This study evaluates the implications of two alcoholism prevention models. The single distribution (log-normal) model posits that the average level of consumption in a society is sufficient to account for the rate of alcoholism; the sociocultural model suggests that variables other than consumption account for alcoholism. Factor analysis and multiple regression were used to assess interstate differences in average alcohol consumption and alcoholism rates. Consumption, controlling for alcoholism rate, was found not wholly to be an economic phenomenon but rather was predicted byurban conditions(a factor measuring unintegrated foreign-born and minority groups and external social control) and two alcohol availability factors. Alcoholism rate was predicted byurban conditionsand a social isolation factor,isolated females.Consumption was not a significant predictor of alcoholism in this multivariable analysis. It was concluded that it is an oversimplification to view alcoholism merely as an extension of heavy drinking. Availability is not a unitary dimension and appears, furthermore, to have little potential utility in controlling consumption of alcoholism. Neither of the two availability factors was related to alcoholism; bootlegging appears to be a compensatory mechanism for offsetting low legal availability. The results imply that alcohol control policies and alcoholism prevention need to be directed toward alleviating anomie and social isolation. A variety of efforts toward these ends are suggested: senior citizens programs, minority employment programs, English enhancement training for the foreign-born, etc.