To actually use this model, you need to convert all data to applicant percentile data, i.e. students with a score of 163 have a score that is at least as good as 81% of applicants and students with a score of 153 have a score that is at least as good as 43% of applicants. You also need to assign values for the probability of admission for applicants and the probability of matriculation for admitted students with these percentile rankings of scores. The model then determines the fraction of applicants that matriculate and the distribution of percentile scores among matriculants. This Demonstration outputs the mean, 25th percentile, 50th percentile (median) and 75th percentile scores.