Importance of Predictive Power in Early Screening Assessments

As schools implement Response-to-Intervention (RTI) to identify and serve students with learning difficulties, it is critical for educators to know how to evaluate screening measures. In the present study, DIBELS oral reading fluency was used to compare the differential decisions that might occur in screening accuracy when predicting two reading comprehension measures (i.e., Stanford Achievement Test – 10th Edition [SAT10] & Gates-McGinitie Reading Test -4th Edition [GMRT]) at the end of second grade. The results showed that the DIBELS oral reading fluency tended to have higher sensitivity and negative predictive power for SAT10, and higher specificity and positive predictive power for GMRT. Furthermore, attempting to achieve a criterion of positive predictive power for a given reading comprehension outcome (SAT10 in this study) appears to render a favorable balance compared to other indices of diagnostic accuracy. These results are discussed in light of trade-offs and a need for considering specific contexts of schools and districts.