Taking a Research Year: When It Helps and When It Does Not

A research year helps when it fixes a specific weakness and produces something you can show, and it does not help when it is a way to postpone a problem you have not solved. NRMP's 2026 data show no difference in median research counts between matched and unmatched non-U.S. citizen IMGs, so more research is not a general cure.

What the evidence shows #

In NRMP's 2026 Charting Outcomes report for non-U.S. citizen IMGs, among those who consented to share data, the median number of research experiences was 3.0 for both matched and unmatched applicants, and the median number of publications was also 3.0 for both. The median Step 2 CK score was 248 for matched and 242 for unmatched applicants, and the median number of contiguous ranks was 5.0 versus 2.0. NRMP notes that research data are self-reported and not verified, and that quality may vary greatly.

For program views, NRMP's 2026 Program Director Survey lists "demonstrated involvement and interest in research" and "evidence of continuous medical education without gaps" as separate factors a program may select. The narrative I read does not give the share of programs that picked each, so I cannot rank them. How programs read research is covered in how programs actually view research, and how they read gaps in gap years and how to explain them.

When a research year can help #

• You target an academic or research-oriented specialty and have no specialty-specific work.
• You need a mentor and a letter from a physician who knows you, and a research role is how you can get one.
• You are far from graduation and need recent, documented activity that shows you are professionally engaged (years since graduation).
• You can finish something: a project with a defined scope, a poster or an accepted abstract, not just "involvement".
• You are changing specialty and need credible evidence in the new field.
• You would otherwise sit idle while waiting for an exam date, visa or next cycle, and the role is structured.

When it does not help #

• An exam problem is unresolved. Research rarely compensates for a missing or weak exam, and programs filter on exam status (program screening filters).
• The role has no supervision or output, so you end the year with nothing to describe.
• You cannot afford it and the debt or loss of income would hurt your application elsewhere.
• Your application's weakness is clinical, such as no US clinical experience, and research does not address it (USCE vs research).
• The cost is a delayed application when your current profile is already competitive for your target programs.
• It rests on a fee-based program that promises authorship, matches or letters. See research fellowships for warning signs.

How programs may read it #

This part is common practice and opinion, not an official rule.

• A year described as "research" with a named supervisor, a project and results looks like activity.
• A year described vaguely, or listed with no output, may look like a gap. Say plainly what you did and why.
• A long sequence of short, unrelated projects looks less committed than one or two substantial ones.
• Clinical currency matters to many programs. A research-only year after a long clinical break may leave a question about patient care, so consider adding clinical exposure (how much USCE programs want).

In MyERAS, research goes in your Experiences as a Research type, with dates, hours, location and setting, and your Scholarly Work section lists outputs. Be accurate about your role and dates (ERAS experiences section).

Paid versus unpaid #

Paid research positions exist but are competitive and often prefer people with work authorization, a research degree or prior output. Many IMG research roles are unpaid or low-paid, particularly remote and volunteer ones. That is common practice, not a standard. Ask:

1. Can you afford the year, including living costs and the cost of delaying a residency salary?
2. Is an unpaid arrangement permitted in your situation? Whether unpaid work counts as employment depends on your status, the host and local rules, so ask the institution's HR or international office.
3. Is the unpaid role backed by a written plan with a mentor, project, hours and authorship expectations (authorship and author order)?

See research fellowships for types of positions and a fuller discussion of paid and unpaid roles, and finding a research position for how to look.

Visa implications, at a high level #

• Your status decides what you may do and be paid for. A visitor visa generally does not allow employment, so do not accept a research job on one on the basis of a mentor's assurance (B-1/B-2 for interviews and exams).
• Remote research from your home country avoids US immigration questions but not local employment and tax rules in your country.
• A research visa is not a clinical-training visa, and time in a research category can affect eligibility for a later exchange visitor program for residency. Ask the sponsor in writing (J-1 visa process, two-year home residency requirement).

A short decision checklist #

1. Name the weakness the year should fix.
2. Confirm exams are done or scheduled and are not the real problem.
3. Get the terms in writing: mentor, project, hours, pay, status, authorship plan.
4. Check your timeline against ERAS dates so output and letters exist in time (ERAS timeline).
5. Plan how to describe the year in your application in two or three sentences.
6. Decide a stop point. If the year is not producing output by a set month, know what you will change.