Most recent reviewers of treatment outcome research agree on a few conclusions. People receiving alcoholism treatment on the average have outcomes a little better than those who don't. However the robustness of current treatment interventions is limited and improvement is transitory. Moreover, all such interventions are about equally (in) effective. Most of the explained variance in outcomes is attributable to patient, rather than treatment factors. This being so the search for greater efficacy is turning away from psychosocial variables and toward biology. An alternative view is that psychosocial interventioins can be robust and durable. The matching hypothesis is that appropriate matching of treatments to patients will reveal enhanced outcomes. The study aims to test the short and long-term efficacy of the Community Reinforcement Approach (CRA) by comparing it with an Individual Focus Approach (IFA), both administered within a standard social learning program in a partial hospital setting. It is hypothesized that the CRA will be most effective with patients with poorer social functioning, while the IFA will be most effective for patients with poorer psychological functioning. It is further hypothesized that the differential effects of the two programs on psychologically dysfunctional patients will diminish over time, while the differential treatment effects for those who are socially more dysfunctional will increase. To test these matching hypotheses patients randomly assigned to the two treatment programs will be followed up for two years after treatment, conducting follow-up interviews at four month intervals, and gathering more extensive information at one and two years after treatment.