Caring Adults Improve Outcomes for Students Experiencing Homelessness: Evidence from New York City Public Schools

• In Event: Cross-Agency Administrative Data Sets for Actionable Intelligence

Abstract

Introduction: Homelessness is a significant barrier to educational success. Students experiencing homelessness are chronically absent on average, change schools frequently, and rarely achieve proficiency. Accordingly, there is considerable interest in understanding the role that schools can play in identifying and addressing the needs of these students. The “Every Child and Family is Known” (ECFIK) initiative, which was launched by the New York City Children’s Cabinet in 2023, is among the most intensive efforts to date to tackle this challenge. A cross-agency program that assigns each homeless student a dedicated “Caring Adult” from their school staff, ECFIK facilitates personalized support through regular check-ins and coordination with students and their families. This study evaluates the impact of the first year of ECFIK on key educational outcomes, including school attendance, academic performance, and school stability.

Method: Using integrated administrative data from the New York City Department of Homeless Services (DHS) and New York City Public Schools (NYCPS), we employ a difference-in-differences design to measure the causal effects of the program.

Results/Implications: Our results suggest that students participating in ECFIK experience significant improvements in school attendance, math test scores, and school stability. Although participant schools are not selected randomly, a comprehensive analysis of pre-trends indicates that non-treated schools are a credible control group. These findings highlight the crucial role that trusted school staff and personalized support can play in improving outcomes for students in shelter, while also emphasizing the importance of interagency collaboration in addressing homelessness.

Authors

• Michael Thomas Cassidy, Icahn School of Medicine at Mount Sinai
  Presenting Author

• Glendalis Perez, Center for Innovation through Data Intelligence
  Non-Presenting Co-Author