The Role of Selection in the College Education-Health Link

Education attainment has a statistical association with various adult health outcomes. This association is among the most robust and replicated in the social sciences (Fletcher 2015). Nowadays many social scientists believe that a key factor in improving adult health may be improving education (Mazumder 2008). But education is an endogenous variable in the health production model. The association between education and health might be a product of selection bias: reverse causation and confounding by unmeasured variables (or omitted variable bias). Reverse causation refers to the possibility that healthier individuals may stay in school longer and be more able to attain higher education. Omitted variable bias may result from the omitted factors that contribute to both education and health, e.g., family background, childhood health and nutrition status, intelligence, personal attitudes or even genes. The issues of reverse causation and omitted variable bias should be completely understood before we can claim that education has a causal effect on health and propose corresponding policies to improve health. Otherwise, policies aimed at increasing education attainment may not necessarily lead to better health. This project proposes two research designs and three analytical approaches to investigate whether the relationship between college education and health is causal.