We employed a nonparametric measure, the randomization test, to evaluate, statistically, a variety of speech understanding measures used to quantify the efficacy of different intervention strategies in five clinical patients. The flexibility of the approach was demonstrated by comparing monaural with binaural amplification, a conventional aid with an assistive listening device, and a cochlear implant alone versus a cochlear implant supplemented by visual cues. The randomization test allows statistical analysis of differences in performance at the level of the individual subject. This approach may be useful in addressing accountability when recommending a specific intervention strategy for a particular hearing impaired individual.
Address for correspondence: Rose Chmiel, 6565 Fannin, NA 200, Houston, TX 77030.
Received March 29, 1994; accepted September 12, 1994.
© Williams & Wilkins 1995. All Rights Reserved.