Is Speech Sound Selection Important for Diagnostic Word-Recognition Tasks?

Abstract

Phonemic balance has been an important consideration in the development of monosyllabic word recognition tasks (e.g., Lehiste & Peterson, 1959). Therefore, efforts have been made in the careful selection of words to ensure the phonemic composition of a word list is representative of English. Some studies show lack of difference in speech recognition performance between tasks using phonemically balanced words and randomly selected words. The purpose of this study was to evaluate the degree that randomly selected monosyllabic words deviate from phonemic balance and whether these deviations have potential consequences for hearing assessment. To assess the clinical consequences, behavioral performance was simulated using percentage of correct responses in listeners with steeply sloping sensorineural hearing loss based on performance described in Dubno, Dirks, and Langhofer (1982). Numerous deviations from phonemic balance were noted in the randomly drawn lists, but, to our surprise, these deviations had no meaningful effect on simulated scores.