Methodological aspects of assessing speech intelligibility among children with impaired hearing

In the assessment of speech intelligibility, procedures that are easy to use, but also valid and reliable, are needed. The aim of this study was to explore the reliability and concurrent validity of two scaling methods for assessing the speech intelligibility of children with impaired hearing.

A total of 51 children aged 4-17 years with a mild to profound hearing impairment performed a naming task comprising 62 single words. Altogether, 85 inexperienced listeners, divided into 17 panels, assessed the speech production of the children. A percent correct score (for identification of the words) was obtained from every listener. The listeners were also asked to judge the overall intelligibility of each child using a visual analogue scale (VAS) and a four-point ordinal level rating scale with verbal descriptors.

The percent correct score correlated significantly both with the VAS and the rating scale with verbal descriptors (Pearsons' coefficient 0.90 and Spearman's coefficient 0.78, respectively). Also, the VAS and the scale with verbal descriptors were significantly intercorrelated (Spearman's rho 0.85). Moderate to high correlations were usually obtained for all grades of hearing impairment, both sexes, different age groups and main communication modes.

Especially for assessing the need for remediation of speech and in monitoring its effectiveness, a quick overall estimate of speech intelligibility can be obtained quite reliably using the above rating scales. Item identification is more time-consuming but is often needed in research and therapy planning, where acoustic and linguistic phenomena that reduce intelligibility need to be traced.