A multivariate strategy for prediction of psychoacoustic performance from electroacoustic characteristics of hearing aids.

Sentence length psychoacoustic material was processed through 16 hearing aids and presented to 220 listeners. Twenty-nine indices of electroacoustic performance were abstracted from a testing program under laboratory conditions. Through a multivariate statistical strategy, psychoacoustic achievement was shown to be positively correlated to and could be predicted by a knowledge of white noise gain, bandwidth below and above 1 kHz, frequency response regularity, and transiet distortion (multiple R = 0.77; F = 2.04; df = 5,10). It was concluded that a multivariate statistical tactic, when used with discretion, can be a useful and powerful tool to predict psychoacoustic performance from a knowledge of hearing aid electroacoustic characteristics.