Evaluation of Utah's Employment Placement Project: Cross-Agency Collaborations and New Data Sources
Ben Peterson, Ph.D., Director of Research and Data,Utah Commission on Criminal and Juvenile Justice (CCJJ) and Utah SAC Director, and Sofia Nystrom, Research Consultant at CCJJ and Ph.D. candidate in Economics at the University of Utah
In 2009, the Utah Commission on Criminal and Juvenile Justice (CCJJ) - which houses both the State Administering Agency and the Statistical Analysis Center (SAC) for Utah - awarded the Utah Department of Corrections (UDC) funding through our Byrne Justice Assistance Grant-American Recovery and Reinvestment Act (JAG-ARRA) grant to promote the integration of parolees and probationers into the labor market. The UDC designed a program to provide offenders with enhanced job readiness workshops and to collaborate directly with potential employers and community partners. This program was named the Employment Placement Project (EPP), and also spurred the development of the Utah Defendant/Offender Workforce Development Taskforce.
The program is voluntary, with many of the EPP participants being referred by their parole agent. The job readiness component of EPP includes an eight-hour workshop in which ex-offenders are taught basic employment skills, including proper etiquette and resume writing, and are given advice on how to explain to a potential employer that they have a felony record and a scattered employment history. After individuals complete the workshop, they schedule a one-on-one session with one of the employment specialists to address their individual needs and concerns. The workshops are held in eight of Utah's counties and are available to all offenders.
A large component of EPP, which is an ongoing program, involves training offender employment specialists who work in partnership with public and private sector employers and Utah agencies, including the Department of Workforce Services, the Utah Office of Rehabilitation, and the Utah State Office of Education. More information about the program can be found at: http://ncjp.org/content/leveraging-partnerships-utah%E2%80%99s-employment-placement-project.
As part of the grant award, the UDC was required to cooperate with CCJJ on a comprehensive outcome evaluation of EPP. In 2012, the Utah SAC began work on this evaluation with the assistance of research partners at the University of Utah Economics Department. While EPP serves both parolees and probationers, the current evaluation focused on the effect of participation in the program for parolees only (there are plans to look at a probationer sample in the future). The evaluators worked with the research team at UDC to identify a cohort of parolees who enrolled in EPP during 2010 and 2011, as well as a comparison pool of offenders who were paroled in a similar time frame for use in creating a comparison group through propensity score matching.
Data Sharing Across Agencies
The unique nature of this program and its evaluation necessitated making connections to new data sources. CCJJ and the Utah SAC often work with offender management data from UDC, as well as data from other criminal justice agencies in the state, including the Administrative Office of the Courts and the Department of Public Safety (which houses the criminal history repository). This evaluation, however, required not only offender management and criminal justice outcome data, but also required employment outcome data and educational background data. In order to gain access to individual-level data on these important variables, CCJJ formed data sharing agreements with the Utah Department of Workforce Services (DWS) and the Utah State Office of Education (USOE). This was not an easy process and this groundwork took much of the first year of the two-year evaluation contract. The difficulties of completing the data sharing agreements were amplified by the obvious sensitive nature of the data, and because of heightened concerns due to an earlier data breach at another Utah state agency that occurred earlier in 2012. Though the process was rather lengthy, it was well worth the effort, as the data sharing agreements lay the foundation for future Utah SAC research projects.
Post-release quarterly earnings by six-digit North American Industry Classification System (NAICS) codes form the basis of the DWS data utilized in this study. These data allowed us to analyze not only the sum of quarterly earnings but also the industries in which the parolees found employment. The latter were then compared to the employment situation for the general Utah population in order to gain a better understanding of how the two populations differ in terms of employment outcomes.
The educational data obtained from USOE include information on the parolees' educational history, including whether they obtained a GED/high school diploma as well as detailed testing information if they chose to participate in adult basic education. Both the earning and educational data were utilized in the propensity score matching, with the employment data as outcome variables and the educational data as important control variables. In addition, the Utah State Tax Commission provided us with information concerning the fraction of parolees that filed income taxes, though this number turned out to be small and a data sharing agreement was hence not pursued.
In addition to the newly explored data sources described above, offender background and criminal history data (for use in propensity score matching) and supervision outcome data were obtained from UDC for both the EPP and comparison groups. This criminal justice information was then merged with the wage and educational data. Merging official wage and educational data with demographical and criminological data had not been done in Utah prior to this study and is by itself an important achievement of this project evaluation. The primary outcome variables of interest included whether an offender recidivated within 18 months and post-release employment outcomes (including whether an offender found employment and his/her average quarterly earnings). Beyond the immediate outcomes used in the evaluation, a rich data source was created through the evaluation process to answer various dynamic questions related to employment and education in parole populations.
As with any evaluation, there were limitations with the data sources and the chosen methodologies, but there were also limitations with the program structure. For example, and as mentioned above, the wage data from DWS is by quarter, which made it infeasible to identify exact employment start and end dates. Moreover, the wage data only include legal reported income; it is important to note that this population often has illegal income, which was not included in the evaluation. In addition, it would have been valuable to have individual-level data on public assistance, though this was only provided on the aggregate level. Nonetheless, data limitations aside, CCJJ is exceedingly thankful to the state agencies that participated in these data sharing agreements, as this serves as an important step in understanding the relationship between employment placement programs, education, employment, and recidivism.
In terms of the program structure, it also created some difficulty. Most notably, the fact that EPP does not have a structured process in which offenders start (and complete) the program after being released from prison made the evaluation process quite challenging.
The final report, which will include the evaluation results, is expected to be released to stakeholders later this summer, and will be posted to CCJJ's website sometime this fall.