Social
20-year income volatility and cognitive function in NLSY79: a replication and extension of CARDIA findings Katrina L. Kezios* Katrina Kezios Calvin L. Colvin Leslie Grasset Catherine dP Duarte Anusha M. Vable M. Maria Glymour Adina Zeki Al Hazzouri
Life course financial circumstances may shape cognitive health and dementia risk. A 2019 study by Grasset et al. reported that income volatility in the CARDIA cohort over a 20-year period (1990-2010) in midlife was associated with worse cognitive function. However, CARDIA lacked a measure of early-life cognition, so reverse causation was a concern. Here, we replicate this study in the National Longitudinal Survey of Youth 1979 (NLSY79), a nationally representative sample nearly identical in age (23-35 to 43-55) to CARDIA participants during 1990-2010 and which includes an early-adulthood measure of cognition (via the Armed Forces Qualifications Test, administered in 1980). Income volatility was assessed as: standard deviation of the percent change in income and number of income drops >25% between ~biannual surveys from 1990-2010. Participants completed immediate and delayed 10-word recall, backwards counting, and serial 7 subtraction assessments during the survey at which they turned at least 48. Our sample (N=2,384) included those who completed all assessments in 2010 or later. Z-scored assessments were averaged to create a global cognition score and subdomain scores for memory and attention. Confounder adjusted linear models with bootstrapped 95% CIs were estimated for global cognition and each subdomain. While associations were smaller, conclusions were consistent with Grasset et al.: higher volatility and more income drops were associated with lower global cognition (e.g., β [95% CI] for 3+ vs 0 drops: -0.15 [-0.28, -0.02]). Early-adulthood cognition was a strong predictor of midlife cognition (ρ=0.39, p<0.001) and findings were stronger prior to its adjustment (e.g., β [95% CI] for 3+ vs 0 drops: -0.23 [-0.37, -0.11], which may partly explain the smaller coefficients in our study compared to Grasset et al. Overall, our findings add to a growing literature demonstrating the cognitive harms of life course financial hardship, including income volatility.