Effects of response format on speech intelligibility in noise: Results obtained from open-set, closed-set, and delayed response tasks

Abstract

Many word recognition studies over the last 40 years have used forced-choice closed-set tasks, based on the assumption that closed-set and open-set tasks differ only in the level of chance performance. However, Sommers, Kirk, and Pisoni (1997) found that lexical competition and talker variability produce robust effects on performance only in open-set tasks, suggesting fundamental differences in the task demands and processing strategies between open- and closed-set tasks. In the present study, listeners were asked to recognize spoken words degraded by a bit-flipping algorithm in three response formats: open-set, closed-set ‘‘before,’’ and closed-set ‘‘after.’’ In the closed-set ‘‘before’’ condition, the six response alternatives were presented 1 s prior to the onset of the auditory signal. In the closed-set ‘‘after’’ condition, the response alternatives were presented 1 s after the auditory signal. Results revealed significant effects of lexical competition and talker variability only in the open-set task. These findings suggest that even a delay of 1 s is not adequate to induce the task demands observed in open-set word recognition tests. [Work supported by NIH.].