Since attrition is often a problem in adult training programs, a study was conducted to investigate the relationship between locus of control and course completion of adults enrolled in a retraining program. Rotter's Social Learning Theory of Personality was used as a starting point for the study. The study population was a sample of 108 English-speaking adults who began the Basic Training for Skill Development (BTSD) program at a community college in Eastern Ontario in the fall of 1983 and spring of 1984. Approximately one-half of the participants were federally sponsored and one-half paid their own fees. Data were collected using a modification of the Internal, Powerful Others and Chance (IPC) scales, the Gates-MacGinitie Reading Test, the Morrison-McCall Spelling Scale, and a geographical profile constructed for the study. Sixty-five percent of the population remained in the program for at least 13 weeks, whereas 79 percent of the noncompleters dropped out during this period. The major reason for dropping out was unexcused absence from class (47 percent), followed by medical reasons (18 percent). A one-way analysis of variance was used to test the hypothesis that there is a significant difference between completers and dropouts on the internal dimension of locus of control, with completers scoring higher on this dimension. No support was found for the hypothesis, and it was theorized that the IPC scales may not measure the locus of control construct effectively. (18 references) (KC)