Effective methods of instruction for learning computer software applications are continually sought by educators in both academic and business settings. Effective training has become more important as computer users interact more directly with software programs in developing and using their own computer-based systems for personal and business uses (Sein and Robey, 1991). End users are now expected to develop their own information systems (e.g., see Rivard and Huff, 1988) without the constant aid of IS professionals. The emergence and increasing importance of end user computing has created a challenge for the IS department, and using effective training to ensure the success of end user computing has been emphasized by several researchers in information systems (Bostrom, Olfman, and Sein, 1990; Davis and Bostrom, 1993). The significance of end user training in practice makes it important for researchers to seek answers to the question, "How should we train end users?" Bostrom et al. (1990) provide a framework for understanding the broad variables involved in developing an answer to this important question. Their framework and research has led to subsequent work (Davis and Bostrom, 1993; Ruble and Stout, 1993;Bostrom, Olfman, and Sein, 1993). Their Prior research into the question of how to train end users reveals an important lesson. To provide useful insights, researchers must study alternative training methods that are based on rich distinctions between realistic training methods. The experiment reported here illustrates this lesson. Lecture-based and computer-based training are experimentally compared for end users with different learning styles (as measured by the Kolb Learning Style Inventory). The end user training outcomes measured were achievement, efficiency, and satisfaction. The end users who participated in the study were college students learning to use a word processing package. The results suggest that computer-based training is more effective than lecture based training except for Assimilators, who appear to learn equally well under either method. Given the limitations of the study and taking into account suggested cautions, these results should be replicated before being widely accepted.research and the research of others have lessons that are critical for researchers to recognize. This research uses the Bostrom et al. (1990) framework to investigate the effect of instructional method and individual differences and their interaction on the effectiveness, efficiency and user satisfaction of end users learning computer software. In this research, learning style is used as the measure of individual differences. The results of the lessons learned from a review of the literature and the lessons learned from this research suggest directions for future research.