Reassessing the effect of education on health

Citation

Esposito, Michael H. (2014). Reassessing the effect of education on health. 2014 Add Health Users Conference. Bethesda, MD.

Abstract

If the effect of education varies across racial groups has been of interest to researchers for some time. Though the empirical work on the subject has yielded valuable insights, much is susceptible to the same criticism directed at research examining education and health more generally. That is, the set of confounding factors which influence both one's health and level of educational attainment is large and often left unaccounted for, and thus, estimates of the effect of education on health (or any variation in it across groups) are biased in that they reflect a great deal of the variation found in factors prior to education. Given these criticisms, this paper seeks to address if the effect of education on health does vary across racial groups while taking special care in accounting for confounders. To achieve this, I utilize Bayesian Additive Regression Trees (a nonparametric modeling procedure which is adept at handling a very large number of covariates) in conjunction with the rich information provided in Add Health. Specifically, I examine if self-rated health, obesity, and hypertension vary by race and highest level of education completed at Wave IV, while accounting for a set of continuous and categorical factors (p = 45) from Wave I which likely served to influence later health and educational attainment. Given past research on the subject, I expect to find differences across racial groups, but for the degree of difference to be less than what has been previously reported.

URL

https://addhealth.cpc.unc.edu/wp-content/uploads/docs/events/20140613_Add_Health_Users_Conference_Abstracts.pdf

Reference Type

Conference proceeding

Book Title

2014 Add Health Users Conference

Author(s)

Esposito, Michael H.

Year Published

2014

City of Publication

Bethesda, MD

Reference ID

6223