Vietnam-era military service: a test of the class-bias thesis.

A number of studies have tested the class-bias thesis that the burden of fighting the Vietnam War was disproportionately born by poor Americans. But findings as well as research methods have been inconsistent, leaving the empirical status of the class-bias thesis unresolved. Following a review of past studies, this paper presents an analysis of national survey data showing that class bias occurred, but it only modestly increased the burden born by low-status young American men while it sharply reduced the burden placed on their high-status peers.

Studies of Class Bias

Vietnam has been called a "class war" that for the most part was avoided by young men of at least middle-class background, by means of student and medical deferments. Extreme versions of the thesis claim involvement was overwhelmingly concentrated among the poor, as in a recent study by Arnold Barnett and his associates. But more often, expressions of the thesis are more moderate, for example that ... "the odds of working-class men going ... to Vietnam were far higher than they were for ... the privileged" in Christian D. Appy's recent book, or simply that Vietnam was not "an equal-opportunity war" as James Fallows recently noted.(1)

Tests of the class-bias thesis have been hampered by difficulty in obtaining data on social class origin for those serving in the military - the Department of Defense has never kept such information.(2) Some studies have inferred soldiers' social class origin from data about their neighborhoods, hometowns, counties, or states. These ecological studies have addressed war casualty rates, and the earliest, conducted by Mayer and Hoult, found that Korean War casualty rates-were roughly three times higher in the poorest compared to the richest Detroit census tracts, suggesting that "Korea was a poor man's war."(3) Comparing states' casualty rates for the Vietnam War, Willis reported that affluent states had lower casualty rates. Badillo and Curry calculated 1964-1972 Vietnam casualty rates along with socioeconomic scores based on income, education, and occupational data for Cook County, Illinois, census tracts. Officers' casualties were higher in wealthier tracts, but for casualties among enlisted men, higher-status tracts had significantly lower casualties. Foust and Botts' correlational analysis of county-level data revealed a modest tendency for white Vietnam war casualties to come from counties with low income and low educational levels. There was a similar but weaker tendency found for black casualties, but none for Hispanic casualties. Barnett, Stanley, and Shore ranked a sample of 1,510 Vietnam casualties according to hometown median family income, comparing this distribution to income data for the nation. Vietnam casualties were about 1.5 times as likely to hail from poorer hometowns than from more affluent communities. Based on data for block-groups in selected cities, casualties were found somewhat more likely to come from the poorer neighborhoods within communities.

There are difficulties with these ecological studies. First, they have addressed only casualties and not the broader question of class bias in military participation generally. Vietnam-era armed forces personnel bore a number of "costs" in addition to the possibility of death in combat, including the threat of serious injury, as well as stress, foregone educational and occupational opportunities, and - among conscripts and reluctant volunteers - the burden of coercion.(4)

Second, ecological data-based findings can not conclusively test the class-bias thesis, because conclusions about one level of analysis (individual soldiers' social class origins) cannot logically be drawn from data for another level of analysis (socioeconomic characteristics of soldiers' census tracts, hometowns, or states).(5) Study authors seem aware of this problem: Foust and Botts warn their readers of it, and Willis and Barnett et al. each specifically discuss it. But conclusions pertaining to the class-bias thesis are drawn from ecological data nonetheless (although Badillo and Curry resist the temptation and scrupulously avoid individual-level conclusions from their census tract-level data). Willis concludes from state-level data …