Differential undercount of Mexican immigrant families in the U.S. Census

Abstract

The U.S. Census is well known to have been compromised by differential undercounting of various populations, in particular, ethnic minorities and immigrants. This study focuses on the undercount in the Mexican immigrant population through a review of both quantitative and qualitative research. It challenges some traditional assumptions, such as that the Hispanic population is homogenous and that Census Bureau estimates of its undercount are valid for all sub-populations of Hispanics. It identifies a key reason for the undercount of the Mexican immigrant population, low-income households living in unusual and/or concealed housing units that are not in the Census Bureau’s Master Address File (MAF) and are unlikely to be enumerated or subsequently captured in the bureau’s Non-Response Follow Up (NRFU). Indeed, it seems probable that non-standard housing conditions in the local community account for one-third to one-half of the total undercount of minorities and immigrants, playing a greater role in undercount than traditional explanations such as lack of respondent motivation or inability to respond due to language or literacy constraints. Recommendations are offered to ameliorate the chronic differential undercount of this population, including employing a multi-variable analysis of census data, using community-based local address canvassing, and revising Census Bureau data processing procedures.