Educational and Health Selection in Ghana–U.S. Migration: Household Comparisons and Duration-Based Health Patterns

Abstract

AbstractThis thesis examines whether migration from Ghana to the United States is associated with patterns consistent with educational and health selection, and whether migrant health varies with duration of residence in the United States. While migration research frequently documents positive selection, empirical evidence on health selection among sub-Saharan African migrants remains limited and methodologically sensitive. This study distinguishes between educational and health dimensions of selection and evaluates whether observable differences are consistent with selective migration. The analysis integrates original household survey data from Ghana with U.S. Current Population Survey (CPS) ASEC data and supplementary evidence from the American Community Survey (ACS). First, individual-level models using Ghana survey data estimate the probability of residing in a household with U.S. relatives to assess cross-household differences in education and health. Second, regression-based comparisons examine differences between Ghana-based individuals and migrants reported by origin households, distinguishing educational from health outcomes. Supplementary pilot analyses assess whether higher educational attainment among migrants was likely achieved prior to migration. Third, weighted cross-sectional models using CPS ASEC data analyze whether self-rated health and work-limiting disability vary with years lived in the United States. Duration measures are constructed directly from reported year of arrival, and functional disability patterns are evaluated using ACS data for external comparison. Results indicate strong and consistent evidence of educational differences between migrant-connected and non-migrant populations, consistent with positive educational selection. Migrants exhibit higher completed education relative to comparable Ghana-based individuals. In contrast, evidence of health selection is limited and sensitive to measurement specification, with differences not consistently statistically significant. In U.S. data, there is no consistent evidence of declining self-rated health with longer duration of residence, and associations between duration and disability are modest and not robust across alternative specifications. Overall, migration from Ghana to the United States appears strongly selective on education but shows limited and measure-dependent evidence of health selection. Health is measured at different time points across data sources—current health among Ghana-based individuals, proxy-reported pre-migration health for migrants, and current self-reported health and disability in U.S. surveys—which warrants cautious interpretation. These findings underscore the importance of analytically separating educational and health dimensions of selection when evaluating migrant outcomes.