Word identification testing in the diagnostic evaluation of dysarthric speech

Publisher Website
Google Scholar
Abstract
This article presents a method for the clinical assessment of segmental contrastivity in dysarthric speech. The method is based on listeners' identifications of single words spoken by the examined subjects. In the form presented here, normal listeners had to identify each word of a 100-item word list among six ‘rhyming’ alternatives which differed from each other in exactly one position. The distribution of target phonemes' within a word list was held invariant over the different lists used for testing and each list in itself consisted of two equivalent halves. This construction proved advantageous from the point of view of psychometric evaluation. It further allowed for the distinction of differential error patterns, thus providing a useful tool in the design and evaluation of therapeutic interventions. Data were collected from a total of 32 dysarthric patients and 10 controls, with responses from a total of 14 (normal) listeners. The most relevant issues of psychometric quality are discussed and applications are presented. The study is particularly focused on an optimization of the materials on the basis of a multiple-choice response format, with the prospect of compensating for the disadvantages of this approach in future versions of the test.