Feedback reduces the metacognitive benefit of tests.

Testing long-term memory has dual benefits: It enhances learning and it helps learners discriminate what they know from what they do not know. The latter benefit, known as delayed judgment of learning (dJOL) effect, has been well documented, but in prior research participants have not been provided with test feedback. Yet when people study they almost universally (a) get feedback and (b) judge their learning subsequent to receiving the feedback. Thus, in the first three experiments, participants made JOLs following tests with feedback. Adding feedback significantly decreased the dJOL effect relative to conditions taking a test without receiving feedback. In Experiment 4, participants made decisions about which items to restudy (without actually restudying); adding feedback also decreased the accuracy of these decisions. These findings suggest that, in realistic situations, tests enhance self-monitoring, but not as much as previously thought. Judging memory based on prior test performance and ignoring the effects of feedback appears to produce an "illusion of not knowing."