Predicting performance on a 3rd grade high-stakes reading assessment

Abstract

In this study, we investigated the predictive value of Dynamic Indicators of Basic Early Literacy Skills (DIBELS) measures for performance on a high-stakes reading assessment at the end of third grade. DIBELS measures were administered to students at the end of third grade, around the same time as their high-stakes tests. Collectively, the measures succeeded in predicting 88.7% of those who would pass the high-stakes test, but only 68.8% of those who did not pass. Importantly, rate, alone, predicted 88.3% of those who would pass and 67.7% of those who did not pass. Multiple regression analyses were next used to test how well the subscales of DIBELS Next predicted students’ scores on Reading Comprehension Achievement Test (RCAT) in both the pass and fail groups. Also, for practical purposes, we examined both the scores and characteristics of students for whom the logistic regression failed to predicted performance. Implications for instruction, assessment, and future research are discussed.