Improving accuracy in predictions about future performance

Abstract

AbstractWithin the academic context, students often expect to perform significantly better on an upcoming exam when compared to their actual performance. Unfortunately, students who are least accurate in their performance predictions are also the most at risk of being underprepared compared to their peers. Thus, efforts to enhance accuracy in performance predictions would benefit the student population that is most at risk of academic shortcomings and failure. Two studies examined whether incentives improve accuracy of academic performance predictions. In Study 1, 126 students in a 200-level undergraduate course provided estimates of how well they would perform on an upcoming exam. Two weeks later, the same students were randomly assigned to receive a reward incentive, punishment incentive, or no incentive for accuracy in performance predictions and their exam performance estimates were reassessed then compared to their actual scores on an exam completed one week later. In Study 2, 144 students from multiple 300-level courses provided performance predictions for an upcoming exam that were then compared to their actual exam scores for accuracy. For a subsequent exam, the same students were randomly assigned to varying levels of reward incentives (no incentive, modest incentive, large incentive) for prediction accuracy and their performance estimates were examined and compared to actual performance on a second exam. Findings from both studies indicated performance estimates are malleable and with the appropriate incentives, accuracy in predictions may be improved and contribute to better actual performance.