Discriminant accuracy of a semantics measure with Latino English-speaking, Spanish-speaking, and English-Spanish bilingual children

Abstract

We explored classification accuracy of English and Spanish versions of an experimental semantic language measure with functional monolingual-bilingual children with and without language impairment. A total of 441 children participated, including 78 balanced bilinguals (15 with language impairment, 63 with typical development); 179 monolingual Spanish (36 with language impairment, 143 with typical development); and 183 monolingual English (49 with language impairment, 134 with typical development) children between 4;0 and 6;11 years. Cut points derived for functionally monolingual children were applied to bilinguals to assess the predictive accuracy of English and Spanish semantics. Correct classification of English monolinguals and Spanish monolinguals was 81%. Discriminant analysis yielded 76% and 90% correct classification for balanced bilingual children in English and Spanish respectively. This semantics-based measure has fair to good classification accuracy for functional monolinguals and for Spanish-English bilingual children when one language is tested.

Learning outcomes: As a result of this study, the reader will describe advantages of lexical-semantic tasks for identification of language impairment. They will be able to describe procedures for conceptual scoring and identify its benefits. Readers will also gain an understanding of similarities and differences in bilingual and monolingual performance on a semantics task in Spanish and English.

Keywords: Bilingual; Diagnostic accuracy; English; Language impairment; Spanish.