PREDICTION OF EARLY SCHOOL PERFORMANCE

Abstract

This study compares observations embedded during a health procedure (START), a developmental screening test (DDST), and an IQ test (Stanford-Binet) as preschool predictors of school performance. The START procedure is a method in which standard vision acuity procedures are altered so that the approach to testing can be modified based on the child's level of understanding and so that behavioral responses can be quantified without prolonging the procedure itself. Our previous research has consistently demonstrated that prediction of concurrent developmental status is as good or better than that achieved by extensive developmental tests designed exclusively for assessing development. We have also demonstrated that the strictly visual acuity component is more cost-effective than standard acuity screening apart from consideration of the developmental data obtained. A county-wide cohort of prekindergarten children (N=382) were screened using START and a parent report (GD-45). A stratified sample, (N=120) based on the GD-45 results with oversampling of low scores, was recalled and tested with the DDST and the Stanford-Binet. Followup school data were obtained for all children in K and 1st grade. START screening was found to correlate with achievement testing (.50 to .58) better than the Stanford-Binet (.35 to .37). Clinical indices predicting poor school outcome (failure and/or special placement and/or achievement test results <20th %ile) were all slightly better from START than from the DDST and substantially better than obtained from IQ scores: (e.g. sensitivity of .64, .57, and .37 respectively).