Effects of a response format modification on the diagnostic rhyme test

Abstract

Detection theory measures of listener performance, as well as distinctive feature scoring, have been shown to provide useful additional information to percent correct scores when testing word recognition. The diagnostic rhyme test (DRT) uses a two‐item multiple choice (MC) response format to provide percent correct and distinctive feature information but does not measure d. This study examined the effects of modifications to the DRT response format that allow a measure of d. The format was modified to a single‐option in which the listener designates yes or no (Y‐N) that the single‐option matches the stimulus. DRT words were mixed in white noise and presented in sound field at 3 and 6 dB signal‐to‐noise ratios at 50 dBHL to eight normal‐hearing subjects. Results showed a significant difference between response formats for percent correct and a significant interaction between signal‐to‐noise ratios and feature errors across formats. At 3 dB, there were more temporal‐based feature errors (sustension and compactness) for the Y‐N response format, but no significant differences between formats for voicing, nasality, sibilation, and graveness. At 6 dB, there were significant differences only for sustension and graveness. The ds for signal‐to‐noise ratios were significantly different for the Y‐N format.