A research mentor is usually a doctoral student or postdoc, not a professor. Free mentors exist through university programmes and cold outreach, but one Stanford lab that publishes its policy takes a high school intern only once every four to five years. Paid mentors cost $150–$200 an hour.
The word covers two very different arrangements: someone who agrees to supervise you as a favour, and someone paid several thousand dollars to do it. Both are called mentorship. They are not the same service and they do not carry the same risks.
What does a research mentor actually do?
Four things, and teaching you facts is not among them.
They stop you choosing an unanswerable question. This is the single most valuable function and the one students least expect. "The effects of social media on teenagers" cannot be researched in four months by anyone. "Whether daily screen-time self-reports in the 2023 Monitoring the Future dataset predict reported sleep duration, controlling for grade and sex" can be.
They tell you when your method will not support your claim. Most first drafts by capable students say more than the evidence permits, and nobody in a school environment is equipped to catch it.
They mark drafts as a journal would, which is more brutal and more useful than school feedback.
And they impose a deadline. A named person expecting a draft on Thursday is most of what separates students who finish from students who do not.
Who qualifies as a research mentor?
Almost always a doctoral candidate or a postdoctoral researcher. Very rarely a professor.
This is worth stating because the marketing in this sector obscures it. The 2023 ProPublica and Chronicle of Higher Education investigation by Daniel Golden and Kunal Purohit found that mentors at paid research programmes were commonly doctoral students paid $150 to $200 an hour — and that at least one company changed its website language from "PhD" to "PhD student" after reporters made contact.
The seniority is not the problem. A fourth-year PhD student in your field is running experiments this week, remembers what confused them three years ago, and has time. A full professor has none of those advantages, and mostly does not want the work.
The problem is only ever misrepresentation. If a programme's page says "Oxford academics" and means "Oxford doctoral candidates," those are different claims, and the second one is still a good offer.
Can you get a mentor without paying?
Yes, through three routes, all harder than paying.
A free university programme. MIT PRIMES pairs students with MIT graduate students and postdocs at no cost; its 2027 deadline is 2 November 2026. RSI places students with scientist mentors at MIT and Boston-area institutions and is free including housing and meals, admitting under 2.5% of roughly 3,100 annual applicants. Stanford SIMR is free. Most of these require US citizenship or permanent residency.
A local lab. Free, and the odds are long. The Goldhaber-Gordon lab at Stanford publishes its own policy, which is the most honest public account of this anyone has written: it receives "tens of such requests per year," takes a high school intern "only once every 4 to 5 years," and tells students plainly, "don't be surprised or offended if you don't get a reply."
There is also a barrier students almost never anticipate. UNC's policy on minors in laboratories prohibits anyone under eighteen from working with BSL-2 agents, radioactive materials or animals, requires constant direct supervision, and requires safety paperwork filed a month in advance. In many wet-lab fields the professor's willingness is irrelevant.
A teacher. Underused. Some high school science and social science teachers hold doctorates and have published. If yours does, that is a mentor you already have access to, and asking costs one conversation.
What does a paid mentor cost?
Between roughly $3,000 and $7,465 for a programme, of which the mentor receives a fraction.
| Programme | Published price | Contact time |
|---|---|---|
| Polygence | From about $3,000 | 10 sessions |
| Pioneer Academics GPSI | $3,100 | 12 weeks |
| Indigo Research | From $3,900 | 10–45 hours |
| Horizon Academic | $5,950 or $6,950 — its own site conflicts | ~100 hours |
| Pioneer Academics PRI | $7,465 | February–July |
| Lumiere Education | Not published | 9–30 sessions |
| Veritas AI | Not published | 8–16 weeks |
Do the arithmetic before you buy. If a ten-session programme costs $5,000 and the mentor is paid $175 an hour, around $1,750 reaches the person teaching you. The rest covers matching, administration, journal support and margin. That may be perfectly reasonable — but you are entitled to know the split, and a programme that will not discuss it is telling you something.
How do you approach a potential mentor?
Short, specific, and about their work rather than your ambitions.
Read one of their recent papers first. You do not need to follow all of it. You need one concrete thing — a result that surprised you, a choice you did not understand, a limitation they mention in the discussion. Name it in the second sentence.
Then say what you can actually offer: hours per week, months available, whether you can physically get to campus, what you already know how to do. "Ten hours a week from June, I can take the train to campus, I have completed AP Statistics and can use R" is a proposal. "I am passionate about neuroscience and would love any opportunity" is not.
The University of Arizona's guidance suggests contacting three to five faculty members rather than one. That is the right order of magnitude for a first round — enough to survive silence, few enough that each message can be genuinely specific.
Four paragraphs. Anyone running a lab is skimming.
What a mentor should never do
Design your study or write your paper.
This is the line the ProPublica investigation was most concerned with, and it found mentors who substantially directed student work. The test is simple and uncomfortable: if your finished paper is more sophisticated than you can defend in conversation, it is not yours. Admissions interviewers and journal reviewers both find this out by asking one follow-up question.
A good mentor makes you produce something slightly worse than they could have produced, and considerably better than you could have produced alone. That gap is the whole value.
How we do it
The Oxford Centre for Advanced Research matches students with PhD candidates and early-career researchers at Oxford, Cambridge, other leading UK universities and Ivy League institutions. We say PhD candidates rather than academics because that is what most of them are.
The Oxford Scholar Programme is £2,000 — roughly $2,600 — for ten contact hours across ten to fourteen weeks. The Oxford Publication Fellowship is £3,400, about $4,400, and supports submission to the Oxford Journal of Student Scholarship, the Journal of Research High School or Vanderbilt's Young Scientist Journal. All three reject student work, ours included.
We accept approximately 60% of applicants. Self-reported and unaudited, like every figure competitors publish.
If MIT PRIMES, RSI or SIMR are open to you, take them instead. They are free and they are better. The case for paying is narrow: you want a deadline and an expert reader, the free routes are closed, and the alternative is doing nothing for another year.
The one question worth asking
Ask a prospective mentor to describe a student project they supervised that did not work.
Anyone who has supervised more than a handful has several. The way they describe those failures — whether they blame the student, the question, or their own supervision — tells you more in ninety seconds than a biography will in a page.
Frequently asked questions
Does a research mentor need to be a professor?
No, and a professor is often the worse choice. Doctoral students and postdocs are closer to live method, more available, and better at explaining why a study design fails. Across paid programmes, mentors are predominantly PhD candidates — the 2023 ProPublica investigation found them commonly paid $150 to $200 an hour.
Can I get a research mentor for free?
Yes, through three routes: a free university programme such as MIT PRIMES or RSI, a local lab that agrees to take you, or a teacher with a research background. All three are harder to secure than paying, which is the entire business model of the paid sector.
What should a mentor actually do for me?
Ask questions you cannot answer yet, tell you when your method will not support your claim, mark drafts properly, and stop you choosing a question that cannot be answered in the time you have. What they should not do is design the study or write the paper — that boundary is the central ethical issue in paid research mentorship.
How many hours do I need with a mentor?
Ten to twelve contact hours is the standard entry format across paid programmes, spread over three to four months so there is time to read and write between sessions. More hours help mainly if the project involves a method you have to learn from scratch, such as regression analysis or lab technique.
How do I know if a mentor is any good?
Ask what they work on and whether they can name a student project they supervised that did not work out. A mentor who describes only successes has either supervised very few students or is selling. The useful ones talk readily about projects that collapsed and why.