Explaining the higher nonresponse rates of immigrant populations: Insights from an analysis of panel attrition

Abstract Many studies have documented lower response rates of first- and second-generation immigrants in surveys. Several theories have been suggested to explain these differences, focusing on socioeconomic differences, differences in residence, and problems with cultural and social integration. Such explanations have been challenging to test, given the lack of information on sample subjects beforehand. In the current contribution, these explanations are tested by analyzing panel attrition. Using two-wave national panel data with a register- based oversample of immigrants in the Netherlands and regression decomposition methods, evidence was found for each of the three explanations. Attrition of immigrants is higher because of differences in how and where immigrants live and because immigrants have a lower socioeconomic status on average than natives. There is some evidence that attrition is increased by lower levels of cultural and social integration, but migrant-specific explanations were less influential than the more ‘generic’ explanations. Biases in descriptive evidence for key integration indicators existed due to selective attrition but were small in magnitude, lessening concerns about underestimations of immigrant integration in survey research.