Identifiers of predominantly Spanish-speaking children with language impairment.

Abstract

The purpose of this study was to identify a set of measures that would discriminate 31 predominantly Spanish-speaking children with normal language (NL children) from 31 children with language impairment (LI children). The LI children were identified as such by experienced, bilingual (Spanish/English), ASHA-certified, speech-language pathologists who were currently seeing the children in their caseloads. Children ranged in age from 5 to 7 years and were matched for age, gender, and school. Additionally, nonverbal cognitive measures assured that they did not differ significantly intellectually. Measures of vocabulary, novel bound-morpheme learning skills, and language form were randomly administered to all children. Further, parents responded to questions about their perceptions of their children's speech and language skills and family history of speech and language problems. A stepwise discriminant analysis indicated that 4 measures discriminated the groups of children with a sensitivity of 91.3% and a specificity of 100% (p < .0001): parental report of the child's speech and language skills, number of errors per T-unit, mean length per T-unit, and family history of speech and language problems. A second discriminant analysis indicated that the sensitivity and specificity could be maintained when only the first 2 measures were included. Confirmatory discriminant analyses of the 2- and 4-measure models indicated that the discriminant accuracy was stable on an independent sample.