Discourse and hearing loss: The effects of a text supplement.
Item
-
Title
-
Discourse and hearing loss: The effects of a text supplement.
-
Identifier
-
AAI3047225
-
identifier
-
3047225
-
Creator
-
Haravon, Anita Beatrice.
-
Contributor
-
Adviser: Loraine K. Obler
-
Date
-
2002
-
Language
-
English
-
Publisher
-
City University of New York.
-
Subject
-
Health Sciences, Audiology | Health Sciences, Rehabilitation and Therapy
-
Abstract
-
Automatic speech recognition (ASR), a technology that converts speech into text in real-time, may prove effective as a method of real-time transcription for people with hearing loss. Although other methods of real-time captioning, such as CART (computer-aided real-time transcription), do exist and are widely used, they require the skills of an additional individual, a highly-trained stenographer or typist. If effective, ASR would free a person with hearing loss from reliance on another individual for communication assistance.;The effectiveness of ASR (Dragon Naturally Speaking, version 4.0) for communication enhancement was evaluated. The Map Task, a collaborative problem-solving task, was used to elicit conversation between six pairs of deaf/hearing communication partners. Deaf participants were all accustomed to producing speech for everyday communication. Participants had conversations under the following conditions: face-to-face, CART-face-to-face, ASR-face-to-face, CART-only and ASR-only. Methods for measuring Communicative Success were drawn from the disciplines of aural rehabilitation and discourse analysis.;Communicative Success was assessed in four ways: Task Outcome, Completion Time, Check Moves and Participant Evaluation. Task Outcome was measured using the Map Task, which was scored for route-reproduction accuracy. Mean map scores across communication conditions were not significantly different. Completion Time was significantly longer in ASR conditions than in the face-to-face-only and CART conditions. Check Moves, which are similar to requests for clarification, were tallied. Significantly more Check Moves were found in the face-to-face-only condition than the CART-face-to-face condition. Participant Evaluation consisted of hearing and deaf participants' ranking the five communication conditions. Rankings were significantly different, with CART-face-to-face ranking first, face-to-face-only, CART-only and ASR-face-to-face tying for second, and ASR-only ranking third. In terms of Transcription Accuracy, ASR was found to be significantly less accurate than CART.;In sum, this study supported the notion that for people with substantial hearing loss, the accurate real-time transcription of conversational speech improved Communicative Success. CART transcription was accurate and was ranked favorably by participants. Accuracy of ASR transcription will need to be improved before it is comparable to CART as a viable as a method of real-time speech-to-text transcription.
-
Type
-
dissertation
-
Source
-
PQT Legacy CUNY.xlsx
-
degree
-
Ph.D.
