Discrimination abilities of impaired listeners compared to the range of variation in performance of normal listeners

Abstract

Discrimination abilities of sensorineural hearing‐impaired listeners have been compared in most studies to average performance of small samples (n≃10) of selected, highly trained, normal listeners. Such comparisons may be biased because: (a) They do not take into account the broad distribution of the performance of normal listeners, that is, the large intersubject differences exhibited by normal listeners on many auditory tasks; (b) unusually poor performers have been exluded from samples of normal listeners; and (c) normal listeners are often college students, impaired listeners are typically more heterogeneous samples. In the present study discrimination performance of 23 impaired listeners was compared to the range of discrimination performance exhibited by large (n > 100), fairly heterogeneous samples of untrained normal listeners. The tasks included discrimination of frequency, duration, intensity, tonal patterns, and nonsense syllables, and were presented at sufficiently high sensation levels (⩾25–30 dB SL). The range of normal performance was defined as the interval between the average psychometric function of the listeners comprising the best 10% of the normal sample and the function for the worst 10%. Performance of normal listeners was distributed over broad ranges in all tasks. Performance of impaired listeners was clearly bracketed by the range of normal performance in all of the nonspeech discrimination tasks. Variability of impaired listeners' performance in the nonspeech tasks was comparable to that of normal listeners, and their psychometric functions were uniformly distributed over the range of those for normal performance.