Example-based learning often uses a design in which learners first receive basic instructional explanations of new principles and concepts and then examples thereof. In this sequence, it is crucial that learners self-explain by using the content of the basic instructional explanations to elaborate on the examples. Typically, learners are not provided with access to the basic instructional explanations while they engage in self-explaining. However, it is reasonable to assume that this established design is suboptimal to some extent. When learners cannot retrieve the required knowledge components of the instructional explanations from memory, they can hardly generate the crucial self-explanations. Against this background, we analyzed the effects of a potential remedy for this suboptimality. Specifically, in 2 experiments with high school students we tested the effects of providing learners with constrained access to the basic instructional explanations while they engaged in self-explaining. The constraints were that learners were instructed to review the instructional explanations only when they could not remember certain required knowledge components or when they needed to check their self-explanations. We found that the effects of the constrained review option depended on learners' academic self-concept. Learners with low academic self-concepts benefitted from the review option, whereas learners with high academic self-concepts were even hindered by it. We conclude that the constrained review option should not be provided to learners with high academic self-concepts. On a more general level, we furthermore conclude that a high academic self-concept can be an obstacle rather than a beneficial resource in certain learning settings.