Goldberg J, Richards M S, Anderson R J, Rodin M B
University of Illinois, School of Public Health, Epidemiology and Biostatistics Program, Chicago 60680.
J Subst Abuse. 1991;3(3):307-13. doi: 10.1016/s0899-3289(10)80014-8.
The randomized natural experiment provided by the military draft lotteries of 1970, 1971, and 1972 is used to compare alcohol consumption in men born in the years 1950, 1951, and 1952. Alcohol consumption data are derived from the National Health Interview Survey supplements for 1977, 1983, and 1985. The analysis follows an intention-to-draft paradigm, where draft eligibility is determined solely by randomly assigned lottery numbers. Draft eligibility status is found to be unassociated with alcohol consumption, despite the fact that eligibility for the draft significantly increased the likelihood of military service and that veterans are more likely to report increased alcohol consumption compared to nonveterans. The randomized draft lottery provides an elegant natural experiment that potentially obviates the problem of confounding by premilitary service factors found in previous studies of alcohol consumption among veterans and nonveterans. However, the indirect association between draft eligibility and any health outcome (where military service is an intervening variable), together with the failure of the draft to produce a higher percent of veterans, prohibits application in all but the largest samples.
