Preference Signaling in Family Medicine: Context Matters

I appreciate the thoughtful engagement by Siddiqi and colleagues¹ and their perspective on our work examining interview formats and preference signaling in family medicine residency recruitment.² Their letter underscores several tensions that many programs continue to navigate as recruitment practices evolve in the post-pandemic era, particularly the balance among accessibility, meaningful assessment, and unintended consequences of virtual interviewing.

My coauthors and I agree that interview modality choice remains highly consequential for both applicants and programs. As the authors note, virtual interviews reduce financial and geographic barriers yet may limit applicants’ ability to assess program culture and complicate programs’ evaluation of interpersonal attributes. Our original article was not intended to advocate for any single interview model, but rather to situate preference signaling within the broader context of application inflation that accompanied the rapid shift to virtual interviews. In that setting, signals (and setting family medicine’s signal count at a relatively low value of 5) were conceptualized as one potential mechanism to restore elements of interest signaling and informational asymmetry that had previously been conveyed through travel and in-person engagement.

Importantly, both our prior commentary and further analysis of the data suggest that preference signals have generally been incorporated as an additive—not determinative—component of holistic review.³ In our follow-up CERA study of family medicine program director perspectives, signals were most commonly considered as a positive factor during application review, but rarely superseded traditional influences such as audition rotations or geographic alignment, and did not independently drive interview offers or rank decisions.³ These findings are consistent with our original framing of preference signaling as a modest triage and alignment tool rather than a primary selection mechanism.

I also concur with the authors that early match outcomes, including fill rates, warrant careful interpretation. Our data suggest that during family medicine’s first year of signal adoption, programs largely exercised restraint in their use of signals—a pattern that may partly explain why broad outcome measures such as fill rates have shown limited immediate change. Studies published since have furthered this assessment, utilizing qualitative analysis of stakeholder perspectives to reinforce our original interpretation, describing preference signaling primarily as an adjunctive tiebreaker rather than a primary driver of interview offers or ranking decisions, and emphasizing persistent uncertainty about its optimal use during early implementation.⁴

Future studies should examine how preference signaling interacts with interview modality, applicant behavior, and program context. Our findings highlight substantial heterogeneity in how programs engaged with signals, including nonuse among a subset of programs, underscoring that signaling is neither monolithic nor uniformly applied. Additional examination—including further qualitative work exploring program intent and applicant interpretation—will be essential to determining how, and for whom, preference signaling adds value.

Thoughtful, data-informed refinement of recruitment practices remains critical for family medicine. Preference signaling, similarly to interview modality, should be evaluated not as an end in itself, but as an evolving component of a complex and interdependent recruitment ecosystem.