Exploring Curriculum-Based Measurement in Elementary Science: Investigating Two Vocabulary-Matching Formats

Abstract

As content literacy intervention is expanded in schools, data-based decision-making practices need to also advance, especially in the areas of science. Vocabulary-matching curriculum-based measures (VM-CBM) may allow educators to identify students needing additional support in science vocabulary to assist with using and comprehending disciplinary language. Typically, VM-CBMs have been given in a one-page format, but there has been little investigation of modified presentation of VM items. Participants were 77 fourth grade students from a U.S. Midwestern rural school district. Students from four different classrooms were administered either typical one-page or multipage VM-CBM forms. Multipage forms produced strong alternate form reliability (r = .92, 95% confidence interval [CI] = [.85, .96]). Typical forms produced a moderate reliability coefficient (r = .58, 95% CI = [.32, .76]). Regarding concurrent validity, multipage forms were also strongly correlated (r = .71, 95% CI = [.50, .84]; r = .72, 95% CI = [.51, .85]) while typical forms were weakly correlated (r = .40, 95% CI = [.10, .64]; r = .47, 95% CI = [.17, .69]) with a standardized state science assessment. The multipage format appears promising; however, further item level analysis is needed to determine the most efficient way to screen and support students in elementary science literacy.