The results of a study which used the regression discontinuity model in the analysis of criterion-referenced test data are presented and inherent drawbacks in the use of the regression discontinuity model with this type of data are discussed. The discussion is based on the results of an analysis of data collected in the Mesa, Arizona School District which uses criterion-referenced testing solely. The problems include celling effects and non-linearity of the data. Problems inherent in implementation of an evaluation design at the school level are also described. (Author/MV)