Welfare Reform/Child Well-Being Administrative Data Linking. Summary of Completed Research and Analysis


A number of research products were initiated. The following lists some of the issues under examination, a description of the process, and the results of the research.

  1. Research Question: Do clients who leave TANF have a higher usage rate of TitleXX and child welfare services?

    Description of the Process: Using the linkage established between the CHIPS system and the TitleXX system, ORS staff refined the information to primarily focus on the CHILD population (children 0-17 years old). Staff proceeded first by sub-setting the data set to children only. After that step, children were separated into two further sub-populations: children who have received AFDC/TANF and children who are in the Food Stamps only population. Two populations were created: "Before TANF" and an "After TANF". These two populations were followed for twelve months (controlling for seasonality) in a longitudinal study to determine by comparison if the "After TANF" population was more likely to incur higher rates of TitleXX services and particularly those related to child abuse and neglect. In order to control bias, both starting populations were controlled for previous TitleXX services experience.

    Results: While initial analysis suggests that the rate for the TitleXX services did not appear to increase for the "After TANF" population, that population needs to be continually studied and refined. The results of some of this work were presented at The National Association for Welfare Research and Statistics held in Chicago from August 2-5, 1998.

  2. Research Question: Are there particular neighborhoods with higher than normal rates of abuse and neglect?

    Description of the Process: From the Child Protective Services database, a 3-4 year trend database of founded cases only was extracted. This trend database was geocoded to the neighborhood level in order to locate "high rates/numbers of abuse and neglect". Maps were produced statewide by zip code showing absolute numbers and rates of children (rates were calculated using a 1996 estimate of children by zip code). In addition, maps at the neighborhood (census block groups) level were produced for Lexington, Richland, and Newberry. For these neighborhood maps, both absolute numbers and rates using census information were included.

    Results: Results do indicate that there are neighborhoods with higher than usual rates of abuse and neglect. This information was also shared with another project "Safe Futures" which is a coalition of state and local agencies in the Midlands area (four-county geographic region) concerned with youth violence. The "Safe Futures" project used the abused and neglected maps along with Department of Juvenile Justice information to target high-risk neighborhoods. In additional several counties have requested to have their TANF caseloads as well as their CPS founded cases be mapped. CHILD LINK has enabled us to respond to those requests.

  3. Research Question: Has Medicaid utilization changed for any member of a TANF household where at least 1 adult member has become employed?

    Description of the Process: In this study, we began with all AFDC/TANF cases that had closed for earned income. In order to provide adequate time frames to detect differences in utilization and to control for seasonality, the study population required 1-year of pre-employment activity and 1-year of post employment activity. Therefore each of these cases must have been an active AFDC/TANF case for at least one year prior to the closure date. In order to be retained, the case must have had at least one adult earning wages in the first and last quarters of the post closure year.

    Because varying Medicaid Eligibility coverage can affect results, the Medicaid Eligibility coverage was next examined for any biases. This study used a working definition of Medicaid Eligibility coverage of 700 or more days of Medicaid eligibility (2 years less 1 month). Claims essentially were divided into three types: HIC (physician, clinic, and laboratory claims), Outpatient (which includes both emergency and non-emergency room visits) and Inpatient hospitalization. For HIC claims, only physician visits were included. Outpatient claims were further sub-categorized as emergency and non-emergency visits. In addition, outpatient and HIC claims were summed thereby providing an index for ambulatory care utilization. For each claim type and combination of claims, the number of paid claims for the pre-employment and post-employment periods were summed. One way Analysis of Variance (ANOVA) with repeated-measures factor was the tool selected to use for the analysis. This tool allowed for the analysis of the number of Medicaid claims per individual filed one year prior to employment versus the number of claims filed one year after employment. In an attempt to control for natural changes in medical care utilization as individuals age, the analysis was stratified by age.

    Results: Our analysis finds virtually no support for employment effects on Medicaid utilization by former TANF recipients. The evidence is weak and inconsistent. The substantive differences among pre- and post – employment means is typically .2 claims per person and never exceeds .45 claims when all claim categories are combined. Few of these differences are statistically significant at the .05 level.

  4. Additional Uses of the Linked Databases: SC conducts a quarterly random sample survey of its TANF Leavers. While the response rate to the survey is excellent (for the interviews conducted during June, July, and August 1998, the response rate was 76%); there is of course concern on the well being of the remaining percentage of clients. To locate these "unavailable" clients, ORS linked all adult household members of the TANF Leavers survey to the ESC Wage match. For Waves 1 and 2 of the TANF Leavers survey, ORS linked overall 67% of the clients to the ESC files where the client had wages after the survey’s interview date. Of those clients who completed the survey, the match rate was 71% (for not completed: the match rate was 55%). The "not completed" category breaks down into three sub-types: no answer, unavailable, and refused. Their corresponding match rates were 69% (no answer), 44% (unavailable), and 81% (refused). In addition to looking at match rates by the various interview result categories, ORS also calculated average number of quarters with wages greater than $0, average wages earned in the last quarter and the average number of jobs in the last quarter. To examine if there were any significant differences, ORS performed several T-Tests on differences of the means. The results showed no significant differences in average wages earned last quarter among the various categories of clients.

    Using the statistical warehouse, SCDSS plans to study two particularly important populations in the near future. One such population is all sanctioned clients. Another population would be the abuse and neglected "founded" cases. In the latter study, we would link to the emergency room and hospital databases to determine if we could have predicted in advance these abuse and neglected cases.