Curriculum-based measurement of oral reading: multi-study evaluation of schedule, duration, and dataset quality on progress monitoring outcomes

Abstract

Curriculum-Based Measurement of Oral Reading (CBM-R) is used to collect time series data, estimate the rate of student achievement, and evaluate program effectiveness. A series of 5 studies were carried out to evaluate the validity, reliability, precision, and diagnostic accuracy of progress monitoring across a variety of progress monitoring durations, schedules, and dataset quality conditions. A sixth study evaluated the relation between the various conditions of progress monitoring (duration, schedule, and dataset quality) and the precision of weekly growth estimates. Model parameters were derived from a large extant progress monitoring dataset of second-grade (n=1517) and third-grade students (n=1561) receiving supplemental reading intervention as part of a Tier II response-to-intervention program. A linear mixed effects regression model was used to simulate true and observed CBM-R progress monitoring data. The validity and reliability of growth estimates were evaluated with squared correlations between true and observed scores along with split-half reliabilities of observed scores. The precision of growth estimates were evaluated with root mean square error between true and observed estimates of growth. Finally, receiver operator curves were used to evaluate the diagnostic accuracy and optimize decision thresholds. Results are interpreted to guide progress monitoring practices and inform future research.