| Summary: | This study identified a set of measures that accurately and efficiently discriminated between predominantly Spanish-speaking children with normal language and with language impairment. Twenty-one 5- to 7-year-old children with normal language and 21 with language impairment, matched for age, gender, and school were studied. Each child responded to a set of verbal and nonverbal measures. The verbal measures assessed vocabulary and bound-morpheme learning skills, spontaneous language form, and responses to a standardized language test. The nonverbal measures assessed nonverbal intelligence, spatial-rotation, and motor-sequential skills. In addition, the children's parents participated in an interview to describe the child's current speech, language, and learning skills; and to report family history of speech, language, and academic problems. The results of a stepwise discriminant analysis indicated that four measures accounted for 79% of the variance of the model (p < .0001). This four-measure discriminant model had a sensitivity of 91.3% and a specificity of 100%. The measures that contributed to the discriminant model were: parental report of the child's current speech, language, and learning problems; number of errors per terminable unit; family history of speech, language, or academic problems; and mean length of terminable unit. An additional discriminant analysis indicated that the same level of discriminant accuracy could be maintained with the two measures that accounted for the most variance in the four-measure model: parental report of speech and language problems, and number of errors per terminable unit. Confirmatory discriminant analyses of the two-measure and four-measure models indicated that the results were stable across an independent sample. This study underscores the need for data-based approaches to the identification of Spanish-speaking children with language impairment, and the contributions of standard evaluations procedures to the identification of these children: parent interview and language-form analysis. In addition, the findings of this study indicate that a language-form deficit characterized by morphosyntactic difficulties and a high prevalence of family history of speech and language problems characterize children with language impairment regardless of the languages they speak.
|
|---|
