Cross Validating a New Preschool Screening Test

Abstract

A new preschool screening battery, consisting of five brief cognitive tasks, was presented to a sample of four-and five-year-old children with, and without mild learning problems. The sample included 177 normally achieving children and 21 children classified as either learning disabled (n = 7) or developmentally delayed (n = 14). Cross-sample test validation was demonstrated when 81% of both educational groups were correctly classified. The White/nonHispanic group achieved significantly higher screening scores than an omnibus minority group even when controlling for testing language. However, Hispanic children tested in English had a significantly higher screening score than Hispanic children tested in Spanish or both English and Spanish, and this impacted the race/ethnicity comparisons. Using percentage of exact matches, the interrater agreement was 80% or greater for all but one of the tasks. It was greater than 90% for all five tasks when the criterion for an agreement between raters was achieving scores within 1 point of each other.