Multiparty speech predominates in children's language environments across diverse cultural settings

Résumé fourni par la source

This dataset reports on 844 children's participation in conversation across 10 countries, based on automated speech metrics derived from 23,000 hours of daylong naturalistic audio recordings. The main analysis focused on quantifying children exposure to different types of conversation, namely: multiparty conversation (2+ interlocutors), adult-child conversation (1 adult interlocutor), child-child conversation (1 other child interlocutor), and overheard conversation (no target child participation). Potential sources of variance in exposure to conversation, including child age, gender, number of siblings, and maternal education, are also examined across countries and individual homes. This resource exists to archive supplementary materials and facilitate the reproduction of the full manuscript, which is currently under review. Submodules in the datasets and input directories exist to preserve dataset structure; additional permission must be sought from the authors for access, detailed in the README.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.