Graduation Year

2023

Document Type

Thesis

Degree

M.S.P.H.

Degree Name

MS in Public Health (M.S.P.H.)

Degree Granting Department

Epidemiology and Biostatistics

Major Professor

Judith Rijnhart, M.Sc., Ph.D.

Co-Major Professor

Jason L. Salemi, M.P.H., Ph.D.

Committee Member

Hongdao Meng, M.D., M.P.H., Ph.D.

Keywords

delayed recall, dementia, immediate recall, mediation analysis, sociodemographic factors

Abstract

In the United States (U.S.), 14 million individuals aged 65+ are expected to be diagnosed with dementia by 2060, and women are nearly twice as likely to be diagnosed. Low education is a well-established risk factor for dementia and is hypothesized to partly explain the gender differences in late-life cognition. However, few studies have investigated education as a mediator of these gender differences. This study aims to investigate education as a mediator of gender differences in episodic memory using longitudinal data from the U.S. Health and Retirement Study (HRS). I used a sample of 35,721 respondents (20,190 women and 15,531 men) with episodic memory data available between 1998 and 2018 and data on self-reported education (median = 12, interquartile range = 12, 15). Analyses were adjusted for confounders (ethnicity, race, childhood SES, childhood immigration status, and birth year) and effect modification by race, ethnicity, childhood SES, and birth year was assessed. Episodic memory was measured as immediate and delayed recall. Linear and linear mixed-effects models were used to estimate the associations between gender and years of education and years of education and episodic memory. Education was a significant mediator of the association between gender and episodic memory. Race was an effect modifier of the association between gender and years of education, and race, ethnicity, and birth year were effect modifiers of the association between education and episodic memory. All direct and total effects were positive, indicating that women had higher episodic memory scores across all racial and ethnic groups for all birth years, both before and after adjustment for education. The indirect effects were negative in those who identified as White and other race, indicating that women recalled fewer words in comparison to men, because they had fewer years of education. The indirect effects were positive in those who identified as Black and African American, indicating that women recalled more words in comparison to men, because they had more years of education. These findings indicate that the gender difference in episodic memory would have been larger if, on average, women had the same educational level as men, and this level of protectiveness varies across levels of race, ethnicity, and birth year.

Download

Included in

Epidemiology Commons

Share

COinS