![Sample our Health and Social Care journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5938/TOC-Subject-Sample-2021-Health-Social-Care.jpg"})
Abstract
Objective: In the clinical practice of speech disorders, trained listeners are often used as judges for scaling procedures in the perceptual assessment of a speech signal. The aim of this paper is to study how a group of experts performs a task of severity judgment compared with a task of intelligibility judgment. Study design: During an off-line experiment, 33 stimuli were presented as icons on a Powerpoint slide. Each icon was associated with a line allowing scoring of each stimulus heard by a simple click on it; each stimulus was judged in terms of voice, resonance, prosody and phonemic qualities. With an analogical scale placed at the bottom of the screen, the judges were asked to put a copy of the icon along the line; the space above the line was used for severity and the one under the line for the deterioration of intelligibility, the arrow being oriented on the right side for the ‘most severe’ or ‘the least intelligible’ and representing a range from 1 to 30. Results: The mean scores of severity and deterioration of intelligibility were 9.8 (+/− 1.72) and 8.2 (+/− 1.59), respectively. Three judges differed in the distribution of the scores, whereas two others gave a similar score distribution. Where a difference exists, the trend is, in the mid-range of severity, to give a lower score for the deterioration of intelligibility. This is associated with a difference as to how the judges use the characteristics of speech analysed just before performing the judgment. Conclusion: There is an argument for measuring intelligibility at the surface code level with a word recognition test or ordinal scales and for allowing the use of interval scales for severity judgment.
Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.