Does schooling attained by adult children affect parent's cognitive function and decline in later life? An evaluation comparing observational and quasi‐experimental approaches.

Background: A nascent body of research suggests that higher adult child educational attainment may have beneficial impacts on older parents' cognitive aging. Children with higher education may offer both financial and non‐financial resources that help support healthy brain aging for older parents. The causality of these relationships is difficult to evaluate; quasi‐experimental approaches are needed to address the confounding bias expected for observational studies. Method: We used data from the Mexican Health and Aging Study (n = 6741), a nationally representative cohort of adults aged ≥50 years (2012), with repeated measures of verbal memory and verbal fluency assessed in 2012, 2015, and 2018. We leveraged a 1993 national compulsory schooling law (CSL), which raised minimum schooling from six to nine years, as a source of quasi‐experimental variation in (adult) children's education that is independent of confounders of the relationship between adult child education and parents' cognition. We generated a binary instrumental variable based on the respondents' oldest child's birth cohort and used two‐stage least squares estimation (2SLS). We compared 2SLS estimates to those generated via ordinary least squares (OLS) regression and mixed linear models adjusted for respondent and family‐level characteristics. Result: At baseline, participants (55.3% female) were on average 60.0 years old. OLS models suggested that more schooling among oldest adult children was associated with better baseline verbal memory (ß(95% CI): 0.017(0.012,0.022)) and verbal fluency z‐scores (ß(95% CI): 0.020(0.014,0.026)). Exposure to the CSL was associated with an increase in average educational attainment by 0.612 years (95% CI: 0.384, 0.842). However, estimates from 2SLS suggested no quasi‐experimental association between increased adult child schooling and older respondents' verbal memory (ß(95% CI): ‐0.039(‐0.102,0.023)) or verbal fluency scores (ß(95% CI): ‐0.005(‐0.080,0.071)). In observational analyses, each year of increased schooling among oldest children was associated with slower decline in verbal memory z‐scores (ß(95% CI): 0.010(0.001,0.019)) per decade, but not with verbal fluency. Conclusion: Our findings present promising, but inconclusive evidence that increases in schooling may have a beneficial effect on parents' levels of cognitive performance. This work has important implications globally, given the massive increases in educational attainment achieved by recent generations in these settings. [ABSTRACT FROM AUTHOR]