University researchers tell Hopkins board peer networks shape college applications and scholarship outcomes

Dr. Anoosha Nath, a researcher working with Hopkins Public Schools, told the board the district’s peer-network analysis shows information that flows through peer groups and school-run services strongly affects where seniors apply to college and whether they receive scholarships. The study links administrative records to a 2024 senior survey and finds students who took additional school services were more likely to apply to higher-quality colleges and 10 percentage points more likely to receive scholarships.

NUT GRAF: The research combines longitudinal student records and a 2024 senior survey to measure how school-provided information (college-prep courses, meetings with college representatives, or meetings with the college-and-career coordinator) — and indirect exposure to that information through peers — is associated with college quality and scholarship receipt. The team said its causal design isolates peer information flows from shared teacher or family influences.

Researchers’ methods and sample: Dr. Nath said the sample included 373 seniors who responded to the 2024 senior survey and that administrative data provided classroom enrollment, GPA and course-taking histories. To separate imitation from information transfer, the study compares outcomes of students who did not take up extra services with outcomes of their peers who did. The investigators used econometric models and predictive tools to control for demographics, prior achievement and other observable factors.

Key findings: Dr. Nath summarized three main findings. First, students who participated in school-provided services were 0.42 standard deviations more likely to attend higher‑quality colleges (quality proxied by the 75th‑percentile ACT of the college’s incoming cohort) and about 10 percentage points more likely to receive a scholarship, controlling for other factors. Second, among students who did not take up extra services, having classmates who did take up services was associated with better application outcomes and scholarship receipt — evidence the researchers interpret as information flowing through peer networks. Third, uptake of services was concentrated among top academic performers and female students: top‑decile GPA and being female predicted higher take‑up in the team’s OLS models.

Questions and limitations: Board members and attendees asked about the survey instrument, the construction of peer links, and potential biases from students who opt out of surveys or tests. Dr. Nath described the primary data sources as administrative classroom rosters (to infer peer ties), longitudinal demographic and GPA records, and the senior survey (which asked which counseling or college supports the student used and whether they received scholarships). She said the team is also exploring Panorama social-emotional data and qualitative follow-up to explain behavioral differences (for example, why female students show higher take-up).

Next steps: The research team said next work will investigate opt‑out patterns for statewide assessments, refine measures of college quality (adding selectivity measures), and use richer qualitative data to test behavioral mechanisms. The district indicated it plans to share the findings with the high school college-and-career office and consider how to broaden access to school-run services.

ENDING: The board’s discussion followed with clarifying questions about measures and equity; researchers said paper materials and follow-up analyses will be provided to the board and the district’s research partners.