A concern of computer-based learning system design is how to accommodate learners' individual differences during learning activities. Previous research suggests that adaptive e-learning systems can effectively address such individual differences and, consequently, they enable more directed tutoring via computer-assisted instruction. In this paper, we explore this assertion, reflecting on the outcomes of two successive experiments that were performed to empirically demonstrate that learners' personality traits might be significant in understanding differences in learning outcomes from using e-learning systems. One hundred and forty-five participants from Computer Science participated in this study. A two-by-two between-subjects factorial study was designed, where the personality traits derived from Myers-Briggs Type Indicator, and the two computer-based learning systems (adaptive vs. non-adaptive) are independent variables. The results suggested that the personality trait that could be indicative of preferred learning styles; in particular extraverted/introverted personal traits significantly influenced learning activity on adaptive e-learning system. A design guideline might be implicative of how the e-learning system suits the learner's personality trait.