Rates of Allocation for Disability Items by Mode in the American Community Survey

Abstract

Knowing the prevalence of disability in a population is seen as informative to policy and public health efforts. The American Community Survey (ACS), a yearly, large scale population-based survey administered by the United States (US) Census Bureau, is the predominant source for estimating disability prevalence in the US population. To provide complete data, procedures such as “allocations” are used by the bureau to fix illogical and missing responses to survey questions. This study investigates the rate of allocation in disability items by using the ACS Public Use Microdata Sample (PUMS) 3-year (2009-2011) file. Rates of allocation are interpreted as a measure of “within-person item-completeness” (WIPC)—where the presences of high allocation rates signal low levels of questionnaire completeness. The highest rates of allocations were in found in mail mode for: males (3.5%); those with no college (4.4%); households who only speak Spanish (5.9%); non-married family households (4.3%); and for Non-Latino-Blacks (6.5%). Researchers should continue to explore how the equilibrium of population estimates from ACS sample data vary as a function of demo- and geo-graphic characteristics.