Bilingual Children’s Performance on Three Nonword Repetition Tasks: The Role of Language Experience and Ability

Abstract

Nonword repetition (NWR) tasks represent one assessment tool for Specific Language Impairment (SLI). The use of such tasks has been established and verified for monolingual children. However, the diagnostic accuracy of NWR tasks for bilingual children has had variable results and must address several unique characteristics of this population. Gaps in research relate to which task characteristics influence task performance. The purpose of the current study was to explore children’s performance on three NWR tasks that differ in design: a norm-referenced task based in English (Comprehensive Test of Phonological Processing, Second Edition; CTOPP-2), a “quasi-universal” task based in the sounds most commonly found in the world’s languages (QU), and a “language neutral” (LN) task tailored to Spanish and English speech sounds. The children’s performance on the three tasks was examined in light of their time spent listening to and speaking English and language ability in Spanish and English, as measured by the Bilingual English Spanish Assessment (BESA). Finally, the different versions of the NWR task were compared in terms of which task is most predictive of a performance on the BESA that is indicative of language impairment. Results indicated that the LN task and the CTOPP-2 task performance may have been influenced by a child’s language experience in English. All three tasks were significantly related to the participants’ language ability, as measured on the BESA. Finally, analyzing the data using quantile regression indicated that for all three tasks, the lower quantiles of the BESA Index score (i.e., participants’ language ability) was more strongly related to the performance on the NWR tasks than the BESA performance at higher quantiles. All three tasks hold promise for diagnosing Spanish-English bilingual children with SLI, with the QU task performance showing more independence in terms of language experience. In conclusion, while LN and the CTOPP tasks revealed participants’ performance based on language experience, the QU task is the most promising task for an assessment tool for bilingual children.