Abstract:
This paper uses a data set consisting of more than 2,000 quiz scores from an introductory statistics class and compares the actual quiz scores with students’ post-test predictions of quiz scores. The main findings are as follows: (1) Predicted scores are higher, on average, than actual scores, (2) better students predict their scores more accurately, (3) self-evaluation accuracy is worse when the material is difficult, and (4) students improve their self-evaluation accuracy over time. While previous research showed only modest improvement in self-evaluation accuracy, this study demonstrates dramatic improvements in accuracy, possibly due to the fact that students in the sample were tested frequently throughout semester.