"This study provides powerful evidence that race and stereotype-consistent traits interact to shape the allocation of punishment at the frontlines of welfare reform," according to Schram.
He noted that African-Americans were more likely to be sanctioned when the recipient had a history of one or more sanctions. According to the study's experimental data, African-American welfare clients with a prior sanction were 29 percent more likely to be sanctioned than previously-sanctioned Caucasians, and 45 percent more likely to be sanctioned than Caucasians without a prior sanction. In the administrative data, the risk of a first-time sanction was 14 percent higher for African-American than for Caucasian clients. When there were prior sanctions, the risk doubled to 28 percent for African-Americans compared to Caucasians who had no sanction history.
"White clients in these experiments suffered no statistically discernible negative effects when linked to characteristics that hold negative meanings in the welfare-to-work context," Schram said. "Minority clients, however, enjoyed no such immunity: their odds of being sanctioned increased in the presence of discrediting markers even when the details of their case did not change a bit."
Despite identical situations posed to welfare case workers in experimental surveys, a pregnant Hispanic with four children was significantly more likely to be sanctioned than a white client with only one child. Although results differed in a similar analysis of administrative data, the lack of consistency between the two data sets confirms previous research that demonstrated a narrower gap between Hispanic and Caucasian stereotypes than between African-American and Caucasian stereotypes.
Experimental data for this study came from a Web-based survey completed during a two-week period at the end of 2006 by 144 case managers with the Florida Welfare Transition program. Case managers were presented with realistic rule-violation scenarios in which key client characteristics were randomly assigned. Case managers then were asked if they would impose a sanction in response to the scenario. The hypothetical scenarios randomized the client's race (Caucasian, African American or Hispanic) and the client's personal situation. In one scenario, the client was a young mother of multiple children and in another she was a repeat welfare recipient with a history of sanctions.
Experimental data were analyzed in conjunction with real-world administrative data from the Florida Department of Children and Families. The data set included monthly, individual-level records for those who received Temporary Assistance for Needy Families (TANF) between January 2000 and April 2004. In an effort to match the scenarios provided in the experimental survey as closely as possible, records were analyzed based on the joint effects of (1) ethnicity and family size and (2) the joint effects of racial status and sanction history. The total administrative sample included more than 6,000 women and nearly 20,000 monthly observation records of participants.
Schram's co-authors for this study included Joe Soss, Cowles Professor for the Study of Public Service at the University of Minnesota, Richard C. Fording, Professor of Political Science at the University of Kentucky, and Linda Houser, a doctoral candidate at Bryn Mawr College.