Corrective feedback on practice tests can sometimes impair later test performance

Retrieval practice is generally good for long-term retention, but multiple-choice (MC) practice tests can introduce problems. For example, the lures gain familiarity during practice and can be erroneously endorsed on later tests (negative testing effect). In this talk, we describe a related problem caused not by lures, but by corrective feedback. Although normally good for learning, we have found that exposure to correct-answer feedback on MC practice tests increases the likelihood that the feedback is erroneously endorsed later on related final-test questions which have a different correct answer. However, at the same time, answering mul- tiple-choice questions during practice causes retrieval of accurate information about the options which opposes this bias. In a series of experiments, we explore the interplay of these opposing memory influences in both experimental and educational contexts. Depending on factors such as practice-test format, final-test format, the degree of false recognition, and the temporal proximity of related questions, final test performance can be facilitated or impaired. We discuss both the theoretical implications of these results as well as offering some practical recommendations