For the past three years, as part of an on-going project funded by the Bill and Melinda Gates Foundation, MDRC researchers have published reports which capitalize on a random component in New York City’s high school application process to estimate effects on student outcomes of enrolling in over a hundred of New York City’s small public high schools of choice (SSCs)[1]. SSCs are defined as small, academically nonselective, high schools that were opened between 2002 and 2008 in the New York City school district. Serving approximately 100 students per grade and open to students at all levels of academic achievement, SSCs were created to serve the district’s most disadvantaged and historically underserved students. Using an unusually large sample of over 20,000 students, the project team has shown that students who enroll in SSCs are roughly eight to ten percentage points more likely to successfully transition into high school and graduate after four years, compared to their control group counterparts. Equally striking is the fact that these SSC effects are experienced by many different types of students – including among others, those that enter high school performing below grade-level on their math and English state standardized tests, those that qualify for free-/reduced-price lunch, and students of all races and genders.

At APPAM this fall we would like to present a new round of findings that dig deeper into the city’s high school reform effort and speak directly to the conference goal of integrating research and public policy decision-making. Specifically, with an additional cohort of students graduating from high school, we will report findings that address the following questions:

• Did New York State’s phase-out of the least rigorous type of high school diploma granted to regular education students - local diplomas – change SSC effects on graduation rates?
• Did SSCs sustain their positive effects in the face of increasing graduation rates, on average, across other NYC high schools against which they are being compared?
• Do special education students and English language learners also experience a positive and statistically significant effect of SSC enrollment? (These subgroups were too small to report separate findings from previous analyses.)
• Once we account for student attrition over time (using a model-based imputation approach plus a naïve and very conservative alternative imputation), is the estimated effect of SSCs on graduation rates still positive and statistically significant?
• Do SSC teachers exert an inordinate amount of influence on their students’ scores on New York State Regents exams (which are part of students’ high school graduation requirements)? And if so, might this explain part of the observed positive effects of SSCs on high school graduation rates?