the month in which the performance evaluation was performed was selected. A window of this size was necessary to establish linkage between performance evaluations completed as early as -Iune lor increases granted in October. Linkage was established between performance evaluations and any proximal salary increases. Finally, an examination ofperformance evaluations ratings associated with selected types of salary increases was performed. The examination of performance evaluations and salary increases was limited to three types of salary increases and adjustments - performance-based increase, salary adjustment, and crir.erion-based adjustment. The granting of these increases is associated with some measure of agency discretion; and separate transaction codes were established in GEMS to record them. Other types of increases, such as across-the-board, assignment of class/job to another pay grade, and teacher salary adjustment were not included in the study because of the general applicabi lity of these increases and lack of agency discretion. - , Performance ! Group High Average Low Table so 1)<2xi.~1Im!!m;g ..(;r(}tm.s PAl-based System Scores PMF-based System Ratings 45 to 50 Exceeds and Far Exceeds 3.0 to 4.4 Meets 1.0 to 29 Does Not Meet In order to facilitate comparison, delineation between High, Average and Low Performance Groups was constructed. For performance evaluations under thePMF-based system,Low was defined as the Does Not Meet rating, and Average was defined as the Meets rating. The ExceNls and Far Exceeds ratings became High Perfor-mance (houp. In addition, the Low Performance Group included ratings of Does not Meet in the Terms & Conditions area regardless of the rat.ing shown for job and individual responsibilitios. By policy, performance-based increases were withheld for ratings in the Low Performance C;roup. October 10. 2000 Page ;')9 Assessment of C;eorgiaGain Goals Data Analysis Report For PAl based performs nee evalu a tions, Low was deli ned as a score of less than :UJ and therefore not eligible for salary increases. High was arbitrarily chosen as 4.5 and above. Selection ofR~!:~(l,r.d~.J!)rAnalysis Table :J 1 displays the number of transactions reported for the three selected types of increases. The numbers in the Total 'I'ransactions column for each year are consistent with expectations, even for the exceptional years ofWH:J and wm;. Salary increases for state employees were reinstated in May UlfJ:) following a two-year freeze. In W!)(i, some employees were eligible for a salary increase under the preC;corgiaGain system on their regular 'anniversary date' in -Ianuary through -Iune as well as another salary increase in October along with all other employees, Also in U)!)(), large numbers of salary adjustments were made before and immediately after Georgieclein implementation to bring salaries for some employees more in line with similarly classified employees. Table :31 Distribution ofSelegted Increases bv Year Criterion- Based Year Adjustment ~t 1995 1996 81 1997 361 1998 1999 1520 1728 Total 3690 Performance- Salary Based Total Adiustment Increase Transactions 337 38710 39047 614 64495 65109 1 152 67532 68.684 10603 103025 113709 9691 67278 77330 10367 11974 67122 66043 79009 79.745 44738 474205 522633 Table ;)2 restates in summary the number of performance evaluations available for analysis from Table 5. Table az Distribution of Perf()rmance Evaluat.ions bv Year Evaluatons under PAl S stem 44771 68193 70096 Evaluations underPMF S stem Total Evaluations 183060 Performance evaluationsfor the period April 1HHa to October HH)f) were linked with the selected salary increase and adjustment actions in a five-month window following performance evaluations. A new dataset was created containing both the performance evaluation and the associated salary increases or adjustments. Page ()O October 10.2000 Data Analysis Report Assessment of GeorgiaGaitl Goals Differences exist between the number of performance-based increases each year show n in Table ;31 and the number ofperformance evaluations shown in Table ;~2. This difference can usually he explained by considering: 1) some performance evaluations documented low performance and. therefore. these employees were not eligible for increases; 2) some employees were above the range of their assigned pay grades and. therefore, were not eligible for increases even though a performance evaluation was p rcpa red; and ;3) performance evaluations may have been prepared for other types of increases and recorded in the system. In W94. U)!)5 W!)7, 1998, and I!)!)!) the difference between the number of performance evaluations and performance-based increases (approximately 1,800 to 4,:300 'extra' performance evaluations each year) can be accounted for under the above rationale. In 19~);L the number of 'extra' performance evaluations was over (),OOO. or nearly 14%) of the total. This difference can be explained by considering that during normal operations the majority of perforrna nee evaluations are dated the month before the effective date of the salary increase (see Figure 2 and Figure :3 for distributions of pe rforrna nce evaluations by period). Consequently, most of the performance evaluations prepared in December t!)!);J were for -Ianuary UHl4 salary increases. As seen in Table G, over 4,800 performance evaluations were prepared in December H)!)a. Because the data for 199:J did not contain twelve months of performance evaluations and twelve months of salary increases (actually nine and eight months respectively), the December performance evaluations constituted excess performance evaluations that were not offset by -lanuary salary increases. InW~)(), the general condition was reversed; over G,OOO more performance-based increases were granted than the number of performance evaluations. Again, some portion of this excess can be explained by considering that -Ianuary IH9(; increases were given based on December WH5 performance evaluations (over 5,000, see Table (;). In addition, the number of perforrnance evaluations reported for -Iune through October IHHG may be u nderstated (see Table 7). Duplicate performance evaluations were eliminated from the analysis database. Some performance evaluations prepared prior to -Iune for salary increases in .Ianuary through -Iune were reused to document performance for the October HHH, salary increase. There is an unexplained difference in the total number of performance-based increases granted during the period (474,205) and the number of increases that can be linked to performance evaluations (4(;7,70(;). This difference of (),4m) is approximately 1.4% of all salary increases. Examination of Relationship between Perfonncm~eand Increases On the following page, Table aa displays the distribution of salary increases and adjustments linked to performance evaluations under the PAl-based system and Table ;)4 displays this data for performance evaluations under the PMF-hased system. Criterionbased adjustments did not exist prior to H)!)(>. The number of salary increases and adjustments reported pel' Performance Group may exceed the number of performance evaluations. Frequently, both salary increase and salary adjustment transactions were processed in the five-month window associated with the same performance evaluation. Also, tabulations are shown of those performance evaluations where no performance-based increases could be identified, even though the performance October 10, 2000 Page ()1 Assessment of Geol'ghtGain Goa ls Data Analysis Report evaluations would otherwise warrant a salary increase. This situation occurs when the employee is already above the maximum of the salary range or when the employee leaves state service between the preparation of the performance evaluation and the increase delivery date. Table :J:J Ped(ml1flnc~:J~.Y1Ih!!I.tions versus lncl'(~!:!l-'es under PAl Svstem CriterionPerformance based Table :34 Perlm:mance Evaluations verB.!!:':i Increases underPMF Sytem CriterionPerformance based Salary Given No Increase Identified Total Table :35 displays the percentage distribution for salary increases and adjustments made under the PAl-based system. Table an displays the percentage distr-ibution fill' salary increases and adjustments made under thePMF-based system. In both cases. the percentages may add to more than 100% because the number of transactions processed exceeded the number of evaluations. Table :)5 Percentage Qj"~tributionof In(~LellSes under 1>;\1 System CriterionPerformance based Page ()2 October 10. 2000 Data Analysis Report Assessment of GeorgiaGain Goals Table ;Hi PerCn,!!!Fl,ge DistributiOll~lfJm~reasest{H' PMJL.S. vstem Criterionbased Salary Total Irnmediately evident is the marked difference between the volume of salary adjustments following Gcot:qiaGain implementation versus before implementation (7.8;~%, versus O"l;Yii.). In addition, st.ricter adherence to the criteria fill' withholding performance-based increases can be seen under the PMF-based system as evidenced by the ~HLHI'X. of Low Performers without performance-based increases versus 8~U)(j'% prior to GeorghlC;;lin. Discretionary increases, such as salary adjustments, were much more frequent fill' employee in the Average Performance Groups than those in the High Performance Group. Employees in the High Performance Group were only 2;",'}f, more likely to receive salary adjustments than those in the Low Performance Group. In the criterion-based adjustment category. employees in the Average Performance Group received more increases than did those in the High Performance Group. 1. The number of discretionary salary adjustments processed by agencies has increased per year over 1000%, following Geot:qiaC;ain implementation (see Table ;n). .lust over 2,000 salary adjustments were processed in WH;~, WH4, and WH5 combined. In each year following Georgtetlein implementation, thefewest number of salary adjustments processed was more than four times the combined pre-implerncntat.ion number. 'J Differences do exist between the proportions of discretionary increases granted to employees in the High Performance Group versus those in the Low Performance Group. However, employees in the Average Performance Group received proportionately more increases than those in the High Performance (~roup (see Table :~(j). A similar distribution of salary adjustments existed prior to GeorgiaGainimplementation. but with considerably smaller numbers of transactions. :3. The actual number of salary adjustments given to reward top performers cannot be discerned. Since ~)2'Yi) of all salary adjustments were given to employees in the Average and Low Performance Groups, this type of action had become a universal salary increase code used for any salary increases other than performance-based increases. October 10, 2000 Assessment of GeOt~giaGain Goals Data Analysis Report Page (;4 October 10, 2000 Data Analysis Report Assessment of Geol:c;iaGain Goals Report :3 (Ceotgietlnin Goal 7) Help Managers Document Poor Performance Performance evaluations under the two evaluation systems were reviewed and those performance evaluations with low scores or rat.ings were identified. Low was defined as a score of 1.0 to 2.!) under the PAIbased system and a rating of Does Not Meet under the PMFbased system. Comparisons of these two groups to their parent groups were performed. Additionally, the internal make-up of the two low performing groups was examined using the fallowing personal and employment characteristics: Age, Ethnic group, Tenure. and Annual compensation range. During the period from April H)na to October HHH). two performance evaluation systems were used. The older Performance Appraisal System using the Performance Appraisal Instrument (PAl) existed until December 19!)5. A total of 18:U)(;O performance evaluations were available for analysis for the period from April H)!);{ to December W!)5. In -Ianuary l!HW, the State implemented a new Performance Management Process that used the Performance Management Form (PMF). A total of :m(;,(>40 performance evaluations were available for analysis for the period from -Ianuary HWG to October ]mw. Performance evaluations with low scores or ratings were extracted. Low was defined as a score of 1.0 to 2.9 under the PAl-based system and a rating of Does NotMeet under the P.MF-based system. This is the same structure used in previous reports (see Table :30 on page 5H). The Low Performance Group also included performance evaluations where Does Not Meet was received on the Terms & Conditions section of the PMF, regardless of the rating on the responsibility section. Table S? displays the number ofperformance evaluations with low score or ratings. Under the PAl-based system, supervisors could use a wide range of scores to indicate the level of unsatisfactory performance. Table Sd displays the detail distribution of low scores within the score ranges used in previous reports in this paper. Cktober 10,2000 Page (;5 Assessment of Gt'orgia(;ain Goals Data Analysis Report Table 87 J!ow Performance Evaluations lw Evaluatio!!...fuJ:::l&nl PAl PMF Total Total low Low Evaluations Evaluations Percentaae 183060 1902 2 1f\qll 1.04% 0,71% 489700 407111 0,83% 'l'able :38 fle t }.DlnfJ~Q,1.Y-.B~~Ol:::"L,P nderr>AI System Discrete Score 2,9 2.8 2,7 26 2.5 2.4 23 2,2 2,1 20 1,9 18 1,7 1.6 15 1.4 1,3 1.2 1,1 10 Total Percent of low Count Scores ::t 16,19% 2103% 241 1? fi7 0j" 195 10,25% 180 9.46% 107 5,630/. 116 6.10% 64 3.36% 52 2.73% 139 7,31% 21 1,~ 22 1.1 17 0,891< 6 0321< 11 0.58% 4 0.21% 7 0.37% 4 0210/. 'J 0,11,.{, 6 0,32% 1902 100,00% Score Grouping Count Percent of low Scores 25 to 1,324 69.61% 2.9 2.0 to 478 25.13% 2.4 1.5 to 77 4.05% 1,9 1,0 to 23 1,21% 1,4 Total 1902 100,00% Examination of Low_.f~EformanceEvaluation Scores and Rating\'! Based on the data presented in Table :37, more low scores proportionately were given on PAl-based performance evaluations than low ratings on PMF-based performance evaluations (1.04'% versus 0.71 (!,{,). Low scores and ratings represented a very small portion of all performance evaluations. Under both performance evaluation systems, the percentage fell below the ideal distribution presented in Table 12. The summary of PAl-based performance evaluations presented in Table 88 indicates that supervisors were capable of discerning performance differences of one-tenth point, even for Low Performers. Also of interest is that over] ()% of the performance evaluations that fell below the minimum required to authorize a salary increase (a score of :LO) did so by only one-tenth of a point. Page (;() October 10, 2000 Data Analysis Report Assessment of Ceotgieciein Goals Exami!lation of Employe~~Profilesof Low ~L<:;~tre and Ratings As evidenced by the data shown in the earlier analysis in Report 1 GeorgielGain Goal ;) ,:\ssure FI.Lir J'.!TI()fml!lg:.<,:j':valuJ.IJiql!$, the influence of some personal and employment characteristics reflected in the distribution of PAl scores appear to have changed under the PMFhased system. This section of the report will revisit some of the personal and employment characteristics, but limit the analysis to those employees receiving low performance evaluations. Of particular interest are the distributions of ratings and scores by ethnic group, tenure, and compensation range. A1)illysis of I.mv Scores and F?iJUngs h,I{ E'thnLcGrollp Using the data displayed in Table l:l and Table 14 on pages :l2 and :38, the following presentation was developed, 1DO"!" 80% 60% I II I PAl Scores PMF Scores. OWhite -Black I:JOther 40% 20% 0% Low PAl Scores Total PAl Scores Low PMF Scores Total PMF Scores Figure 2:3 Comparison of Low Scores and Ratings by Ethnic Group Between lHHa and 19HH, the proportion of black employees and other employees increased in the overall population of state employees (see Appendix B) and their representation in the total number of performance evaluations also increased. However, the proportion of performance evaluations with ratings received by black employees decreased more than ten percent (from 5:t.(>8lJ() of low evaluations to 44.55%). l)espite the decreased numbers of white employees in the overall population of state employees, performance evaluations with low ratings received by white employees now comprise the plurality of low performance evaluations. Analysis of Low Scores cHId Ret ir:JJJs by Tenure Using the data displayed in Table Ltl and Table 20 on pages :m and 40, the presentation shown in Table 24 was developed. October 10, 2000 Page (;7 Assessment of Georgieileir: Goals Data Analysis Report Performance evaluations with low ratings were prepared on a much broader basis under the PMFbased system than under thePAI-hased system. The representation of more senior employees, particularly those with more than 10 years of service, increased by as much as 7tl~) in some groups. 40% 30'1'0 20% r\ddlll'-ltml 'urpp<.)r!iog di~L:lJl for this KTaph nm ~H:< ftlttn,_t ,)"() i,og(' IGO PAl Scores I PMF Scores --'_.'--~.'--'-""""---""I. -<2 112-4 05-9 010-19 020-29 113G-39 -40+ 10% 0% low PAl Score low PMF Rating Figure 24 ~ Comparison of Low Scores and Ratings by Tenure Analysis of Low Scores and I?atings by Salary I?ange Using the data displayed in Table 28 and Table 24 on pages 45 and 4G, the following presentation was developed. 60% PAl Scor I PMF Scores -<$20K -S20K. C$30K C$40K 40% 0$50K D$60K B$70K+ 20% 0% low PAl Scores low PMF Ralings Figure 25 Comparison of Low Scores and Ratings by Salary Range Page G8 October 10, 2000 Data Analysis Report Assessment of Georgi,JGain Goals A noticeable reduction was seen in the proportion of performance evaluations with low ratings prepared for employees earning less than 820,000 annually. During this same period, the average salary of state employees increased to just under $:m,ooo as the resu It of annual salary increases and extraordinary steps taken by agencies to increase the salaries of the lowest paid employees. Obs(~rvati()ns 1. Table 38 presents an interesting counter to one of the arguments raised initially to the rapid development of a pay-for-performance compensation system. Concerns were expressed that managers could not discern tine levels of performance sufficiently to support variable pay increases. However, it appears managers were able to distinguish very fine degrees of difference between employees receiving performance evaluations with low ratings. 2, The very small proportion of performance evaluations with low scores and ratings may indicate that mangers had some difficulty preparing such performance evaluations. This appears to be more true under the PMFbased system developed as part of GeOT)-!iaGaill. The proportion of performance evaluations with low scores fell from 1.04(% to 0.71%). The 'ideal' proportion of low ratings or scores is 2% to :3%. :3. The assignment of performance evaluations with low scores or ratings was more equitable under the new Pl\-1F system. The proportion of performance evaluations with low scores or ratings received by black employees dropped from 54'% to 44(yil and was more in line with their representation in the overall employee population, 4. Under the PMFbased system, managers assigned more low ratings for senior employees, particularly those with 10 to :3H years of service. The proportion of low ratings assigned to performance evaluations for employees with 20 to 2~) years of service increased from H% to 14(%. o. Although managers assigned more low ratings to performance evaluations for employees with higher salaries. this shift was probably due more to general increases in salaries than to changes in manager behavior. October 10, 2000 Page es Assessment of Georgietletn Goals Data Analysis Report Page 70 October 10, 2000 Data Analysis Report Assessment of Merit Svstein Reiortn (joa)s VI. Merit System Reform Goals Measurable with GMS Data October 10, 20f)O Page 71 Assessment of Meri! System Reiorm Goals Data Analysis Report Page 72 October 10, 2000 Data Analysis Report Assessment of Merit System Reiorm Goals Report 4 (Mett: System Reionii Goal 2) Advance Employees on the Basis of Abilities, Knowledge and Skills Within the Georgia Employment Management Systems jEMS), promotions were recorded using six transaction codes to indicate the type of promotion. Use of multiple codes for a single general transaction type was particularly common prior to 11 coding simplification effort begun in HH)(; as an adjunct to Merit ,System Reiorm. This reduced the number of promotional transaction codes to two. Tahle :J!) Promotionatj),uthority (~odes USe~U!LGJ~MS Promotional Personnel Transaction Authority Codes CPROMO Conditional Promotion ...."'"U II~ Licensure PROMO -10100 10102 12405P Regular Promotion _.. 00,....01'" Promotion Conditional Promotion ~ '~"'1:' Licensure Promotional Move to Unclassified Service 9700P Pr0r.:t:.?~~~~~!.AppOi ntment..~~.~C?.~.:gE:?r:ti.ticate Position Code developed or retained after 1996 code Simplification effort ~ .. Last Used September. 1996 " ' _ ' _ " _ ' _ " '___ _ _ ' _ _ ~~"'~'~' ~~N~_ September, 1996 September, 1996 ~,,-"'" " August. 1996 ~."., .,__ ~.~.w._,w~._ _ _ ".~~_~ N>_~_'" Personnel transaction histories stored in the GEMS database were reviewed and relevant personnel transactions between -January 1. Im)a and September :30, U)!H) were extracted. With the exception of promotion upon reappointment, all types of promotions were considered including promotions within the employing agency as well as promotions concurrent with a transfer to another agency. Employees in both the classified and unclassified service were considered. Performance evaluation scores and ratings were used as a measure of the "ability, knowledge and skills" of the employee. Higher scores and ratings were considered indicative of higher levels of "ability. knowledge, and skills". Linkage was established between each promotion personnel transaction and the performance evaluation score or rating that immediately preceded it, if any. Promotional patterns were examined using the combined data of promotions and performance evaluations. Further examination was conducted within the High. Average, and Low Performance Groups described on page 59. October 10,2000 Page 7:3 Assessment of Merit Svstem Reiottu Goals Data Analysis Report Selection of Records for Analysis Promotions can be processed by agencies at any time limited only by the agency's personal services budget. Table 40 displays the number of personnel transactions for promotion processed between WHa and HHm. In cases where multiple promotions were processed in the same month for the same individual, the latest act.ion was retained fill' analysis. 'fable 40 Number of Promotions P~.:LX!:!l!: Year 1993 1994 1 1 1997 1998 1999 Total Promotion Transaction Types -,.-.------j To Non- Unclassified Certificate Total Service Conditional Position Promotions 175 45 8 8,090 175 12 4 9,079 196 10 161 .,~XL 7-,69]5042 8,103 5.711 17 54.126 For purpose of this table, both codes fill' Regular Promotions (10100 and PROMO) and both codes for Conditional Promotions 00102 and CPROMO) have been combined. The number of promotions each year appears to be consistent with prevailing trends at those times. The sma II number of promotions for l~)HH represents transactions for the period from January to September (annualized to 7,6(15). The relatively low number of promotions in UH>7 is unexplained but may be due, in large part, to two results of Georgieciein implementation. The new job structures eliminated many purely promotional jobs and agencies may have reported some promotions under a variety of 'adjustments transactions' made availahle to revise the individual effect of Ceorgietlnin implementation. Table 41d isplays the availability of performance evaluation data for the selected promotions. Table 41 }'erfi.mnanee Evaluation Data\vailable fbr All Types of Promotions _ _._>H ......_ " _.. Matching r - - - -... No Matchfng- Performance Performance Year Evaluations Evaluations Total 1993 2.264 5,826 8,090 1994 r-' 1995 7.516 8.089 1.563 798 9,079 8.877 1996 6~.?Z~. ,-", "'1997 .,,_._.__._.__?-.~ 83~ .. .< '-" , ,,~,~~,.-,,~-,~>'-'~-'-'-'~ "'--"14.16890" 7.954 -.._.. -..............._6."30-2- 1998 7,502 601 8.103 _____rl,_<,','ij-' "~~.~,,,. 1999 Total 5,268 _._._----~?_,~~ .. 443 5,711 _ 10,880 ....... ...5...4_,.1_2-6- w'~~m","~'~.~ ""~ Page 74 October 10,2000 Data Ana lysis Report Assessment of Merit System Reiorm Goals All promotions were linked to the performance evaluation database described in Hl?l1PIL! {;!-~Q!gL'!f:!.(]JTI Goal;) Assure EgjrJ~(:rformanceEvahlgtiQ.J1;? Before implementation of Ceorgi;,G,lin, performance evaluations were generally prepared based on the employee's anniversary date (see Table (j on page 2G). For pre-GcorgiaGain promotions, performance evaluations prepared in the month of the promotion or in the 11 months prior to the promotion were considered. After GeorgiaG(lin implementation, performance evaluations were generally conducted during a narrow window from -Iuly to October covering performance in the preceding fiscal year (see Table 7 on page 27). For promotions after GeorgiaGain implementation, performance evaluations prepared in the nine-month period prior to the promotion were considered. Additionally, performance evaluations prepared in -luly through October were considered for promotions occurring in -Iuly to October even if the performance evaluation was dated after the promotion. During periods of relative stability (lHHij. I!HJ7, UHJ8, and IH!}!), the matching technique used here located corresponding performance evaluations for Hl% to !)2'% of promotions. Much lower matching rates are seen in HHJ:) and IH!)4 where performance evaluations were recorded beginning in April l!)!):J. Therefore, no performance evaluation data exists for promotions early in 199:) and this trend continues through mid-l!)!).1 where many promotions occurred before employees' first performance evaluations were recorded. The following examinations of promotions will only {()CUS on regula r promotions because of the relatively small number of other types of promotion. E~~mlinationof Pn~mQtionsfor High Perfon:n(!r~ Using the same High, Average, and Low Performance Group developed for other analyses (see Table :lO on page 59), the combined database of promotions and performance evaluations can be displayed as follows: Table 42 Number of RegulaI' J?romotions bv Perl())'mance GnillQ Regular 7 7862 22 8888 7 8681 8 4260 4 3526 10 6299 24 8101 14 96 53324 Because l'vferit System Reform occurred in the middle of Um(i, data for that year was split into two parts. Tbe row labeled' HHH>a' contains those promotions that occurred p rior to the effective date of Merit System Reiottu and the row labeled' l!)!)6b' contains those promotions that occurred after Meri: System Reiorm. October 10, 2000 Page 75 Assessment of Merit System Reionn Goals Data Analysis Report Due to the variability in the number of promotions each year and the number of promotions without performance data, Table 4:3 displays this information as a percentage using only those promotions for which performance evaluation data could be located. Table 4:3 E"U:~,s:!1tageDistribution of Regular Promotions ~2.LI)--Evaluatlons - . - Promotions 80% 70% 1993 1994 1995 1996 1997 1998 1999 Figure 27 Comparison of Evaluations and Promotions for Average Performers For 1!J!)(i, the percentage ofpost-Merit System Reiotm promotions for employees in the Average Performance Group was plotted against all performance evaluations in the Average Performance Group rather than only performance evaluations prepared after Merit System Retortu. Observations 1. The trend of fewer promotions may be part of the continuing impact of Geotgieclein particularly when the growth in the number of employees (see Table 4 on page 18) is compared with the continuing smaller number of promotions. 2. In the years W9(i through]!)!)!> following Mertt SYstem Reiorm, the percentage of promotions given to employees in the High Performance Group (l;~.20%, 14.1:3%), 18Aa'%, and 20.570;(, respectively) was higher than their representation in the general body ofperformance evaluations (8.{iO%, 11.54%, 14';31%, and 1:').;38'Yc. respectively). Further analysis (not shown here) confirms that a similar trend also existed for the very highest performers those with performance evaluation ratings of Far Exceeds. a. In the years UH)(i through l!H}H following Merit System Reform, the percentage of promotions given to employees in the Average Performance Group (8(Ui6(X., 85.70%, 81.25%, and 7!). H;'}-() respectively) was lower than their representation in the general body of evaluations (90.92%, 87.;7%, 87.!)2(Yo, and 8:3.82% respectively). 4. The increasing percentage of promotions for employees in the High Performance Group is consistent with an overall increase in higher ratings (see Figure 5 on page ;$0. Page 78 October 10. 2000 Data Analysis Report Assessment of Merit Systen, J..YJ>erfonnanceGroup Using the same Performance Croups described in Table ao on page 5n as the controlling variable. the salary lines tor the various Performance Groups are shown below. 'I'able 47 _. SalarY.ll1crease Rates bv Performance G.rOUPB Discreet Increase Rates Nominal Increase All EmDlovees Hiah Performers Averaae Performers Low Performers 1993-94 1994-95 1995- 1996-97 1997-98 1998-99 1999..Q0 250(X 4.00% 5.00<>! 4.00% 4.00<>! 400;, 3.00% 4.27% 5.95% 6.29% 574% 5.81':>;0 5.98% 478"lt 4.06% 609/0 6.26% 628% 7.80% 8.03% 6.72% 439% 595% 6.34% 573% 558,{ 5.63% 4.42% 0.46% 1.47% 0.34% -0.11% 0.08% 060% 055"lt Cumul ative Increase Rates Nominal Increase All Emolovees Hiah Performers Averaae Performers Low Performers 1993-94 1994-95 1995-96 1996-97 1997-98 1998-99 1999-00 2.50% 650;' 11.50% 1550% 19.50% 2350% 26.50% 427% 10.22% 16.51% 22.25% 28.06/< 34.04;' 38,82/< 406;' 10.15% 16.41% 2269"1< 30.49/< 38.52,{ 45.24% 439% 1034% 1668'Yo 2241% 27.99% 33.62% 3804% 046% 193% 2,27% 2.16% 2.24% 2.84% 3,390/. The first part of Table 47 deals with the annual rate of salary change for employees. The numbers displayed on the Nominal Increase line are taken from Table 4() using only the Normal or Meets percentages. This establishes the lowest increase rate for an adequately performing employee. The All Employees line displays the average percentage increase for every employee present at. the beginning and end of the year and who received a performance evaluation in the year. The number of employees varies from year to year as separations were replaced by new hires. The High Performers, Average Performers. and Low Performers lines display the actual average percentage salary increase received by employees whose performance evaluation in that year meet that. lines criteria. Employees were reported in different lines each year based on that year's performance evaluation. Notice that on all lines, except Low Performers. the actual average increase is higher than the Nominal Increase. Because the percentage was calculated based on the actual salaries paid to employees, higher actual rates are expected. Actual Halaries included other types of increases given throughout the year including salary adjustments, promotions, and permanent salary supplements. For Low Performers in fiscal yearHH}()B7, the actual salar-ias of enough employees declined to cause a negative average salary increase rate. The second part of Table 47 deals with the cumulative effect of the discreet increases shown in the first part. These numbers were calculated by adding the increase rates for the corresponding line in the first part. These cumulative numbers are very important, however, because they show the combined effect of slight differences in rates year-afteryear. Figure 28 graphically displays these cumulative salary lines. October 10, 2000 Page 81 Assessment of Merit System Refbrm Goals 50% 40% 30% ox Nominal --+--High Perf Emps --0 - Avg Perf Emps --D--low Perf Emps 20% Data Analysis Heport ....... 0 . 10% 0% 1993-94 1994-95 1995-96 1996-97 199798 1998-99 Figure 28 Cumulative Salary Lines by Performance Group 1999-00 There is very close alignment of the salary lines for High Performers and Average Performers in 199:394 through lH!)(j97. Divergence of the line for High Performers begins in Hm7-!)8. In l!W7!)8, the first variable salary increases were awarded to High Performers. By the lH!}92000 fiscal year, that difference had reached over 7%). The salary line for Low Performers is not zero. Salary adjustments and promotions are given to some Low Performers in years in which they are not eligible for a performancebased salary increase. Page 82 October 10. 2000 Data Analysis Report Assessment of Merit System Reiorm Goals Examination of Sa)a!Y~Lln~~by Compensation Gro!l~ In an earlier examination of pe rforma nee evaluations by salary ranges beginning on page 45, there is a marked difference in the distribution of performance evaluations by salary level. Although this difference was reduced in later years, it was always present in the distribution of performance evaluations. Table 48 begins an examination of any difference in salary lines for employees in two compensation groups. For the purpose of this examination, employees were grouped according to their salary shown in the last employee extract of August :11, Imm or an earlier extract if the employee separated before that date, Employees with a last salary of less than $aO,OOO were placed in the Low Compensation group and those with last salaries of $:J(J, 000 and above were placed in the High Compensation group. 'fa ble 48 Salary Jpcrease Rates by Compensation (~n)lill Discreet Increase Rates Nominal Increase All Emnlovees Hiah Comoensation Low Cornnensation =iscal YA.,r 1993-94 1994-95 1995-96 1996-97 1997-98 1998-99 1999-00 250'* 4.00% 500";' 4.00% 4.000/. 4.00;' 3.00% 4.2JOr~ 5.95'7'" 629";' 5.74;' 581% 5980/< 478% 4.31% 5.92% 6,31% 586% 586% 5.79% 4.84% 4.25% 5.98% 627% 5.65% 5.77";' 6.16% 4.69% Cumulative Increase Rates Nominal Increase All Ernolovees Hiah Compensation Low Comoensation 1993-94 1994-95 1995-96 1996-97 1997-98 1998-99 1999-00 250% 6.50% 1150% 15 23.50% 2650% 4.27% 10,22% 16.51% 2225;', 28.06% 34.04% 38.82% 4.31% 10.23% 16.54"1< 2240% 28.26% 34.05% 38.890/< 4.25% 10.23% 1650% 2215% 2792% 34.08% 3877% The first part of Table 48 dea Is with the annual rate of salary change for employees. The numbers displayed on the Nominal Increase line are taken from Table 4t) using only the Normal or Meets percentages. This establishes the lowest increase rate for an adequately performing employee. The All Employees line displays the average percentage increase for every employee present at the beginning and end of the year and who received a performance evaluation in the year. The number of employees varies from year to year as separations were replaced by new hires. The High Compensation and Low Compensation lines display the actual average percentage increase received by employees whose income meets the criteria described above. Employees were reported in only one category even if their salaries moved between the two groups. The last salary determines the category for reporting. The second part ofTable 48 deals with the cumulative effect of the discreet increases shown in the first part. These numbers were calculated by adding the increase ratesfor t.he corresponding line in the first part. These cumulative numbers are very important, however, because they show the comhined effect of slight differences in rates year-afteryear. Remarkably, the salary lines for all three groups show very little difference and are within 0.12'% after seven years (;~8.82%1. :38.8!)%. and as.7!'}(}) despite the wide variance in October 10, 2000 Page 8;~ Assessment of Merit Svstem Reform Goals Data Ana lysis Report performance evaluation scores and ratings shown in Figure H; on page 4;") and Figure 17 on page 'W. Figure 2B graphically displays these cumulative salary lines by Compensation Group. 50% .. .;(.. Nominal Increase 40'Yo ,-A-High Compensation :- - 0 - Low Compensation Merit System Reform 30% . ..... 20% . .;:K" . ,.::+(-: 10% .. . 0% 1993-94 1994-95 1995-96 1996-97 100798 1998-99 1999-00 Figure 2H - Cumulative Salary Lines by Compensation Group Page 84 October 10, 2000 Data Analysis Report Assessment of Merit System Reiorm Goals I;:~aminationof Salary Lil!~~d!y~EthnicGroup In an earlier examination ofperformance evaluations by ethnic group on page :32, differences were seen in that white employees received performance evaluation ratings of Exceeds at nearly twice the rate of black employees and ratings of Far Exceeds at nearly three times the rate of black employees. Table 4!) below displays the salary lines for white, black. and other employees. Table H) W;.!-lary I ncrease Rates byJ~.thflic C;roup Discreet Increase Rates Nominal Increase All Emolovees White Emolovees Black Ernolovees Other Emolovees nlitr.al Yeal S 1993-94 1994-95 1995-96 1996-97 1997-98 199899 1999-00 2.50"/0 4,00<1" 500% 400% 4,00% 4,00% 3,00% f=4.27% 4,30% 595% 6.05% 629% 630% 574% 5,75"1., 5,81% 5.91% 5,98% 607% 4,78/" 4,65% 4,170/, 576% 6,26% 5.71% 5,65% 581% 4,95% 568"1. 676% 671"'\ 6,24% 594% 644",{ 5,27% Cumulative Increase Rates Nominal Increase All Emotoveos White Emolovees Black Emolovees Other Emolovees 1993-94 1994-95 1995-96 1996-97 199798 1998-99 1999-00 250% 650% 11.50% 1550% 19,50% 4.27% 10,22% 1651% 22.25% 28,06% 34.04% 38.82% 4.30% 1035% 16,65% 2240% 2831% 3438% 39,03% 417% 993"1. 16,19"1. 2190% 27550/, 33.36"1. 38,31k, 568%, 12.44% 1915% 2539% 31.331. 37.77k 4304% The first part of Table ,H) deals with the annual rate of salary change for employees. The numbers displayed on the Nominal Increase line are taken from Table 4(; on page 80 using only the Normal or Meets percentages. This establishes the lowest increase rate for a fully performing employee. The All Ernp loyees line displays the actual average percentage increase for every employee present at the beginning and end of the year and who received a performance evaluation in the year. The number or employees varies from year to year as separations were replaced by new hires The second part of Table 49 deals with the cumulative effect of the discreet increases shown in the first part. These numbers were calculated by adding the increase rates for the corresponding line in the first part. These cumulative numbers are very important, however, because they show the combined effect of slight differences in rates year-afteryear. Although there are slight variations in the increase rates, after seven years the cumulative salary rates are remarkably similar. White employees have salary lines that are 0.72%) higher than black employees on average. The salary line for other employees, however, is more than 4(% high than that of white employees. Figure :lO graphically displays these cumulative salary lines by ethnic group. October 10, 2000 Page 85 Assessment of Merit Systetn Reform Goals 50% 40% -::t:::-Nominal --O--Whites --a-Blacks ~Others 30% Data Analysis Report 20% 10% 0% 1993-94 1994-95 1995-96 1996-97 1997-98 1998-99 1999-00 Figure ao - Cumulative Salary I . ines by Ethnic Group Page 8f) October 10, 2000 Data Analysis Report Assessment of Merit Svstetn Reiotrn Goals Examinat!!>..n of Sa)aryJ:1!nes by G~nder In an earlier examination of performance evaluations by gender on page i35, slight differences were Seen in the distribution of performance evaluation ratings of male and female employees. Under the older PAIbased system, female employees consistently received the very highest rating nearly twice as frequently as male employees. Under the newer PMFbased system, female employees consistently received slightly more of the higher ratings. Table 50 below displays the salary lines for male and female employees. Discreet Increase Rates Nominal Increase All Emolovees Males Females Table 50 Salary Ir:g;rease Rates by (lender 1993-94 250% 427% 446% 4,13% 1994-95 400% 5,95% 6.04% 5,88% F i~cal VAal ~ 1995-96 1996-97 1997-98 5,00,{, 4.00'Vo 4,00% 629% 5,74% 581% 6.51% 6,35% 6,04% 6.12% 5,30% 5,65% 1998-99 4,00% 5,98% 6,13% 5.87% 1999-00 3,00% 4,78% 533% 439% Cumulative Increase Rates Nominal Increase All Emolovees Males Females 1993-94 250",{ 4,:n1. 4.46% 4, 13,{. 1994-95 650% 10.22% 10,50% 10.01% 1995-96 11,50,{ 165101< 17.01% 1613% 1996-97 15.50",{ 22'2~ 2336 21.43 1997-98 1950",{ 2806,{ 2940% 27,08% 1998-99 2350% 34.04% 35,53% 32,95% 1999-00 26501. 38,82% 4086% 37,34% The first part of Table 50 deals with the annual rate of salary change for employees. The numbers displayed on the Nominal Increase line are taken from Table 4(j using only the Normal or Meets percentages. This establishes the lowest. increase rate for 11 fully performing employee. The l\1I Employees line displays the actual average percentage increase for every employee present at the beginning and end of the year and who received a performance evaluation in the year. The number of employees varies from year to year as separations were replaced by new hires Although the differences are slight, there is a consistent and growing difference between the salary lines for male and female employees. Despite the fact that female employees conaistently received slightly more of the higher evaluations, salaries for male employees are now increasing faster than salaries for female employees at nearly 1% per year (5.:liV% versus 4.im'X, in HHl9-2000). The second part of Table 50 deals with the cumulative effect of the discreet increases shown in the first part. These numbers were calculated by adding the increase rates for the corresponding line in the first part. These cumulative numbers are very important, however. because they show the combined effect of slight difference in rates year-after-year. Figure ill graphically displays these cumulative salary lines by gender. The cumulative effect of higher increase rates for male employees can be seen in that the salary line for male employees is over :l.;'')'1() higher than that for female employees. October 10, 2000 Page 87 Assessment of Merit System Reiorm Goals Data Analysis Report 50% 40% 30% - ; ( - Nominal :--ll. -Males '--O--Females 20% 10% 0% 1993-94 1994-95 1995-96 1996-97 1997-98 1998-99 Figure in - Cumulative Salary Lines by Gender 1999-00 Page 88 October 10, 2000 Data Analysis Report Assessment of Merit Svstem Reioim Goals The topic of "date of hire" has not been examined in earlier reports. However, with the increased latitude given to managers under Merit. System Reform, it is appropriate to examine the salary line fin' 'new hires' versus that of employees hired under the traditional rules of the Georgia Merit System. In addition, many managers are concerned about the lingering impact of underfunding implementing Georgieilein and the lack of a formal plan to consistently raise entrance salaries fill' state jobs. -Ianuary 1, umo was picked as an arbitrary marker separating 'old' employees from 'new' employees. The date itself is insignificant. However, employees hired before this date were hired in an era of 'classified service' protections and planned salary increases. After this date, new hires were immediately faced with no increases for nearly two years, followed by small. uncertain increases, reclasaificat.ions, movement to new jobs and a new pay structure, a nd removal of 'classified service' protections upon promotion. Table 51 displays the salary lines for employees hired before and after -Ianuary 1, l!)!)O. Table 51 SJllary Increase Rates bv Hire Date Discreet Increase Rates Nominal Increase All Ernolovees Hired Since 1990 Hired Before 1990 Fiscal Veal's 199394 199495 1995-96 1996-97 1997-98 1998-99 199900 2.50% 400% 5.00% 4.00% 4.00% 400% 300% +=1:27 .82% 595% 684% 6.29% 7.14% 574% 6.73% 5.81% 6.68% 598% 6,62% 4.78% 5.68% 3.74"/0 5.55% 579% 5.07% 5.14/0 538% 3.89% Cumu lative Increase Rates Nominal Increase All Emolovees Hired Since 1990 Hired Before 1990 1993-94 199495 1995-96 1996-97 1997-98 199899 1999-00 250ex 6.50% 1150;; 15501< 19.50% 23.50;; 26.50% 4.27% 1022% 1651% 22.25% 28.06% 34.04% 3882% 5.82% 12.661(. 1980% 2653% 33.21% 3983% 45,51% 3.74% 9.29% 15.08% 2015% 25.29% 30.67% 34.56% The first part of Table 51 deals with the annual rate of salary change for employees. The numbers displayed on the Nominal Increase line are taken from Table 4G using only the Normal or Meets percentages. This establishes the lowest increase rate for a fully performing employee. The all employees line displays the actual average percentage increase for every employee present at the beginning and end of the year and who received a performance evaluation in the year. The number of employees varies from year to year as separations were replaced by new hires The second part of Table 5(l deals with the cumulative effect of the discreet increases shown in the first part. These numbers were calculated by adding the increase rates for the corresponding line in the first part. These cumulative numbers are very important, however, because they show the combined effect of slight differences in rates year-afteryear. Any salary difference prior to H)9:3 has been ignored to focus on the impact of Merit System Reiorin in the period from wn:3 to WH!). October 10, 2000 Page 8!) Assessment of Merit Svstem Retortn (Joals Data Analysis Report The salary lines examined here have the widest divergence of any studied. After seven years. the cumulative salary rates are increasingly different. Employees hired after -January l , 19HO have a salary line that is nearty Ll % higher than those hired before l!)!)O. The difference does not appeal' to be reducing; the annual difference of over] .5'}-(1 continues to expand the difference in cumulat.ive salary line. Figure an graphically displays these cumulative salary lines by hire date. 50'Vo 40% 30% - ; ( - Nominal :- -<> - After 1990 --0-- Before 1990 Merit System Reform . . 20% 10% 0% 1993-94 1994-95 1995-96 1996-97 1997-98 1998-99 Figure i32 ~ Cumulative Salary Lines by Hire Date 1999-00 Page ~)O October 10.2000 Data Analysis Report Assessment of Merit System [>?efbrm Goals Oh;"!ervat ions 1. Examination of salary lines by Performance Group shows very dose alignment of the salary lines for High Performers and Average Performersin fiscal yearsUm;3~)4 through Um(;!)7. Divergence of the line for High Performers begins in Imn !)8. In W!)7 !)8, the first variable salary increases were awarded. IJy 1~)~)!), that difference had reached over 7'%. 2. There is very little difference in the salary lines of employees earning under $;~(),OOO per year a nd those earning $;30,000 and above. The salary lines are within 0.12% after seven years despite the wide variance in performance evaluation scores and ratings shown in Figure If) on page 45 and Figure 17 on page 4(). ;t Although there are slight variations in the increase rates for black, white, and other employees, after seven years the cumulative salary lines are remarkable sirnilarv White employees have averaged 0.72'X. higher salary lines than black employees. The salary line for other employees, however. is more than 4'!;,', high than that of white employees. 4. Despite the filet that female omployees consistently received slightly more of the highest evaluations. salaries for male employees are now increasing faster than salaries for female employees at nearly It}(') per year (i).;~;J% versus 4.:J9o/c, in ]!)992000). The cumulative effect of higher increase rates for male employees can be seen in that the salary line for male employees is over :15'){) higher than that for female employees. 5. The salary line for new hires since HH)() compared with more senior employees has the widest divergence of any studied. Employees hired after -Ianuary I,W90 have It salary line that is nearly 11% higher than employee hired before W!)O. An a nnual difference of over 1.5% continues to expand the difference in cumulative salary line. (). Within the data used in this report. approximately 25,000 employees are present in all seven fisca 1years. They represent a unique group of employees and are worthy of further study. October 10, 2000 Page 91 Assessment of Merit Svstetn Reiorm Goals Data Analysis Report Page H2 October 10, 2000 Data Analysis Report Assessment of Merit System Retortn Goals Report G (A/eril System Reiotm Goal 5) Retain Employees on Basis of their Performance Approach This report combines personnel transactions for separations with the performance evaluation data to create a new database. Unavoidable separations were excluded from the data. as were separations of employees who usually do not receive performance evaluations. All other separations, including dismissals and resignations, were part of the database. The database of separations was examined in three ways: Separations were grouped to show the number of separations each year for High, Average, and Low Performers (page !)5); A comparison was made by Perforrna nee Group of the number of separations each year and the number of evaluations received by employees (page ~l7); and The tenure of active and separated employees was compared (page ml). Selection of Reco~4.~ for Analysis For each separation, the performance evaluation, if any, that occurred in the preceding twelve months was identified. Unavoidable separations, such as deaths and retirements, were excluded from the data, as were separations of temporary and seasonal employees who usually do not receive performance evaluations. Transfers from one agency to another that were processed as separations and rehires in the Georgia Employment Management Systems (fiEMS) were also eliminated. All other separations, including dismissals and resignations, were included in the database. Despite attempts to reduce the dataset to only separations of regular employees, this examination lacks rigor because of the nature of state employment and the timing of performance evaluations. It is not uncommon for new hires to separate from state employment prior to receiving their first performance evaluation. No secondary performance data exist, such as exit interviews to indicate the reason fill' such separations or managers' evaluations to indicate performance prior to separation. Consequently, a large proportion of separations could not be matched with any data indicative of performance. In addition, the difference in timing of separations and performance evaluations makes definitive statements or conclusions difficult to make. T'his is particularly true after GeorgiaGain implementation when all performance evaluations were prepared in a narrow window from -Iuly to October but separations occurred throughout the year. October 10, :WOO Assessment of Merit System Reiorm Goals Data Analysis Report 1993 _ _n""""""",,- 1994 1995 ~m 1996 1997 -",-""'" 199B 1999 Tolal Table 52 8l1I1Jm:l!:LQf Sepa rr!!jQ~L12{lc!!! Total I Non-Status Separations ! Separations ,~~- Total Regular Separations ", ,,-~'-"~""" .~.~.~~,~ Performance Performance Data Missing Data Found 4.916 10,012 13 __ """,",,~.w.__ 3.006 4,899 "","'-' 3,199 1,910 5.113 5.~~~" j 1 323 3.235 B.OB8 814 4,063 4,733 5.621 11 338 11 )40 2.867 3,247 7.971 7.793 _. 1.268 ' . " """_~~"",._"_,,m,,w~~~ 1,801 6,703 5.992 7.296 1,550 5.746 1.498 4.248 65.562 .. 23.003 42.550 1 10,385 32.174 Caution should be used 'when interpreting the data for Wf);~ and 19!)!). In 19!);), most separations could not be> matched with performance evaluations because of the delayed start of performance evaluation data collection in April 199;~. The data for many employees who separated in W9:3 was purged from GEMS prior to this analysis. In 1999, only D months of separations had been reported before the HRMS was deactivated. Consequently, the number of sepa rations for I!)!)!) is underreported by 25%. Page !)4 October 10,2000 Data Analysis Report Assessment of Merit Svstetn Reionn Goals Examination ()fS~"~nltionsby Perform~n~eGroup One approach to determining whether the best performers were being retained is to examine the performance evaluations of employees who have separated. If the State is retaining the best performers, the expectation is that the performance evaluations of employees who separated would be lower than the average of performance evaluations overall. Using the Performance Groups described in Table :30 on page 5~) as a controlling variable. the separations fix regular employees (classified and unclassified) can be displayed as follows: Table f>:; 135~!?arations by Perf(Jl'mang&roup 1- Hl93 1994 1995 1996 1997 1998 1999 Total Total 1 910 5 113 5938 8088 7971 7793 5.746 42559 Unknown nf High Performers bv Average Performers ' l'::rnlln Low Performers 1096 1050 1 205 2467 1268 1 801 1498 10385 154 668 682 513 298 419 316 3050 614 3150 3812 49:23 6239 5381 3844 27963 46 245 239 185 166 192 88 1 161 Approximately 25% of separations each year were for employees who did not receive a performance evaluation within the preceding twelve months. Some of these employees were new hires who had not yet completed a performance evaluation cycle. .._--.'., -- 100% ,-----_._..,_ AddltJ'.nal f'UppnrtHlg- dd<:H! fOf t.hw };twph c;:m lw found PrJ l".gl' 16~) ... 10,000 80% 8.000 60% 6.000 40% 4.000 2.000 0% 1993 1994 1995 1996 1997 1998 o 1999 Figure:W Percentage of Separations by Performance Group October 10, 2000 Page U5 Assessment of Metl! System Reiorm Goals Data Analysis Report Figure ;3:> displays the number of separations along with the percentage of separations in each Performance Group. There was a sharp increase in the number of separations for regular employees beginning in 1!H)() and continuing through W9!). The number for H)!)9 represents only those separations processed through September 199!). The projected number of separations for all of Imw is 7,t;Cll, which is consistent with this elevated trend. The proportion of separations for High Performers fell from the 151}(,- U)I}(. range before Merit System Reiorn: to the 5%8(~1., range after Merit System Retorm. During this same period. the proportion of separations for LowPerformers also declined. These declines were offset by an increase in the proportion of separations for Average Performers. Page !)6 October 10. :WOO Data Analysis Report Assessment of Merit System Reionu Goals Examination of PerformaJ!_(:'eEy~J1H!tLonsand Separations by Y~~r A second approach to examining whether the best performers are being retained is to examine the relationship between the performance evaluations prepared and the percentage of employees who separate. The difference between this and the previous examination is that this examination focuses on performance evaluations prepared for all employees (both active and those that will separate) while the previous examination focused only on the performance evaluations received by employees who separated. This type of examination is somewhat suspect because of the very weak linkage between the gross number of performance evaluations and the gross number of separations. Only the most general of conclusions should be drawn. Due to the difference in timing between performance evaluations and separations, no conclusions should be drawn about trends in a single year. Performance evaluations after implementation of Merit System Reiorm and GeorgiaGah) are prepared in a narrow window from -Iu ly to October while separations occur throughout the year. Also, data for 19!):J is incomplete due to routine purges of data for separated employees. For any separations that exist in I!)D:~, only a small number ofperforrnance evaluations are available. Table 54 Approximat(d~!:X(~~~Ilt:!ge of Performance Groups Separati!}Z.J:l!lnuallv 1993 1994 1995 1996 1997 HI~R 1999 Low Performers 1011"/0 31.94!<, 3515% 3298!. 3057% 38.17% 15,66% Average Performers 1.88% 6,14% 677j(, 5.67% 10,60% 9.37Ir, 6.81% High Performers 1,32% 4.13'% 4.22% 6.22% 374% 4.16% 2.92% Although only general conclusions should be drawn from this data, there is a marked difference in the separation rate for employees receiving low performance evaluations. Particular attention should be paid to the Im)(i data. As seen in Table :~2 on page fiO, the number of performance evaluations in 1!)~)(; is considerably larger than in other years. This should tend to reduce the percentages displayed here for l!HH;, because although an employee may have received two performance evaluations in IHH6, those employees that separated only separated once. However, the proportion of LO\v Performers is consistently in the ;30'% range. More importantly, the proportion of High Performers separating in lmw increases to over G% despite the deflating effect of multiple performance eva luations. 'I'his G'X, rate probably corresponds to an 8% rate in other years. This data is presented graphically in Figure :34 on the following page. October 10. 2000 Page ~)7 Assessment of Merit System Reioriu Goals Data Analysis Report 1993 1994 1995 1996 1997 1998 1999 Figure :34 Percentage of Performance Groups Separating (approx.) While the rate at which High Performers separate remained fairly constant at 4%), the rates for Average and Low Performers moved slightly, though erratically, upward. Page H8 October 10. 2000 Data Analysis Report Assessment of Merit System Refimn Goals Examin!t!Qn of Tenure of AcJiY~,and SepalC!t~_c!.Emnloyees Another aspect of retention deals with length of tenure. Using the 'date hired' in the GEMS database. the tenure of an employee (in months) can be calculated for any date or event. Measuring tenure for active employees was accomplished by choosing a date for each year. The date of -Iu ne :m was chosen to determine tenure for active cmployeea. For separated employees, tenure was determined as of the date of separation. Table ;")f) displays tenure data for active and separated employees. 199~ 1994 1995 1996 1997 199R 1999 Total Table 55 Tenure in MO(l.1rt? of Separated l<~mployees bv , GrnUD All High Average Low Performers Performers Performers 47 80 64 62 57 82 59 61 56 85 52 64 61 85 59 74 61 76 61 81 58 71 58 92 56 73 57 77 58 80 58 73 II Active 112 110 110 111 114 113 112 112 The difference in tenure between active employees and separated employees requires some comment. I n general. managers have always suspected that there is a threshold to longterm retention of employees. In other words. if an employee has been with the State-for a certain number of years, he/she will tend to stay with the State until retirement. death. or some extraordinary opportunity presents itself. It was further suspected that this number was approximately ten years, This may be because at ten years of service employees become vested in the retirement system and begin earning the maximum amount of time off. Conversely, ernployees with less than five years and particularly those with less than two years of service are seen as very likely to voluntarily separate state service. These two suppositions of early voluntary separations and continued retention past 10 years of service can be seen in the separation data presented here. One might expect the average tenure for separated employees to mirror that of all active employees. Within the data, however. the average tenure for all active employees was between nine and ten years (112 months) and the average tenure of separated employees was slightly under five years (58 months). As would be expected. the average tenure of separated Average Performers is close to that. of all separations, This is due in great part to the sheer number of Average Performers (approximately 80%,) as seen in Table 10 and Table lLon page :30. Although not shown III this study. the average tenure fOI' active employees in this group was 4D months. October 10.2000 Page !)!) Assessment of Merit System Reiottn Goals Data Analysis Report The slight increase in tenure for active employees following Merit System Relorm can be traced to the increase in separations beginning in U)9(). More employees with short tenures separated, thereby increasing the average tenure of those that remained. 120 108 c 0 :;::> c~o 96 0. (J) (f) 84 <0 (l) 72 () .~ (l) ro 60 (f) III U5 "6 48 s:t/) C 36 0 2 ~ 24 12 0 Ad\lllij~n,ll 17l --0-- High Perf Terms --0-- Average Perf Terms .--a - Low Perf Tarms .--:t--Active Employees 1993 1994 1995 1996 1997 1998 1999 Figure Sfi Trend Lines for Service at Separations Observations 1. Approximately 25% of separations each year were for employees who had not received performance evaluations within the preceding twelve months. Some of these employees were new hires who had not yet completed a performance evaluation cycle. 2. The proportion of separations for High Performers fell from the 15%)19% range before Merit System Reiorm to the ;",%)8%, range after Merit System Reform. During this same period, the proportion of separations for Low Performers also declined. These declines were offset by increases in the proportion of separations for Average Performers. :J. There was a sharp increase of approximately :JO'% in the number of separations beginning in H)!)6 and continuing through HHH). 4. While the rate at which High Performers separate has remained fairly constant at 41~(), the rates for Average and Low Performers moved slightly, though erratically, upward. Separation rates for Average Performers moved from just over ()% to approximately ~)% and rates for Low Performers moved very erratically within the ao% bracket. 5. The average tenure for all active employees was between nine and ten years (l J2 months), while the average tenure of separated employees was slightly less than five years (58 months). There was 11 slight increase in tenure for active employees following Page 100 October 10,2000 Data Analysis Report Assessment of Merit Svstein Rciorui Goals Merit System Reiottn that can be traced to the increase in separations beginning in tsss. October 10.2000 Page 10 I Assessment of Merit System Reform Goals Data Analysis Report Page 102 October 10, 2000 Data Analysis Report Assessment of Merit S}istem Reform Goals Report 7 (Merit System Reiorm Goal G) Take Action to Address Inadequate Performance Approach This report examines a sub-group of employees -- those receiving performance evaluations with low scores or ratings and examines results of actions taken by managers to address the performance of such individuals. The effectiveness of managers in addressing low performs nee was examined through a review of personnel activity for a period of one year following the assessment. Selection ofE~~Qr!~Jor Analysis Using the definitions shown in Table ao on page ;")~), all Low Performers were identified. In addition, the definition was expanded to include those PMF-based performance evaluations that include a rating of Does Not Meet on the terms and conditions section. Because performance evaluations are a recurring activity, a one-year window beginning with the date of the eva luat.ion was selected in order to identify activity related to addressing previously identified low performance. Within this window, all personnel transactions were identified. Also, any performance evaluations prepared with la months of the original performance evaluation were selected. This larger window was chosen to maximize the opportunity of capturing the employee's next annual performance evaluation in an attempt to gauge performance improvement From Table 10 and Ta hie ll on page :Kl and the number of Does Not Meet rating for terms and conditions from Table 2~J on page f)(i, the number of low performance evaluations was determined to be 4.:JH!J as shown below. Table S() Number of Low Ped{)rm!!x!~e Evaluations per Year 1993 455 1994 767 1995 680 1996 656 1997 638 1998 561 1999 642 Total 4,399 The distinction of performance evaluation year was maintained throughout the examination to facilitate comparison of results pre- and post-lvlerit System Reiottn. The data for 19~)!J is included only to give an indication of trends in preparing low performance evaluations. The State's human resources management system (HRMS) was abandoned immediately following issuance of the 1mm performance evaluations. No subsequent datil, either in the form of performance evaluations or personnel transactions. was available. Consequently, Imm data is not shown in the following analysis and the Low Performers for October 10, 2000 Page 10:3 Assessment of Merit. System Reiorm (30als Data Analysis Report l!)!)!) were removed from the examination. The total number of [,0\\1 Performers used below is :1,7;")7 (4,:m!) less (;42). The relatively small size of the remaining database makes it possible to examine the activity in detail. As a benchmark, the personnel transactions for all employees were also captured to facilitate a comparison between the activity levels for Low Performers and all employees. Examination of C(!!!~l!veA(~tions Tak~J! The number of performance evaluations with low scores or ratings has consistently been 1'% or less of the total number of evaluations or employees. This is a somewhat lower rate than anticipated in traditional models of performance distributions (see Table 12 on pageS'l). Table 57 Ic&w performers as a P.~L!:5;entage of li:mplovees:..~!n.Q Evaluations 1993 1994 1995 1996 1997 1998 Total Low 455 767 680 656 638 561 3.757 As%of 0.66% 1.06% 0.91% 0.86% 0.85% 074o/r 0.85% As%of 1.02% 112% 0.97% 0.68% 093% 080o/r 0.90% The number of Low Performers in W!);3 is understated because performance evaluation data was captured for only April to December. Also, HRMS data for IHH:3 is incomplete due to purges of old transactions for some separated employees. The slight increase in 1!)!)4 of both the percentage of employees and the percentage of performance evaluations is unexplained. The decrease shown in the percentage ofperformance evaluations for HH)() can be reconciled with of the larger than average number of performance evaluations reported that year. Those additional performance evaluations were associated with some employees receiving two evaluations and two increases in 1H~)(>' All personnel transactions for Low Performers were identified within a one-year window of the performance evaluation. As expected, some portion of this group separated during that year; these separations are examined further in Table 60. Other Low Performers received personnel transactions while remaining in the work unit or transferring to other agencies. These actions are examined further in Table (i2. Employees who did not separate are examined below in Table 58. No follow-up data could be located 'within a one-year window for approximately HJ'% of Low Performers. This number varies by year and the missing data rate before Merit System Reform (12'% to 14%,) is considerably lower than the missing data rate after lvlerit System Reiorm (2()'~i. to :34%). This is most likely due to delays in processing subsequent I The number of employees is reported in Table 4. Those number of employees for all agencies and authorities are shown there including those agencies that did not prepare performance evaluations (see page 52). The number of performance evaluations is reported in Table 10 and Table 11. Page lO4 October 10. 2000 Data Analysis Report Assessment of Merit Svstem Reiortn Goals performance evaluations. The large window from Average )( zll) Low 1993 1994 1995 1996 1997 1998 Actual 455 767 680 656 638 561 3757 1.32% 1,30% 0.44%. 061% 1.25% 0.890/. 0,96;;, 54.73% 5176% 4794% 44.36% 44.04% 34.40% 46.23% 945% 6.26% 5.29,{, 11.28% 6.11% 6.95% 7.43% Terminated 2242% 28.68% 33.24% 24.24% 25.24% 24.~~ 26.80% Total 87.91% 88.01% 8691% 80.49% 76.65"/0 67.0 'II 42% Because of the missing data and the extreme variability across years, few conclusions should be drawn from this data. Table 5!) p[~~=.J.!nd 1)()st}xleLiLSX':!L(UJ..Bef(JUJ.J. i\(~tjY!tyj9!_I,<0w. ..perfQrml}!' Number of Low Evaluations e .2 High iii :::l iii -w> )( Average zG> Low Terminated Pre-Reform Post-Reform 1.902 1.855 1.20% 1.40% 61.90% 61.70% 8.10% 12.20/(, 28,80% 24.70% Total 100,00% 100.00% Once adjusted for the missing data, the management of Low Performers pre- and post-Merit System Reiotrn is remarkably similar. Nearly two-thirds of all Low Performers were subsequently evaluated as adequately performing employees. The major difference lies in the percentage of Low Performers separating and those being re-assessed as Low Performers at the end of one year. Contrary to expectations, more LO\v Performers were reassessed as Low Performers following Merit System Relorm by 4(%. Although not presented here, within the data can be seen the occurrences of those employees who continually repeat as Low Performers. The number is very small (less than October 10. 2000 Page 105 Assessment of Merit System Reiortn Goals Data Analysis Report 25), however, several individuals were consistently evaluated as Low Performers for as many as five out of the seven years studied. LO\\' Performers who separated within one year of evaluation did so in a variety of ways. Approximately (j2%) of Low Performers who separated were reported as resigning. The degree of autonomy in taking this action is subject to question. Table (;2 shows the number of Low Performers who left the agency by way of transfer to other agencies. The major differences pre- and post- Merit System Reform deal with dismissals and resignations. Post,\1erit System Reform managers are slightly more inclined to dismiss a Low Performer. Resignation rates for Low Performers are down more than 5% after Merit System Reiorm. 'I'able 60 Separation Types f{H' Low J?~~l:!!lrmers 1993 1994 1995 1996 1997 1998 Total IOA;ath from - . from - Other T durinG nf Nnn.St;atuiOt ions Tnt;al - OA4,{ 0.00% 5.051.., 5.87% 0.88% 0,65% 044% 0.13 0,{ 1451% 19.17% Test 0,22% 1.43% ~'04~ 015% 6.62% 1.62% 0.29% 21180/1 1.18% 1.62% o SClOj" 0.300/1 7.01% 1.37% 0.15% 1463% 0.15% 0.30% n 0.31% 9.09% 0.63% 1.10% 13.95% 0.00% 000% 0.36 7.47 089 0.71 Ok 14.95A. 0.36A. OOOA. 0000/; 2.42% 28.68% 33.24/0 24.24% 25.24% 24.73"'{ 0,240/1 689% 1.01,{ 0,45,{ 16.660/1 0.61% 0.64% n ?Q0t. 2680% For comparison, Table (; I presents the separation rates for the same periods for' all employees including Low Performers. Table s: Separation Types for All Employees The overall separation rate for Low Performers was nearly three times greater (2H.80'!11 versus H.8H%) than for all employees. Resignations comprised a greater proportion of separations for all employees (nearly 80%) than for Low Performers. Low Performers were nearly nine times more likely to be dismissed and eight times mote likely to simply abandon their positions without. notice (presumptive resignation), The overall separation rate for Low Performers was lower in the years following Merit System Reiottn. However, the overall separation rate for all employees was higher following Merit System Reform. This included employees who separated prior to receiving a performance evaluation. Page lOG Octobe r 10, 2000 Data Analysis Report Assessment of Merit System Reiottn Goals Table G2 displays some of the other types of personnel transactions processed fill' Low Performers. This selection of transaction types includes major disciplinary actions as well as discretionary actions affecting compensation. Table ()2 Personne I '1'ra nsaction Till'S f~xr.l!.~lw PerfOrr~l(Tii Approximately 18'h, of personnel transactions IIII' Low Performers were demotions of various types while 10'% of transactions were promotions. This latter figure should be considered a clear indication of performance improvement. Also reported here are Low Performers who transferred to another agency. These employees should have received subsequent evaluations in their new agencies, which would have been reported in Table f>8 if completed within the one-year window. There are some differences in the types of transactions processed for Low Performers preand post-JMeril System Reiortti. The rate at which Low Performers transferred to other agencies was lower following Merit System Reiottn. Low Performers were also less likely to tr-ansfer to other positions in the same agency. For comparison, Table t):3 presents the transaction rates for the same periods for all employees including Low Performers. Table ():3 J~~rsonnel Transaction Ratesfllr All Employees ()ctober 10. 2000 Page 107 Assessment of Merit System Reform Goals Data Analysis Report The overall transaction rate for Low Performers was one-third lower (22.:l8% versus ao. 71(){l) than that for all employees. Promotions comprised 40(}';' of personnel transactions for all employees but only 1()ll/i. of personnel transactions for Low Performers. Even so, subsequent promotions for Low Performers must be seen as positive indication that managers took action to improve performance. Observat.ions 1. The number of low performance evaluations has consistently been 1% or less of total evaluations. This is a somewhat lower rate than anticipated in traditional models of performance distributions 2. Once adjusted fell' missing data, the results of managing Low Performers pre- and postMerit System Reiortn is remarkably similar. Nearly two-thirds of all Low Performers were subsequently evaluated as adequately performing employees. The major difference lies in the percentage of Low Performers separating and the percentage being assessed as Low Performers at the end of one year. Contrary to expectations, more Low Performers are re-assessed as Low Performers following Merit System Retortn by 4(X). :3. Although the number is very small (less than 25), several employees were consistently evaluated as Low Performera for as many as five out of the seven years studied. 4. Approximately (;2()'ll ofLow Performers who separated were reported as resigning. 5. Following Merit. System Retoriu. managers were slightly more inclined to dismiss a Low Performer. Resignation rates for Low Performers decreased more than i')% after Merit System Reiorm. ti. The overall separation rate for Low Performers was nearly three times greater (2(;,80':);)% versus 4.:1H% in H)!)!)2000), The cumulative effect of higher increase rates for male employees can he seen in that the salary line for male employees is over :L5'% higher than that for female employees. 24. The salary line for new hires since 1990 compared with more senior employees has the widest divergence of any studied. Employees hired after -Ianuary 1, 1~)HO have a salary line that is nearly 11% higher than employee hired before HmO. An annual difference of over 1.5% continues to expand the difference in cumulative salary line. 25. Within the data used in this report. approximately 25.000 employees are present in all seven fiscal years. They represent a unique group of employees and are worthy of further study. 2(). Approxima.te ly 25'% of separations each year were for employees who had not received performance evaluations within the preceding twelve months. Some of these employees were new hires who had not yet completed a performance evaluation cycle. 27. The proportion of separations for High Performers fell from the 15%-1!ltil, range before Merit System Relorm to the 5'%8%1 range after Meri; System Reiotm. During this same period, the proportion of separations for Low Performers also declined. These declines were offset by increases in the proportion of separations for Average Performers. 28. There was a sharp increase of approximately :W'% in the number of separations beginning in In!)(; and continuing through 1!l99. October 10. 2000 Page 111 Summary of Observations Data Analysis Report 2D. While the rate at whichHigh Performers separate has remained fairly constant at 41Yi" the rates for Average and Low Performers moved slightly, though erratically, upward. Separation rates for Average Performers moved from just over ()%, to approximately ~J'.\f, and rates for Low Performers moved very erratically within the ;30%, bracket. ;3(L The average tenure for all active employees was between nine and ten years (] 12 months), while the average tenure of separated employees was slightly less than five years (58 months). There was a slight increase in tenure for active employees following Meri! System Reform that can be traced to the increase in separations beginning in 1D!JG. a1. The number of 10\\/ performance evaluations has consistently been 1% 01' less of total evaluations. This is a somewhat lower rate than anticipated in traditional models of performance distributions :32. Once adjusted for missing data, the results of managing Low Performers pre- and post- Mei'i/ System ,) Reioru: is remarkabl. y simi. lar. N, early two-thirds of all Low Performers were subsequently evaluated as adequately performing employees. The major difference lies in the percentage of Low Performers separating and the percentage being assessed as Low Performers at the end of one year. Contrary to expectations. more Low Performers are re-assessed as Low Performers following Meri! System Reform by 4%,. ;);L Although the number is very small (less than 25), several employees were consistently evaluated as LowPerformers fill' as many as five out of the seven years studied. ;)4. Approximately (;2%, of Low Performers who separated were reported as resigning. :35. Following Merit. System Reiotm, managers were slightly more inclined to dismiss a Low Performer. Resignation rates for Low Performers decreased more than 5%) after Merit System Reiortn. :3(;. The overall separation rate for Low Performers was nearly three times greater (2(;'80%, versus ~).;)~)(%) than for all employees. Resignations comprised a greater proportion of separations for all employees (nearly 801%) than fill' LO\v Performers. Low performers were nearly nine times more likely to be dismissed and eight, times more likely to simply abandon their positions without notice (presumptive resignation). :{7. The overall separation ra te for LO\v Performers was lower in the years following Merit System Reiorm. However, the overall separation rate for all employees was higher following Merit System Reform and this included those employees separating prior to receiving a performance evaluation. :38. Approximately 18'?h of personnel transactions for Low Performers wen) demotions of various types while 10% of transactions were promotions. :39. The rate at which I..ow Performers transferred to other agencies was lower following Merit System Relortn. Low performers were also less likely to transfer to other positions in the same agency. Page 112 October 10, 2000 Data Analysis Report Conclusion and Recommendations VIII. Conclusion and Recommendations Conclusion From the preceding analysis, it is clear that very little information is available to provide a definitive answer to the simple question .- "Have Ceorgietleit: and Meri: System f?efonn met the goals established for them?" Although some of the goals set forth for these endeavors a re clearly outside the abilities of any data analysis effort, more than half the goals could have been addressed in this study if the State had planned this research concurrent with the implementation of these projects. Despite the efforts of thousands of individuals in the development and implementation of these projects, only smallest amount of data was serendipitously available through the State's human resources management system (HRMS). Considering only the limited data and the analysis performed here, the answer to the question is a very tentative "yea'. There are changes in the data that are indicative of the kinds of changes in behavior and practices desired by GeorgiaGain and Merit. System Reiottn. It is impossible to say, however, that any single goal has been fully attained. Rather, there have been some changes in trends and data patterns that are indicative of goal attainment. For example, the distribution of performance evaluation ratings postGeorgiaGain and Merit System Reiorm is more equitable and there are indeed differences in the salary lines for high performers and average perforrnera. However, without more data, and more tools to analyze that data, general indications are all that can be shown. Recommendations The inability of this study to obtain definitive answers is not based on a failure that occurred recently the failure occurred when these projects were proposed. No in-depth plan for measuring goa I attainment was developed. In fact, no measure of the problem being addressed was presented other than anecdotal stories and generalizations. In order to measure success. one must have not only a goal, but also a staring point. In order to avoid similar failures attempting to measure goal attainment of future projects, the following recommendations should be considered. By implementing these recommendations, the State will be able to gauge the success of any project that depends on data from the HRMS. Since the cost of people is one of the largest components in the State's budget, it is safe to anticipate that many future projects will need l:IRMS data to measure success. These recommendations are not revolutionary; they have been espoused by successful project managers and project methodologies for decades. These recommendations are common sense and sound management. Prudent managers probably assume that the steps proposed here are being taken for every project .. but, based on the results of this study, that is not always the Case. Begin data analysis prior to implementing the project Ensure the HRMS can capture necessary data Acquire data analysis tool8 October 10,2000 Conclusion and Recommendations Data Analyais Report 1i~"nJlata Analysis Prior t.o Implement.i~g tht~ Project In the best of all possible worlds, the data analysis effort to measure project success would begin before the project is funded and staffed. The very first analysis would be aimed at measuring the problem, or defining the current state. 'This information is crucial to dearly defining the problem. Early analysis is also indispensable in assuring that the project addresses the problem and is not sidetracked by other issues. The measurable goals of every project should be clearly identified before the project is implemented. These goals should be stated in general terms that can be understood by everyone particularly those directly affected. For each goal, a statement should be developed that describes the problem being addressed and any baseline measures of the problem. This statement should also put forth how success is defined and measured. By defining how success will be measured in advance, the data systems can be ready to provide the necessary data whenever it is needed. Early definition of data needed for measuring goal attainment also presents the opportunity to review how the data will be collected and organized and who will be able to access it. Many future problems can be avoided by documenting project goals and measures of success in writing. Although it may be tempting to assume that everyone involved with the project fully understands goals and measures, the written documentation will he available when everyone associated with the project bas moved on to other areas. Future researchers and those charged with measuring goal attainment will find the documentation an invaluable tool. Finally, periodic measurement of goal attainment should take place on a routine basis. The actual process of extracting and reporting the data should be tested before the results become critical. Early testing can also help detect problems in data collection or in project implementation. Ensurej,lle HRMS Can Capture Nece~sary Data Defining the analysis needed to measure success is only the first step. The data must be collected, stored, and organized so that when it is needed it can be accessed. Two main avenues exist to capturing data. First, the data can he captured directly through special programs or procedures customized to the specific needs of the project. The data can also be captured indirectly hy drawing data from existing programs or processes used as part of normal transaction processing. Although the former is technically the easiest to develop and may provide the cleanest data, in many cases the latter approach will ensure that the data is captured consistently for every transaction. Whenever possible, data collection should be a part of routine daily activity and not a special activity. Frequently, special HRMS activities are the first to be stopped when there are staffing shortages. In a crisis, the only transactions that will be proceseed are those necessary to generate paychecks. This simple fact dictates that to ensure data is collected every time It particular event occurs, the data to be collected should he as closely associated to a payroll transaction as possible. This may require modification to critical programs and Page 114 October 10, 2000 Datil Analysis Report Conclusion and Recommendations data stores to add new data fields. Of course, some transactions and business processes cannot be directly associated with a payroll process. In these eases, the data collection process should be closely associated with the most critical transaction in the appropriate business area. Fina lly, all the codes or data values needed to measure project success should be identified as early as possible. The meaning and use of these data items should he documented, and training materials developed to ensure they are used correctly. As seen in this study, too often cod ing structures were simplified without regard to future reporting needs, or users were instructed to use codes incorrectly because the original meaning and use of each code was not fully understood. Acquire_I>'~tt!LAnalysisTools The State should acquire a set of data analysis tools that are compatible with its new HRMS and flexible enough to be used with other databases. Unfortunately, acquiring such tools is much more complicated that purchasing a software package for a personal computer. To fully support future data analysis needs, the tools will need theresources to store large amounts of data and process that data quickly. Access to data is frequently needed at high summary levels as were generated in this study; at other times access to detail individual transactions is needed. Several vendors have data warehousing, data mart, and data mining tools available that could easily provide the data storage. access, and analysis support needed. Proper installation and use of these tools will require a thorough understanding of the existing data and may require development of an enterprise data model. Data modeling tasks are complex and should not. to be taken lightly. Acquisition and installation of these tools will he quite expensive. A core group of data analysts should be established and fully trained in the use of the data analysis tools. Training and retention of these analysts will be critical to the successful use of the tool~. These analysts should be involved from the beginning in any project that will require data analysis to measure attainment of project goals. Their role should be to document the data required to measure project success and begin development of data models and queries. Of course, the business and software analysts supporting theHRMS should also be involved; together they can define how the data needed to measure project success can be captured by the HHl\lS. Finally, consideration should be given to making the data analysis tools available to other agencies so they can perform their own analysis of HllMS data for their agency. Agency data analysts might prove to be an effective recruitment source for the central HRMS data analysis group. October 10. 2000 Page 115 Conclusion and Recommendations Data Analysis Report Page llG October 10, 2000 Data Analysis Report Appendices IX. Appendices October 10.2000 Page 117 Appendices Data Analysis Report Page 118 October 10, 2000 Data Analysis Report Appendix A Hue Appendices 17202 92()(iN 121DC i l)i'P!\Ii;sglt'!,II!-!1 EQ1l'l2.'::!:l.~I.~..r.!.t........ .. .. Dismissal Of Permanent Status Em ilovee DlsmlssalDrug Related Convictwn ""_~ ....._ .. I}\,:,m ~ssa I HandoH!..I?EI!fL?creerlll}.K .. ~I~ducational Le,lvc Without.,1,.),aocY. ~._~~~~_ Emergency /\pPOtr:l;;;;;tn~l~e;;:n:.:.I~~~~~~~~_ ...t~.!I(ID. C. Ihstl'ibution ...__.~~~~~~~+-:~.::.:.:.~-'-_+.:.:.::.~:::...:::::...._ Ec u iva lent Pa Y Grade (2.:.:trc,.:v;;.p.:;rs:;:.:.i:;:.:o;;:n:.- ---l-:..:.:....::.:-: October 10. 2000 Page] H) Appendices Data Analysis Report I ~-_.~- '~-""'~-"l I7~1');a~ittei.~~ Personnel Transadion U3:'ill.b~fltion Of Employment 91)-10-01 L~2!J!J:!~.!1Of Elllployrm'nt ~_.~~,=".'~'::.::=- 78-07-01 Et:~~~~e Date 99-09-:)0 96-09-:J0 ~(~:'i~:l Code! EXPIHf-'l '12;IO~ Extension Of Conditional AllImintment . , Of ( M M _ " Of e 740H0 I !1().(/8-:J 1 !)ZO!)B .~l.:H}7.{!1 !)!}-O!)-:lO :)1021 I 92-IZ01 'M~:JI022 l~xtensionOfC~ovel~ . _ !H-07,OI 97-07-ln al022 _ Extension Of Leave Without Pay Traininu 92-10-0 I 96-08-:3I I 840f)Z Extension Of Leave Without Pay.Dis;l~ility 92-10-01 94-11-0:3 184051 ~---". . .. --~'----"""~'f-"";,.,..c..;..c..;..--f-"--c..;..--l- ''';;'';;'';;';'';;;'_--1 Extension Of ProvlslOJlalAppointment 7:3-08-0I 96-0S-a 1 Hf>OZ i Fniluro To Return FrOl1l Leave Of Absence 7S-07-01 H!).(l!):JO r-'-F'a~~;IY Leave WiO Pav 9H-IO01 99-09-:30 r~;lyT;;;;~;;'\V/O-I);;~:i3;I:lh'()r A{l0p.~!2EL"""_"M . Ha-O!:OI...... H(j-09-:30 1220a FA1VIWC)P 2an ,... ,I ! .~.c.,::~~i~: i.;~l.~'Mt:.,~ ;l.~ )~.;-.;~J'.~".'.~.'l~l:.:.~)~;.er.eD.isabilitL...1.-.0.1... ~~:~~:~~~ Family Leave \V!O Pay Parental. Care 9(j-09,44 , GeorgiaGain Correction !)()IO-OI !)I.I~ (,AC,NC GeotuiaGain ImlJltmwntation _ _. 9()10-01 9710-02 GACiN GeotOia(;ain Review 9G-IO0 I .. 99-09-:l0 GAG N-R Hourly Appointment !lli07-01 !}909:l0 HOURIi,'C_, Initial ]AJadOf Non - PACS Agencies 7S-07-01 !)!)-m):30 9999S Interim Anoointment 7['-Oa01 H4-0()-:30 :H02 .J!lt~;E!!!)tt0n!~:llm2IntnlPnt . __ ,. n-OS-01 !)(;-~;}Q'M 94(){).,_ . j Involuntarv Demotion I Involuntarv Demotion - 9(j-lOOl 7:l0S() I !}!)-09:J0 96-09:30 IDEMO 10:304 Involuntary (RIP') 9]'()8-()1 mJ-09-:10 lOa04Z ~JOb~.~~.;::::i~I.'() 1{(.~dllcti(.H1 In Force off Due To HJ~duetion In Force Tb 01' Hppatitls , Leav() [)~~~:_To Injury .. [A~av~~ Due To Injurv ._I!t..0~=:~~ ~YJ.th.I'.l:ly Due To T(:!.r:.1E()r_lIJ:yI?I"'~11)iJil:..\' ~_ WIthout Pay DIW To Lack Of Work ~:~:~~~:~~ 78-0701 92- ro-ot !}H-IO-Ol !/2lO01 92-10-01 H2-1 0-0 1 ;~;~:~;::;:~: it~~K 9()-0~la() IZ600 9!1-0!J:lO 18a051 U!)O!):m 9G-OU-:J0 I IN,JURY --18:3055 9.tlJ~::!Q._ 1880~~....J 9!lOU-;lO.. I8404:l I Militarv Leave Without Pay !)(;1O-01 !lHOU:JO MILTLV , Militarv [A'llve Without Pav 92 9HO!l-:JO Uvlodification Of Adverse..A(~li~'1l By SPf3 . ==+m=+2-10.01 !I.9.m)..ao No PosltJon For H,~turnFrol11 Contingent !.:_.._._ ~}9-0~);iO 18500 2040a 12204 Non-Certificate Almointmel}.t !l608-:H I 9700 ..tl(~I1:M<;!:~L.~)I:H'Y Ad,l11St!11.i?:.!.1.t. . . . ()ther [A'a\!(' WIthout Pav Authori'lA{l ._~-0l'2:!)J..... 92-10-01 !}H-09;JO 17:3OlN 184042 I 1~!ler Leave Without Pay Unauthorized 92-] 0-01 !}()09:l0 184041 .Partial Salary lrl(:rpa",~;_. Pav Plan AdJustment .. Across The Board Pav Plan Adj~ultmellt .. Across The Board . ... M'... .!!.~!:970J ,., ~J()- 100 I H208()] 96-08] I !lHOH-SO ao 1?a05P ATB 17:100 Page 120 October 10, 2000 Data Analysis Report Appendices ..!:telnstotement Bv Pprsonnel HOlml Decision H2-10()] !)!)O!l-aO 20404 _.,. I:RR~,e;i~n~ts;:;att~e:~m;;e:nI;ttiF~r'oromm l..ayoff Layoff '" 7:)08-0 1 +~:::.H:::.)...:0..:9~-::.:.:)(:.:.:)__4~~:::.1!::::.}:(:::.)(.:':) j 9610-01 !!!l-OH-ilO 19900 Uleins(~lten~!:'.!.)LOf A,clive EmployedHIF ~)(;-10-01 99-m)-:l0 1HH01 Reinstatement Of Active l~r!mlovee/RIF 80, H)-O I ~)(j-OH-;;O 9!JOOI ;.-:.R:.;:e.:.l<::;':')<;::ls:.;;'e. ;;l:. ;o'l:.;:'o:.:n.:.l. :;E::. ;;I::. ;l1I:J,;:).:.lt""lv:::.:'n.;;;;l.;;;.e.:.n;. t ,l--"~).:.7_-;(;;.;):..::.)-_O;::.:.1 99,OH-:10 RELEAS Release From Ernnlovment Removed From S1> Provision A i ReSIgnation 84,05,01 7f)-OiJ-0l +!:.:..:lG:-IO-Ol H9-ml-:lO 1::;2::;2,::,0.:.7_ _--1 ~)9-0!1-:!0 ,,- -lO:!05 9909:30 RESIGN I Rpsignatlon ,l~t.::.,ignation'l'o Accept Unclassified POSit . f-~<:stora~itl.r.!QL~__ ~_ B!l( OU> H~IO();)() TRaO I Transfer And Annt From Unclassified 82-04-16 !Hj08-;\1 9205'1' Transfer And Annt From Unclassified 8209-0( 96-08 a 1 9700'1' 'l'rnnsfer An(!J:!)~1~~J~!!!~~:111!:'!:'iQ::A........ ~. 82-04-l() 9()0!)a.~t....__ J.?1Qr1I Transfer And Change To Unclassified With 'Transf~r From Interim To Career 92-lO01 78-07-01 9()08-;\) 94-06aO 1840aT :n02X .Tt:i!t\'fl'EJ:')'())l]"E():,isional Appointment . _ ,....lfl07:Q.L .. ~ ~~1:09::~!t ..~~Q(~~~_~_._. Transfer Of Non-Status Emnlovee 7!)0 1-01 H(j08-(31 1020 I Transfer Of Provisional Emnlovce n08-0 I 9()08-;H 10204 Transfer To A Difft'rent Class ................- - I To .Job Trnnsfer ;;..;;;;.:c;;.~=.:.:...;;:.::.:.:.::.::.. !\ Different ()8-01 H()09-(30 _ 1020a j;J. _ l - ' - = - = --I_'_:::....:.:.::::..=_--I....::.:::....:.~=~ 96-lO01 !J!}O!};)O .JOBCHG Transferred From Another Aaencv I~fl!.!~:!:!..I!.g)!LIt?J~!l~>.I!ll:r_I:il'yX!~l1U ni t Transl~rred To NonMerii With Leave Right i Undassified Snecial Function rrr~{;~~f:\ppointment.:J>-tll~" )Jndassifll'd Consultant Appointment Unclassified Position Bv Law Unclilsslfied '!!::;~lI.!.~)rarv Appointment Unclassified Trades Annointment Voluntm'y Demotion Voluntary Demotion iLY(W)l!o:!r!k!~i~nlrg..YTl1;;s~t(lSrfeepltallnr!~.~!.l(oJLnI~l!DY..rua Tesnng !)()lOOI !)909-;lO TRANIN 78 07 0 l~_ .99.Q9:!~~..._....!.g1.Q.~.j 9210-01 9609-:JO 1840:3 89-01-01 9 9 . 0 H . a O . _ .....120SPF:...._.J 8!)02-01 HgO!);JO 120DNH i 89-01-01 8901-01 SHOl-01 _ 9H09';'Q._. 99-09-aO 9~)'()HaO 120CST 120 i! 120TEM ... S!)OIOI 9H09;J0 120TRD .. !)()lOOI !)!)-O~)ao VDEMO 7:3.08.01L96.09:aO IO;)O:l ._ ~..?).)J}:,._.:.~(... )LI).... 0'" T"99.9~1-'(0)'9,I,:.a..,o()_ =.~.~"'.87."4.t!)_;,.'..}.).:=....~~ ..~.j Page 122 October Hl, 2000 Data Analysis Report Appendices Appendix B Summary of Employees !n':-Pm::,~~~!!!!JCharacteristics Employmen t Snapshot Date nf ... 81 Fthir. Grn un White Black Other under Studv 8v"'" Male Female ,for Period 3-31-1993 6-30-1993 9-30-1993 12-31-1993 3-31-1994 6-30-1994 9-30-1994 12-31-1994 3-31-1995 6-30-1995 9-30-1995 12-31-1995 3-31-1996 64,970/. 64,91% 64.76% 6445!0 64,20% 6400% 6365% 6333% 63,05% 62,77% 62,66% 62.44% . 62,08% 6-30-1996t:f ,70/< 8-31-1996 1,88% 33,89% 3391% 3402% 3430% 3452% 34,72% 35,07% 3536'Yo 35,61% 35,84% 35,91% 3610% 3647/< 36,85% 36,70% 115'X 1,18% 1.23% 1,24% 1,28% 1,28% 1.28% 1,31% 1,34% 139% 1.43% 145% 145% 145% 142% 44,610/. 44,58% 44.58% 44,44% 44,39% 44,37% 4442% 4417% 43,90% 43,78% 43.42% 43,27% 4332% 43,21% 43.21 % 5539/< 5542% 5542% White 5 5 Black 55,63% 5558% Other 55,83% 56,10% Male 56,22% 56,58% Female 56,73% 5668% 56,79% 56,79% 6334% 3533% 1,33% 4396% 56,04% 12-31-1996 3-31-1997 6-30-1997 9-30-1997 12-31-1997 3-31-1998 6-30-1998 9-30-1998 12-31-1998 3-31-1999 6-30-1999 9-30-1999 61,67% 6179% 61,53% 61,53;' 61,12% 6090"/< 60,60% 60,33% 60,03% 5973% 5940% 59,02% 36,82% 3669% 36,95% 36,92% 37,30% 3750% 37,76% 3800% 3827% 38,54% 38,83% 39,14% 1,50% 1,52% 1,52% 1.55% 158"/< 1,61% 165% 167% 1,70% 173'7< 1,77/< 1,84'YQ 42,88!<:, 42,60% 42,62% 4246% 42,16% 41,99% 41,73% 4142% 4115% 41,04% 40,96% 40.57% 57,12% 5740% 5738% White 5754% 57,84% Black 58,01% 58~Other 58 58,85% Male 58961. 5904% Female 5943% 60,62% 37,74% 1,64% 4178% 58.22% October io, 2000 Page 12:3 Appendices Data Analysis Report Page 124 October 10, 2000 Data Analysis Report Appendices Appendix C P.AI Scon-s for Classes with 1QO+~y~J~~.tions _-,- _ _'~'M~'" 1.0 1.5 20 2.5 :>.0 :>.5 H) 4.f) to to to to to to to to Class T1tle 1A U) 2.4 2.9 :~A a.9 4.4 4.!) 5.0 Totals Accountant __ _ _ " " w , ~ ~ ~ ~ Accountant. Prin.t.:!Q~~l....... ,.< .~.,~~ Accountant. Senior AccountInu Clerk Accountina Tech 1 ~!l!d!!.lt1'!eh 2 Act Ther. 81' . Act Therapist. ,,--~'''_. n " w w w _ " " m , , __ vc ~yitv Leader Admin Asst f-~~ztmn Svcs tarv Mar I 1 1 ;) II 2 "'16 ail f)8 40 ww, .. ,. 2(; -4_,. a9 a9 18 ....,.J~ 10 ];12 :~ 2,1 25 so I 50 24 176 .~- 1 .J.L _J..:lJ:L 19,1 262 Ias 29 76a .- s 2 9 201 a .-82 :H)5 647 l:lf) 259 11I I I:{i >'.", 17l; 76 180G . n:.>.. f) .. .. ~,.,~" ~ ao :u 58 46 II .. 183 1 I 2 46 76 140 89 19 a74 4 ,, I 24 :n 5a ~.~"._-~ a1 7 ._.._... 155 ~ 18 4!) 114 1O:~ 58 :H:l ~." i 1 2 18 a9 125 211 98 194 " !) 17 58 52 la 144 Admin Svcs 1\lg-r a Admin Svc" Supv Administrative Asst Administrative Clerk Ag Sanitarian - ., ...._"'~,-- I J:l 2(; :)7 58 40 175 I 7 ~ :18 2 :32 -~ 4l;2 89 96 ..."" :39 2ia 77 ,,..-- S!) 57 : 282 729 581 221 2:.109 ._~.:- I 2.5 46 7] I 144 Agriculture Inspect Assistant Director Assoc Counselor ,,-'" .. . l}J 5f) 8a 2 151 i, 7 .. . 7 :31 4f) 21 I II 2 79 79 is 1 ISO Assoc (~.r1:'i.i~l~il}~"t . . , ..... ,..,.. ~ Asst Spec Aat-In-Cha Auditor. Senior .,,- Behav Spec. Sr Behavior Specialist Benefits Counselor ..Casework Supervisor Caseworker .. Caseworker --- Chf Forest Ranger Sr Chf Probation Off I __ ___ "_"" __' _~_m " , , _ ' _ ' ~ ... ~... Chief Counselor Chief Drv Exam 1 Chief Forest Ranzer --- _.1------ ~~__w __ ! - ... I 2 I :~ I I -"'-""'~ _. 27 so I 2 16 I 14 20 1 27 4li ~!I f-._ 8O. 5 4 a :~7 90 :1 4 119 HI ,,,_. 107 -,) I .. 2 IE) 107 I il7 1 - .__ _ }7 . .....;.... ... 28.. I ~6a 48 21 87 :31 61 ao 60 zs 12a 74 ];lG ss 22 a2 207 15:J IS6 90 ... ___H_c9_ ; - - -6_1 60 :J !l7 4:J 42 I:J :J2 I IE) Ia 20H _. ~,~ I 10 120 _ -29 ...... ilO:3 22 aTI 91 154 42 5a5 W 11 , 528 sto i ISH a 181 I 102 I la lS5 Civil Eng Tech I I 2 !in 94 1H 22 29:J Claims Examiner - ..Clerk Clerk/Trans I Clerkf!'rans 2 Corum Epidcm. Sr Corum Tech Ii ,__ ,___v__ ... _ __,~,mm,_ Communitv Worker __, _~~m ""0"""'" I a aa 84 if) a1 :I 2ao I a 71 so 6;) 55 a2 :lOS 2 I f) f) 10:1 I 11 40 19 :1:J4 1 i ......,,"',.f-., I42 101 2" ~~ 1ss as 7 50 22 I 25 IS :{i8 I I!J 165 I 2 5 1:19 . :.. :157 L~8 i}9 B8:l COl1lP Com pIx Opel', SI' (i2 zs IG ;) I 109 S2!!~2E~~gr. . f'l:t.~;1" --','.".,,- Conservation C~.__... Conservation Sat Consultant Coordinator ______,_."_""''''V ~, ,. I ,-,--~- , 82 126 4d 28 70 :J 15 54 S 10 88 } 90 a:J 97 26..- 262 51 Hi! 4:1 us ss 26 ISf) 88 75 I :J8f) October 10. 2000 Page 125 Appendices Data Analysis Report 2.5 a.o :1.5 1.0 4.5 to to 10 to 2.9 ;IA a.9 s.u la lOa (il 69 :l52 221 1 8 a4 Ii) :J9 1)7 2 (:sr Agent. Sr Cswk Sunvr Sr Custodian J!!:t'.I . T~;IJ!~,.!.._ ... Data Transcriber Dental Assistant Dis Adjud, Sr Dis Adjudicator Division Director Drafter III Driver Examiuer Education Sunervisor Emn Interviewer Emil Proz Ilep ~~:~~~; Eng Tech 4 ~Ted~ En!!ineering Tech I Envirn Health Snec 2 Fiscal Analvst., Sr Food Serv Emp.:...;;S.;;t Food Serv Employee Food Service Aide Food .~!>..l.l,gy.... For' Svcs Tech For Svcs Tech Sr .+._ 22 191) lao 46 :m 51 10 2 7 2 54 11 15 50 Ia ao Ila 4 14 171 452 104 2 15 79 :~5 86 as 20('> 247 ~~ 15 1 19;~ 2 167 520 +_----j-___ 27 268 80 1 224 9.::.. ;..1+-----=a-+--"-"''''-1 ss 12 I a2 55 10('> 84 H7 55 79 1Ia 78(i 7 28 1) 47 9 al 5 15 56 42 22 141 -+_ _........_ _.t_.-.".~--_'_+-..;;;;..;;;1;.0.1;;;;..;;1;;3t_6-"--"-+_-s..a;;;;..;;;;_+---44:...+_--~::...;-a--9-5:.=:...j Cia4 40S 199 45 UH7 19 7 aao + a:;:,. I:+ 1') t 2;:-f,,):.0::: { 2 :172 104 Page 12H October 10, 2000 Data Analysis Report Appendices 1,0 Lfl 2.0 25 :\.O s.s 1.0 U} to to to to to (:lass Title J--.-~."".'''' Forester. :'~'l1lor -~,"" 1.4 UJ 2.4 2,!) ...-,,,. a..l 1 18 .,"'~, \-Ol'n Trades Foreman ---"'..',,,,--'.'-- Health Services Tech ____ __ '~'~ ~~ ~~'_.~."m."_'_<'" m v_y,,~'~"' 1 20 4 10 .~} 749 t-'H-:-i-w'-"a-~v: .."l.\1aint Forman 2 ," Hlth Svcs Tech. Lil ; :J 1 ., ~<._ 1 15 ;)9 Z40 '~'.'.,~. Hlt.h Svcs Tech. Prin ~ . " " r " " ,--~~""-"""",, 2 1 29 filth Svcs Tech, Sr 15 5:1 ." ...,-_.1:l1!l ~1n Svc Pl'og Spec II ""'m", HousekeeJ)('[. i I..-"~".,, ... ." 2 14 12 18 (i07 ~l}'il::ke('ping Mgr ,-~. 1 2 25 HousekeI'PlT~!:L Sunv 70 ~ ",.!Jousepal'ent .tII)t.l.f'<':lla1'<'n t __......". HouseparenL'?upvr ---'--'-..~.., ...--.~. Human Serv Provit~:E.. Human Srv Prov f-'-._..>.__ ....... > ,..... .. c-~-----_ -, 2 4 , I 5 ~J:tl}man Srv Provo Sr_...~__ 1 9 :ll 24 542 4 :31 2 22 8 . 2:tl 5 77 ~an ~vcs.Tech. Human Svcs.Tech ..Pr. rl:!_uman Svcs.Tech..S_f.' ---"""""-,,., I :l I :l 12 ~.- 2 ".-,'>'- :J 18 281 57 ;)40 Instructional Aide Instructiona I Aid I I 2 2 8 ,, lfi 7 .___1!~~7 Instructor ~!'stlgator (J)HlO _ . L P N ,,- LPN, Sf ",~,~_~_~,_~_c ,_.w_ .. w",,~,,_ Lab Sc~f'ntist, Sr >+ 1 1:.1 4:i9 2 2 rn ---~- ..--_...:.6..... -....:I!=) t=i4&02 to 10 ;l.!) 4.4 'W'A 88 50 ",_~!L ;J4 474 iJ,i 9') ;l/;!J :)4;) ilL 15Hi 6a2 I'}') "~mm,, 1757 _...?l; 101 406 44G I 7 , ,,,">'- 44 79 8!) :l2 ti7 '-;i()2' a07 ,~,,--~ :31 29 _. 35 52 ;n2 li21 129 :lO7 ;H;7 444 9] 22fi tel UJ s.o TOlals ,~,,,'A. 21 '(;1 161 a:n I:H; ".'~"~' HZ!) .'" 119 158 :n 5 "".W!2lI.. 19 I 170G ,,,. mn-n 10 70G i 102 I()(i!l a9 :m)'" 156 5475 .~"",,- 2(; ais >m""'v"" 2:1 ....1,,;;6;.,7;..;0;.... 1j 121 ;J8 45 12f) -- 8 41 I2 2 8 8 . , , ; ; . . . . , . . . , . . 47 2:1:1 -21 . ,,,.:::.:.. 1427 ~____o 7 ..l.!.lL 14 167 ~ 8(; 57 154 77 I til0 487 22 262 -l:36 Ii "> laO 650 4 15 8!) I 200:l ->. 22 -_rI: 25_. 1 is 215 5() 16 87:J G2 ~l~ (i!) ss 18:10 I 7:J 167 ss 19 508 464 ;)2 760 5f) 451 44 _._. _ 187 _2_ 289 --+~_m. """'''~".,.,~''' as t IH2 Lab Tech. Sr .. ~i~lor;V SCientist ~._~~- ..I.~llmdrv Worker , Lead Ntu::",e(Comm/Cln) ..\-. ~':'HI}"urse(Inpt Srv) .......'",._._.. Maint EqUIp Opel'.. .. ~, ~- Maintenance Enaineer -_. Maintenance WOl'ker -- r-~."N"'" Nurse ~ Nurs...e. (Comm/Cln) ..' ,~ I ! Nurse (lnnt Srvcs) 21 1 :l 14 , '> 1 :l 128 I 1I 1 2 - a7 ss ;) -If) 59 I lCl 4 14Cl 15 4 52 1 Hi 2 19 ;l 2:.1 2a :1 :15 4 5~) I :i6 , .. :l~l f}4 84 70 144 :.10 20 14 128 44 a9 ;li :17 :12 4:l 2~) () :;0 a4 :l8 6 257 248 I:J9 I 48 :321 1:.If> 45 18 74 2 ,H; 47 271; 55 1la 70 ~['!...j... 124 9\) 74 87 44 59 27 64 2(; 4 22 201 69 ts , ... 14 I, a is 7 18f)z[ zs 40 ! 16 ! 2 104 la!l I:W 8a2 479 srs 121 260 128 ...~ 2G2 a:l;1 257 2:11 I 72 201 ~Ncu_ rs. e Pract I t io IW I' HU"'" """ Office n Office Sunv. SI' _OperatIOns Analy~t 0pl1S.!\:!l"""~ 6 :12 -24 :H ~iL 41 ~ f)f) 54 4 laO 188 105 October] 0, 2000 Page 127 Appendices Data Analysis Report -_. .".~~. 1.0 to Class Title 1.. 4 Physician Physician (Brd Cert) PoultrvJXrader ._--"...,,,.~ ......,.~",,,, _!:!:jnCswk S l;!E.~~.____ .... _ Prin Instructor 1.5 2.0 2.') to to to i.s .... 2.4 , 2.!> I +---; -."" ;} I :\.O to :3.4 ~ -rI: fl I:J 2 16 :1.5 '1.0 to to :J.n 4.4 25 50 15 41 41; :JJ -.""., El I 29 -~_.-.~ 18 ,12 ~;~ll'- -- . 4.9 41 !)(; 2:3 55 29 5 () 1 ..- Totals 9 15a aOd I 11 :18.. I~f) . 14 12:3 r-P.rin Nurset'Comm/Cln) Prin (hms An~~... ,~._._ ..., Prin, Caseworker """.-_.". Principal Agent Principal Clerk ---,,~.~"-"~ r1?_rin(:lpl!L(;If:}l!.i;~. ! 4 ;) "",,-"'" 2 ~ s l1a 5 10 i :If) ..,,,.,,, 54:.! I 14 20 IS :Hl(; - 6!). f)~1')- so 54 40 !)4;} 2aO 807 4(; G!) ao .~"~,,,~,,, 1187 f) I 429 12 81 15 :HI I')') 228 109 2920 aEl(} 2618 12:1 Principal Secretary - Prn Nurset Inn! Srvs) I ~ll'Ole Aide '" -- arole Ofcr 2 Prob/Parole Ofcr ;1 I -~.'-'.", .. Prob/Parole Offcer I I :.l (i 104 2 21 I '2:1 ;) xj" 257 II 5 11 :l 186 250 27 50 8(i7 I El4 127 525 [,5 92 121:1 200 6S 419 ,8~I 54 .. ZEII 82 8 207 1514 40 2:12 f) ...........2:2: G~ 6 ZG22 1 ;146 a9'1 Procurement Ofcr I Proa/Analvst I 2 II a 5 al 28 18 44 2!) 4a 12 5 126 9 104 Prou/Analvst. Adv I ... i .!?t:()g!j~!.1alys t . Prin ~-~>~~. Proa/Analvst. Sr I I 9 35 47 Z '~l ;>7 90 7 22 a:l 71 a5 .. 4 202 57 as I 225 as 22 :.l IZ6 Program Manager Project )'(~'!.~~~L.. Project Manazer ~--~-~,-_._~ Pron & Sup SUPy I Prop & SUP SUPy 2 ._ Psvcholozist Psvchtrst (Brd F:lg) Pt/Ot Technician Qual Cnt.rl Reviewer Radio Operator Radio Op('r:}tor, Sr .. !~!!)g(~r I Ret Sup VI' m"_,__ ,,.~__..,_,, Recoverv ARent .... Recreation Director Rehab Counselor -, l{(~hab Counselor. Sr Right-OfWav SI)('C II .. Rizht-Of- Wa)l,tiE!Z~~ III Sales Manaaer I Secretary Typist . (~(;!:f:tary- _ _ _,, __ mm.~' ' Senior Ag Inspector Senior Aaent _ -.-.... .". I -~~,,,""~ i 2 I 9 15 22 f) 61 I 20 la 61 20 r=m= <10 H) za a2 56 9, 4fl ~. 52 :JfJ 29 I I7 ;~1~4 42 :I4 aiJ 28 ,I 29 :10 78 54 Ij 4 la 11 ao an ~~._.- -"",.".- .. _~.- ~>>>, ----.~._- 1 2 2 180 15 2 .....TI8 4 68 14 15a ... 17 I ;! ,..~,,,,,, ~ 1 4 15 :1(i 8 5 2 :51 "',.~'. i j 5 zao ""'~~- I 2 17 27 748 1 ao ".~.< 20 2:J4 ii4 II 6Z 54 9 505 .._._1,H., .8...~-" 81) 44 1 210 a:l8 212 so il7 8 ;l(; (i4 16 117 297 a4 ss 2 ss ,"""" 55 ,15 .. '...)..'")"'') -< a 2 -,....,~ Z9 liZ 18G 98 I I a 5 175!) .._ _ .... > ~c" _"ww""~.,, HO 70 !)!12 10 49 5a Z 12 I 15 2 laa 12 ..~ 124 (i Z7:l 5 147 2a 120 15 105 15 Z07 ...7.r t. ....""410952 I 140 IOa7 20t F)8 !J87 2 I 14 s 171 1;0 755 14:l ~, 71 702 :170 504H I cI:.~!.'..)=... 124 Senior Caseworker' 1 , 10 26 5fl 1094 1290 1768 !)66 204 ()414 Senior Clerk :1 I 2 lil 20 402 480 77f) 414 155 22(j5 Senior Counselor -- I Senior Craftsman Senior Secretary 1 1:1 Sergeant (DPS) . ~=F Sergeant FIrst GIs .- S ~: ervi --- ce -- : C-o.....o....r...d... i,>n'>' I Skilled Craftsman >' , I Soe Ser Spec I 5 IS 7 I 10 aoo 2 147 ss ...~.~ 554 10 225 1004 67 7 ()O 2 1)7 '_m __ s 201 mwmmm 4H 261 as 4!l!L "."-,,,,.,. 7:17 20H zoa unl t)i 6S ,,~,w GI 28:3 1.Hi8 44 ra 176a 10 ia az lOt; 590 6 674 a2 (i82 , , !1:J2L 6)'...). ~- Z-: !I -+-11- 4540~ 5 217 !: l ~44 Page 128 October 10, 2000 Data Analysis Report 1.5 2.5 to i.s 2.!) [) :1 .j :1 2 ----~I----+-- .~-+---+-.--+-- , I 2 -2 I 2 1 ;) Appendices s.o 23 7 41 Hi 102 22 107 8 lOa 1 28 2 52 84 47(; 8;) I 1 2 II 12 HS :\H 30 44 :)0 (j02 73 27 2:14 G 1D8 78 J21i9 G4 ;)8 is: 14408 :l229S October 10, 2000 Page 129 Appendices Data Analysis Report Page iao October IO, 2000 Data Analysis Report Appendices Appendix D PMF Ratin~l(,J,Lh)bswith 100+ Evalm~J!!ln~ Does Not ." .,,~.,~,~~ Fa ,Job Title Accountant Account ant 11. Prof. ~~~"m~,'_<""""~ ~~('()_1:!~t;]rlt..JII, . I).1:5l.L~ _____~...... ( :I 4 4 Meet I I Meets -- EXCN'ds 8a :~2 -,--~, 19 "' 55 HI 2!) Exceeds 4 :1 -\ " 123 IS!) 1:29 f\(x~ountant, Paraprof .__.,.....,.,,".,. Accounting Tech '" "I" Clerk u' "' I 'I'hcramst " Acnvitv Thel'anv, Ldr Admin AS81 Admin Operation."~!gT ~!Ein Opn8 CoordJI Admin. Opns Coord [ . ~._,.".~,--,,, Adrninistrativo Asst- A1'1'. I nspcctor, Sr. Aar. "n ,"'........, _._-,-,_v.>."~",,,", Aide .-- Area Enc, Asst-Maint Area EE!giE!~'.l't"....~_~____.._,, __mm~_mm ASSIstant Director Asst Area Eng ...,,..,,,-.- Asst Sunt ,_.....- Behavior S )ec ~~~" ----. --,....,,.,. BiUmQ' Clerk II .... ,-,-,~~ J~I(lgrv!()nitor/l st&~.1~51.~ .. "'W"A" Budget AnalV':L.... Bus Mar (Ctr/Bc) A""lord 59 38 I:l a 5 12 :J 8 f) I) _~o_" 8 18 I s ..." 5 1 1 2 I 2 1 I ---2 I .- ] 2 12 4 G 1 --",,,,,,,," I 2 4 1 ~-"..., ]:1 I 2045 no .....~.,. 72H :H4 126 '-'-~~_.- 168 ;J:lf) .,.,,..,,--. 66 250 181 272 157 1~"I" ._~m_m 101 68 HI 91 -~ ..-,- ii 8:1 284 11:1 129 14:J -] 10 Ga 80 514 22 8 I 107 8 f,O - - .. .,,-.- I 2 IG 217 :18 -, 40 7 139 23 -~ - :l 'if 12 4 f) 14 48 --- cIa 48 IH 24 2 ,,,.,.,,..,.. 26 2 50 ;) 12 ]1) 2 :l _,_._ ,m,,~,~, 25 :1 If) 2 ''''''~''''- ]8 2 2,H58 100 884 :~8:l ~ 1;03 I is 422 240 :IH!) IH2 1:J2 124 I l(; 124 If,7 105 I 1:J :H,;) iaz 147 147 .~~ 140 1 19 114 'ad ( :l t ----~,~.. ~.< ~" 41 108 2 ... If)2 f) 482 19 ~.. 24 :38 2j ]29 ..~ 5H6 Cert Nursz Asst., Ld CNI Nursma Assist 4 __ Charge Nurxe (liS) Chf Exm 2 ->.-'. ..-- 4 12 "-' 7 - ,.....'<.~._~- I .1 :) ~.~w~< 1 121 ... 799 :)60 67 I 45 4:1 --.~."., 12H 812 ..~..._-4--1-5--112 Chief Counselor ~iefDrivel' Examine 2 ,> 2 278 .. ;>2 2 Chief Of" <.n__", 5 ] ..11;4 51 Chief Parole Office,' ~fpo Civil l~n2' Teeh C;INk ,_....." C;lerk 1. General ,,,---~< '",-,- -","'">,~"'~ I 2 5 ..-~ 2:1 ~.. 97 :1 "'V'~""<_""-' ;) al Ull ]:31 ~."",-_.~-~~ 245 400 2.0;)H ~W' _ _ 2:1 sa 21 :n ]59 2 2 Ii J] Clerk 2 (1)0'1') -~""'"''''''''''''''''''''''"''''''''''''''-'''''''~-''''''' Clerk 2, Gpner1l1 ... Clinical Lab Tech ,.~.o.,,~. 1 G 199 91 12 2,185 ."' 1 IIH 50 5 :121 . . 19 ->,,~--_ 4:3 :la'1 ~ 457 2.;):)4 2(;4 2H28 211 Corum EqulP Ofcr. Sr 2 91 42 5 140 October] 0, 20()() Page I:ll Appendices -_. ,"'"-.-"'''',,,, f-'-.~_". -Iob Title Corum I~qtup pffi~~el' Corum. Tech. II ~mm,\lni.cahle Dis Cornu Onns Tech l~l.sr - .retv ~,~v_' . ,,-~~,..~, C ,~m" "--~- --- ~.~ Meets I :) r'''''' 464 s I ",,-_.114 8 2 141 6 1 ]20 ,1 ... 96 n Data Analysis Report Exceeds 48 8 10 a 15 40 Far Exceeds ",w,,, 4 Total 518 128 IGI iao 1I I 12] Cons Rtr 15t CIs Consorv Corporal --""''''",,,,.,-.,,.,, H HI7 46 a 158 :35 2. 254 IHG ....(.;~!r.~J:l(\t:y;r.i!.()n ~.__.... ....,- ,., m ,,.,,,.,' ,.",.",., Const lrisncctor 2 ~!:'ll Inspector.Sr 2 " Const Proj Msrr I Const Pro) Mgr 2 w--_- Consultant - " Corporal Correctional Officer --,--_ ,--_ .... ... Cottage Life Supv (;l.l.~.!.r!,!?.r!12,b. mw,_,_, ,.,- .~.~ Counselor . """"'~""V ~r (OTP) s-rcr: County Director II . ~~_.. ,Conoty Director III Countv Nurse Manl},:'?l:_ I -1 f> -=".",. . l.2[>-1 4 ., 19 42 7 8 ><>,. .""~, I I .. , I .. J() Craftsman 25 ,"""" Craftsman (GDC) 1!l Cl'ime Lab Scic.T'rin "'~" Cse Agent 38 I 72 2 251 147 122 1::f--'--'248 Hi2 :3 a57 2 III 6:) -. HI,568 1()1 :3 255 7 G4G 2 155 1 102 2 a77 2 I')') I 101 2 2G5 f> 728 I :m7 6 9() 17 8aa :16 71 ao 21 4H 4:3 99 sa - Ui85 7 42 .. 9(, 12 I ., 74 42 :31 8tl 75 Gil 35 107 ios 5 i ~ I ..u ........, aal 179 14a aoa 4 .. ,, ..la 7 2Ja 177 18a 2 22,274 ]~I9~ I a20 2 79;> 2 .._1-78- 112 4(,4 l(i7 ];33 8 374 J() 849 2 482 un !!!lo Cse Manager esc Review/Modificat ~,~~~~w",v.':""w""""""",ww~,,>~~,,,,,~,,~~~,~,_~_~,",n 0' (A,l~ Supervisor .. Custodian Data Transcriber I I 4 4 .I 15 51 52 J()7 4 _)I~ I 14(, 54 I 1:30 4 9 54!! a9 2 106 I 1!l!! I -~ 4 140 5 617 1~8n=~,,, fluent Tax Colle -,,-- 11 Asst ;omm Res Sued )rog"ram Cons Dir. Dev Dis Trng/W .. w.~ Dis AdiudAssoc II 5 H; a m_, _ _ 6 1 J() 2 I ,,~-,.,- I 2t> a m,,,. :37 HI ;322 117 144 iss m",,,,,,,,,nn'm_"m'" 17G n3 _." [17:3 '~N_ 24!! fJ5 6 ,\5 21 28 ai 51 ._ - - I 60 :1 1;3 1 178 aa 28 :H}!J 17:J 182 2:17 -~ 108 812 aaa Dis Ad 14 2()5 84 "__ v""" alia Division Director a 1 84 20 17 12.'> DOL Srv Spec, Adv ~-_.- ,_.. 2 210 18 I 2;10 DOL SI'V Spc aa 42 1,.162 95 2j l.lia4 ~~"-- Driver Examiner .. , ~;:tRep. Sr I 4 23a 2:> I :117 91 15 I 18a a8 261 2 441 2 2an min ...",,, ,,,, DwSedC&T Education n Sp Electrician Erun M!!!' (Vor Hh})) _ _" " W _'_~' Emn Svcs Case Manage N_'''"'~' Emnlov. Svcs. SUDV. I I'll 34 I I 17 I'> 1 li4 8-1--_. ,\ 217 I ,. 110 28 I 140 2 109 In I Ial 1 1!l I s -" 78 411 54 41 1:34 480 I 2 8:1 ;)1 I .I.-18 Page] :32 October 10, 2000 Data Analysis Report ,lob Title Does Not c Mept 8 Appendices Far Excepds " IGO 60(; I a02 7 I 2 252 ;1 ;1 s 4 I I :1 8 la5 ;1 s 187 4 22 Total 1 679 18 1 97 5 1.210 2 125 25 5 149 22 4i 2 2 5 '":::..:~=:.:..::- .__. _ + -+ .7 2 !lO 7 2. --"-'''-'-:--------.--.t-----+------+.-.----=.::....:..+-----'-]--.- -----+-----'---1 1 ..., --_......._ - - - - + - - - ,IIlV Prob/Par Spec 2 ~.!:'!:()l)ll)_. . ~.~ -Iob Title I (CDC) C om ,.,"" AIde Procurement Officer a Program ASSistant 159 "",- Program ts Program Coodinator 5 Imgram Manager ". ... ,~,,,,,.,,, 11 Data Analysis Report ",,,,,,, Does Not F; Meet Meets (;82 1"'"""""'"" 208 5H2 Exceeds ." I 15 aa aa Exceeds ."~.",,.,.,.. 1 a 8 l,f>27 40 :l lfil 21 2 m ~_.".~~ a a """"""', " J?:: 8 185 "~.,,, !;54 7;1 124 4:15 ,190 :10 104 27 a7 59 8G ,~,~ ~ ",.", " si 4I 11 4 I .. .~ ,-.".<-~ :i I +,"".". 8 112 92 148 Gal ias 1 02a 8 24 m 41 52 __ ~m'"',, 10 52 _ .. ... ~.. 2 I 2 , :l 104 47 2 a60 G5 17!) 9""r ~, , I 77 ss _m '''''''' '. 2 lOG 17 1 160 - 10 1 7 S81 59 I fl 498 186 19 Hl6 >,-,.,.., 5a a 95 80 2 ,_.""".,..,.. 9a 14 188 12 I W,',_V_ _ _ _.. 18 sao 88 4 1 69 1 2;n 24 ~.. :l 40 1 2 99 ;~2 1 75 85 4 m_ 1 (),") 45 a 102 :~6 4 ""~- 2 a28 149 t 2 G4:1 5 .....1,(;12. 1 .. 125 1 :181 (; 2!;2 a a87 41 109 ."w. _ _ _,_ 54 .. so 44 1.7 1f a - I 1 a 5 116 '_)I- 2 95 12 84 20 I 6 8(;4 49 4 7i '....,., 68 24 I 149 47 () .. 17 1,741 221 7 21a 7,1 5 l:n 28 2 a I;~2 ;12 5 a4 4.f~12 SIO 12 1 170 47 2 "'" -~,,~.~- 78 - 88 4 ... ~~.~.~h., ,~ 93 a5 8 Total "'~_.",804 2fJ4 G;'lr) "m_ 1.59fi "'"',.~ 220 76~) 109 176 ' 50G ,,""mm~ 120 122 21 I 710 15:") 1,)i)! 1 )1 4: 214 107 la.l,,- 188 sai ," 716 2'_)"I" 1as ...._- 12~ 5 201 HaG 101 279... I4a 168 ,,-.11JL 150 49fl 70(; 1 765 185 _._--34G128 .... 440 147 109 lOS Ial 1ss 204 2.050 296 _ _ _ 161 ~_m~,~. rrs ~~ 2aG 12G 147 Page I:~4 October 10. 2000 Data Analysis Report Appendices ", ,Job Tttle Dol': Not Far c: Meet Mcet.s Exceeds ,',,",' Total Program I{tpr. _ _ _ _ m~~~>~~_" Prol.'TnmfAnah:IiLIL., <'--~-~' Proaram/Analvst III Proura rnmer!A).l.~'b'l'.L._ J:!gJ~~t Manager Prop & Supplv Supv Psvcholouist ~,.~,,""""-- 1 76 a(i 1'1 :l I 128 19') - 49 47 4 I 96 If> ,~,,,",,,,_,"'''hY~'_'_' 4 __ " ___,cn 91 4 a ..,~.y--~~,~~, 224 10 10 7 I lOl 27 :3 1 Hi 4 11'19 :l ; ._,- 176 2 1 18 4 I :I!l 271 Ia6 Pt/(lt Tech Public Health I<:d ,-"~---,>~,,, ,"._"._~,~ '1A Spec (IWCS) -,-~--- Hanger I I{pceptlOotst Reereation Director .- I{(hab Co un se loI' Right OfWav ~.~:,L. ".,,,-Sales Manager Secretarv _~M_,~._.~m S(>(~ re t arv .~C1'e!3!.!yI Secretarv II -,,,",,,,,,-,.--.- ~. 4 f ._-_..., 14 f) 14 I is H:l I 5!l I8 - - _ . " . " ..,...,.>._, I I 12 16 104 9 I 77 2 ... I :35 1 la :l .- ......J,oao 104 6 121 10 2 U:l9 I 144 !l6 20 I 120 ""-'''"-'''-'''''''' 7 G7n lOS .'m. ....-.m_ 808 I 148 18 H,2 :\ 142 2 ._- 17!l 18 ._. 168 40 I f>aG I J() IW 2:3 .. 2. H,7 288 25 2.764 870 12 ... 2&~!fi.. 70 a._~ Secretarv II c 1II Sergeant .~~I];~'.~[lt(tP~D.... .. ---_.,., ,.St'rgeant First Class Service Coordinator Services Generalis! Soc Serv Coord I. .. --,.,,, ~,~- .~"._,-"" 26 I .. 42 ..~~I-' I a5 .) I 114 .. 6:17 1 lO4 8 1,;142 ;J 108 1 :no 259 2 --75 20 :l 1:18 ais 1!) ..J,Qi3..!L (;2 12 180 20() I 1.{J.~!'~). 55 () I ~.) I~ I,t ''''''-''''--' 17 , 420 :Ja I 111 Soc Se rv Coord 11 Soc Sery Dav Coord Soc Serv Spec ._"".--,_.~~~- Soc SYCS Case Mgr ,.'. , .~.--.,- 'W__ >-_w_' ~_.~ 2 .) 1I 122 I 152 . 45 () In a8 a 148 Hl 76 :l,t:37 47B 12 7. I 212 22(i lSI 4.10!) Soc Svcs Sunervisor Social Svcs Pl'OY Ii , ~ ~ . ~ _ m " m m m m _ _ '- Social Svcs Prv/Hosp SOCIal SYCS TchfH2:".e.. SOCial 5vs Coord . 1I 2B .. s 2 9 8 601 6 492 a 272 I I(i!) .",..,. a 260 178 I 799 94 12 627 a6 ,._-- ." ~_..~ :JHI B2 204 80 !l :Hil Social Svs Prov I 66 50 Ui77 21 1 17 2,021 Social Svs Tech Prill '~-~-~-"~~~ SOCIal Svs Tech Sr ---, , ...-....~- S(~i,~LSvs Tech. --------- ~.........- Spec Surveyor (ORS) 8 .t:J as ,~~~f> Special Agent Princi J"pci ~.lJ~g<~D.!,$t'rll().. Sr StalT DevfTrng Cd -,.. _ ... ,_ _-" ,.. ....~_ ..~ ....... ... ... _--,,--~ _ 10 58 Case Met Assoc. 42 Staff Dev/'Trng Cord a Steamplant Operator I 10 4a 18 2 ,',-- 1 >W,v"'_'_, __ I ".."~.,....",,.--.- (i8S 1,755 l.41 1 Sf) 281 151 64 H)!l 9,1 I lO la9 I f>:l 108 20 71 21 45 lG as 2 11 856 I,If) 2.00!) 1,581 I 12 is ;372 4 176 :J 122 528 2 la2 I 1;~ Storekeeper 10 2 ;300 l(i :328 Substance Abuse Coun -;..,;,:tllt'..r~l1t-" "_"_,~---~~_.<.~ c..:[~~. Examiner II (B) '. 16 89 28 2 I:V) October 10, 2000 Page 1:35 Appendices m, _ _ _,_ Does Not ,lob Tide C Meet Meets - 'rtix Fidd Agent III ~lireci~llislIII 12 2 225 2 82 Tax Snecialist IV Teacher .. Teachel' (Blind/Deaf) Teacher, Spec Ed Tech 1 ",--,-_._~. __ ""AA.o~~~ la 8 9 80 I 257 178 -, ss 1 208 tf;'Chmcal Instructor eo Trainee 1'1' Sergeant Trades Supervisor Trainine Inst. Sr 13a " 8 .- 102 4 201 ""~---'" :J 12(; """,,,"'''''' 98 Training Instructor ." Transfer Officer TrlmsJ)ortal~Jl1 Corp Trooner .- Trooper Cadet Trooper First Class .. ..,~l!J;.!.~liln.,.t~.~,l;l!ll,ineL~~ _ _...,. LiI Fld Tax Audit II .. ",." Utility Wrkr, SkId Vehicle O[)erfCour 9 6 ,., 1 2 >_,,_......... ~.h 1 1 ,~-~~ 1 1 2a 5 4 2 2 la-;:- . .,.~ i 14 _ _, , " , , " , , " , - " < 0 f) 280 246 szs 208 I:n 1 164 ](;;) la6 -,~.~~--_.~- 141 54(; Veteran Benefits COli 2 90 ViCI" President Warden .~"""",,,-...,...- 4 120 7:) Wildlife Tech :I 2 Il7 Work Prep Tech ,. fm~~~,JOb Titles h5 2 Ll21 :125 __JU:E!.,.l" .".'........... J.60!J 141 :HA5!1 172,814 Data Analysis Report -- F; K T: Total ,....""",,"',-,,, 60 '-' -~ ~---"""""', 28 15 9i <1 8 4 :>08 lOa us 290 2 180 8 :l 112 ~~ 11 229 4 I:n 4a --~~. 44 4' ao '1 i _.~._. ao 2 24 .. I 22....... ,,~.~~_. 4 .9...~1 5 lJO 254 177 1:,2 :326 ,) ........ _II a59 24;) a 144 aa2 28 1,5:>8 4 175 f>7 la . 218 155 as 612 48 140 ;) 12:> sa 14 144 29 4 ,. 22 8,214 B7G 152 165 ]-\,sss 26.mW .. ~~m~_~~ m m m m m " " UI:3:3 _...zO!),817 Page 18G October 10, 2000 Data Analysis Report Appendices Appendix E Reprint of .complex Charts ~iU1JLc:ataValues This Appendix redisplays the graphs from the seven analytical reports and provides the associated data values. Detail Figure 2 ~ PAl Performance Evaluations by Month }:38 Detail Figure S PMF Evaluations by Month 1:3H Detail Figure t PAl Scores by Year 140 Detail Figure 5 - PMF Ratings by Year 14 1 Detail Figure (; PAl Scores by Ethnic Group 142 Detail Figure 7 PMF Ratings by Ethnic Group H:3 Detail Figure 8 PAl Scores by Gender. 144 Detail Figure H P.NtF Ratings by Gender 145 Detail Figure 10 --PAl Scores by Age 14H Detail Figure 11 PMFRatings by Age 147 Detail Figure 12 PAl Scores by Tenure 148 Detail Figure }:3 - PMF Ratings by Tenure )4!) Detail Figure 14 PAl Scores by Pay Grade 150 Detail Figure 15 PMF Ratings by Pay Grade ){)I Detail Figure 1(, - PAl Scores by Salary Range 152 Detail Figure 17 PMI" Ratings by Salary Range 15:3 Detail Figure 18 PAl Scores for 2;j Most Frequently Evaluated Class Titles 154 Detail Figure IH PMF Scores for 25 Mostly Frequently Evaluated ,Jobs Titles 155 Detail Figure 20 PAl Scores for Larger Agencies 15G Detail Figure 21 PMFRatings for Larger Agencies 157 Detail Figure 22 Terms & Conditions by Responsihility Rating 158 Detail Figure 2:3 .- Compa rison of Low Scores and Ratings by Ethnic Group 1,j!) Detail Figure 24- Comparison of Low Scores and Ratings by Tenure )(;0 Detail Figure 25 Comparison of Low Scores and Ratings by Salary Range WI Detail Figure 2() Comparison of Evaluations and Promotions fill' High Performers W2 Detail Figure 27 Comparison of Evaluations and Promotions for Average Performers W;) Detail Figure 28 Cumulative Salary Lines by Performance Group IG4 Detail Figure 2H - Cumulative Salary Lines by Compensation Group W5 Detail Figure:W Cumulative Salary Lines by Ethnic Group H,(; Detail Figure:H Cumulative Salary Lines by Gender W7 Detail Figure Sz Cumulative Salary Lines by Hire Date 1()8 Detail Figure :3:3 Percentage of Separations by Performance Group W!) Detail Figure :34 ~ Percentage of Performance Groups Separating (approx.) 170 Detail Figure ;);j ~ Trend Lines for Service at Separations 17] October 10. 2000 Page 1:J7 Appendices Data Analysis Report 15,000 Detail Figure 2 PAl Performance Evaluations by Month 01993 111994 -1995 5.000 o Page 188 October 10, 2000 Data Analysis Report 40,000 30,000 20,000 Detail Figure B PMI<' Evaluations by Month -1996 01997 81998 -1999 10,000 -1996 3,989 01997 18 I 81998 14 -1999 : 248 11 14 Appendices October 10. 2000 Pagel:J!) Appendices 40% 30% 01993 .1994 .1995 Detail Figure :1 PAl Scores by Year Data Analysis Report 20% 10% I ... 0% 1.0 to 1.4 Page ]40 October 10, 2000 [lata Analysis Report 100% 80% Detail Figure;') P1\fF Ratings by Year -1996 01997 &11998 -1999 Oldeal 40% 20% Appendices October 10, 2000 Page 141 Appendices Data Analysis Report Detail Figure fi PAl Scores by Ethnic Group 30% 20% 10% 1.5 to 1.9 2.0 to 2.4 16.68% 27.00% 20.95% 5.0 7.64% Page 142 October 10,2000 Data Analysis Report 100% 80% 60% Detail Figure 7 PMF Ratings by Ethnic Group OWhlte IIilSlack Other Appendices 40% 20% 0% DWhite ! ill Black Other C 2.06% 3.31% 4.29% DNM 0.59% 0.88% 078% 82.35% 8811% 83.89% 13.19% 7.21% 10.01% Far Exceeds 1.22% 0.42 % 0.97% October 10. 2000 Page 14:> Appendices 40% Data Analysis Report Detail Figure 8 PAl Scores by Gender 20% 10% 0% 1.0 to 1.5 to 20 to 2.5 to 1.4 1.9 2.4 2.9 0.05% 0.24% 0.04% 029% 079% Page 144 October 10. 2000 Data Analysis Report 100% 80% Detail Figure ~) - PMF Ratings by Gender III Female -Male 60% 40% 20% 0% II Female -Male C 2.75% 2.34% DNM 0.68% 0.75% 84.04% 85.42% 11.63% 10.59% Appendices Far Exceeds 0.91% 0.90% October 10.2000 Page 145 Appendices Data Analysis Report Detail Figure 10- PAl Scores by Age 100% ...----..--- 1]4.5 to 4.9 04.0 to 4.4 -35 to 3.9 1]3.0 to 3.4 112.5 to 2.9 112.0 to 24 -1.5 to 1.9 01.0 to 1.4 80% 60% 20% 25.43% 22.20% 0,81% 0.27% 0,04% 0.02% 32.63% 21.74% 0.22% 0.03% 0.01% 32.23% 0,66% 0.25% 0.07% 0.01% 10.01% 20.07% 31.48% 0,33% 0.04% 0,00% 14.59% 22.18% 28.99% 15.76% 17.90% 039% 0,19% 0,00% 0.00% Page 14f:i October 10. 2000 Data Analysis Report Appendices Detail Figure 11 -- PM F Ratings by Age 100% .---.--.---- - Far Exceeds 80% flExceeds -DNM E1c 60% 40% 20% 0% Far Exceeds II Exceeds o Meets i,DNM lie 0.55% 7692% 1.37% 21.15% 7.20% 85.62% 0.55% 6.18% 11.11% 8445% 0.68% 2.80% 1285% 83.59% 0.71% 179% SO's 094% 1198% 84.99% 0.82% 1.26% 0.73% 8.78% 88.74% 0.80% 0.95% 0.38% 6.34%, 90.24% 1.65% 1.39% October 10. 2000 Page 147 Appendices 100% 114.5 to 49 04.010 4.4 .3.5 to 3.9 83.0 to 3.4 112.5 to 29 11I2.0 to 2.4 .1.5 to 1.9 01.0 to 1.4 80% 60% 20% 0% Data Analysis Report Detail Figure 12 PAl Scores by Tenure 4.15% 12.85% 28.65% 27.98% 25.05% 6.07% 17.04% 32.91% 24.68% 18.21% 8.48% 21.95% 33.62% 21.37% 0.06'>jo 0.01% 0.05% 0.02% 0.04% 0.01% 0.03% 0.02% 000% 0.00% 6.90% 6.90% 3.45% 0.00% 0.00% 0.00% Page 148 October 10, 2000 Data Analysis Report Appendices Detail Figure 1:J PMF Ratings by Tenure 100% - Far Exceeds II Exceeds o Meets -DNM Be 80% 60% 40% 20% 0% - Far Exceeds BExceeds ,0 Meets ;-DNM 'Be 11.57% 85.90% 0.70% 1.01'K. 14.89% 81.90%, 0.93% 0.28% 6.25% 90.63% 0.00% 3.13% October 10, 2000 Page 14~) Appendices 100% 80% 60% 40% 20% 0% Data Analysis Report Detail Figure 14-PAI Scores by Pay Grade 114.5 to 4.9 04.0 to 44 .3.5 to 3.9 113.0 to 3.4 112.5 to 2,9 112.0 to 2.4 .1,510 1.9 01.0 to 1.4 1,0 to 1,5 to 2,010 25 to 30 to 35 to 4,0 to 4,5 to 1.4 1,9 24 2,9 34 39 4.4 49 5.0 Page 150 October 10,2000 Data Analysis Report Appendices 100% Detail Figure Jfi PMF Ratings by Pay Grade 80% 60% 40% 20% 0% C DNM Meets Exceeds Far Exceeds OA 4.76% 0.69% 85.44% 8.40% 0.71% OB 2.43% 0.57% 73.40% 20.08% 3.52% 05 1.49% 0.77% 94.26% 3.37% 0.11% 06 2.670/. 0.73"; 93.01% 3.410/. 0.180/. 07 2.69% 0.75"; 93.40% 3.06% 0.10% 08 266% 0.91% 88.61% 7.52% 0.30% 09 2.61% 0.76% 8411% 11.87% 0.65% 10 1.66% 0.52% 82.30% 1436% 1.16% 11 4.81% 066% 84.37% 9.77% 039% 12 307% 1160/. 80690/. 14.13";; 0.95% 13 2.66% 1.03% 8115;' 14.49% 068";; 14 2.07% 079% 7897% 17.32% 0.86% 15 2.16% 0.58% 7401;' 21.62% 1.64% 16 1.75% 0.55% 7140% 24.12% 2.18% 17 178% 064% 71.08";; 24.56% 193% 18 1.49% 0.72% 68.01;' 26.84% 2.94";; 19 1.86% 0.50% 80;' 30.34% 5.51"; 20 3.77% 050;' 07% 28.64% 4.02% 21 2.15% 021 % 2918% 4.08% 22 5.07% 0.72% 67.39% 19.57% 725% 23 0.00% 0.00% 000% 100.00% 0.00% 25 0.00% 0.00% 000% 6667% 33.33% 26 0.00";; 0.00";; 000% 10000;' 0.00";; October 10, 2000 Page 15] Append ices Data Analysis Report Detail Figure Hi PAl Scores by Salary Range Page 152 October 10. 2000 Data Analysis Report Appendices Detail Figure 17 PMF Ratings by Salary Range 100% 80% 60% 40% 20% 0% --e--<$20K -'.$- A $30K -A-$40K I '-&-$50K =""=$60K ,--+-- $70K+ C 5,23% 1,88% 1,39% 1.20% 1,16% 1.75% 2,68% 0,67% 0,74% 0,78% 0,63% 0.53% 0,34% 0,58% Meets 89,04% 86.44% 8092% 76.58% 72,77% 68,73% 70.28% 10,35% 1562% 19,63% 22,47% 24,97% 21.43% 0,28% 0,60% 1,30% 1,97% 3,07% 4,21% 5,03% October 10, 2000 Page 153 Appendices Data Analysis Report Detail Figure 18 - PAl Scores for 25 Most Frequently Evaluated Class Titles Accounting Tech 1 Caseworker, Prin Caseworker, Sr Clerk, Admll! Clerk, Prin Clerk, Sr Correctional Officer 1 Correctonal Officer 2 Equipment Operator 2 Health Svcs Tech Health Svcs Tech Sr Health Svcs Tech, Lead Housekeeper Houseparent Human Services Provider Human Services Tech, Instructor LPN, Sr Nurse, Sr(Commleln) ProbationlParole Ofc 2 Secretary, Prin Secretary, Sr Secretaryrrypist Social Service Spec 1 Techmcal Instructor Other JobTitles Total o t.o to 1,4 1.1,5to 1.9 :112,0 to 2.4 112.5 to 2,9 :113.0 to 3.4 '-3,5to 3,9 04.0 to 4.4 1114.5 to 4.9 i-s.o 0% 20% 40% 60% 80% 100% Total Other Job Titles Technical Instructor Social Service soec 1 ISecretarvITvnist lsecretarv s- Secretarv Prin Probation/Parole Ofc 2 Nurse Sr IComm/Cln\ LPN Sr =$P'OV;".' Instructor Human Services Tech Sr arent Health Svcs Tech Lead Health Svcs Tech Sr Health Svcs Tech Ecuiornent Ooerator 2 Correctional Officer 2 Correctional Officer 1 Clerk Sr Clerk Prin Clerk Admin Caseworker Sr Caseworker Prin Accounnno Tech 1 1,0 to 14 0.01% 0,01% 0.00% 0.00% 0,02% 0.02% 0.00% 000% 0.00% 0.00% O.OO/', O,OO'X 0,00% 0.141< 0.00% 000% O,OO'X 0.00% 000% 0,03% 0.00% 0.18% 0.00% O.OooA 002'X 0.10'X 0,00% 1,5to 1.9 0,04% 0,04% O,OO'X 0.16% 0.04% 0.06% 0001< 004% 0.00% 0000/. 0.050/. 0.05% 0.06/0 000% 0.00% 0.18% O,OO'X o.n% 0.00 0.00 0.01' 0,09% 000% O.OO'X 0,180/. 007% 0.00% z.o ro 24 0.26% 0.26% 000% 0.57'X 0.34% 0.21 'X 0.20% 0,11% 0,15% 0.26% 000'X 0.15'X 0,30% 0.28% 0.72% 0.06% 0,27'X 0,59% 0.36% 020'X 004% 0.571< 0.38% O,~O% 0,481< 034% 033% 2.5 to 3,0 to 3.5 to 4.0 to 4.5 to 2,9 3.4 3,9 4.4 4,9 50 0,72% 20.39% 23.52% 31,02% 17,64% 6.39% 0.63'X 15,73% 23.23% 33.51'X 19.48% 7.11% 0,00% 38.79'X 7.00'X 11.28% 20.28% 1,21% 15,781< 23.44% 3715% 18,77% 2.931< 0.541< 1472'X 22.49% 34.861< 19.66'X 7,331< 0,56'X 8.83% 15.9904 31.40% 28.09% 14,851< 040% 6.87% 16.51% 34.68'X 27.68% 13,670/. 0.88% 980% 33.07'X 4630% 9.57% 0.23% 0.15% 9.64% 18.26% 3440% 26.59% 1081% 083% 17.56% 20.27% 33.20% 19,70% 8.17% 071% 23,991< 2383% ~3.281< 14,37% 3.77'X 0,90% 16971< 24,31% 32450/. 20.72% 4,44"... 048% 13.991< 22.34% 37.30% 21.020/. 4.50% 1,68<>;0 37.98% 28.17% 21.51% 876% 147% 1,08% 36.35% 26.71% 9.46% 1381< 0.90% 14,38% 20551< 37.87% 19.95% 6,11% 0.971< 24.64% 27.69% 32090/. 11.491< 2,85'X 3.??0/n 43.90% 27.78% 19.581< 3.58'X 1.111< 0.21% 34.74% 9.49'X 51.56% 2,21% 1.440/. 1.651< 8394% 1347% 0.64% 0.08% 0.00% 0,82% 52.70% 40.39% 560% 0.42% 002% 0,88% 17.75% 21.19% 34,2204 18.28% 6,840/. 134% 20.74% 2567% 30,83% 16.39% 4.66% 0391< 12,99'X 20,01% 31.57% 25160/. 9.570/. 1.021< 20.21 % 23.831< 32,661< 17,84% 3.77;' 0.51% 4.90% 10.48% 32.29% 40.65'X 10.65% 0.50% 11.13% 21,87% 3583% 22.76% 7,59% Page 154 October 10, 2000 Data Analysis Report Appendices Detail Figure 1!) PMF Scores for 25 Mostly Frequently Evaluated -Iobs Titles Acct. Paraprofessional Clerk 1 General Clerk 2,General Correcbonal Officer DOL Services Spec EqUIp Operator 2 Equip Operator 3 Fam lndCase Mgr 1 Health Svcs Tech Housekeeper Houseparent ......- - - - - - - - - - - - - - - - - - - - - - - =====================] Instructor LPNJu(vInCpaartireOntffSicvecrs1) Probalion Officer 112 Program Assistant Public Health Nurse Secretary 1 Secretary 2 Sergaent (GDG) Social Svcs Case Mgr Social Svcs Prov 1 Social svcsTech Social Svcs Tech, Sr Trooper tst Class Other Job Titles Total 20'7'" 40% 60% 80% 100% Total Other Job Titles Troooer 1st Class Social SVC5 Tech Sr Social svcs Tech Social Svcs Prov 1 Social Svcs Case Mar Seraaent (GOC) Secretarv 2 Secretarv 1 Public Health Nurse Proaram Assistant Probation Officer 1/2 LPN (tnoatient Svcs) Juv Corr Officer 1 Instructor Houseoarent Housekeeoer Health Svcs Tech Fam Ind Case Mar 1 Ecuin Ooerator 3 Eouio Ooerator 2 DOL Services Soec Correctional Officer ~ ~ General 1c~IMk 1 General Acet Paraorotessional C 3,11";' 2,53% 0.07% 2.09% 241% 3,27% 297";' 2.63% 244% 449% 2,04% 2.81;' 3,12";; 1.32% 16,18% 1,14% 3,78% 1.28% 3.56"1. 4.32"1. 007% 0.18% 2.02% 5.62% 346"1. 4.16";; 2.22% ONM 0,77/" 0,75% 085% 2.14% 1.14"1. 2.48% 1.85";' 050% 0,65% 0.88% 0.28% 0.60% 0,83% 0,50% 0,16% 0.65% 1.36% 0.88% 051% 1,74% 021% 0.37"1. 257% 029% 046% 1.29;' 0,68% Meets 8236"1. 79.80% 75.68% 8740;' 8924";' 82,96% 83.64% 83.92% ~ 91,32% 8205% 84,93"1. 95.67% 81,64% 9348% 92.32% 9541 % 94.29"1. 85.81"1. 8633% 9259% 89.47% 87,85% 83.14"1. 8725% 7693% Exceeds Far Exceeds 12,84% 0,92% 15,59% 1.34% 21,59% 1.82% 7.62% 0,75% 6.84% 0.38% 10.450;' 0,84"1. 11.52"1. 0,02"1. 1289% 0,060/ 0 22,78% 183% 11,06"1. 0.46% 6.19% 0.17% 1433"1. 0.21% 10.78% 0.34"1. 2,51% 0.00% 183% 0,20% 473";; 0.00% 2.25% 0.30% 2.29% 0.13% 1.64;' 0,00"1. 8,07% 0,06"1. 13,11% 028% 6,81;; 0.05% 5.81% 0.12% 622% 001'Y. 1222% 072% 6.83;' 047% 1935% 0.83% OctoberIO, 2000 Page 155 A p p e n d i c e s D a t a Analysis Report 100% -5.0 113 4.5 to 4.9 80% 04.0 to 4.4 -3.5 to 3.9 113.0 to 3.4 60% 1125 to 2.9 112.0 to 2.4 _1.5t01.9 40% 01.0101.4 20% Detail Figure 20 PAl Scores for Larger Agencies 0% Agic CSB DHRDHR- DJJ DNRDOASDOC DOE DOL DOR DOT DPS TA- GFC P&P OtherTotal LO Sch Acre CSB DHR DHR-LO DJJ DNR InOAS DOC DOE DOL DOR DOT DPS TA-Sch GFC P&P Other Total 10 to 14 0.050/. 0.03% 0.01% 0.02% O.OooA 0.00% 0.000/, 0.01% 0.04% 000% 0.06;' 0.D1o/. 0,00% 0.00% 000% 0.000/, 0.01% 0.010/. 1.5 to 1.9 0.00;' 0.050/, 0.05% 0.08% 0.02% 0.00% 0.12% 001% 0.11% 0.14% 0.12% 0.02/... 0.020/. 0.000/, 0.00% 004% 0.00% 0.040/. 2.0 to r-;.510 3.0 to 3.510 4.0to 4.5 to 2.4 2.9 34 3.9 44 49 50 0.19% 0.43% 17.71;' 32.92;' 40.90% 7.461.: 0.34% 0.25% 0.77% 19.11% 22.550/, 31.52% 19.02% 6.69% 0.31% 0.87% 16.76% 18.87/0 30.53% 22.71% 990% 0.30% 0.64% 1346% 1846% 34.03% 24.040/, 8.98% 0.28% 0.54% 13.740/, 24.970/, 34.90% 20.80% 476% 0.21% 0.71;' 7,64;' 24.370/, 44.90% 21.13% 1,05% 0.36% 1.52% 36,15;' 39110/, 16040/, 5.59;' 1.12;' 012;' 0.87% 40,02% 35.07;' 19.12;' 4.54;' 0.25% 0.60% 0.75% 7,74% 11.69% 23.600/, 29.500/~ 0.42/0 092% 12,38% 2445% 3645% 20.66% 0.67% 1.44% 16.45% 25.21% 37~16330;' 226% OA3'7l 0.38% 15.95% 16.61% 46 17.17% 3,350/, 0.06% 040% 19.140/, 44.62% 26.310/. 7.340/. 2.11% 0.04% 0.02;' 38,76% 6.64% 23.410/, 13,500/, 17.62% 0.00% 0.27% 22.46% 48.06% 24.21;' 4.68% 032% 0040/. 0.750/. 11.72% 23.92;' 43.99'7l 17.29% 2.25% 0.15% 0.68% 13.02% 23.27;' 32.21% 22.69% 7.97% 0.26% 0.72% 20.39% 23.52% 31.02;' 17.64'7l 6.39% Page 15() October 10. 2000 Data Analysis Report Appendices Detail Figure 21 - PMF Ratings for Larger Agencies -Far Exceeds II Exceeds Meets -DNM IIC Agri CSBDHR DHR- DJJ DNRDOASDOC DOE DOL DOR DOT DPS TA- GFC P&POtherTotal LO Sch Aari CSB DHR DHR-LO IDJJ DNR DOAS DOC DOE DOL DOR DOT DPS TA-Sch GFC P&P Other Total C 085% 198% 2.24% 2.77% 846% 126% 453% 4.01% 043% 2.08% 5.67"/" 031";( 0.14% 000% 0.06% 0.14% 110% 2.58'7< DNM 0.97% 1.11% 0.60 0.96 038 122% 0.84% 0.34% 0.50% 146% 067";; 0.59";; 0.73% 0.09% 0.38% 0.98% 0.71'\\ 0.71% Meets 90.24% 85.95'\\ 85.54% 85.06% 82.32% 76.19% 81.07% 87.01% 8746% 83.56% 7857% 84.65% 7823% 95.28% 83.70/" 89.26% 7815% 8462% Exceeds FarExceeds 7.69% 0.25% 9.89% 1.06% 1107% 0.54% 1067'\\ 054'\\ 7.72% 1.11'\\ 19.69% 164<>/" 12.71% 0.84% 8.40<>;; 0.25";; 1074% 0.86% 12.41% 0.50% 1349% 13.30<>j( 17.73% 1.60'7< 1~ 3. 428% 0.35% 1554% 0.32% 9.07% 0.55% 1670'7< 3.35% 11.20% 0.90% October 10, 2000 Page 157 Appendices Data Analysis Report 100% Detail Figure 22 Terms & Conditions by Responsibility Rating 80% 60% 40% 20% 0% Jilt-Meets t-NI I-t-DNM IIt-C 0.00% 0.00% 0.00% 100.00% 57.77% 13.55% 28.68% 0.00% 99.37% 0.51% 0.13% 0.00% 99.96% 0.04% 0.00% 0.00% 99.96% 0.04% 0.00% 0.00% Page 158 October 10, 2000 Data Analysis Report Appendices IJetail Figure 2;~ Comparison of Low Scores and Ratings by Ethnic Group 100% 80% 60% I PAl l"jPMF Scores DWhite BJBlack .Other 40% 20% October 10,2000 Page 15!) Appendices Data Analysis Report Detail Figure 24 ~. Comparison of'Low Scores and Ratings by Tenure PAl Scores 30% 20% .<2 D2-4 05-9 .10-19 1120-29 1130-39 1140+ 10% 0% .<2 1112-4 I ICl5-9 .10-19 112l)..29 113l)..39 ! 11140+ Low PAl SCore 21.35% 2387% 25.66% 19.24% 9.67% 0.16% 0.05% Low PMF Rating 16.08% 20.65% 20.74% 26.36% 14.52% 1.66% 0.00% Page l(;0 October 10. 2000 Data Analysis Report Appendices Detail Figure 25 ~ Comparison of Low Scores and Ratings by Salary Range 40% 20% .<$20K !1I$20K D$30K !1I$40K 1l$50K 1l$60K 1l$70K+ 0% .<$20K 1I$20K O$30K .$40K 1l$50K 1I$60K 1IiI$70K+ Low PAl Scores 50.21% 35.54% 10.78% 2.84% 0.42% 0.05% 016% Low PMF Ratings 24.06% 46.27% 19.54% 6.54% 207% 0.51% 1.01% October 10, 2000 Page Hi} Appendices Data Analysis Report Detail Figure 2() Comparison of Evaluatione and Promotions for High Performers 30% 20% 10% 0% 1994 1995 0.237 0.2303 0.2162 0.1907 1998 0.1431 0.1841 1999 0.1538 0.2057 Page 162 October 10, 2000 Data Analysis Report Appendices Detail Figure 27 Comparison of Evaluations and Promotions for Average Performers 100% - Promotions 90% October 10, 2000 Appendices Data Analysis Report Detail Figure 28 Cumulative Salary Lines by Performance Group 50% -X-Nominal - - H i g h Perf Emps --o-Avg Perf Emps --O--Low Perf Em s 40% 30% 20% 10% 0% 0.46% 10.15% 10.34% 1.93% 11.50% 16.41% 16.68% 227% 2.16% 3.39% Page Hi4 October 10. 2000 Data Analysis Report Appendices Detail Figure 2n ~ Cumulative Salary Lines by Compensation (}roup 50% 20% 10% OctoberIO, 2000 Page }(;5 Appendices Data Analysis Report Detail Figure :30 Cumulative Salary Lines by Ethnic: Group 50% - ; l - Nominal Whiles -"~Blacks --0-- Others 40% 30% 20% 10% 0% Blacks 2.50% 4.30% 4.17% 5.68% 1997-98 1998-99 1999-00 22.40% 2190% 25.39% 31.33% Page }(;() October 10,2000 Data Analysis Report Appendices Detail Figure :31 Cumulative Salary Lines by Gender 50')10 Merit System Reform 40% 30% 20% 10% 0% -J;.-Nominal -A-Males 1993-94 2.50% 4.46% 4.13% 1994-95 650% 10.50'% 10.01 % 1995-96 11.501<. 17.01% 16.13% 1996-97 15.50% 23.36% 21.43% 1997-98 19.50% 29.40% 27.08% 37.34% October 10, 2000 Page 167 Appendices Data Analysis Report Detail Figure ;32 - Cumulative Salary Lines by Hire Date 50% Page 168 October 10, 2000 Data Analysis Report Appendices Detail Figure a:3 Percentage of Separations by Performance C}roup 100% c.=JTotal Terrmnations High Performers -::'-Avg Performers -o-Low Performers 80% 60% 10.000 8.000 6,000 40% 4.000 20% 2.000 0% 'c::JTotal Terminations. 1.910 - - HIgh Performers 18.92% ......r~Avg Performers 75.43% -0- Low Performers 5,65'7'0 603% 80.54% 8758% 93.08% 5.05% 3.29% 248% 89.80% 3.20% 7.441..\ 90.49% 2,07% October 10, 2000 Page l()9 Appendices Data Analysis Report Detail Figure :34 - Percentage of Performance Groups Separating (approx.) --O--Low Performers --c) - Average Performers - . - .-High Performers 31.94% 614% 4.13% 35.15% 6.77% 1996 32.98% 5.67% Page 170 October 10,2000 Data Analysis Report Appendices Detail Figure ;~5 Trend Lines for Service at Separations 120 108 96 High 84 Perf ---e-Avg 72 Perf 60 --Cl-Low Perf 48 -:g_.- Active 36 24 12 o - - - - High Perf ---e-Avg Perf .--Cl-Low Perf --.1-Active 1993 80 64 62 112 1994 82 59 61 110 1995 85 52 64 110 1996 1997 76 61 81 114 1998 71 58 92 113 1999 73 57 77 112 October 10, 2000 Page 171 Data Analysis Report October 10, 2000