Data from 79 male alcoholics who were randomly assigned to either coping skills training or interactional group psychotherapy were used to replicate a multidimensional, empirically derived typology and to evaluate the typology's usefulness in matching patients to treatment. Consistent with previous cluster analysis research, indicators of risk for alcoholism, alcohol dependence, drinking history, and psychopathological impairment distinguished alcoholics along two broad dimensions of vulnerability and severity, with one subtype (type B alcoholics) manifesting an earlier onset of problem drinking, more familial alcoholism, greater dependence on alcohol, and more symptoms of antisocial personality than the other subtype (type A alcoholics). Analyses of outcome indicated that type A alcoholics fared better in interactional treatment and more poorly with coping skills training. Conversely, type B alcoholics had better outcomes with the coping skills treatment and worse outcomes with interactional therapy. Differences in treatment response were maintained for 2 years from the beginning of aftercare treatment.