In 2007, the Center for Economic Opportunity in the Mayor’s Office of the City of New York mounted “Opportunity NYC: Family Rewards” (FR), the first holistic CCT initiative in an economically advanced, services-rich country. The initiative uses cash incentives to promote families’ connection and engagement with three systems –health care, education, and employment.  Evaluations of the program show that in the first two years of the program FR reduced family economic hardship, increased use of preventive healthcare, improved educational outcomes for a subset of high school students who were academically proficient at baseline, and reduced the rate of high school students’ problem behaviors such as delinquency and substance use.

 This paper analyzes impact variation in FR on educational processes and outcomes based on high school children’s school quality, measured at baseline with administrative data. At the end of 11^th grade, approximately 30 months after the start of implementation of the three year program, a survey was administered to children and their parents and school records were collected.  Multivariate OLS regression models are used to estimate the relationship between treatment status, school quality, and their interaction for each outcome independently.  All regressions control for covariates and used Huber-White standard errors to adjust for the modest clustering of children within schools (on average, there were 1.5 children/school).

 Findings show that treatment impacts varied by school quality on five of the nine outcomes assessed including children’s intrinsic and extrinsic academic motivation, academic time-use, academic performance and graduation rates (but not school engagement, parents’ spending and saving on children or grade retention).  The moderation analysis reveals that the intervention increased children’s motivation in low-quality schools but decreased children’s motivation in relatively higher quality schools.  In addition, children in lower quality schools changed their time use to focus more on academic activities.  Finally, the intervention increased academic performance in the third year of the program and graduation rates one year after the program ended for children in low-quality schools, and these effects decreased with school quality.  Results indicate that school quality affected children’s response to the intervention such that FR was successful in changing approaches to schooling and academic performance in lower quality schools but not in medium or relatively higher quality schools.  Implications for future adaptations and evaluation of CCT programs are discussed.