The Power of a Signal: The Impact of Preference Signaling on Matching in the Top 10 Most Competitive Specialties

Abstract Purpose Given varying preference signal numbers and structures across residency specialties, this study investigates the impact of preference signaling on match outcomes in highly competitive medical specialties. Method Data were from University of Texas Southwestern Medical School’s Texas Seeking Transparency in Application to Residency survey of applicants to the top 10 most competitive specialties using signaling between 2021 and 2024. Bivariate statistical testing compared groups across categorical and continuous variables. Multivariate logistic regression compared outcomes between 10 or fewer and 20 or more signals. Results The dataset contained 4,469 applications from 4,391 unique students. Number of signals used did not affect number of overall matches (2,458 of 2,908 84.5% for 3-5 signals, 94 of 112 93.9% for 6-10 signals, 178 of 203 86.7% for 21-25 signals, and 585 of 662 88.4% for 26-30 signals; P = .08). Higher signal numbers were associated with significantly higher matching rates at signaled institutions (916 of 2,098 37.2% vs 525 of 662 89.6%, P .001). Away rotations (odds ratio OR, 9.25; 95% CI, 6.37-13.43; P .001), signaling gold (OR, 7.74; 95% CI, 3.85-15.55; P .001), geographic connections (OR, 4.12; 95% CI, 3.01-5.64; P .001), and signaling programs (OR, 3.38; 95% CI, 2.43-4.68; P .001) were significantly associated with matching. Away rotations were ranked as most important (β = 2.23) followed by gold signals (β = 2.05), geographic connection (β = 1.42), and program signals (β = 1.22). Program signals had a stronger impact for applicants with 10 signals or fewer vs 20 signals or more (OR, 5.99 95% CI, 3.96-9.08 vs 3.00 95% CI, 1.33-6.77; P .001). Conclusions Specialties with more signals favor successful matching to signaled programs, but signal effectiveness diminishes as quantity increases. Applicants should prioritize impactful strategies to improve their chances of matching.