Comparison of selected methods of assessing intelligibility of misarticulated speech

Speech intelligibility is one of the most frequently employed parameters used to describe the effectiveness of treatment for articulation disorders. Intelligibility is measured by evaluating the percentage of agreement between the speaker's intended message and the listener's response to the message. Speech-language pathologists rely on indirect measures of speech intelligibility, making inferences about an individual's overall communicative abilities based on subjective evaluation, or on indirect measures of intelligibility derived from commercial tests of articulation.

The primary purpose of this investigation was to compare several methods of measuring intelligibility, including direct measures, subjective measures, and indirect measures derived from commercial tests of articulation. The exact nature of the relationships among measures of intelligibility has not been previously examined, making it difficult to compare and interpret results based on different measures of intelligibility. The efficacy with which indirect indices of intelligibility derived from commercial tests of articulation predict objective intelligibility scores based on the rate of correct identification of stimulus materials by listeners, as well as prediction of subjective impressions of intelligibility, was determined.