Interactions of 18 demographic factors with alcoholism treatment outcome and with aftercare participation were studied. For 2 years subsequent to discharge from a 30-day inpatient program, 1210 alcoholic men veterans were followed up and grouped with regard to (1) posttreatment status (abstainers, improved, unimproved, unclassified, deceased) and (2) in terms of responders (abstainers and improved) and nonresponders (unimproved). Classifications were based on comparing each subject's 2-year posthospital drinking status with his 2-year prehospital drinking pattern. Seven variables were found to discriminate between the groups in both analyses. Those responding to treatment tended to be older, married and employed at admission, had more days of prehospital abstinence, were less likely to have had prior hospitalizations, and were more likely to participate in aftercare and to visit more frequently. In multivariate analyses, however, only the last two variables showed promising predictive ability, whereas the variables of days sober, age and married contributed only slightly to the prediction of treatment outcome. An effort to find variables that might predict aftercare participation was unsuccessful; the combined contribution of four factors amounted to only 5% of the total variance.