A Similarity Scale to Evaluate the “Sub-Intelligible” Speech Patterns of Young Deaf Children

Abstract

The speech of young deaf children is usually unintelligible and difficult to evaluate with standardized speech tests. The observer will usually have difficulty assigning conventional phonetic symbols to the unintelligible response of the child. However, with the appropriate speech training, some deaf children improve and develop speech that is intelligible. Thus, it is desirable to have one procedure that can evaluate the entire continuum from subintelligible to intelligible levels. To accomplish this, a similarity scale was used to evaluate improvement as a function of therapy time. Deaf children were tested with 27 words in four stimulus—response conditions. The teacher provided the oral stimulus (word) and the child responded by attempting to immitate the stimulus. Normal-hearing listeners used a nine-point similarity scale to judge the response as it related to the stimulus. A score of 1 indicated that the response was similar to the stimulus and a score of 9 indicated that it was dissimilar. A group of deaf children demonstrated a significant improvement over testing time. Then, the responses that represented different points on the similarity scale were judged for intelligibility. It appears that changes in similarity values between 9 and 5 represent the subintelligible range and can be partially attributed to changes in segmental features. From 2.5 to 1 there is a rapid improvement in intelligibility. Perceptual changes on the similarity scale will be related to corresponding changes in spectrograms. [Work supplemented by U. S. Office of Education.]