Sixty-one learning disabled (LD) adolescents in four junior high schools were randomly assigned to experimental or control groups as part of an effort to teach LD students to set realistic goals so they might experience success and satisfaction in school. Ss in the experimental group made achievement contracts and predicted their performance in either math or spelling. Pre- and posttest scores on the Intellectual Achievement Responsibility Questionnaire, the Task Attribution Questionnaire, the Test of Goal Setting Strategies, and the Evaluation of Treatment Inventory revealed that the intervention was successful in teaching strategies for setting realistic goals, and that experimental Ss used effort attributions to explain their performance significantly more often than they used ability, luck, or task difficulty attributions. Three case studies illustrate strategies used by LD students in the study. (CL)