AN INVESTIGATION OF CONTRASTIVE STRESS PRODUCTION IN HEARING-IMPAIRED AND NORMAL-HEARING CHILDREN
Abstract
Results of studies employing oral reading and other formal tasks have reported that the prosodic features of hearing-impaired children's speech are often aberrant and contribute to decreased intelligibility. This study attempted to elicit hearing-impaired children's use of one prosodic feature, stress, as it occurred meaningfully in discourse to denote differences in given and new information. Another purpose of the study was to compare the patterns of contrastive stress production for the hearing-impaired children with those of normal-hearing children. Twenty hearing-impaired children were selected who demonstrated pure-tone averages of 55-90 dB in the better ear, and were matched with 20 normal-hearing children on the basis of Mean Length of Utterance. The 40 children described pairs of pictures to the examiner where one semantic component of the second picture, the agent, action or complement, changed from the first to second picture. A stressed production of the changed element contrasted new from given information. All of the recorded responses were played to a group of eight listeners who judged whether contrastive stress had been produced. Analyses of the listeners' responses yielded no significant differences for hearing ability. That is, the hearing-impaired children were capable of utilizing contrastive stress, where appropriate. This finding was probably a function of the positive correlation between stress production of MLU, a subject selection criterion, demonstrated by both of these groups. Similar patterns of stress production were observed. Stress was produced significantly more often for second utterances than for first utterances, while stress production for the changed element occurred significantly more often than would be predicted by chance. Complements were stressed half as frequently as agent or action components, a pattern already reported for normal children. This latter finding may suggest differences in acoustic features for these components. Future research should include analyses of spontaneous conversational data produced by hearing-impaired speakers. In addition, data collected from more severely hearing-impaired and less intelligible speakers should be compared with that reported in this study.
Degree
Ph.D.
Subject Area
Speech therapy
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.