Modell, Harold I., Joel A. Michael, Tom Adamson, and Barbara Horwitz. Enhancing active learning in the student laboratory. We previously examined how three approaches to directing students in a laboratory setting impacted their ability to repair a faulty mental model in respiratory physiology (Modell, HI, Michael JA, Adamson T, Goldberg J, Horwitz BA, Bruce DS, Hudson ML, Whitescarver SA, and Williams S. Adv Physiol Educ 23: 82–90, 2000). This study addresses issues raised by the results of that work. In one group, a written protocol directed students to predict what would happen to frequency and depth of breathing during exercise on a bicycle ergometer, run the experiment, and compare their results to their predictions (“predictor without verification”). In a “predictor with verification” group, students followed the same written protocol but were also required to show the instructor their predictions before running the experiment. Students in a third group reported their predictions verbally to an instructor immediately before exercise and reviewed their results with that instructor immediately after exercise (“instructor intervention group”). Results of this study were consistent with our earlier work. The predictor with verification and predictor without verification protocols yielded similar results. The instructor intervention protocol yielded higher success rates in repairing students’ mental models. We subsequently assessed the efficacy of a prediction period at the beginning of the lab session and a wrap-up period at the end to compare predictions and results. This predict and wrap-up protocol was more effective than the predictor without verification protocol, but it was not as effective as the instructor intervention protocol. Although these results may reflect multiple factors impacting learning in the student laboratory, we believe that a major factor is a mismatch between students’ approaches to learning and the intended learning outcomes of the experience.