We selected the three study states based on the availability of data collected for other research studies, which could be used to examine the use of TANF sanctions.(1) In each state, we supplemented the existing data with additional data collected specifically for the present study.
Administrative Data. In each of the study states, we use administrative data on single- parent families (excluding child-only cases) to examine how often TANF sanctions are imposed and how the rate of return to the welfare system compares between sanctioned and nonsanctioned families. Our analysis examines the use of TANF work sanctions among a cohort of recipients on the TANF caseload whom we follow over time. In all three states, the administrative data include information on basic demographic characteristics as well as welfare receipt and sanctioning status over time. The time at which the administrative data sample was selected for the study varies across the states, but in all cases it occurred several years after major reforms were implemented and after substantial caseload declines had already occurred.
In New Jersey, the data come from administrative records on all 51,539 single-parent families that received TANF benefits at any time between July 2000 and June 2001. In Illinois, the data come from administrative records for the 33,495 single-parent cases that were authorized to receive a TANF grant in November 2001. Also included in Illinois are a small number of "zero-benefit cases," which include some fully sanctioned families whose TANF grants had not yet been closed. In South Carolina, the data come from administrative records for the 10,852 single-parent cases that received a TANF grant in June 2002.
Survey Data. Survey data are available for a randomly selected subsample of recipients in all three states, and comparable data are available in South Carolina and Illinois. South Carolina and Illinois both fielded a telephone survey of a subsample of recipients to examine the assets and liabilities of the "current" TANF caseload. While each survey included some state-specific questions, most questions were identical, ensuring comparability across the states. In Illinois, a sample of 416 single parents was interviewed between November 2001 and March 2002. In South Carolina, a sample of 1,128 families was interviewed between August and November 2002. In New Jersey, a follow-up survey of 1,219 families conducted between April and August 2002 as part of the Work First New Jersey evaluation gathered detailed information on a wide variety of topics, including timelines of employment and earnings. In this study, we analyze information for a subset of 126 single parent cases from the Work First New Jersey survey who received a full family sanction during the follow-up period and had a year of follow-up data after receiving a sanction.(2)
Case Studies. For purposes of the present study, we conducted case studies of the implementation of TANF work sanctions in two local offices in each of the three states. The states selected the local sites, although we asked the states to select at least one local site that illustrates innovative approaches to implementing sanctions and demonstrates success in overcoming implementation challenges. We asked for the second site to be located near the first and, if possible, for that site to have followed a different approach to implementing sanctions. We conducted site visits to each study site in winter and spring 2003. A two-person team, made up of a researcher from Mathematica Policy Research, Inc. (MPR) and a research analyst from our subcontractor, AFYA, Inc., conducted the visits, which lasted about three days per state. During the visits, we interviewed TANF administrators, case managers, eligibility workers, and employment service providers. We also reviewed a small number of cases with workers and obtained written reports and copies of sanction notices and other relevant materials.