CampusBreeze

Grad School Rankings Are a Different Animal From Undergraduate Ones

Grad school rankings measure a department, not a university, and most of them run on unaudited reputation surveys rather than outcome data. A programme can sit far above or below its own institution's overall name. Read the methodology before letting any list order your applications: what follows is how each list is actually built.

Undergraduate ranking

Ranks an institution

One number for the whole university: the school, not any single department inside it. Endowment, admit rate, class size and graduation rate are all measured at the institution level and blended into a single score.

Graduate ranking

Ranks a programme

A separate number for each department in each field. A university's PhD programme in one field can sit in the top 15 while a different department at the same school does not crack the top 100. The two numbers do not move together.

Key numbers
April Best Graduate Schools publishes each spring. The 2026 edition came out April 6, 2026.
6 professional fields ranked yearly on outcome data plus surveys: business, education, law, engineering, medicine, nursing.
3-4 yrs refresh cycle for the peer-survey-only doctoral science fields: last done in 2022 or 2023 depending on the field, then again in 2026.
2022 / 2023 years law schools, then medical schools, began publicly withholding ranking data.

What Does a Grad School Ranking Actually Rank?

Undergraduate lists like Best Colleges score one thing: the university as a whole. Every applicant to that school, in every major, sees the same number. Graduate rankings do not work that way. A university does not have one graduate rank; it has one for its MBA program, a different one for its law school, a different one again for its PhD program in chemistry, and so on for every field it offers. Those numbers are computed independently, from different data, on different schedules, and they routinely disagree with each other and with the university's undergraduate reputation.

This is why a public university can sit outside the top 60 on a general "best national universities" list while one specific doctoral department at that same school ranks in the top 15 in its field, and why a university with a famous undergraduate brand can have a doctoral program in a specific discipline that is not ranked competitively at all. Neither number is wrong. They are answers to different questions: how strong is this institution overall, versus how strong is this one department, in this one field, according to a survey of other people in that field.

If you want the mechanics of how the undergraduate-level list is put together, including the metrics that get blended into that single institutional score, see how college rankings work. Everything below is specific to the graduate and professional layer, where the unit being measured shrinks from a whole university down to one program.

How Does the US News Best Graduate Schools Methodology Work?

Each spring, U.S. News invites schools in six disciplines, business, education, law, engineering, medicine and nursing, to complete a statistical survey. Programs that decline are not ranked in most of these fields. For the 2026 edition, that survey ran in fall 2025 and early 2026, and the rankings published on April 6, 2026. These six are the only graduate fields ranked from a mix of hard outcome data (job placement, salaries, bar passage, board pass rates) and reputation surveys of academics and, in some fields, working professionals.

The weights differ by field, and they are public. Two worked examples, from U.S. News's own 2026 methodology pages:

Ranking factor categoryMBALawMedicine (research)
Output / outcome measures (jobs, salary, bar passage, research activity)50%58%60%
Peer & professional reputation survey25%25%0%
Admissions selectivity (test scores, GPA, acceptance rate)25%10%20%
Faculty & library resources0%7%20%

For the MBA ranking, the 50% outcomes block breaks down into employment at graduation (7%), employment three months out (13%), starting salary and bonus (20%) and salary by profession (10%); the 25% reputation block splits evenly between a peer survey of business school deans and a survey of corporate recruiters. Law's 58% outcomes block is job placement 10 months after graduation (33%), first-time bar passage (18%) and ultimate bar passage within two years (7%); its 25% reputation block splits evenly between academic peers and practicing lawyers and judges. Medicine's research-track formula shows 0% for reputation because, as the next section explains, U.S. News removed that component entirely. None of this is a secret formula: every number above comes from U.S. News's own published methodology pages for the 2026 edition, linked in Sources below.

How Are PhD Program Rankings Built?

Outside those six professional fields, the picture changes completely. Doctoral programs in biological sciences, biostatistics, chemistry, computer science, earth sciences, mathematics, physics and statistics, plus the social sciences and humanities fields and most health, library and fine arts programs, are ranked solely from a peer assessment survey. There is no salary data, no placement data, no funding data and no resource data anywhere in these numbers. U.S. News sends a questionnaire to two people per program, typically the department head and the director of graduate studies, who rate other programs in their field on a 1-to-5 scale. Each school's score is a trimmed mean: the two highest and two lowest ratings are dropped, and the rest are averaged. Response rates for the 2026 science surveys ranged from 5.8% (biological sciences) to 49% (biostatistics).

These lists are also not refreshed every year. Biological sciences, statistics and biostatistics were last resurveyed for the 2022 edition and did not get new survey data again until fall 2025 and early 2026, a four-year gap. Chemistry, earth science, mathematics and physics were last resurveyed for the 2023 edition and went three years before their next survey, also landing in fall 2025 and early 2026. Computer science is the exception among the sciences: it was resurveyed every single year in between, including the years the other six fields were still running on older data. Social sciences and humanities fields (economics, psychology, English, history and similar) run on the same peer-assessment-only method but their own separate, non-annual schedule, so check the specific field's own methodology page for the year its current numbers were actually collected before treating them as current.

A PhD program rank, in other words, is a popularity poll among a small number of surveyed academics, not a measurement of anything happening to actual students.

What Did the Law School Ranking Boycott Change?

In November 2022, Yale Law School announced it would stop submitting data to U.S. News, calling the methodology "profoundly flawed." Harvard Law School followed within days, and dozens of other law schools joined over the following weeks. Medical schools made the same move starting in early 2023, led by Harvard Medical School. Neither ranking disappeared. U.S. News changed how it fills the gap instead:

  • Law: for the 2026 edition, nearly 200 accredited law schools were surveyed and 157 responded (up from 154 the prior year). For schools that do not submit the full statistical survey, U.S. News now uses the same figures those schools already report to their accreditor, the American Bar Association, and it cross-checks every school's self-reported numbers against ABA filings, emailing schools about discrepancies before publication.
  • Medicine: the response was more structural. U.S. News replaced numbered ranks (No. 1, No. 2...) with four evaluative percentile tiers: Tier 1 (85th to 99th percentile), Tier 2 (50th to 84th), Tier 3 (15th to 49th) and Tier 4 (1st to 14th), with schools inside a tier listed alphabetically rather than by score. It also dropped the peer and residency-director reputation survey from both the research and primary-care formulas entirely. As of the 2026 edition this is the third consecutive year with no reputation component in medical school scoring, which is why the table above shows 0% for it.

The practical upshot: current law rankings still use the same weighted categories as before the boycott, just with more ABA-sourced data filling the gaps left by non-participating schools. Current medical school output is not an ordinal ranking at all, it is a percentile band built entirely from outcome and resource statistics, with zero input from what other deans think of the program.

Are Grad School Rankings Reliable Predictors of Outcome?

For the six professional fields, the outcome data is real and worth reading. For every doctoral field ranking, the answer is weaker, and the reason is mechanical, not conspiratorial: a peer-assessment score measures a department's aggregate reputation among a handful of outside academics, and that reputation is generated by a department's most visible, usually most senior, faculty. It says nothing about the specific advisor who would actually supervise you, the specific lab's funding situation, or that lab's placement record.

Two students admitted to the same PhD program in the same year, working with two different advisors, can end up with completely different funding security, completely different publication opportunities and completely different placement outcomes after graduation, sometimes more different from each other than either is from a student at a different department several rank positions away. The ranking cannot see any of this because it was never built to look inside a department. It looks at the department from the outside, once, through a handful of surveyed colleagues.

How Do You Evaluate a PhD Program Without a Ranking?

Four things predict a good doctoral outcome better than a department's peer-assessment score, because they operate at the level where your actual experience happens: the advisor and the lab, not the department average. None of these appear in any ranking.

Signal 1

Advisor-specific placement, last decade

Where this specific faculty member's own former PhD students landed, not the department's aggregate placement page. A strong advisor in a middling-ranked department can out-place a weak advisor in a top-ranked one.

Source: the advisor's own CV or lab website (most list former students and their current positions), or ask directly during an admitted-students visit. See how to choose a thesis advisor for the full question list.

Signal 2

Guaranteed funding years and conditions

How many years of funding are contractually guaranteed in writing, not verbally implied, what happens if you have not finished by then, and whether the funding is fellowship, TA or RA money with different teaching-load implications.

Source: the funding offer letter itself. Record every offer's numbers in the funding offer comparison calculator so you compare guaranteed years, not headline stipend figures.

Signal 3

Time-to-degree distribution

The spread matters more than the average. Ask for the range across a lab's last five to ten graduates, not just the mean: a long tail usually signals underfunded years, an advisor who left mid-program, or unclear milestones.

Source: current fourth- and fifth-year students in that specific lab, asked directly. The department's own graduate handbook sometimes publishes cohort-level data if you ask for it.

Signal 4

Cohort attrition

What fraction of students who entered a given year are still enrolled or have finished five years later. Programs rarely volunteer this number, and a vague non-answer is itself informative.

Source: ask the graduate program coordinator directly, "how many students started in each of the last three cohorts, and how many are still here or graduated." Compare the answer across the programs on your list.

Which Grad School Rankings Are Worth Checking Anyway?

None of the above means ignore every list. A few narrow cases are still worth ten minutes:

  • Fields with rank-gated recruiting. Large law firm on-campus interviewing and federal judicial clerkship hiring both concentrate heavily on a narrow band of law schools, and elite consulting and banking recruiting works the same way for a fixed list of "core" MBA programs. If your target career runs through one of these pipelines, checking where a school sits in the professional ranking (not the doctoral one) tells you something real about your odds of getting screened in, independent of teaching quality.
  • Carnegie's research-activity designation. This is not a ranking at all, it is a formulaic, non-survey classification of an institution's research capacity: Research 1 status requires at least $50 million in annual research and development spending and at least 70 research doctorates awarded per year, averaged over recent years. It is a blunter signal than a department-level number, but it is current and it is not built from opinion. See the Carnegie classification methodology for the full tier definitions.

Both of these describe institution-level infrastructure or field-level hiring norms. Neither tells you anything about your specific advisor, funding or placement odds, which is why they belong alongside the four signals above, not instead of them.

Does the Ranking Question Differ for a Master's vs a PhD?

Yes, and the difference is about who is paying for what. A funded PhD is closer to a five-to-eight-year job: someone else is paying you to do research, and your outcome depends on your advisor, your lab and your funding letter, none of which a department's aggregate reputation score captures.

A self-funded professional master's is closer to a purchase. You are paying tuition for a credential that a future employer, who has likely never heard of your specific cohort, will read cold off a résumé. Here the outcome data behind the six annually surveyed professional fields is genuinely more relevant, because it is checkable: starting salaries, employment rates and bar or licensure passage are reported figures, not survey opinion. For a broader, federally sourced check on completion rates, debt and post-graduation earnings by field of study across many more institutions than any single ranking covers, the U.S. Department of Education's College Scorecard is a useful cross-check, listed in Sources below. If you are deciding between the two paths in the first place rather than comparing offers within one, see PhD vs master's for the fuller trade-off.

Testing policy is a separate question from ranking methodology, but it affects what you can compare across offers. GRE requirements vary by field and by individual program and have not been uniformly reinstated after most programs waived them; check each program's current requirement directly, and see do you need to take the GRE before assuming a rule that holds for one field holds for another.

Where to Go Next

A methodology explains what a number can and cannot tell you. It does not tell you which offer to accept. Once you have narrowed a shortlist using the signals above rather than a single ranking column, the next real decision is financial: compare the actual guaranteed money, years and conditions across your offers, not their department's peer-assessment score. Use the funding offer comparison calculator to lay stipend, tuition remission, guaranteed years and fees side by side, and the grad school application timeline to plan the months of due diligence, advisor emails and campus visits that a ranking cannot substitute for.

Frequently Asked Questions

Are grad school rankings reliable?

Only for what they actually measure. The six professional fields ranked annually with outcome and survey data (business, education, law, engineering, medicine and nursing) track real placement and salary figures you can check yourself. The doctoral field rankings in sciences, social sciences and humanities are built entirely from a reputation survey of two people per programme, refreshed every few years, and carry no outcome data at all. Treat the first group as evidence and the second as one input among several.

What is the US News Best Graduate Schools methodology in one sentence?

Six professional disciplines (business, education, law, engineering, medicine and nursing) combine a statistical survey of student and faculty outcomes with a peer and professional reputation survey, while doctoral programs in the sciences, social sciences and humanities, plus most health, library and fine arts fields, are ranked solely from a peer assessment survey administered on a rotating, non-annual schedule.

Why did law schools boycott the US News rankings?

Yale Law School announced in November 2022 that it would stop submitting survey data, calling the methodology profoundly flawed, and Harvard Law School followed within days; dozens more law schools joined over the following weeks. Medical schools made the same move starting in early 2023. US News kept publishing both rankings: for law schools that do not submit data it now uses figures reported to the American Bar Association instead, and for medical schools it replaced numbered ranks with four evaluative percentile tiers and dropped the reputation survey component entirely.

How often are PhD program rankings updated?

Not every year. Biological sciences, statistics and biostatistics were refreshed in 2026 for the first time since 2022, a four-year gap. Chemistry, earth science, mathematics and physics were refreshed in 2026 for the first time since 2023, a three-year gap. Computer science is the exception among the sciences and is updated annually. Social sciences and humanities fields run on their own rotating schedule, so check each field's own methodology page for its most recent survey year before assuming the number in front of you is current.

How do I evaluate a PhD program if I can't trust the ranking?

Look at four things a ranking cannot show you: the specific advisor's own placement record over roughly the last decade, the number of guaranteed funding years and what happens if you run past them, the spread (not just the average) of time-to-degree for that lab's recent graduates, and the fraction of a recent cohort still enrolled or graduated five years later. All four live in places a ranking never touches: faculty CVs, the funding offer letter itself, and direct questions to current students.

Does a grad school ranking matter more for a master's degree than a PhD?

Somewhat, yes. A self-funded professional master's is closer to a product you buy: the reported salary and employment data behind the annually ranked professional fields (business, education, law, engineering, medicine and nursing) is directly checkable and relevant. A funded PhD is closer to a five-to-eight-year job: the department's aggregate reputation score matters far less than your specific advisor's funding letter and placement record, which the ranking does not measure at all.

Sources