The instructional effectiveness of learning programs derived from Gagne-type task analysis, ordering theory analysis, and random sequenced presentation of complex intellectual skills were investigated. Fifty-seven high school students completed a self-instructional program derived from one of the three sequences. No significant differences were found among the three groups in number of objectives mastered. The two groups receiving hierarchical presentations required significantly less time to complete the program than the random sequence group. The study seems to support in part the assumption that an optimal sequence of successful instruction exists. It is concluded that it is unwise to consider the learning of intellectual skills only in terms of prerequisite skills, but that interactions with other variables must be examined. For learning hierarchies to have a profound effect in the classroom, more efficient procedures for task analysis or hierarchy generation are needed. Improvements upon the ordering theory analysis may provide promise in this direction. (Author/LBH)