The Mastery-Learning test model is extended. Methods for estimating prior probabilities are described. The use of an adjustment matrix to transform a probability of mastery measure and empirical methods for estimating adjustment matrix parameters are derived. Adjustment matrices are interpreted as indicators of instructional effectiveness and as evidence of the existence of learning hierarchies. Two decision variables are considered: probability of mastery for an individual and proportion in mastery for an instructional group. Discussion of the reliability, complexity, and interpretability of these decision variables and comparison with decision variables for other test models is also included. (Author/DEP)