ABSTRACT: Analogies play significant roles in communication as well as in problem solving and model building in science domains. Analogies have also been incorporated into several different instructional strategies--most notably in science domains where the concepts and principles to be learned are abstract or complex. Although several instructional models for teaching with analogies have been developed, the role of non-examples or contrasting analogies has not been incorporated into instructional models. Research on contrasting analogies as well as the effect of non-examples in concept learning supports the incorporation of contrasting analogies into instructional strategies using analogies. In addition, research on conceptual change has shown that effective instructional strategies in inducing conceptual change produce some form of cognitive conflict in the learner. Taken together, these findings suggest that a conceptual change strategy integrating contrasting analogies (non-examples) that promote cognitive conflict could be effective in promoting conceptual change. The purpose of this study was to examine the effects of contrasting analogies on learning and conceptual change. Specifically, this study investigated the effect of instruction including refutational contrasting analogies (situations which are analogous to common misconceptions) on understanding of and reasoning about natural selection. Instruction including refutational contrasting analogies was compared to instruction incorporating non-refutational contrasting analogies (analogies which are neither analogous to natural selection or a common misconception of natural selection) and no contrasting analogies (stated analogies which are analogous to natural selection). Participants were students enrolled in undergraduate education courses and were randomly assigned to one of three conditions: refutational contrasting analogies, non-refutational contrasting analogies, or no contrasting analogies. Participants worked through a self-instructional unit on natural selection--an area for which misconceptions have been well-documented. The effects of the analogies were evaluated in terms of student understanding of and reasoning about natural selection. Reasoning about natural selection was identified as transformational reasoning (a common but incorrect view of natural selection) or variational reasoning (consistent with correct understanding of natural selection) and was measured by performance on a posttest. Understanding of natural selection was evaluated by performance on a posttest designed to measure understanding of ten concepts related to natural selection. Analysis of variance showed no differences among groups on either understanding of or reasoning about natural selection as measured by the posttests. However, there were significant differences between groups on the analogy portion of the instructional materials, with the refutational contrasting analogies group answering fewer questions correctly than the non-refutational contrasting analogies group. This suggests that the instruction did not allow participants to reach a level of mastery that may have influenced performance on the posttests. Considerations for the design of instruction and directions for future research are discussed.