Language Use During Naturalistic Caregiver-child Interactions in 6-year-old Spanish-English Sequential Bilinguals

Publication Date


Document Type



Child and Adolescent Development

Publication Title

Society for Research in Child Development 2021 Virtual Biennial Meeting

Conference Location



Many children in the US live in primarily Spanish-speaking households with few opportunities to learn English until they enter school. While increased exposure to English supports English-language development in these sequential bilingual children, this is often at the cost of proficiency in their first language, Spanish. Achieving high levels of proficiency in both languages has significant socio-cultural as well as economic benefits. In this ongoing longitudinal study, we track exposure to Spanish and English from toddlerhood to the early elementary years in sequential bilingual children. We ask: Does exposure shape the choice of language use (i.e., English, Spanish or mixed) by children and caregivers during naturalistic interactions? Is language choice associated with patterns of children’s English- and Spanish-language proficiency at 6 years?
Participants (n=65, 40F, 25M) were Spanish-speaking families primarily of Mexican descent (~88%). Relative exposure to English and Spanish from all sources (e.g., caregivers, siblings, school) was documented via caregiver interview when the children were 2 years and 6 years of age. At 2 years, caregivers completed the Spanish-language MacArthur-Bates Communicative Development Inventory to estimate vocabulary size. At 6 years, each caregiver and child engaged in a Structured Observational Play Assessment (SOPA), which was designed to elicit natural conversation in a semi-structured, developmentally- and culturally-appropriate, pretend play activity (i.e., making pretend soup). We transcribed sessions (~10 mins) following standard CLAN protocols. For each caregiver and child, we derived numbers of utterances, word tokens, types, and mean length of utterance (MLU-words) in English, Spanish and mixed-language utterances. At 6 years, we assessed child proficiencies in receptive vocabulary and expressive language in both English and Spanish using standardized tests (PPVT/TVIP; CELF-4/CELF-4-Spanish).
We found that, at 2 years, children were exposed primarily to Spanish overall; exposure proportions from caregivers were >95% Spanish (Table 1). By 6 years, exposure to English increased overall, however, exposure from caregivers remained primarily Spanish (>85%). During SOPA, interactions occurred primarily in Spanish, but children produced significantly more English utterances, tokens, and types, than caregivers, all ps<.01. Children who were exposed to more Spanish produced a greater proportion of Spanish vs. English utterances, tokens, and types (rs =.32-.42) and produced longer Spanish utterances (r=.42), compared to children who were exposed to less Spanish. Mixed utterances were rare. A k-means clustering analysis revealed that children fell into three outcome groups based on standardized test performance: Hi Spanish/Hi English (HI/HI), Hi Spanish/Low English (HI/LOW), and Low Spanish/Low English (LOW/LOW). Children in the HI/HI and HI/LOW groups had caregivers with significantly longer Spanish MLU-words compared to children in the LOW/LOW group, F(2,47)=4.5, p<.02 (Figure 1), controlling for SES, Spanish exposure, and child vocabulary size. No relations were observed between caregivers’ use of English and child English-language scores, suggesting that other sources (e.g., siblings) support children’s English outcomes in this population.
This ongoing project provides new insights into the language(s) in which primarily Spanish-speaking caregivers and their emerging bilingual children engage during naturalistic interactions. Further analyses will continue to explore factors that support successful bilingual language development in this growing population.