註釋 Meta-analyses and systematic reviews of literature comparing the use of virtual patients (VPs) to traditional educational methods support the efficacy of VPs (Cook, Erwin, & Triola, 2010; Cook & Triola, 2009; McGaghie, Issenberg, Cohen, Barsuk, & Wayne, 2011). However, VP design research has produced a variety of design features (Bateman, Allen, Samani, Kidd, & Davies, 2013; Botezatu, Hult, & Fors, 2010a; Huwendiek & De Leng, 2010), frameworks (Huwendiek et al., 2009b) and principles (Huwendiek et al., 2009a) that are similar in nature, but appear to lack consensus. Consequently, researchers are not sure which VP design principles to apply and few validated guidelines are available. To address this situation, Huwendiek et al. (2014) validated an instrument to evaluate the design of VP simulations that focuses on fostering clinical reasoning. This dissertation examines the predictive validity of one instrument proposed by Huwendiek et al. (2014) that examines VP design features. Empirical research provides evidence for the reliability and validity of the VP design effectiveness measure. However, the relationship between the design features evaluated by the instrument to criterion-referenced measures of student learning and performance remains to be examined. This study examines the predictive validity of Huwendiek et al.'s (2014) VP design effectiveness measurement instrument by determining if the design factors evaluated by the instrument are correlated to medical students' performance in: (a) quizzes and VP cases embedded in Neurological Examination Rehearsal Virtual Environment (NERVE), and (b) NERVE-assisted virtual patient/standardized patient (VP/SP) differential diagnosis and SP checklists. 