The Impact of Data Collection Frequency and Trend Estimation Method on the Consistency of Growth Estimates From Two Computer-Adaptive Tests

Abstract

ABSTRACTWithin multitiered systems of support, assessment practices that limit the amount of time students miss instruction should be prioritized. At the same time, decisions about student response to intervention need to be based upon technically adequate data. We evaluated the impact of data collection frequency and trend estimation method on the magnitude of average rates of growth as well as the stability of individual estimates of growth from a computer-adaptive test. Students in Grades 2 through 5 (n > 2,000) were progress monitored once a month across the 2015–2016 school year with Star Reading or Star Math assessments. Results suggest that using ordinary least squares regression to estimate growth from triannual screening periods is generally sufficient to make program evaluation decisions about response to instruction across a school year. To make decisions about individual student progress, data should be collected at a minimum bimonthly but preferably once a month.