Although many students benefit from evidence-based reading comprehension interventions, not all students will exhibit adequate response. Moderation analysis provides a statistical approach to examine for whom and under what conditions interventions are most effective. Conducted within a parent project, which investigated the effects of an inferential reading comprehension intervention, the current study examined factors related to the deployment of students' attention as well as language status that might be associated with differential response to intervention. Sixty-six struggling middle school readers were randomly assigned to a computerized version of the intervention, a teacher-led version, or business-as-usual (BaU) control instruction. The influence of language status (i.e., English Learner status) and pre-intervention levels of mind-wandering, anxiety, and mindset on the effects of the inferential reading comprehension intervention were examined. There were no moderator effects for the teacher-led group compared to the BaU control. Conversely, anxiety, mind-wandering, and language status moderated the effects of the computer-led intervention for some reading and inference-making outcomes. The computer-led intervention was associated with improved inference-making for students with higher levels of self-reported anxiety and mind-wandering. In contrast, the computer-led intervention was less beneficial than BaU instruction for English learners. Findings are discussed with respect to how these factors might be relevant for interpreting the effects of interventions for struggling middle school readers in general, and for English learners in particular. The findings also point to the importance of considering the characteristics of both student and instructional features in the creation and testing of reading comprehension interventions.