Abstract
Oral reading fluency (ORF) assessments are commonly administered triannually in schools across the United States. Students read grade-level passages, and the number of words read correctly per minute (WCPM) serves as a universal screening measure. Observed WCPM scores have several psychometric limitations, including a global error of measurement and dependence on specific passages read by students. A model-based approach to ORF assessment, along with derived estimates of the WCPM scores, overcomes these limitations. This paper demonstrates the advantages of using a model-based approach to measure ORF from two perspectives: the availability of conditional standard error of measurement and efficient ORF score equating with a calibrated passage pool.