The authors experimentally investigated the effect of crib cards on exam performance and student learning. Fifty-one students expected to use their prepared crib cards during an exam. However, they first completed an unexpected pretest without their crib card. Students performed significantly worse on the pretest than on identical questions when crib cards were allowed during the exam. Students prepared for a later exam by making a crib card but expected that they could not use it during testing. Following the exam, they completed an unexpected posttest containing identical questions with their crib cards. Performance did not differ. Students' expectations of crib card use fostered dependency and reduced learning, while expectations of non-use did not.