Validating discourse production performance of two sequential picture descriptions by Cantonese-speaking population aged 50 or above

Abstract

Available norm-referenced discourse analysis tools in Cantonese-speaking population are scarce. Pictorial stimuli are often randomly selected by speech-language pathologists (SLPs) without standardization to evaluate discourse. Discourse production targets are also predicted by SLPs without validation against normative performance. This study carried out normative data collection based on two culturally appropriate picture sequences among local adults aged 50 or above, to establish age-stratified quantitative discourse norms on narration content and efficiency. The results indicated that only cognitive abilities of normal speakers were significantly associated with content richness. When controlling the effect of cognition, there was an insignificant effect of age and stimuli used on the narration content. On the other hand, results showed a significant effect of the stimuli used on narration efficiency, but insignificant effects of age and cognition on the same measure. Validation results also showed that both SLPs and students generated MCs targets that deviate significantly from the normative MCs production. When different stimuli were used as discourse tasks, clinicians’ MC predictions varied significantly while normative performance on narration content remained consistent. Referring to established norms as well as taking cognitive abilities into consideration when evaluating discourse performance are implied.