Federal Data CollectionAgencies' Use of Consistent Race and Ethnic Definitions Gao ID: GGD-93-25 December 15, 1992
Race and ethnic background is a sensitive subject, and making determinations of such identity for statistical purposes is an imprecise effort at best. Federal guidelines aim at ensuring consistency in federal data collection. Although the agencies GAO reviewed appear to follow these policies, they do not verify that data collectors comply with the definitions during surveys because of the substantial resources that would be required. This failure to verify compliance, combined with limited review by the Office of Management and Budget, could be a source of error in reporting, as seen in an infant mortality study by the Centers for Disease Control. In addition, inconsistency can become a problem when agencies use data that are not collected according to federal standards because these sources may have their own rules that may differ from federal guidelines. The agencies GAO reviewed appear to understand this situation and took it into account when publishing reports.
GAO found that: (1) agencies used standard definitions in compliance with applicable laws and regulations, but the Office of Management and Budget (OMB) only monitored the agencies' methodologies for data collection and not the results of the data collection; (2) agencies' extensive use of Census Bureau support and products, which complied with OMB guidance, helped ensure uniformity; (3) the Departments of Justice and Education had consistency problems in their statistical reports because they used data collected by nonfederal entities which were not governed by federal guidance; (4) Justice compensated for the inconsistent reporting by state and local governments by noting deviations in its reports and by periodically surveying some populations to measure those unaccounted for; (5) Education had problems with states not completely reporting the racial and ethnic makeup of their student populations, particularly for those of mixed parentage; and (6) there were inconsistency problems when observers classified an individual's race or ethnicity, rather than having the individual classify its race.