Language proficiency across tasks in sequential bilingual and monolingual children

A number of different language measures are used in child language acquisition studies. This raises the issue of comparability across tasks, and whether this comparability diverges depending on the specific language domain or the language population (e.g., monolinguals versus bilinguals). The current study investigates the comparability across tasks in the domains of vocabulary, morphology, and syntax in primary-school-aged sequential bilingual children with L1 Arabic/L2 English (N = 40, 5;7–12;2) and age-matched monolinguals (N = 40). We collected narrated speech samples to produce measures across language domains, and additional measures from separate vocabulary, morphology, and syntax assessments. Using a logistic regression analysis, we find a correspondence between syntax measures in monolinguals; however, we find no further correspondences in the other domains for monolinguals, and no correspondences at all in bilinguals. This suggests that assessment measures are highly task-dependent, and that a given assessment measure is not necessarily indicative of language as a whole, or even of language within a domain. We also find selective effects of age for monolinguals and both age and length of exposure (LOE) for bilinguals; in particular, while LOE predicts variation between the first and second language, age effects reflect underlying similarity across languages. We consider the implications of these effects for language assessments across populations.