This article examines the extent to which impacts of a conditional cash transfer (CCT) program on children's academic outcomes vary by key characteristics associated with families' propensity to earn the rewards offered by the program. We utilize an experimental study of Opportunity NYC-Family Rewards, a comprehensive CCT program in New York City in which low-income families were offered rewards for outcomes in health, education, and work. Building from a technique in propensity score research but used here to assess subgroup impacts in experimental studies, we use multiple baseline characteristics to predict the amount of rewards earned as a means to identify theoretically important, multivariate-defined groups of children for whom program effects might be more concentrated. Fourth-grade children in families with higher likelihood to earn rewards experienced more positive impacts of the program on academic outcomes. By contrast, no program impacts were found among 7th- or 9th-graders whose families were most likely to earn rewards. Differences across age groups are discussed in terms of their different developmental periods and the differing way the children experienced the intervention.