Over the last few days I’ve gotten a number of questions about how to distinguish between language impairment and normal development in bilingual children probably due to the feature story on the UT Home page last week. It’s been great to hear from bilingual speech pathologists from around Texas and other areas of the country. I think that many are struggling to deal with the same questions that I’ve been pursuing with my colleagues. That is, how do we know what disorder looks like in bilinguals; and what can we do to document these distinctions?
In the absence of standardized tests (so far anyway) language sampling is often recommended. For example, an article by Gutierrez-Clellen and colleagues reviews some of the issues involved in using language samples with children. Rojas and Iglesias give guidelines for using language samples over time to distinguish between difference and disorder. We have a new article that additionally addresses this question. We focus on samples in BOTH languages from Spanish-English speaking kindergarten children.
One of the innovative things we did in this study is to test children in both of their languages. We also conducted language sampling (using narrative samples) in both Spanish and English. Another innovative thing we did was with the test data. We used our Bilingual English Spanish Assessment (Peña, Gutierrez-Clellen, Iglesias, Goldstein, & Bedore, in development). For the morphosyntax and semantics subtests, we compared children’s scores in Spanish and English and took the higher score for each. We then averaged this score to obtain a composite. The composite score then represents the best semantics score, the best cloze (morphosyntax) score, and the best sentence repetition (morphosyntax) score. An interesting observation we made here was that every possible combination of “best score composite” that could occur did. We we had kids who were better in English on semantics, but better in Spanish in morphosyntax; better in Spanish on semantics, but better in English morphosyntax. Within morphosyntax, sometimes kids did better on the cloze in Spanish but better on the sentence repetition in English. Other kids did best in Spanish across the board; and others who did best in English for all sections. The advantage of taking the best score for each is that we give kids maximum credit for any configuration of best performance. These score composites correlated with a combination of grammaticality in BOTH languages plus MLU in English.
I wonder if these same composite approaches could be used for assessment. Say, compare the EOWPVT-2000 in each language and take the best score; and then compare the PLS in Spanish and English and take the best score. It’s similar to conceptual scoring but at the test (or subtest) level. I’m not sure exactly how this would work out, but this idea can provide speech-language pathologists who work with bilingual children another tool to make clinical decisions.