RESUMO
BACKGROUND/AIMS: Studies about the reliability and validity of intelligibility measures are fundamental to direct their use and interpretation in both clinical and research settings. The present study assessed inter-rater reliability, internal consistency and criterion validity of scores for the sentence and word subtests of the Protocol for the Evaluation of Speech Intelligibility (PESI). METHODS: This study included 180 adult volunteers who were divided into three groups: 30 speakers with dysarthria, 30 control speakers and 120 listeners. Intelligibility scores were calculated according to the percentage of words correctly transcribed in each subtest: words in sentences and single words. RESULTS: The intelligibility measurements exhibited a high level of inter-rater reliability for sentences [intraclass correlation coefficient (ICC) = 0.94] and for single words (ICC = 0.96). The items from the single words subtest of the PESI revealed high internal consistency (Cronbach's alpha = 0.93). The measurement of intelligibility of sentences and single words had a high discriminatory power (0.82 and 0.95, respectively), with high sensitivity and specificity, particularly for scores in the single words subtest. CONCLUSION: This study provides preliminary evidence of the reliability and validity of the PESI.