Regular quizzing that requires retrieval (e.g., fill-in-the-blank and open-ended questions) has been found to enhance the retention of information relative to quizzing that requires recognition (i.e., multiple-choice questions). This phenomenon, called the testing effect, has been demonstrated in a variety of laboratory and classroom studies. In past semesters, students in an upper-level psychology course who took fill-in-the blank quizzes performed significantly better on multiple-choice exams than students who took multiple-choice quizzes covering the same material. More recently, students have been provided with information about the testing effect, including data from earlier semesters of the same course, and allowed to individually choose their quiz format. While many students initially chose fill-in-the-blank quizzes, the majority switched to multiple-choice quizzes when allowed to do so one quarter of the way through the semester. Students in three sections of the course have exhibited this same pattern, despite evidence from their own sections that, on average, students taking fill-in quizzes earned higher grades on the first exam. The implications of this behavior, as well as its potential as a "teachable moment," will be considered.