Classification Agreement for Gated Screening in Mathematics: Subskill Mastery Measurement and Classwide Intervention

Abstract

This study examined the classification accuracy for subskill mastery measures administered in mathematics for students in kindergarten and Grades 1, 3, 5, and 7 at fall (n = 564) and winter (n = 602) screening. In addition, response to classwide math intervention was examined as another layer of screening for students in kindergarten and Grades 1, 3, and 5 (n = 387). Criterion measures were the year-end accountability measure for math in Grades 3, 5, and 7. For kindergarten and Grade 1 a researcher-constructed composite measure was used. For all measures (screening and criterion), 20th percentile performance was the risk criterion. Authors hypothesized that classwide intervention response would provide more accurate identification of risk than fall or winter screening. Receiver operating characteristic (ROC)–generated sensitivities and specificities supported this hypothesis. Data indicated that intervention dosage was adequate and did reduce the base rate of risk in each screening group. Posttest probabilities indicated that subskill mastery measures and response to classwide intervention could be used to accurately indicate math risk and the need for more intensive intervention for students in a multitier system of supports (MTSS) model.