Speech perception in a communicative context: an investigation using question/answer pairs

By diminishing the role of communicative context, traditional tests of speech perception may underestimate or misrepresent the actual speech perception abilities of adults with a hearing impairment. This study investigates this contention by devising an assessment that may better simulate some aspects of "real-life" speech perception. A group of 31 participants with a severe-to-profound hearing impairment took part in a series of speech perception tests while wearing their hearing aids. The tests used question/answer or adjacency pairs, where the stimulus sentence was preceded by a question spoken by the participant. Four conditions were included: (a) where there was no initiating sentence, as in a traditional open-set speech perception test; (b) where the initiating question was neutral (e.g. "Why?"); (c) where there was a disruptive semantic relationship between the question and answer; and (d) where there was a strong contextual relationship between the question and answer. The time delay between the question and answer was also varied. Results demonstrated that in all conditions where there was a preceding question speech perception improved, and increasing the cohesion between the question and the reply improved speech perception scores. Additionally, time delay and the relatedness of the reply interacted. The effects of semantic context appeared to diminish over a 10-s period while other linguistic effects remained more constant. These results indicate the utility of simulating communicative environments within speech perception tests.