Assessing Self‐assessment in Anatomy

Abstract

IntroductionFormative self‐assessments are frequently utilized in anatomy courses to gauge learning. Self‐assessment tools have been demonstrated to improve student learning and performance on exams. However, few studies have assessed how specific elements, such as the inclusion of visual components, Bloom's taxonomy level, and the degree of topic integration, influence the efficacy of self‐assessments in an anatomy course.AimThis study retrospectively examined the relationships between self‐assessment construction characteristics, exam construction characteristics, and student exam performance. Specifically, 1) What elements of exam question construction correlated with performance? and 2) What elements of self‐assessment questions correlated with performance on corresponding exam questions?MethodsSelf‐assessment and exam questions administered during the Clinical Anatomy Block at the University of Arizona College of Medicine ‐ Phoenix from 2014–2018 for the Thorax and Abdomen and Pelvis and Lower Extremity units were examined. Self‐assessment questions were distributed to students via email, while exams took place in the anatomy laboratory, with a majority of questions incorporating structures on the donors, images, or anatomical models. For all questions, the constituent learning objectives, use of visual components, Bloom's taxonomy level, and degree of topic integration were noted. Exam performance data were collected for each class as the percentage correct for each question. These variables were entered into linear multiple regression models to assess their relationship to student performance for individual learning objectives. Because no individually identifiable data were utilized in this study, IRB approval was waived by the University of Arizona Office of Research, Discovery & Innovation (Protocol #: 1810045386).ResultsRegression analysis indicates that having visual components in exam questions is negatively associated with performance (t = −5.53, p <0.00). Conversely, a higher average Bloom's level for targeted learning objectives in self‐assessments (t = 2.02, p = 0.04), and a greater number of image‐based questions (t = 2.33, p = 0.02) are positively correlated with performance on corresponding exam questions.ConclusionsThese findings provide evidence that elements of self‐assessment construction, such as a visual component and a higher cognitive level, influence student exam performance when the exam format also incorporates these elements.This abstract is from the Experimental Biology 2019 Meeting. There is no full text article associated with this abstract published in The FASEB Journal.