The aim of this study was to determine whether problem-based writing with peer review (PW-PR) improves undergraduate student performance on physiology exams. Didactic lectures were replaced with assignments to give students practice explaining their reasoning while solving qualitative problems, thus transferring the responsibility for abstraction and generalization to the students. Performance on exam items about concepts taught using PW-PR was compared with performance on concepts taught using didactic lectures followed by group work. Calibrated Peer Review™, a Web-delivered program, was used to collect student essays and to manage anonymous peer review after students “passed” three calibration peer reviews. Results show that the students had difficulty relating concepts. Relationship errors were categorized as 1) problems recognizing levels of organization, 2) problems with cause/effect, and 3) overgeneralizations. For example, some described cells as molecules; others thought that vesicles transport materials through the extracellular fluid. With PW-PR, class discussion was used to confront and resolve such difficulties. Both multiple-choice and essay exam results were better with PW-PR instead of lecture.