Administrative records from TANF and many other programs can be used for tracking applicant and client outcomes, evaluating the effectiveness of intervention strategies, monitoring caseload dynamics, managing programs, and tracking indicators of a state's social and economic well-being over time. They can help policymakers and program administrators better understand the intended and unintended consequences of actions taken, including possible cost shifts among programs. In addition, administrative records provide the data needed to meet federal reporting requirements and to compete for the TANF high performance bonus. For example, Maryland has been using administrative records to track client outcomes of its welfare reform program.
As with the other data collection methods, it is important to clearly identify the questions to be answered, set priorities among competing interests, and determine which administrative data can most likely answer those questions. Most states conducting research on their welfare reform programs are supplementing administrative data with surveys, case studies, and other data collection methods to address questions that administrative data cannot address. Potential administrative data sources include unemployment insurance wage records; Food Stamp program files; Medicaid eligibility files; new hire registries; vital records, which contain information on births, receipt of prenatal care, complications in birth, and infant birthweight; child welfare/foster care records; drug and alcohol abuse data; Supplemental Security Income and refugee data; juvenile justice records; general assistance records; public housing information; and school records. In addition, national data from the Survey of Income and Program Participation can supplement state administrative data. Researchers are also beginning to use demographic and other data available from telemarketing and similar firms to complement other data sources.
Using administrative data is not a panacea. These records have many limitations, both technical and administrative. However, compared with other data collection approaches, using administrative data is inexpensive. Moreover, analyses may be completed more quickly and can be done retroactively because data already exist. States must proceed with caution when linking welfare and child welfare data, despite the relative ease with which it can be done, because of the volatile nature of child welfare data. For example, reports of child abuse increase after a highly publicized death of a child in the child welfare system. It is therefore better to track the status of all children as they move in and out of different parts of the states' human service systems over time. This is the approach that University of Chicago's Chapin Hall is taking in a study of welfare reform programs in California, Illinois, Massachusetts, and North Carolina.