The Benefits of Phoning it in: Phonemic Diversity in Teacher-Child Conversations and Language Development

• In Event: Poster Session 03
  In Poster Session: PS 03 Topic - Language, Communication

Abstract

Individual variation in the quantity and quality of children’s early language input and production, during play with caregivers at home and during laboratory tasks, predicts future language development and academic performance (Hoff, 2013). Much less is known about children’s language experiences in early education programs. Participation in high-quality inclusive education programs typically improves outcomes for children with hearing loss who often experience inconsistent or distorted access to acoustical information. However, no study to date has compared indices of the quantity and quality of language experiences in classrooms including children with hearing loss and their typically hearing peers. Understanding which features of classroom language input and production are most strongly associated with developing language abilities will help to shed light on the mechanisms that support language learning for both sets of children. Here, we employed automated measurement in inclusion classrooms to examine the relative contributions of the quantity (rate) and quality (phonemic diversity) of children’s language experiences to their language abilities.

Participants included 29 children enrolled in three English dominant oral language inclusion classrooms. Twenty-one children have hearing loss and wear cochlear implants or hearing aids and eight children have typical hearing. Vocalization data were collected once every two weeks using LENA audio recorders over 34 total observations. Using LENA pattern recognition software, audio recordings were segmented into language input from teachers and children’s own speech-related vocalizations. From the segmented speech stream, we calculated the rate per minute of teachers’ and children’s vocalizations. Sphinx speech recognition software was used to calculate phonemic diversity, the number of unique phonemes per vocalization. Language abilities were measured at the end of the school year using the PLS-5.

There was no effect of hearing status on the quantity or quality of language input that children received. However, children with hearing loss produced a lower rate of vocalizations (χ2=4.26, p<.01) and less phonemically diverse vocalizations (χ2=13.95, p<.01) than typically hearing children. The rate and phonemic diversity of teacher speech positively predicted the rate and phonemic diversity of child speech. The phonemic diversity of children’s vocalizations was a stronger unique predictor of children’s end-of-year language abilities than the rate of their vocalizations. Children who produced more phonemically diverse vocalizations exhibited higher end-of-year receptive (p=.015) and expressive language abilities (p=.026) even when controlling for differences in phonemic diversity related to children’s hearing status (Figure 1). Mediation analyses indicated that children who received more phonemically diverse language input in the first half of the school year produced more phonemically diverse vocalizations during the second half of the school year, which was in turn associated with higher end-of-year receptive and expressive language abilities (Table 1).

The current investigation utilized automated measurement to capture children’s moment-to-moment language experiences in their classroom and highlighted a specific feature of the classroom language environment, phonemic diversity, as a promising correlate of children’s developing language capacities. These findings highlight both the importance of phonemically diverse adult speech and the role of children’s own production of phonemically diverse speech in children’s language development.

Authors

• Samantha Mitsven, University of Miami
  Presenting Author

• Lynn K Perry, University of Miami
  Non-Presenting Author

• Yudong Tao, University of Miami
  Non-Presenting Author

• Neil Johnson, The George Washington University
  Non-Presenting Author

• Daniel Messinger, University of Miami
  Non-Presenting Author