The Predictive Utility of Kindergarten Screening for Math Difficulty: How, When, and With Respect to What Outcome Should It Occur?

Abstract

The purpose of this study was to examine the reliability, validity, and predictive utility of 3 measures for screening kindergarten students for risk for math difficulty (MD). The screening measures assessed number sense and computational fluency, constructs central to typical early mathematical development. Conceptual and operational outcomes were assessed at the end of 1st grade, with MD operationalized as below the 16th percentile. The study evaluated how single-skill vs. multiple-skill screeners compared in terms of predictive efficiency; the accuracy of fall vs. spring kindergarten screening; and whether development could be forecast more precisely in terms of conceptual or operational outcomes. Students (n = 196) were tested during the fall and spring of the kindergarten year; follow-up data collection occurred during the spring of 1st grade. Alternate form and internal consistency reliability, concurrent and predictive validity, and predictive utility for the screening measures were examined. Logistic regression and ROC analyses indicated that the multiple-skill screeners produced good and similar fits as the single-skill quantity discrimination measure at both the fall and spring screening occasions in forecasting conceptual mathematics outcome at the end of 1st grade. To forecast operational outcome, the multiple- and single-skill screening measures produced similar but less accurate fits.