The application of Information Integration Theory (Anderson, 1981), a general model of social judgment, overcomes shortcomings in the evaluation of moral development by offering a clear distinction between moral values and reasoning. To test the applicability of Anderson's theory to moral development research, two experiments were conducted using high school freshmen who were given eight written standard moral dilemmas to respond to. In experiment 1, students who had just completed a moral decision making training program (N=64) were compared to controls (N=64). In experiment 2, students (N=32) were subjected to a cognitive reasoning training program designed to increase use of relativistic weighting with nonmoral social judgment tasks. The effect of training on moral reasoning was determined for these students, as compared to a control group (N=32) that had not had the training program. In experiment 1, trained students responded more strongly toward the moralistic choice and showed less configuration or relativity in considering rationales than nontrained students. Trained subjects were not more influenced by postconventional rationales and were not less influenced by conventional or preconventional rationales. In experiment 2, trained subjects weighted the cost of helping more after discussions than before. There was no tendency for trained subjects to choose more moral alternatives. In general the information integration approach provided a powerful means of separating reasoning from values. (JAC)