The ability to recognize and diagnose acid-base disorders is of the utmost importance in the clinical setting. However, it has been the experience of the authors that medical students often have difficulties learning the basic principles of acid-base physiology in the respiratory physiology curriculum, particularly when applying this knowledge to the analysis of arterial blood gases. Collaborative teaching strategies may enhance student performance in quizzes in undergraduate physiology courses. In particular, the so-called peer instruction technique has been reported to enhance both the transfer and retention of learned material in a time-efficient fashion. The Harvard professor Eric Mazur originally developed peer instruction for teaching physics to undergraduate students. In this teaching approach, specific questions are integrated in the conventional lecture format; in an initial "individual phase," students mark down their answer and rate how confident they are about the correctness of their answer, and this is followed by a "pair phase," where students are asked to convince their neighbor of their answer. After this, they may revise the answer and again rate their confidence in their second answer. This teaching approach may differ substantially from conventional group work, where students typically discuss questions in small groups and mark down their answer during the lecture. The present study attempts to implement and compare conventional group work and Mazur's peer instruction technique for teaching second-year medical students to diagnose common acid-base disorders by arterial blood gas analysis.