The idea of learner control over instruction has enjoyed increasing popularity as a result of the growth of computer-assisted instruction in the schools. Some research results indicate that individuals learn more when given control over their instruction; other research indicates that individuals learn less effectively. This study investigates the effects of learner control and program control in full and lean computer-assisted instructional programs on the achievement of higher-ability and lower-ability university students. Subjects were 200 undergraduate education majors enrolled in their first semester of a professional teacher preparation program. The chapters from a required course textbook that were adapted into a computer-assisted program contained a total of 13 learning objectives. The program was developed in four versions that represent the four different treatment conditions (program-control lean and full and learner-control lean and full). Program control subjects were required to respond to all 174 screens in their version of the program; the full and lean versions varied in amount of practice-with-feedback screens (66 and 22, respectively). In the learner-control treatments, the learner began with practice-with feedback items for specific objectives. The lean version differed from the full in the manner in which students chose to add practice or move on to the next objective. A paper-and-pencil posttest covers the 13 objectives in the instructional program. A 13-item attitude questionnaire assesses subjects' satisfaction with the material, their perceived effort, desire for more information, continuing motivation and their confidence in their posttest performance. (Contains 25 references.) (AEF)