This report describes the development and pilot testing of a low-cost, generalizable, quality-assured, peer-instructional model suitable to the training needs of men of varying measured aptitude. The report presents a brief overview of the project, followed by a detailed description of the APSTRAT model and the considerations that led to its development. The model is discussed in terms of the instructional principles incorporated and the practical constraints accommodated. The data comparing the performance proficiency, academic attrition and recycles, and costs of the conventional and APSTRAT systems indicate that APSTRAT students achieve greater proficiency with a reduction in the rate of academic attrition and a considerable savings in cost. (Author)