New York City's high school admissions process requires all eighth graders to submit a rank-ordered list of up to 12 schools, with assignments made through a deferred acceptance algorithm. Despite the strategy-proof design of this system, students from low-income backgrounds and those with lower academic achievement disproportionately apply to and enroll in schools with low graduation rates — an outcome driven largely by application behavior rather than eligibility constraints. Prior research documents widespread misunderstanding of the admissions process, including mistaken beliefs that listing fewer schools improves one's chances of admission, as well as informational overload and unequal access to guidance from parents and counselors.
Sarah R. Cohodes, Sean P. Corcoran, Jennifer L. Jennings, and Carolyn Sattin-Bajaj investigate whether and how informational interventions can improve school choices at scale by conducting a school-level randomized controlled trial across 473 NYC middle schools serving over 115,000 eighth graders. They evaluate three counselor-delivered interventions that vary in their level of customization and mode of delivery: a simplified paper list of nearby schools with graduation rates above the city median (Fast Facts), a personalized online recommendation app, and a publicly available online school-search tool. The interventions were distributed through school counselors rather than directly by the research team, deliberately approximating how a district would implement such tools in practice.
Every intervention reduced the likelihood of application to and enrollment in schools with graduation rates below the city median of 75%, with the most effective arms reducing low-graduation-rate enrollment by 5 to 6 percentage points — a 13 to 15 percent reduction. A key mechanism was displacing students' default first choices: many students had been defaulting to schools with guaranteed admission and low graduation rates, and the interventions redirected them toward higher graduation-rate alternatives with nonzero admission probabilities. Simplified paper interventions outperformed digital formats, and produced more uniform effects across student subgroups. English learners showed the strongest response across all treatment arms. Putting the same information online as in the paper arm was ineffective, suggesting that engagement — not merely information access — drives impact. The authors find no evidence of academic mismatch: students steered into higher graduation-rate schools performed similarly to peers in control schools. They conclude that large-scale informational interventions can be effectively delivered through school counselors, but that intervention design matters for both average effectiveness and distributional consequences.
- Journal Article: https://journals.sagepub.com/h...
- Working Paper: https://www.nber.org/system/fi...