Comparative Effectiveness of Correction Strategies in Connected Discourse Tracking

Abstract

The effectiveness of four correction strategies commonly used in connected discourse tracking was investigated in the present study. The strategies were 1) verbatim repetition of a word or phrase; 2) use of antonyms or synonyms as cues; 3) use of phonemic cues, with no whole word repetition; and 4) going back or ahead in the text, with no repetition of the missed segment. Four normal-hearing adults served as listeners. Live-voice presentation of text by two female talkers was employed for all conditions. Listeners were tested in two stimulus presentation modes, speechreading alone and speechreading plus a multichannel tactile aid. Results indicated that strategy 1, repetition of the missed segment, produced the highest tracking rates, significantly higher than any of the other strategies. Strategy 2 produced the lowest tracking rates. Strategies 1 and 3 yielded the lowest percentage of initially missed words, or blockages, and strategy 4 the highest percentage. Significantly higher tracking rates were found under the speechreading plus tactile aid presentation mode, compared with speechreading alone. Further, tracking rates increased significantly from the beginning to the end of training. Data were compared with a more typical CDT task, in which all correction strategies were operative, and results showed little difference in tracking rates between this task and the constrained CDT employing only strategy 1. Overall, results suggest that simple repetition of missed segments is an effective correction strategy for CDT and argue for its inclusion in computer-assisted tracking implementations.