Assessing Children’s Home Language Environments Using Automatic Speech Recognition Technology

Abstract

The purpose of this research was to replicate and extend some of the findings of Hart and Risley using automatic speech processing instead of human transcription of language samples. The long-term goal of this work is to make the current approach to speech processing possible by researchers and clinicians working on a daily basis with families and young children. Twelve hour-long, digital audio recordings were obtained repeatedly in the homes of middle to upper SES families for a sample of typically developing infants and toddlers (N = 30). These recordings were processed automatically using a measurement framework based on the work of Hart and Risley. Like Hart and Risley, the current findings indicated vast differences in individual children’s home language environments (i.e., adult word count), children’s vocalizations, and conversational turns. Automated processing compared favorably to the original Hart and Risley estimates that were based on transcription. Adding to Hart and Risley’s findings were new descriptions of patterns of daily talk and relationships to widely used outcome measures, among others. Implications for research and practice are discussed.