The Case for Measuring and Reporting Bilingualism in Developmental Research

Abstract

Many children around the world grow up bilingual, learning and using two or more languages in everyday life. Currently, however, children's language backgrounds are not always reported in developmental studies. There is mounting evidence that bilingualism interacts with a wide array of processes including language, cognitive, perceptual, brain, and social development, as well as educational outcomes. As such, bilingualism may be a hidden moderator that obscures developmental patterns, and limits the replicability of developmental research and the efficacy of psychological and educational interventions. Here, we argue that bilingualism and language experience in general should be routinely documented in all studies of infant and child development regardless of the research questions pursued, and provide suggestions for measuring and reporting children's language exposure, proficiency, and use.

Automatized analysis of children’s exposure to child-directed speech in reschool settings: Validation and application

“…Decoding these classroom tasks is relevant to explain the higher heterogeneity in interactions and CDS during child-selected activities, compared to largegroup activities [21]. While studies analyzing talk in home environments employ audio recordings with longer duration for ecological validity and reliability of speech processing tools [16,39], our recording times are consistent with observational studies analyzing interactions and CDS in preschool classrooms [40][41][42].…”

“…Both services are offered in a pay-per-use format, which, at the time of this publication, is $0.10 per minute of video files, and $0.00004 per second of audio. While there exist open-source alternatives for audio and video featurization that could replace AWS (see references in [16,25]), the rationale behind the AWS choice included ease of use, low cost, satisfactory performance results on preliminary analyses with our dataset, and data security measures provided via institutional access to the cloud computing resources. AWS deep neural network models are trained with Amazon's proprietary data, and their online tutorials enable their deployment by non-experts.…”

“…Second, the system does not provide nuanced measures of linguistic complexity, such as number of different words and mean length of utterance, which represent important aspects of CDS that correlate with children's language development [6]. Third, with the rapid pace of developments and deep-learning applications in the automatic speech recognition (ASR) community, it is unclear whether LENA has adopted these methods, and being a closed-source software hinders the possibility of adapting the system to custom acoustic environments and new developments in automatic speech recognition [16]. Fourth, recent reviews raise concerns about the validity of the tool [17].…”

Documenting bilingual experiences in the early years: Using the CECER-DLL Child and Family and Teacher Questionnaires

“…Generally, studies do not provide information about the age at which children were exposed to each of their languages or the amount of their exposure to and use of these languages. This information is important because language exposure and use may explain the variability in children's abilities in each language (Byers-Heinlein, Esposito, Winsler, Marian, Castro & Luk, 2019). The notion that bilingual children come from monolingual non-English speaking homes is a common misconception, which may lead to misleading interpretations of study findings.…”

Bilingualism in Döver Mahallesi in Defne District of Hatay

“…Bilingualism is common and perhaps it is increasing in many parts of the world as one in three people is either bilingual or multilingual (Wei, 2000). Many children in the world grow up with two languages, and use two or more languages in daily life (Grosjean, 1982;Alladina, 1995;Byers-Heinlein et al, 2019). Researchers trying to define bilingualism act from different points.…”