Rank research programmes by what you hold at the end, not by the name attached. On that test the field splits four ways: free selective placements, structured mentorship that produces a paper, qualification-linked projects you are already doing, and paid courses that produce a transcript line. The fifteen below are sorted by output, with every figure taken from the programme's own pages.
Most "best research programme" lists rank by prestige. That tells you nothing useful, because the prestigious names are mostly closed to you and the available ones are mostly unranked.
A better test: what do you hold when it ends?
On that test the field splits four ways, and the fifteen below are sorted accordingly.
| Type | What you leave with | Cost | Selection signal |
|---|---|---|---|
| Free selective placement | Original work, a named supervisor | £0 | Strong |
| Structured mentorship | A finished paper | £1,500–£5,000 | Moderate |
| Qualification-linked project | A marked, examined paper | £0 | None, but universal |
| Paid course | A transcript line | £3,000–£13,000 | None |
Free and selective (1–6)
1. Research Science Institute. The benchmark, and the only one that publishes its own selectivity. CEE's application FAQ states over 2,500 US and 600 international applications a year with fewer than 2.5% accepted. Free, including room and board. From 2027 it runs at Stanford as well as MIT — 100 places at MIT and 50 at Stanford. Juniors only; seniors are ineligible. Our full guide to RSI covers the application.
2. Summer Science Program. The strongest option genuinely open to international students. Around 700 participants admitted in 2026, need-blind admission, and $3,000 stipends for selected participants who would otherwise need to work. Four disciplines: astrophysics, bacterial genomics, biochemistry, cell biology. If you are choosing between this and RSI, see our RSI versus SSP comparison.
3. Simons Summer Research Program. No programme fees or tuition at Stony Brook. The unusual requirement is nomination by a school official, with a maximum of three students per school — so there is a selection round before the external one. US citizens or permanent residents only; visa holders without a green card are explicitly ineligible.
4. MITES Summer. Six weeks at MIT, free of charge with room and board covered. Juniors, US citizens or permanent residents.
5. Anson L. Clark Scholars Program. Twelve places at Texas Tech, seven weeks, a $750 stipend on completion of a research report, no participation fee. Must be 17 by the start date. US only, "NO EXCEPTIONS".
6. SIMR. Stanford's medical research programme, about eight weeks. US citizens or permanent residents living in the US, 16 by the start date, with preference for Bay Area students. Stipends are needs-based, not universal.
Free and UK-only (7–11)
Under-applied relative to how good they are, largely because they are not marketed.
7. Institute for Research in Schools. The one most lists miss entirely. IRIS runs research projects free for UK schools, and its 2025 impact report records 1,809 students across 93 schools completing 257 research projects in 2024/25, with around 10,000 students and 1,200 projects since 2016. Ninety-two per cent of participants were at state-funded schools. Nine active projects, from ATLAS big data to earth observation.
One honest caveat: IRIS output is overwhelmingly posters and conference presentations — 243 posters and 690 students at conferences in Exeter, London and Manchester in 2024/25 — and the report records no journal publication count. If your aim is a published paper, treat IRIS as the research experience and publication as a separate step.
8. Nuffield Research Placements, 9. Sutton Trust UK Summer Schools, 10. UNIQ, 11. In2STEM. All free, all UK-only, all with eligibility criteria that are the part students get wrong. We have separate guides to Nuffield, Sutton Trust, UNIQ and In2STEM.
Qualification-linked (12–13)
The two most overlooked research programmes are ones you may already be enrolled in.
12. The EPQ or IB extended essay. A supervised, examined, several-thousand-word research project with a teacher assigned to you at no cost. Students treat it as a hoop and then pay for a mentorship programme to do the same thing. Our guides to the EPQ and the extended essay cover using it properly, and turning an EPQ into a published paper covers the conversion.
13. AP Research. A 4,000–5,000 word paper with original data, externally marked, with a teacher supervising across eight months. Our AP Research rubric breakdown covers what scores.
Neither is glamorous. Both produce exactly the artefact a paid programme sells.
Paid, with a research output (14)
14. University-run paid placements. BU RISE, the Garcia Center at Stony Brook and UT Austin's High School Research Academy charge between $4,000 and $10,735 and put you on an actual project. Pricing and eligibility are in our pre-college programmes comparison — note that UT Austin is Texas-only and Garcia's entry bar is a 95/100 unweighted GPA.
Structured mentorship programmes sit here too. The field has grown fast and quality varies, so apply the four questions below rather than the brand. Our reviews of Polygence, Pioneer Academics, RISE Education and Immerse Education go through them individually, and free versus paid covers the decision.
Free, and nobody organises it (15)
15. A placement you arrange yourself. Email faculty at a university near you.
Two constraints to price in. Reply rates are low — the University of Arizona's careers service recommends approaching three to five faculty members. And the honest arithmetic from one Stanford lab's public page on high school placements: it receives "tens of such requests per year" and takes an intern roughly once every four to five years. That is not hostility; it is capacity.
Computational, archival and data-based groups face neither that constraint nor the minors-in-laboratories rules that stop many wet labs hosting under-18s. Start there. Our guide to emailing a professor covers the message.
The four questions
Whatever the programme, these four separate the real ones:
Who reads your drafts, and how many times? One read is a courtesy; three is supervision. Written feedback on a full draft is the intervention that improves student work most, and the one that costs a provider real money — which is why it is the honest test.
Was anyone turned away? If not, the programme is a purchase, not a selection. That is fine, as long as you are not expecting it to function as a credential.
What do I hold at the end? A paper, a dataset, a presentation, a transcript line — or a certificate. Ask before you pay.
Whose name goes on the output? Supervision does not by itself earn authorship. The ICMJE's criteria state plainly that general supervision of a research group does not qualify someone for authorship on its own. A programme that expects its mentor's name on your paper is telling you something about what it thinks it sold you. Our guide to student research paper authorship covers how to handle it.
The ranking, honestly
If you are eligible for the free selective tier, apply there and apply early — the deadlines fall six to eight months ahead and that is where most students lose, not at the merit stage.
If you are not, the best value in the whole field is the qualification you are already taking, supervised properly, finished well, and then revised for publication. It costs nothing, it is externally marked, and it produces the same artefact.
Paid mentorship is worth money when it buys sustained written feedback on your drafts. It is worth nothing when it buys a logo. Our guide to which programme fits which student works through the cases.
Frequently asked questions
What is the best research programme for a high school student?
The one that produces an output you can show and that you are actually eligible for. Those two filters remove most of the field before quality matters — the most selective US programmes exclude international students and rising seniors, and the UK's free schemes are means-tested. Work out what you can apply to, then ask what each one leaves you holding.
Are paid research programmes worth it?
They can be, when what you are buying is a supervisor who reads your drafts and a deadline structure that gets a paper finished. They are not worth it when what you are buying is a brand name, because open-enrolment programmes carry no selection signal. The test is whether anyone was turned away.
What is the most selective research programme?
The Research Science Institute publishes its own figures: over 2,500 US and 600 international applications a year with fewer than 2.5% accepted. That is unusual transparency — most selective programmes publish no acceptance rate at all, so any percentage you see for them is an estimate circulated between content sites.
Can international students apply to the top research programmes?
To some. RSI accepts international applicants from participating countries, and the Summer Science Program is open to them with need-blind admission. MITES, Simons, SIMR and Clark Scholars all require US citizenship or permanent residency with no exceptions. Check each programme's own eligibility page rather than a list.
Do UK students have good free options?
Yes, and they are under-applied. Nuffield Research Placements, Sutton Trust summer schools, UNIQ and In2STEM are free and UK-only, and the Institute for Research in Schools runs research projects free for UK schools — 1,809 students across 93 schools in 2024/25. The catch is that most UK schemes are means-tested or widening-participation targeted.
How early should I apply?
Six to eight months before the summer concerned. RSI's 2027 applications closed on 11 December 2026. MITES and Simons deadlines fall in early February. SIMR's was 21 February. Students who start looking in spring have already missed the selective tier for that year.