There is keen interest in whether the American Community Survey (ACS) is successful in collecting complete data from all population groups and if, therefore, the ACS estimates are representative of these populations. While the ACS has demonstrated consistently high levels of total survey response at both national and sub-national levels, the Race and Ethnic Advisory Committees have requested similar measures of survey response by race and ethnicity. It is straightforward to produce survey response rates for specific geographic areas and geographic stratifications but we cannot produce response rates by demographics such as race and ethnicity in the ACS because the race and ethnicity of nonrespondents is unknown.
This paper provides an alternative measure that is not a response rate but provides important information about the inclusion of racial and ethnic groups in the ACS. We call these new measures, "inclusion rates" because they compare the demographic characteristics of interviewed ACS households (those that are directly included) with the expected demographic characteristics of all households.