This study examined the effects on performance of online access to peer-generated questions during question-generation activities. Two eighth grade classes (N = 63) participated in six weekly question-generation sessions to support English learning. An online student question-generation learning system was adopted. In contrast to expectations based on the literature on observational learning and scaffolding, no significant differences in English academic performance and weekly question-generation performance were found between the groups with and without online access to peer-generated questions. Nonetheless, the question-viewing function of the system used in this work induced a significant and immediate increase in QG performance for the group with access. Possible reasons for the unexpected non-significant results are provided with reference to limited attention theory and cognitive load theory, along with the significance of this study, as well as suggestions for instructional implementations and future research.