Choosing a Specialty

Part of Medical School. This is the decision everyone asks you about from the day you get in, and almost nobody makes it when they think they will. It happens in third year, from a small sample, under time pressure. What you do before then is widen the sample and understand what the word "competitive" is actually measuring, and this page is about both.

Looking for the argument about whether debt should decide your field? That is Debt, Specialty & the Life You Want, which used to live at this address. This page is the practical half.

The short version

You will mostly decide in third year, from the six to eight core clerkships your school happens to schedule you into, in whatever order it happens to schedule them.

That is a small and slightly arbitrary sample, which is the single best reason to get exposure to things outside it before third year rather than to pick early.

"Competitive" is four separate things, and only some of them are still in your control by the time you find out they matter.

The profiles are for eliminating and for surprising yourself, not for choosing. A field you have never seen described is worth more to you right now than a ranking of the ones you have.

The decision gets made in third year, from a sample of about eight

Most of your first two years is classroom work. Then clerkships start, and you rotate through the core specialties: internal medicine, surgery, pediatrics, obstetrics and gynecology, psychiatry, family medicine, neurology, and often emergency medicine. That set varies a little by school and the order varies a lot.

What that schedule does to you is the practical heart of this page.

The sample is small. There are far more than eight specialties, and most of the ones people end up loving are not core rotations. Anesthesiology, radiology, pathology, dermatology, ophthalmology, physical medicine and rehabilitation, and every subspecialty of every field above are not on the standard third-year list. If you leave your exposure entirely to the schedule, the fields you consider are the fields your registrar picked.

The order matters more than it should. People fall for whichever specialty they meet on a good rotation with a good team, and that's not a character flaw, it's how exposure works on anybody. Knowing it is happening is most of the defense. If you loved surgery and it was your first rotation with a resident who taught you well, ask yourself honestly how much of that was the field and how much was the six weeks.

Neither of those is a reason to decide early. It's a reason to build a wider sample before the schedule builds a narrow one for you.

What you can actually do before then

Join the interest groups, and go to the ones you have no interest in. Every school has specialty interest groups, they cost nothing, and the useful ones are the fields you know nothing about. An hour with a pathologist is the cheapest possible way to find out that pathology is not what you assumed it was.

Find one physician in a field you have never seen up close and ask to spend a day. This is the same skill as Shadowing, used for a different purpose. Then you were proving interest in medicine. Now you are sampling within it.

Do research in something you might want, but hold it loosely. Research in a field is genuinely useful if you end up applying there, and pre-clinical research is when you have the most time you will have for years. Just don't let a lab you joined at 23 quietly become the reason you apply to a specialty at 26.

Notice what you're drawn to about the work rather than the field. Continuity or episodes. Procedures or reasoning. Adults or children. Fast or deep. Working alone or on a team. Every specialty is a bundle of those, and the bundles cut across the names in ways that surprise people. This is the axis the profiles are actually organized to help you think along.

"Competitive" is four separate things

When people say a specialty is competitive, they're compressing at least four different mechanisms, and it's worth separating them because they have different deadlines.

Board scores. Step 1 has been pass/fail since 2022, which moved weight onto Step 2 CK. That shift matters most for anyone who was counting on a strong Step 1 to offset something else, because the exam that now carries the number comes later, after clerkships, when you have less room to maneuver.¹

Research output. In the most competitive fields the expected number of abstracts, posters, and publications has risen a great deal, and it is the requirement that most obviously advantages students who arrived with connections and time. It is also the one you can start earliest.

Signaling and geography. Most specialties now use program signals in ERAS, where you flag a limited number of programs as genuinely preferred, and many weigh geographic preference explicitly. This is a strategy problem rather than a merit problem, and it rewards knowing the system, which is exactly the kind of thing the hidden curriculum hands to some students and not others.²

The raw number of positions. Some fields are hard because there are very few spots, full stop. No amount of preparation changes the arithmetic, and it is the reason a parallel plan is normal rather than pessimistic in those fields.

The NRMP publishes what actually happened, by specialty, in Charting Outcomes in the Match.³ Read it when the question becomes real for you, and read it instead of forum consensus, because the forums systematically overstate what a competitive applicant looks like.

Away rotations, and what they cost

In several fields, mostly the surgical subspecialties and dermatology, an away rotation at a program you want is close to expected. You spend four weeks working at another institution, essentially auditioning.

The financial version of that sentence is: you pay for housing in another city, for travel, and for application fees on top of it, during a year when you are not earning. Then interview season adds its own costs. For a student without family money this is one of the sharpest points on the entire path, and it arrives late, when most of the aid conversations are behind you.

Three things to have early rather than in fourth year.

Many specialty societies run away-rotation and visiting-student scholarships aimed specifically at students who could not otherwise go, and they are consistently undersubscribed. Your school's student affairs office knows which ones exist for your field.

Your school's financial aid office can often increase your cost-of-attendance budget for documented away-rotation and interview expenses, which raises how much you are permitted to borrow. That is more debt rather than free money, and it is the difference between going and not going.

And if the money genuinely will not stretch, fields differ enormously in how much an away rotation matters. It is close to required in a few and close to irrelevant in most. Find out which one yours is before you assume you are excluded.

Fourth Year covers this cost block properly, and The Money covers the rest.

How to use the specialty profiles without letting them decide

The Specialty Explorer holds a profile for every specialty and subspecialty, with what the work is, what training looks like, what the pay and the hours and the burnout data say, and what people in the field say about the culture. It exists because this information is otherwise scattered across forums, salary surveys, and whatever a single attending told you once.

Three ways to use it that work, and one that doesn't.

Read the fields you have never heard of. This is the highest-value use by a wide margin. Most people can name fifteen specialties and there are far more than that, and the reason someone ends up happy in interventional pulmonology or medical genetics is almost always that they encountered it at all.

Read for the shape of the day, not the title. Every profile describes what the work actually consists of. Compare that against the axes in the section above, because "I want continuity with patients over years" eliminates and includes more fields than any specialty name does.

Use the hard numbers to eliminate, gently. If a field's training is seven years and you know with certainty that your life cannot absorb seven years, that is real information. Use it to narrow, and hold it loosely enough that a rotation can still change your mind.

What doesn't work is treating them as a ranking. They are not ordered, on purpose. Every profile carries an exploratory-not-prescriptive note at the top for the same reason. No amount of reading substitutes for four weeks on the service, and the reading is there to make sure the four weeks happen in the right places.

What belongs to other pages

Whether the debt should get a vote belongs to Debt, Specialty & the Life You Want, which argues that specialty choice is a smaller lever on your finances than you fear and that letting it drive the decision is how people end up in fields they resent.

What the Match actually is, and the mechanics of ERAS, rank lists, and SOAP, belongs to The Match, and What Happens If You Don't.

What the specialties pay belongs to What Doctors Actually Make, and the compensation section inside each profile.

Whether any of this changes because of AI belongs to Medicine in the AI Era, which is the page most likely to be out of date and says so.

Where this fits on the Road

This is the last decision the Road covers, and it is the seam where the site hands you over to The Specialty Explorer, which is where the actual work of this decision happens. The useful thing to carry across is that you are not looking for the right answer yet. You are widening the set of things you have seen, so that when third year narrows it again, it narrows from somewhere honest.

Match mechanics, signaling rules, board-exam formats, and specialty competitiveness all change between cycles. Verify current requirements with the NRMP, the AAMC, and your own school's student affairs office before planning around anything here. — Last reviewed: 2026-08-03

References

USMLE Step 1 moved to pass/fail score reporting for examinations taken from 26 January 2022, shifting numeric-score weight in residency selection onto Step 2 CK. https://www.usmle.org/usmle-step-1-transition-passfail-only-score-reporting ↩
AAMC ERAS program signaling: most participating specialties allow applicants a limited number of signals indicating genuine interest in specific programs, with geographic preference collected separately. The number of signals and the participating specialties change every cycle, and the AAMC publishes them per season. Verified 2026-08-03 from the ERAS hub, which carries the current season's signaling details, timelines, and tools: https://students-residents.aamc.org/applying-residencies-eras/apply-residencies-eras-system ↩
NRMP, Charting Outcomes in the Match. Verified 2026-08-03, most recent edition published 28 July 2026. Reports are organized by applicant group (U.S. MD seniors, U.S. DO seniors, U.S.-citizen and non-U.S.-citizen IMGs) rather than as one combined specialty table, and examine eleven to thirteen measures including Step 2 CK score, the number of contiguous ranks in the preferred specialty, the number of distinct specialties ranked, and the counts of research, work, and volunteer experiences. Separate interactive reports let you filter score baselines by specialty. This is the source to use in place of forum consensus, which systematically overstates the profile of a successful applicant. https://www.nrmp.org/data-topic/applicant-competitiveness/?post_type=match-data ↩