We conducted descriptive quantitative analyses of administrative data on program participation in two study sites El Paso County, Colorado and Utah to examine the extent to which sites are succeeding in engaging all or nearly all TANF recipients in work or work-related activities. To the extent the data were available, in each site, we collected data on the number and types of activities to which recipients are assigned, the number of hours they are assigned to participate, the extent to which recipients actually participated in assigned activities, and recipients' progress over time. This appendix describes the administrative data sources and the samples for our analysis in each site.
A. El Paso County, Colorado
The administrative data for El Paso County come from the Colorado Works state management information system (MIS). We obtained the data that are used to create the state's "Numerator/Denominator" report for each county in the state and for the state as a whole. This report contains information on participation in up to five different activities for all adult recipients with an open TANF case, including a category for county-defined activities (e.g., substance abuse or mental health treatment) that are not considered in the federal participation rate calculation. Scheduled and actual hours of participation are reported for all activities that are considered in the federal participation rate calculation, but generally neither scheduled nor actual hours are reported for activities that are not considered. An indicator identifies recipients who are in job search or education activities for longer periods than allowable under federal work requirements. The report also identifies the case manager for each recipient, making it possible to distinguish in-house DHS case managers who handle single-parent recipients with multiple and/or serious personal and family challenges that interfere with work from contracted case managers from Goodwill Industries who handle single-parent job-ready recipients and all recipients in two-parent families. In addition, the report indicates whether each recipient was counted in the numerator of the federal participation rate calculation and provides a reason why a recipient was not counted.
Colorado's MIS stores five months of historical data on program participation. We were able to obtain data for adult recipients in El Paso Countyfor August - December 2003. We conducted most analyses for recipients on TANF in August 2003, which can be considered a snapshot of the El Paso caseload in a typical month. In August, there were 1,204 adult recipients on the caseload. We also tracked the cohort of recipients that was on TANF in August over the ensuing four months.
Administrative data for TANF recipients in Utahare stored in two systems: (1) the PACMIS system, which stores information on program eligibility and benefits, and (2) the U-WORKS system, which stores information on assigned activities and hours of participation in each and every available program activity. Case managers are instructed to record assigned hours of participation in activities in U-WORKS and to overwrite assigned hours with actual hours if the two vary, however the extent to which they actually overwrite assigned hours varies with the frequency with which they meet with their clients, their ability to closely monitor and track client participation, and competing caseload demands. Both systems store current as well as historical data, and the U-WORKS system indicates whether each recipient was counted in the numerator of the federal participation rate calculation and provides a reason why a recipient was not counted.
We collected data on all 7,114 adult recipients on TANF in Utahin May 2003. From PACMIS we obtained demographic characteristics for recipient as of May 2003, and information on case openings and closings over time. From U-Works we obtained the start and end dates and assigned hours of participation for each program activity to which each adult recipient was ever assigned. We obtained data as far back as the early 1980s through October 2003. We conducted most analyses for recipients on TANF in May 2003, which can be considered a snapshot of the Utahcaseload in a typical month. In May, there were 6,187 adult recipients on the caseload who were subject to program requirements (we excluded from our analysis 927 adult recipients who were ineligible aliens, disqualified for fraud, deemed parents, or receiving Supplemental Security Income). We also tracked the cohort of recipients that was on TANF in May over the ensuing five months and conducted some analyses spanning first case opening through October 2003.