A computer simulation procedure was developed to reproduce the overall pattern of results obtained in the Educational Testing Service Growth Study. Then simulated data for seven sets of 10,000 to 15,000 cases were analyzed, and findings compared on the basis of correlations between estimated and true growth scores. Findings showed that growth was estimated more accurately by the difference between pretest and posttest scores. The measure was said to be accurate and easier to compute, and meaningful to the outsider. It was concluded that more complex assessment procedures should not be used until better results can be drawn. (Author/SK)