Heterogeneous Effects of Education on Mortality in the United States: Evidence from Within-Twin Pair Comparisons

Fabio Bolz, University of Minnesota
John Robert Warren, University of Minnesota
Andrew Halpern-Manners, Indiana University
Jonas Helgertz, University of Minnesota/Lund University
Evan Roberts, University of Minnesota

There are large educational disparities in mortality in the United States. The association between education and mortality may either be partly or entirely due to a causal effect of education on longevity or may merely be the result of selection on unobserved confounders. Several recent studies identify an effect of education on mortality using relatively robust designs consistent with a causal interpretation of the education-mortality association. Little is known, however, about how the effect of education on mortality varies across subgroups of the U.S. population and the few previous studies on this topic do not employ robust methodological designs. Leveraging within-twin pair variation in educational attainment we examine whether the effect of education on mortality in the United States varies by gender, race, family socioeconomic background, and region. We identify all twins born between 1910 and 1920 in linked full count U.S. census data which we subsequently link to administrative death record data and estimate twin fixed-effects models interacting educational attainment with the respective moderator. The findings will have important implications for our understanding of the processes underlying between-group disparities in mortality in the United States as well as for policies targeted at reducing lifespan inequality.

Keywords: Mortality and Longevity, Inequality, Disadvantage and Discrimination, Linked data sets , Census data

See extended abstract.