Do the benefits of education for late‐life cognition change across birth cohorts? Assessing period effects in the Health and Retirement Study

Abstract

AbstractBackgroundEducation consistently predicts late‐life cognition, but studies rarely consider the transformation of education that current older adults experienced across the 20th century. We assess whether the relationship between education and late‐life cognition remains constant across birth cohorts spanning over 40 years of changing social investments and progress.MethodUS Health and Retirement Study participants (age 50+) were enrolled at multiple points between 1998 and 2018, allowing for the study of participants of different birth years at the same age (n = 28,299 participants and 151,514 visits across all available data collection waves). US‐born participants with at least one memory measurement were stratified into eight birth cohorts of approximately five years each, spanning more than 40 years. Mixed models with random intercepts for each participant investigated the presence of effect modification across birth cohorts for the education‐dementia relationship. Age was not evenly distributed across birth cohorts; therefore models were additionally run restricted to ten‐year age groups. Secondary analyses investigated whether period effects further existed by race and gender. All models adjusted for age, quadratic age, gender, race, cognitive practice effects, childhood census region, and parental education.ResultEarlier birth cohorts showed larger associations between education and late‐life cognition, with each additional year increase in schooling associated with a 0.10 SD (95% CI: 0.09, 0.11) higher memory score in the earliest cohort compared to the latest (interaction p‐value <.0001). Findings were consistent when restricting analyses to age groups. Effects of education across cohorts also slightly differed by gender, with increased education being associated with larger gender disparities in memory of 0.04 SD (95% CI: 0.02, 0.06) more per year of schooling in the latest cohort compared to the earliest (interaction p‐value <.0001), and by race, with a narrowing of racial disparities in later cohorts for the cognitive benefits of increased education (interaction p‐value <.0001).ConclusionEach year increase in attained education had overall larger effects among earlier cohorts, though this may differ across sociodemographic and age groups. Future research is needed to examine whether differences arise from historical transformations of the educational experience or from changes in downstream economic opportunities across time periods.