The abuse of alcohol is a complex behavior pattern exhibited by approximately nine million people in this country (Tarter & Sugarman, 1976) and has become a field of interest to science, industry, and the helping professions. Estimates of spontaneous recovery among alcoholics vary drastically. In 1971 the National Institute on Alcohol Abuse and Alcoholism began sponsorship of a comprehensive alcoholism treatment program located in 45 community centers throughout the nation. Using a multiple criterion measure of improvement, their report (Rand Corporation, 1976) estimated the rate of spontaneous improvement for alcoholics to be as high as 53%. In reviewing the rates of spontaneous improvement from a number of studies, Baekland (1977) states, "It thus appears that depending on the patient's personal and social assets, there is a 2-15% spontaneous improvement rate in alcoholics who do not receive formal treatment" (P 390). Unlike the Rand Report, Baekland used total abstinence ·as the measure of spontaneous improvement. The present study sought to compare the relative effectiveness of group administered covert sensitization with traditional insight-oriented group therapy in treating alcoholism. To control for the effects of relaxation training, therapist contact, favorable outcome expectancy, and the act of imaging (variables inherent in the covert sensitization procedure but not controlled for in the previously cited outcome studies), a relaxation placebo control group was employed which was empirically evaluated for its credibility. It was hypothesized that subjects in all three groups would show significant improvement over time on each of the dependent measures, and that subjects receiving covert sensitization would show significantly greater improvements than subjects receiving traditional group therapy or the relaxation placebo treatment. No differences were expected between group therapy and the relaxation placebo treatment.