When students are learning for a test, most of them will think about how well they will do, and consider if they have learned enough and can stop. In other words, students try to predict how well they will do on tests. Likewise, immediately after a test students might discuss together how difficult the test was, and swap discuss which answers they gave. Trying to guess afterwards how well you performed is called postdiction.
The question is, how accurately are the predictions and postdictions of students? This was investigated in a study by Hacker et al, Ninety-nine undergraduate students participated during a semester-length course in which the relation between self-assessment and performance was stressed. High-performing students were accurate in their pre/postdictions, with accuracy improving over multiple exams. In other words, they tend to calibrate their estimations to be fairly accurate. Low-performing students showed moderate prediction accuracy but good postdiction accuracy. Lowest performing students showed gross overconfidence in predictions and postdictions. Interestingly, the students' judgments of their own performance were influenced by their previous judgments, but not their prior performance. Performance and judgments of performance had little influence on subsequent test preparation behavior. Students tend to be set in their ways, and appear to make little use of feedback (in the form of test performance) or discrepancies between their own judgments and their actual performance.
Hacker, D. J., Bol, L., Horgan, D. D., & Rakow, E. A. (2000). Test prediction and performance in a classroom context. Journal of Educational Psychology, 92(1), 160.