Search for: All records

Abstract Metacognitive calibration—the capacity to accurately self-assess one’s performance—forms the basis for error detection and self-monitoring and is a potential catalyst for conceptual change. Limited brain imaging research on authentic learning tasks implicates the lateral prefrontal and anterior cingulate brain regions in expert scientific reasoning. This study aimed to determine how variation in undergraduate life sciences students’ metacognitive calibration relates to their brain activity when evaluating the accuracy of biological models. Fifty undergraduate students enrolled in an introductory life sciences course completed a biology model error detection task during fMRI. Students with higher metacognitive calibration recruited lateral prefrontal regions linked in prior research to expert STEM reasoning to a greater extent than those with lower metacognitive calibration. Findings suggest that metacognition relates to important individual differences in undergraduate students’ use of neural resources during an authentic educational task and underscore the importance of fostering metacognitive calibration in the classroom. 

IntroductionModels are a primary mode of science communication and preparing university students to evaluate models will allow students to better construct models and predict phenomena. Model evaluation relies on students’ subject-specific knowledge, perception of model characteristics, and confidence in their knowledge structures. MethodsFifty first-year college biology students evaluated models of concepts from varying biology subject areas with and without intentionally introduced errors. Students responded with ‘error’ or ‘no error’ and ‘confident’ or ‘not confident’ in their response. ResultsOverall, students accurately evaluated 65% of models and were confident in 67% of their responses. Students were more likely to respond accurately when models were drawn or schematic (as opposed to a box-and-arrow format), when models had no intentional errors, and when they expressed confidence. Subject area did not affect the accuracy of responses. DiscussionVariation in response patterns to specific models reflects variation in model evaluation abilities and suggests ways that pedagogy can support student metacognitive monitoring during model-based reasoning. Error detection is a necessary step towards modeling competence that will facilitate student evaluation of scientific models and support their transition from novice to expert scientists.