Estimating the Characteristics of Unauthorized Immigrants Using U.S. Census Data: Combined Sample Multiple Imputation

Abstract

Contemporary U.S. immigration policy debates would be better informed by more accurate data about how many unauthorized immigrants reside in the country, where they reside, and the conditions in which they live. Researchers use demographic methods to generate aggregated information about the number and demographic composition of the unauthorized immigrant population. But understanding their social and economic characteristics (e.g., educational attainment, occupations) often requires identifying likely unauthorized immigrants at the individual level. We describe a new method that pools data from the Survey of Income and Program Participation (SIPP), which identifies unauthorized immigrants, with data from the American Community Survey (ACS), which does not. This method treats unauthorized status as missing data to be imputed by multiple imputation techniques. Likely unauthorized immigrants in the ACS are identified based on similarities to self-reported unauthorized immigrants in the SIPP. This process allows state and local disaggregation of unauthorized immigrant populations and analysis of subpopulations such as Deferred Action for Childhood Arrivals (DACA) applicants.