Predicting Reading Problems at Kindergarten for Children in Second Grade: SEARCH as a Screen

Abstract

In this study, 668 kindergarten children were screened and followed through second grade to compare the 2-year predictive validity of SEARCH (a test designed to assess delays in the acquisition of spatial and temporal information) with the predictive validity of academic achievement tests as screens for identifying children at-risk for developing reading problems. Six different cutoff scores on SEARCH were used to establish risk status, and reading scores at or below the 30th percentile were used to establish both at-risk status and inadequate reading at second grade. The use of SEARCH scores indicating major risk resulted in false negative error rates in the 52% to 69% range and false positive error rates in the 56% to 70% range. The use of academic tests to predict reading problems resulted in error rates of 65% for false negative and 62 % for false positive. The results are discussed in the context of issues in measuring perceptual and cognitive performance in an age range where normal and deviant variation are difficult to assess except in extreme cases.