Thresholds and psychometric functions of the individual spondaic words

Two experiments were conducted concerning the relative intelligibility of spondaic words. In the first experiment, the 36 spondaic words typically used in clinical settings to determine the Speech-Recognition-Threshold (SRT) were presented to 20 subjects at levels ranging from 10 to 28 dB SPL. The percentage of persons able to repeat each word was plotted as a function of decibel sound-pressure level to determine (a) the sound-pressure level necessary to yield a 50% response point for each word, and (b) the rate at which each word became intelligible, for example, the slope of the psychometric function. By these two criteria 15 words were found to be homogeneous. In the second experiment, SRTs were obtained on 24 subjects with the full list of 36 words and with the 15 spondaic words that were homogeneous with respect to the sound-pressure level necessary for a 50% correct response level and the rate of intelligibility growth. The results demonstrate that the two groups of spondaic words yield equivalent SRTs. The data suggest that an SRT may be determined with 15 spondaic words, which is as valid and accurate as the SRT determined with the entire list of 36 words.