In this study, we examined whether think-aloud procedures would uncover differences in the kinds of inferences generated by average and below-average readers. Participants were 40 third-grade children who were divided into groups of average and below-average readers. All participants completed measures of nonverbal IQ, reading, language, and working memory, and a story comprehension task that consisted of two conditions: listen through and think aloud. The major findings in this study were that (a) average readers generated significantly more explanatory inferences than below-average readers, and (b) comprehension performance as measured by story recall was significantly better for both groups in the think-aloud condition than in the listen-through condition. The discussion addresses the implications of these findings.