The effects of contingency training, instructor expressiveness, and student incentives on student achievement and attributions were investigated in a simulated college classroom. The following conditions were involved: a contingency manipulation resembling an aptitude test; an instructor lecture; two levels of student incentive; and an achievement test based on the lecture material. A total of 296 University of Manitoba students participated. Following an incentive manipulation, students wrote an aptitude test providing contingent, noncontingent, or no feedback, and responded to an attribution questionnaire. All students then observed a low or high expressive instructor, and completed an achievement test and an attribution questionnaire. After the contingency manipulation, noncontingent students reported less perceived control and made less internal attributions to their performance on an aptitude test. Post-lecture results indicated that the highly expressive instructor increased achievement and self-confidence in contingent compared to noncontingent students in low incentive conditions. The findings suggest that exposure to noncontingent outcomes can impair some aspects of a student's academic development. (Author/SW)