Discrepancies in Measures of Racial and Ethnic Disparities: Implications for Research, Policy, and Practice

Abstract Estimation of racial/ethnic disparity in imprisonment rates in the United States is complicated by differences in data collection, definitions, and methods of counting such persons across data sources. This paper describes these sources of differences and their impacts on assessments of imprisonment disparity. Estimates of imprisonment rate disparities derived from state corrections administrative data diverge from those derived from self-report survey data. Changes in how survey data measure race and ethnicity affect inferences about changes in disparity. Administrative data show less Hispanic-White and Black-White disparity than self-report data, and the differences between them are increasing. Measures of disparity based on single-race classifications show larger Black-White disparity than those that are based on multiracial classifications. Various approaches to measuring race and ethnicity can affect conclusions about fairness in imprisonment. Therefore, disparities analysts should consider the multiple data sources and methods available to measure state imprisonment disparities, understand the assumptions behind the measures, and arbitrate the differences and similarities before drawing inferences about the extent of disparity in U.S. penal systems.