Geographic Signaling and the Match Funnel: How Programs Read It and How IMGs Should Use It

Geographic Signaling and the Match Funnel

How the modifier programs read alongside your signal works — and how IMGs should deploy it.

Program signals get all the attention. Applicants ask which programs to signal, how many gold versus silver, whether to signal home programs. Geographic preferences sit quietly underneath — three U.S. Census divisions, an optional 300-character narrative, no badges, no tiers. Most IMGs treat the section as an afterthought. The evidence says they shouldn’t.

Geographic Signaling is a modifier — and for IMGs, it is one of the few geographic levers structurally available. Understanding where it sits in the Match Funnel, what programs do with it, and how to deploy it deliberately is the difference between strategy and default.

Where Geographic Signaling sits in the Match Funnel

The IMGPrep Match Funnel positions Geographic Signals as a dashed-line input feeding the Selective Screen — not a binary filter, not a primary signal. The placement is deliberate.

The binary screens at the Selective Screen stage belong to SUVY: Scores, USCE, Visa, and Year of Graduation. These are the structural filters every program applies, often through automated criteria in the Program Director’s WorkStation (PDWS), before any human reads an application. Geographic preference operates downstream of SUVY — a modifier on the screen, not the screen itself.

Programs are instructed by AAMC’s official guidance to use geographic preferences as “a ‘plus factor,’ not a screening tool.” The intent is to weight geographic alignment within holistic review, not to filter on it. Programs that treat it as a hard filter are operating off-label.

The dashed line in the Funnel diagram represents exactly this: a modifier of the screen, not a filter.

How much each lever moves the needle

The most rigorous analysis to date is Romanoski et al., published in the Journal of Graduate Medical Education in February 2025. The study used AAMC’s own internal records — actual interview invitations pulled directly from PDWS — to measure how strongly each application lever influences a program’s decision to invite an applicant to interview. It is the canonical study of how program signaling and geographic preferences affect interview invitations for U.S. MD and DO applicants.

The findings are reported as odds ratios. An odds ratio is a multiplier: a value of 2.0 means twice as likely, 3.0 means three times as likely. The table below reads as a direct measurement of how much each lever raised the odds of an interview invitation in the 2023 application cycle.

*In-state status is structurally unavailable to IMG applicants; included for comparative context. Source: Romanoski NL, Morgan HK, Kerlek A, et al. J Grad Med Educ. 2025;17(1):56–62.

Three findings read off the table directly.

Aligned geographic preference raises interview odds. Across every specialty analyzed, applicants whose selected census division matched a program’s location received interview invitations at 1.83 to 2.75 times the rate of applicants whose preferences did not align.

“No preference” underperforms aligned preference in every specialty. Selecting no geographic preference is better than selecting a preference that does not match the program — but consistently worse than aligned preference.

Program signals are the dominant interview lever. Sending a program signal raised interview odds by 2.71 to 9.07 times — substantially more than geographic preference alone.

The DO-to-IMG bridge — why these numbers matter for IMGs

Romanoski’s analysis included U.S. MD and U.S. DO applicants. It did not include applicants from international medical schools. That exclusion is a real limit on the data — but it does not eliminate the relevance of the findings for IMGs.

The reason: many of the programs that interview and rank DOs are the same programs that interview and rank IMGs. Community-based Internal Medicine programs, Family Medicine programs across the country, and Psychiatry, Pediatrics, and other specialties in a range of regions all draw mixed applicant pools that include MDs, DOs, and IMGs. These programs use the same PDWS software, apply the same filter logic, and read the same signal and geographic preference fields. The mechanism Romanoski measured is the same mechanism operating on IMG applications at these programs.

What this means practically: the direction of effect is preserved. Aligned geographic preference helps IMGs at IMG-receptive programs. “No preference” is suboptimal. Program signals dominate.

What it doesn’t mean: that the magnitudes port over identically. IMG applicant pools have structural differences DOs do not face — visa screens, year-of-graduation thresholds, ECFMG certification requirements — and IMGs cannot claim in-state status, which Romanoski found is one of the strongest levers for U.S. seniors. The lever works for IMGs. How much it moves the needle at any specific program depends on that program’s IMG track record and selection patterns, which is where program-by-program analysis matters more than population-level odds ratios.

The combination effect — signals and geography compound

Two studies confirm that program signaling and geographic alignment work together rather than substituting for each other.

Benjamin et al. (2024, Internal Medicine) found applicants who both signaled a program and aligned geographically had an odds ratio of 3.2 for interview invitation and 6.4 for matching — versus applicants with neither. Abramowicz et al. (2025, Anesthesiology) found the strongest single predictor of interview invitation was a program signal that aligned with the applicant’s stated geographic preferences.

A signal landing inside the applicant’s preferred region is empirically stronger than the same signal landing outside it. The two levers are designed to compound, and the published evidence shows they do.

PDWS — the operational backbone

PDWS — the AAMC’s Program Director’s WorkStation — is the software programs actually use to receive and process applications. Every ERAS application lands here on the program side. PDWS displays signals and geographic preferences, applies structured filters (USMLE scores, year of graduation, visa status, ECFMG certification), tracks review status, and generates interview invitations.

It is also the system AAMC uses to collect program-side research data. Romanoski’s analysis pulled its interview-invitation outcomes directly from PDWS records.

The operational reality: PDWS makes binary SUVY filtering trivial at scale. A coordinator can set USMLE Step 2 CK ≥ 230, year of graduation within three years, and visa-sponsorship requirements, and narrow a pool of 3,000 applications to a reviewable subset before a human reads anything. This is the structural reason geographic preference operates as a modifier rather than a filter — by the time it enters the calculation, the binary screens have already executed.

The IMGPrep operating model

The IMGPrep Match Funnel Model is a structural representation of how the Match actually executes — anchored in NRMP outcome data, AAMC Program Director surveys, peer-reviewed signaling research, and the operational logic of the software programs use.

For Geographic Signaling specifically, the operating logic is four points:

What Geographic Signaling means for IMG applicants

Three operating principles follow.

Do not default to “No Preference.” AAMC’s own descriptive data shows IMG applicants select “No Preference” more often than U.S. MD or DO applicants in many specialties. Within the framework Romanoski measured, this is suboptimal — “no preference” sits below aligned preference in every specialty analyzed. If your USCE sites, family ties, or LOR geography support a region, select the aligned divisions.

Build defensible regional coherence. The 300-character geographic narrative is read alongside USCE, LORs, and personal statement. A regional claim that contradicts the rest of the application reads as performative. A regional claim that the application substantiates reads as authentic — and the application as a whole reads as coherent. This is the layer where Goodness of Fit either holds together or doesn’t.

Treat geographic preference as a multiplier on signal strategy, not a separate exercise. Signals belong inside preferred regions. Geographic preferences belong inside the application’s broader story. The two strategies are one strategy.

Related Resources