Traditional courses in statistics generally approach inference from a theoretical probability based perspective. Since the mathematical backgrounds of students is often not strong, many courses use computer based simulations to empirically justify ideas which are too complex or too abstract for most students. However, eventually, students must move from the empirical to the theoretical understanding of the concept if they are to apply these ideas to traditional inference methodologies. This paper questions the effectiveness of some of these strategies, and discusses how computer based technologies may be used effectively to bring together these theoretical and empirical perspectives.