Response Errors in Females’ and Males’ Sentence Lipreading Necessitate Structurally Different Models for Predicting Lipreading Accuracy
The research was supported by the U.S. National Institutes of Health (DC014523 and DC012634).
Abstract
Lipreaders recognize words with phonetically impoverished stimuli, an ability that varies widely in normal‐hearing adults. Lipreading accuracy for sentence stimuli was modeled with data from 339 normal‐hearing adults. Models used measures of phonemic perceptual errors, insertion of text that was not in the stimulus, gender, and auditory speech perception in noise thresholds to predict lipreading accuracy of 10,170 responses. Interactions of the lipreading predictors with gender necessitated different models for males’ versus females’ lipreading. Females’ lipreading accuracy was significantly predicted by their auditory speech in noise thresholds and an interaction between the magnitude of their perceptual errors and the number of nonstimulus phonemes in their responses. Males’ lipreading accuracy was a function of their auditory speech in noise thresholds in interaction with the magnitude of their perceptual errors. The predictor coefficients of the two models suggest the possibility of different mechanisms influencing lipreading accuracy in males versus females.
