Communication Dans Un Congrès Année : 2024

Wearable sensors in the DyLNet project: a methodology to study social interactions and language produced by children and adults at preschool

Résumé

The DyLNet project (Language Dynamics, Linguistic Learning, and Sociability at Preschool: Benefits of Wireless Proximity Sensors in Collecting Big Data, ANR-16-CE28-0013) is a multidisciplinary collaborative project bringing together linguists (sociolinguists and psycholinguists), ethologists and network science researchers, the main aim of which is to observe and characterize the relationships between child socialization and oral language acquisition during the preschool period. The project involved monitoring the children and teaching staff of a French socially mixed preschool (≈ 200 individuals) for two and a half years. Each month, for one week, participants wore small boxes containing two types of collection devices: (1) RFID (Radio Frequency Identification) sensors recording the proximity between individuals (via RSSI1 measurements taken every 5 seconds); (2) microphones recording their speech and surrounding sounds. In addition, at various points during the longitudinal follow-up, language tests (receptive lexicon and syntax, sociolinguistic judgements and repetitions) were administred to the children. We also collected sociodemographic and sociolinguistic information for each child by means of a questionnaire completed by their family. In the end, we recorded 1,705,126 social interactions and have ≈ 35,000 hours of audio recordings, of which just over 800 hours have been transcribed. During the seminar we will present the aims of the project and focus on the methods used to record and process both the social interactions and the speech collected. We will conclude with a discussion of future challenges and the interdisciplinary framework we have created.

Fichier non déposé

Dates et versions

hal-04815391 , version 1 (02-12-2024)

Identifiants

  • HAL Id : hal-04815391 , version 1

Citer

Aurélie Nardy, Jean-Pierre Chevrot, Màrton Karsai. Wearable sensors in the DyLNet project: a methodology to study social interactions and language produced by children and adults at preschool. International group DARCLE (Daylong Audio Recordings of Children's Linguistic Environments), Dr Rana Abu Zhaya, UCL, Oct 2024, London, United Kingdom. ⟨hal-04815391⟩
73 Consultations
0 Téléchargements

Partager

  • More