Intercohort Variations in the Education-Health Gradient: Sociohistorical Changes in Early-Life Selection Mechanisms in the United States

Objectives: Most prior studies on cohort-specific changes in the education gradient relative to health treat the distribution of education within a particular cohort as a "starting place" for understanding later-life health disparities. This premise has obfuscated the role that sociohistorical changes in early-life selection mechanisms play in the widening of education-based inequalities in functional limitations across birth cohorts. Methods: Drawing from the Health and Retirement Survey (1992-2016; n = 20,920), this study employs inverse probability weight (IPW) to account for early-life selection mechanisms that are likely to affect both educational attainment and functional limitations. IPW-adjusted generalized linear mixed-effects models were used to estimate the total effect of education on functional limitations across birth cohorts (born 1924-1959). Results: A significant linear decline in the negative effects of childhood socioeconomic disadvantage on education (beta = 0.005, p < .01) over the birth year was documented. By contrast, the same variable's negative effect on functional health increased significantly (beta = 0.006, p < .001) across cohorts. Adjustment for childhood socioeconomic status did yield narrower education-based inequalities in functional limitations, but the difference between IPW-adjusted and unadjusted results was not statistically significant. The pattern of significant widening of education-based inequalities (beta = -0.05, p < .001) in functional limitations across birth cohorts was maintained. Discussion: This study underscores the role that sociohistorical changes in early-life selection mechanisms play in modifying patterns of education-based inequalities in health across cohorts.