Abstract
Gaps in life expectancy between Americans with and without a college degree have widened markedly over the past 3 decades. One explanation points to increasing educational attainment changing the type of people with and without a degree. If preexisting health condition in the 2 education groups changes as the fraction with a degree changes, health selection might explain the widening mortality gap. We examined this explanation using (1) education and mortality in each birth cohort of men and women from 1940 to 1988, and (2) the natural experiment caused by the Vietnam War, which increased the fractions of men with a degree in affected birth cohorts. For each cohort, we examined the relationship between the mortality gap and the fraction of the study population with a degree. We found no relationship between the fraction of a birth cohort with a degree and the corresponding mortality gap. For men, the large increase in college attendance spurred by the Vietnam War has no perceptible counterpart in the mortality gap. The evidence from the natural experiment induced by the Vietnam War does not support a health-selection explanation for the widening mortality gap. This article is part of a Special Collection on Methods in Social Epidemiology.