This article explores the effectiveness of intervention discussion sections for a college general chemistry course designed to apply research on student preconceptions, knowledge integration, and student explanation. Two interventions, on bond energy and spontaneity, were tested and intervention student performance was compared with that of a control group that did not use the experimental pedagogy. Results indicate that this instruction, which identifies students' initial conceptions and integrates those ideas into class discussion, leads to enhanced conceptual understanding. The intervention group outperformed the control group on a written course midterm, the thermodynamics portion of a standardized American Chemical Society examination, and an in-depth interview. In interviews, the intervention group students explained the energetics of bond breaking and formation at a more sophisticated level than did the control students. In contrast, control students were more tenuous in their thinking, tended to contradict themselves more when discussing bond energy, and harbored more misconceptions about spontaneity. © 2002 Wiley Periodicals, Inc. J Res Sci Teach 39: 464–496, 2002