Predicting Reading Problems In At-Risk Children

Abstract

This study was designed to determine early predictors of reading problems in children at risk for such problems. Three groups of children participated in the study: those with a specific language impairment; those who presumably had a language delay or disorder early in life and had no or a mild disorder at present; and a group of premature children. The data collected were standard speech and language test measures, given as the children entered the study, measures of language meta-processing abilities on an experimental battery, given 6 months after they entered the study; and standard measures of reading, given when the children were aged 80 to 96 months. Many significant relations were found between measures of oral language ability and meta-processing ability at an earlier age and reading ability in first and second grade. The language processing battery scores accounted for a somewhat greater amount of variance on the reading tests than did the scores on the standard language tests. Three ability groups, comparatively high, middle, and low, were found in the population as a result of cluster analysis. Some premature children and some children with early language disorder or mild language disorder at entry into the study, as well as most SLI children, were members of the low language-ability group. These data were then examined to see if membership in the low language-ability group, as measured by either standard speech and language tests or the language meta-processing battery predicted at-risk reading performance on the WRAT. Forty-six children were found to be at risk by this test. Twenty-one of the children were identified by either set of measures, an additional 10 were identified by the language meta-processing measures alone, an additional 3 by the intake measures alone, and 12 of the children were not identified by either set of measures as potential problem readers. The results indicate that early measures of language awareness are good predictors of later reading performance but that different measures of this awareness are good predictors for different children.