摘要：Simulated conversations (SC) with trained actors are a performance-oriented method for assessing communicative competences in authentic task situations. This study evaluated the psychometric properties of parallel designed SC in a cross-professional setting: In teacher-parent and physician-patient conversations. Specifically, we addressed three research questions regarding the reliability and construct validity of the SC: (1) whether trained observers reach a satisfactory interrater reliability in rating examinees performance; (2) whether correlations among three types of ratings (external observers’, SC partners’, and students’ self-ratings) correspond to expectations; and (3) whether hypothesized correlations with external criteria (prior communication training, semester of study, high school grade point average) could be found. To answer these questions, n = 72 undergraduate medical students and n = 96 pre-service teachers conducted SC. Results showed sufficient interrater reliability (ICC = 0.71). Moreover, the pattern of correlations among the observer ratings, the other two types of ratings, and external criteria emerged as expected. These results provide evidence for the reliability and validity of the developed SC assessment.
