An attributional approach to social behavior traces problems in personal adjustment back to the assumptions individuals formulate about the causes of behaviors and events. Attributional information presented during counseling may have therapeutically beneficial consequences. The effectiveness of attribution therapy was investigated in a factorial experiment which varied controllability and internality of causal factors. Pairs of subjects (N=82), who had previously completed Rotter's Internal-External Locus of Control Scale were given negative interpersonal evaluations followed by one of four types of therapy (internal/controllable, internal/uncontrollable, external/controllable, or external/uncontrollable) or no therapy at all. Results indicated that internal/controllable therapy produced more positive affective reactions and performance evaluations for internal locus of control respondents; externals were more variable in their responses. Findings suggest that the negative effects of stressful life events may be reduced by modifying individuals' attributional inferences. (Author/NRB)