Product dissection, or the systematic disassembly of design products, has been utilized in engineering education in order to better prepare students for industry. Despite the common use of product dissection in engineering classrooms, knowledge is lacking about how effective different methods of dissection are for encouraging learning and student engineering self-efficacy. This is problematic because without this knowledge, we do not know what components of product dissection impact (positively or negatively) learning. Therefore, the purpose of this study was to identify the impact of dissection virtuality (physical and virtual), product power source (electrical and manual), and product complexity (simple and complex) on efficiency, learning, and engineering self-efficacy through a factorial experiment with 30 engineering students. The results of the study show that virtual dissection is more efficient than its physical counterpart and also maintains the same learning benefits as physical practices. These results are used to develop recommendations for the use of product dissection in education and propel future research that investigates relationships between example-based design practices and student learning outcomes.