Individual and peer calibration in team-based testing
MetadataShow full item record
The purpose of this study was to investigate whether students on permanent teams taking repeated Team-Based Tests (TBTs) can develop over time more accurate performance reputations in the eyes of their team-mates. The ability to estimate the expertise of one's self and one's team mates is important in team-based testing because it informs which position one may try to assume in team discussions as teams pursue consensus on test questions. Simply put, if I can increase my ability to assess my team-mates' general mastery of the content, then over time I will be able to make more effective decisions about whom I should try to learn from and whom I should try to teach. To make this judgment about relative levels of mastery, students must appraise both their own levels of mastery as well as those of their team mates. Team-Based Tests are part of an instructional strategy called Team-Based Learning (TBL) and this developmental effect is a pillar upon which TBL practitioners build their argument about the instructional value of the method. However, only indirect evidence has previously existed to support this claim: this study attempted to more directly document the reputation formation effect by comparing individual and team mate performance estimates to actual levels of performance over time. Participants in this study included 49 students in an undergraduate educational psychology class. Students in permanent teams taking repeated team-based tests developed increasingly accurate reputations within their teams, achieving statistical significance on the fourth team-based test. However, test difficulty played an important and potentially confounding role: the fifth and final team-based test was the most difficult of the semester, and on it reputation accuracy fell back to original levels. Individual metacognitive monitoring accuracy across tests did not improve over time and introverts' reputations were more accurate than extraverts' reputation on every test. Implications of these findings and directions for future research are discussed.