Causal Effects of Schooling on Memory at Older Ages in Six Low-and-Middle Income Countries: Nonparametric Evidence with Harmonized Datasets

Abstract

Objective

Higher schooling attainment is associated with better cognitive function at older ages, but it remains unclear whether the relationship is causal. We estimated causal effects of schooling on performances on the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) word-recall (memory) test at older ages in China, Ghana, India, Mexico, Russia, and South Africa.

Methods

We used harmonized data (n = 30,896) on older adults (≥50 years) from the World Health Organization Study on Global Ageing and Adult Health. We applied an established nonparametric partial-identification approach that bounds causal effects of increasing schooling attainment at different parts of the schooling distributions under relatively weak assumptions.

Results

An additional year of schooling increased word-recall scores by between 0.01-0.13 standard deviations (SDs) in China, 0.01-0.06 SDs in Ghana, 0.02-0.09 SDs in India, 0.02-0.12 SDs in Mexico, and 0-0.07 SDs in South Africa when increasing schooling from never attended to primary. No results were obtained for Russia at this margin due to the low proportion of older adults with primary schooling or lower. At higher parts of the schooling distributions (e.g., high school or university completion) the bounds cannot statistically reject null effects.

Discussion

Our results indicate that increasing schooling from never attended to primary had long-lasting effects on memory decades later in life for older adults in five diverse low-and-middle-income countries.

Cognition, Partial-Identification, Bounds, Causal Inference

Accepted manuscripts

Accepted manuscripts are PDF versions of the author’s final manuscript, as accepted for publication by the journal but prior to copyediting or typesetting. They can be cited using the author(s), article title, journal title, year of online publication, and DOI. They will be replaced by the final typeset articles, which may therefore contain changes. The DOI will remain the same throughout.