Quantifying multilingual children’s language exposure through parental report: More information is not always better

Abstract How much information do we need when estimating multilingual children’s relative language exposure? In the current study, we compared three different estimates at varying levels of detail: (i) global estimates per language, (ii) the average of per-speaker estimates, weighted for the time the child spends with each speaker, and (iii) estimates obtained using the Experience Sampling Method, which consisted of five surveys a day across 7 days. Data were collected from 102 multilingual children (ages 3–9 years) in the Netherlands. We found that the three exposure estimates were highly correlated and that there were only small differences in how well they correlated with children’s vocabulary knowledge. Discrepancies between estimates were largely unrelated to participant characteristics such as children’s age or the number of languages spoken at home. We conclude that the simplest estimates (i.e. global estimates) may be sufficiently reliable as a measure of multilingual children’s language exposure at home.