The Speech Efficiency Score (SES): A time-domain measure of speech fluency

This study is a preliminary attempt to evaluate a new speech fluency measure, the Speech Efficiency Score (SES), in comparison with subjective stuttering severity rating scales and stuttered syllable counts (%SS).

277 listeners (92 naïve, 39 speech-language pathology (SLP) students, 124 practicing SLPs, and 22 SLPs who specialize in stuttering) evaluated short recordings of speech on an 11-point scale. Recordings were obtained from 56 adults, of whom 20 were people who stutter, 16 were people who stutter who were using fluency-shaping techniques, and 20 were speakers who do not stutter. In addition, %SS and the SES measure were obtained for each recording.

The four listener groups rated stuttering severity similarly, with no statistically significant between-group differences. Listeners' responses on the stuttering severity rating scales and the SES yielded significant differences between all three speaker groups. The %SS measure yielded a significant difference only between the stuttering group and the other two groups but not between the fluency-shaping and the control groups. A very strong positive correlation was found between the SES and the subjective stuttering-severity rating scales (r = 0.92). The correlation between %SS and the perceptual evaluation, as well as the correlation between %SS and the SES, were lower, though they still reached significance.

Results suggest that speech efficiency scores, which are based on a time-domain analysis, closely match subjective stuttering severity ratings and could ultimately provide a more objective way to measure speech fluency.