Transcribing toddler vocalizations: comparison of a modified naturalistic listening approach with phonetic transcription

Abstract

ABSTRACT Previous investigations have demonstrated that naturalistic listening in real time (NLRT) can be used reliability to assess prelinguistic vocalizations and is less time-consuming than phonetic transcription. The current investigation was performed to examine the correspondence of syllable and consonant information obtained using a modified naturalistic listening in real time (NLRT) methodology compared to broad phonetic transcription for 20 toddlers with repaired cleft palate. A forty-minute vocalization sample was obtained for each toddler. Two listeners listened to each recording independently and coded the number of syllables and consonant types produced by each toddler using NLRT. Each recording was phonetically transcribed by each listener 2 to 16 months following the NLRT analysis. High reliability was evident between the modified NSRT method and phonetic transcription for identification of both syllables and consonants. Differences were evident in the actual number of syllables and consonants identified between the two types of assessment across the 20 participants. Possible explanations for those differences are addressed. The results of this investigation indicate that the modified NLRT procedure is a reliable method for determining consonant/inventories of young children. It is less labor intensive than traditional phonetic transcription and may be useful clinically when documentation of early sound development is needed to support early intervention decisions.