Testing the Importance of Individual Growth in Predicting State-Level Outcomes Beyond Status Measures

Abstract

The literature reports mixed findings on whether measuring individual change over time on an interim progress monitoring assessment adds value to understanding student differences in future performance on an assessment. This study examines the relations among descriptive measures of growth (simple difference and average difference) and inferential measures [ordinary least squares (OLS) and empirical Bayes] for 800,000 students in grades 4, 8, and 10 and considers how well such measures statistically explain differences in end-of-year reading comprehension after controlling for student performance on a mid-year status assessment. Student differences in their reading comprehension performance were explained by the four growth estimates (simple difference, average difference, OLS, and empirical Bayes) and differed by status variable used (i.e., performance on the fall, winter, or spring benchmark assessment). The four growth estimates examined in the study all contributed significantly to predicting end-of-year reading comprehension when initial, fall performance was used as a covariate. The simple difference growth estimate was the best predictor when controlling for mid-year (winter) status, and all but the simple difference estimate contributed significantly when controlling for final (spring) status.