Sibylla Leon Guerrero, Laura Mesite, Sarah Surrain & Gigi Luk (Harvard Graduate School of Education)
Cognitive flexibility; Executive function; Spanish-English bilinguals; Home language use; Trails task
In recent decades, US public school classrooms have become increasingly diverse, both linguistically and culturally . While English is the primary medium of instruction in U.S. schools, many children have diverse language experience outside of school. Traditional group comparison may not capture the heterogeneity in bilingual groups . We examined variability in home language usage and executive functions (EF) in children. We expected (1) no group difference in EF between monolinguals and broadly defined bilinguals; (2) differential outcomes in EF between bilingual subgroups with different proportions of bilingual home language use.
Eighty-five (85) monolingual and bilingual fourth-grade children (female=45) between the ages of 10-13 yrs. (m=11.01) attending public elementary schools participated in both a home language survey and an individual task battery as part of a larger study of reading and EF. Parents completed a detailed questionnaire about children’s daily language use and home language environment. Non-English language use at home was coded as a polynomial treatment factor using quartiles of non-English language use at home (i.e. 0%, 25%, 50%, 75%, and 100% of other language use). Children completed an experimental EF task, Trail-making. Cost of switching, a measure of cognitive flexibility, was computed by subtracting accuracy on the non-switching (number only) trials from switching (letter-number) trials of this task.
Our sample, particularly the bilinguals, had heterogeneous demographic backgrounds: groups with differences in non-English language use also displayed significant differences in maternal education (F(9,75)=2.62, p= 0.011). We therefore conducted analyses with both non-matching and matching multinomial propensity scores to reduce covariate effects. First, when monolinguals and bilinguals were compared as dichotomous groups without distinguishing proportions of non-English language usage at home, no significant differences between groups emerged, both before (b=-0.10, s.e.=0.07, t=-1.41, p=0.164) and after binomial covariate matching (b=-0.05, s.e.=0.07, 95%C.I.:[ -0.1990, 0.0984]). Then, we considered the bilinguals divided into five subgroups based on proportion of non-English language use at home. A polynomial regression model revealed that individuals who spoke a more balanced mixture of languages at home (e.g. 50% English and 50% another language) incurred lower switching costs in accuracy, while children who spoke more of only one language at home (e.g. 100% English or 100% another language) displayed greater switching costs (R2= 0.13, F=2.91, p=.02). After covariate rebalancing, a propensity score-weighted model continued to find a significant effect of home language group with the balanced home language group (50% English) showing the lowest cost of switching (b=-0.21, s.e.=0.06, t =-3.52, p<0.001).
Results suggest a multidimensional, rather than categorical, characterization of bilingual language use can aid in understanding the allocation of cognitive resources during childhood. Proportion of language use, a dimension of bilingual experience, may interact with EF development, creating distinct cognitive profiles that affect classroom learning.
 United States Census Bureau / American FactFinder. America’s children: key national indicators of well-being, 2013. U.S. Census Bureau, 2013. http://www.childstats.gov/americaschildren13/famsoc5.asp
 Luk, G. & Bialystok, E. Bilingualism is not a categorical variable: Interaction between language proficiency and usage. Journal of Cognitive Psychology, 25(5), 605–621, 2013.