This study assessed whether learner control of feedback has a beneficial effect in student performance and attitude, focusing on whether the degree of elaboration feedback needed for high performance could be determined. Program and learner control were used at three levels of feedback, verification, correct answer, and elaboration. Subjects were 126 undergraduate education majors at a southwestern university, divided into 6 experimental groups. This study and a previous study demonstrate that verification feedback alone is insufficient for subjects to reach high levels of performance on a posttest. This study shows only a minimum improvement when elaboration was added to the correct answer feedback. The additional time and expense needed to design the instruction with elaboration feedback is probably not warranted, since students perform at almost the same level with only the correct answer. The pattern of improved performance from verification to correct answer was seen in program and learner-control conditions. Program control subjects always outperformed learner control subjects, and learner control subjects spent less time studying feedback than did program control subjects. Results suggest that program control with correct answer feedback is the most effective and efficient way to present instruction on the computer. (Contains 4 references.) (SLD)