The purpose of this set of analyses, conducted on a randomized control trial examining the effects of a fourth-grade fraction intervention, was to assess the initial academic deficit severity hypothesis. With this hypothesis, at-risk students with more severe initial academic deficits are expected to profit less from intervention than do students with less severe initial academic deficits. Moderation analyses indicated that students with varying degrees of initial academic deficits benefited comparably from the intervention, such that effect sizes comparing intervention against control students were similar across the range of initial academic deficits. In a similar way, across the range of initial academic deficits, intervention students' posttest (spring) calculation performance was normalized (one standard error of measurement above the 25th percentile of a not-at-risk normative group's spring performance). On the most distal fractions outcome, however, normalized performance was achieved for intervention students with less severe initial academic deficits. Findings are discussed in terms of methods for judging intervention efficacy and for making individual decisions about when students should exit intervention.