Abstract The state of the art in the evaluation of alcoholism treatment programs for the years 1968–1971 was surveyed by reviewing journal articles reporting treatment evaluations. A comprehensive computer-based search of the literature yielded 40 such studies. These were analyzed along 10 methodological dimensions. Serious inadequacies in evaluation methodology were revealed in a large proportion of the studies. Comparisons with two previous comprehensive reviews of program evaluations indicate modest improvements in description and methodology. These include greater use of control groups and random assignment of subjects; the use of multiple criteria of success and more adequate descriptions of follow-up techniques.