Demobilizing a Generation : The Behavioral Effects of the Vietnam Draft Lottery

A generation of men were assigned lottery numbers based on birth dates to determine draft status during the Vietnam War. This natural experiment offers an ideal setting to assess the behavioral and psychological effects of differential exposure to the risk of serving and dying in a politically controversial military conflict. However, a conventional analysis of these effects may be biased by out-of-sample attrition associated with lottery numbers and behavioral outcomes. We develop a novel methodological approach, maximum entropy imputation (MEI) to correct for potential attrition bias in estimating causal effects in experiments. MEI reweights observations to construct a synthetic sample from the group of ‘non-attritioners’ that near-perfectly resemble actual ‘attritioners’ in observable characteristics. We discuss the assumptions of MEI for identifying causal inferences, and present the results from a series of validation tests that assess the reasonableness of these assumptions in the context of the Vietnam draft lottery. We then employ MEI to estimate the effects of the lottery on important political outcomes in the Youth-Parent Political Socialization Study, and show that the draft lottery may have had the overall effect of disengaging the political participation of a generation of men, while simultaneously inducing more conservative political attitudes