The role of mixed dominance in bilingual assessment

The Bilingual English Spanish Assessment (BESA) is a new standardized test normed on young bilingual children who speak Spanish and English. This bilingual standardized measure is unique because it considers the possibility that children might have differential semantic and grammatical abilities across their two languages. The BESA's scoring scheme accounts for mixed dominance by allowing children to demonstrate best performance. Information on how this new measure aligns with other indices of underlying language ability, such as parent questionnaires and nonword repetition (NWR) is needed. Information of this type could lead to the development of new screening procedures. Twenty-six Hispanic Spanish/English-speaking children were recruited from Salt Lake City, Utah. Parents reported on children's input/output in both languages and rated their Spanish and English abilities using the Speech Language Assessment Scale (SLAS). Participants were administered three nonword repetition tasks and the BESA. An English NWR task and a Spanish NWR task were administered to each child. The highest achieved score from the Spanish and English NWR tasks was used to allow for mixed dominance across languages (NWR-best). A recently developed framework was used to create a quasi-universal NWR task with quasineutral prosody with syllable patterns and phonemes in English and Spanish (NWR-U). Results of correlational and regression analyses indicated that performances across the NWR tasks were highly correlated with participants' performance on the BESA. The parent ratings from the SLAS were moderately correlated with performance on the BESA. A linear regression analysis including the SLAS, NWR-best, and NWR-U accounted for 82% of the variability in children's BESA scores. A second linear regression analysis including NWR-best and NWR-U indicated that on its own, NWR-U accounted for 85% of the variability in children's BESA scores. These findings suggest that NWR measures in general and the created quasiuniversal NWR measure in particular, show promise as potential screeners for young Spanish/English bilinguals. Parent questionnaires continue to be useful in collecting current information regarding bilingual children's language abilities. Using NWR-U with bilingual children might represent an appealing alternative because it can account for both languages and potentially maintain high levels of diagnostic accuracy.