In developing a model to utilize simulation techniques for assigning remedial reading to children, 100 third graders in a St. Paul, Minnesota, school district were placed in 16 remedial classes at 12 schools under the supervision of seven experienced remedial reading teachers. All of the subjects had scored 90 or above on a group IQ test and were designated as reading one grade below placement. The classes were divided for 8 months of training into three groups, each taught by different remedial methods: one visual, one kinesthetic, and one auditory-phonic approach. Pretest batteries to dermine skills in perceptual, perceptual-motor, and verbal areas and three reading tests were administered. The average of two reading tests was used as a post-test criterion. Prediction equations in all three methods were applied to the post-test scores in order to compare simulated reactions to the two methods not used with a pupil to his actual reaction to the method used. More than half the total sample predictively would have achieved a tested reading level of from 7 months to 1 year higher had they been placed in methods other than the one to which they were assigned. This suggested that children's reactions to a remedial treatment can be predicted by their individual characteristics. Tables and references are included. (BT)