Ask for one bounded thing, not a placement. A Stanford lab that publishes its policy receives tens of requests a year and takes a high school intern once every four to five years — so a request that can be answered in ten minutes gets answered far more often than one that would cost someone a summer. Send five, personalised, and expect four to go unanswered.
The published numbers are worse than students assume, and knowing them changes what you should write.
What the reply rate actually is
One lab publishes its policy, and it is the most useful document in this whole area.
The Goldhaber-Gordon lab at Stanford states that it receives "tens of such requests per year," takes a high school intern "only once every 4 to 5 years," and tells students "don't be surprised or offended if you don't get a reply."
Tens of requests annually. One place every four to five years. That is roughly a one-in-a-hundred-and-fifty proposition, and there is no reason to think that lab is unusually selective — it is unusually honest.
So if you have sent thirty emails and heard nothing, you have observed the base rate rather than a verdict on your ability.
Why the answer is usually no before anyone reads your email
Two reasons, and only the second has anything to do with you.
Policy. UNC's laboratory safety policy bars under-eighteens from BSL-2 agents, radioactive materials and animal work, requires constant direct supervision rather than the ordinary autonomy a lab runs on, and requires safety paperwork filed roughly a month ahead. Most research universities have comparable rules. A professor who would genuinely like to take you frequently cannot.
Cost. Even where it is permitted, supervising a sixteen-year-old is weeks of someone's attention against an uncertain return, usually from a postdoc already behind on their own work.
The strategic consequence: ask for something that does not trigger either problem. A question about a paper costs ten minutes and no paperwork. A summer placement costs a term and a liability review.
The four emails that do not work
"I am passionate about your field." Every email says this. It carries no information.
"Can I work in your lab this summer?" You are asking for the expensive thing first, and for most labs the answer is set by policy before your name is read.
The obvious template. Academics spot a mail-merge instantly. A paragraph of generic admiration followed by a request reads as one of thirty, and is treated as one of thirty.
The five-paragraph life story. Your email is being read on a phone between meetings. Anything past the first screen is not read.
The email that works
Short, specific, and answerable in ten minutes.
Subject: Question about your 2025 paper on [specific thing]
Dear Professor [Name],
I am a Year 12 student at [school] working on an independent research project on [one clause]. I read your paper on [specific title] and was particularly interested in [one genuine, specific detail — a method choice, a limitation they named, a result that surprised you].
I am using [dataset] to look at whether [variable A] predicts [variable B], controlling for [variable C]. I have a draft methods section of about one page.
Would you be willing to read it and tell me whether the design is sound? I would be grateful for ten minutes of your time, and I am not asking for a placement or supervision.
Either way, thank you for the paper — it was useful.
[Name]
[School]
Why each part does work:
| Element | What it does |
|---|---|
| Specific subject line | Proves you read something before it is opened |
| One genuine detail from the paper | The only part that cannot be mass-produced |
| Your actual question, with variables named | Shows there is a real project, not a wish |
| "A draft methods section of about one page" | Bounds the ask to something readable |
| "I am not asking for a placement" | Removes the expensive request they were bracing for |
| A thank-you that does not require a reply | Makes silence costless for both of you |
The hard part is the detail from their paper. There is no shortcut: read the abstract and introduction properly, find the thing that genuinely puzzled or interested you, and name it.
Who to email
Graduate students and postdocs first. They say yes far more often than professors, they read methods sections more carefully, and they remember what confused them at sixteen. They are usually listed on a lab's people page.
Academics whose paper you have actually read. Not the famous one in the field — the one whose specific work overlaps your question.
Someone at a university near you, if the field permits. Proximity removes an objection even when you are not asking to visit.
The University of Arizona suggests approaching three to five faculty rather than one. Five personalised emails will outperform thirty copied ones by a wide margin.
What to do when they do not reply
Nothing, mostly. One follow-up after two weeks, one short paragraph, and then stop.
Silence is not a judgement — the Stanford lab says so explicitly. Chasing further costs you goodwill in a small field.
What matters more is that the emails are not your project. If five go out and nothing comes back, you still need a question and a dataset, and a week inside one public dataset's documentation reliably produces both. Our guide to research without a lab covers the fields where a laptop is the whole apparatus, and how to secure a research opportunity covers the structured routes.
If they say yes to something smaller
Take it, immediately, and make it easy.
Send exactly what you promised — one page, not six. Reply within a day. Ask one follow-up question, not five. A person who gives you ten minutes and finds it painless often gives you another twenty later; a person who gives you ten minutes and receives a dissertation does not.
This is the same conversation a paid mentor would have in the first three sessions, and getting it free from someone who agreed to one bounded read is the best outcome available.
Where we fit
The Oxford Centre for Advanced Research pairs students with PhD candidates and early-career researchers at Oxford, Cambridge, other leading UK universities and Ivy League institutions. The Oxford Scholar Programme is £2,000 — about $2,600 — for ten contact hours over ten to fourteen weeks.
We exist largely because the email route has the reply rate described above. That is an honest account of our market rather than a pitch: if a teacher with a doctorate or a graduate student will read your drafts, you do not need to pay anyone, and you should try that first.
This week
Pick five people whose work you have actually read. Write five different emails. Ask each for one bounded thing.
Then go and open a dataset, because four of the five will not reply and the project still has to happen.
Frequently asked questions
What should I actually ask for?
One specific, bounded thing. 'Would you read my methods section and tell me whether the design is sound?' is answerable in ten minutes and gets answered. 'Can I work in your lab this summer?' asks for a term of supervision, safety paperwork and liability, and is usually declined by policy rather than by choice.
How many should I send?
The University of Arizona suggests approaching three to five faculty rather than one. Five personalised emails beats thirty copied ones — academics recognise a template instantly, and a visibly mass-sent request signals that you have not read their work.
Why do I never get a reply?
Usually volume and policy rather than judgement. The Goldhaber-Gordon lab at Stanford receives 'tens of such requests per year', takes a high school intern 'only once every 4 to 5 years', and tells students not to be surprised or offended by silence. Separately, most universities bar under-eighteens from the lab work students are asking to do.
Should I attach my CV?
Not on the first email. A CV signals you are applying for a position, which raises the question of whether they have one. Attach the thing that is actually relevant — a one-page summary of your project, or nothing at all, with an offer to send more if useful.
Is it worth emailing at all?
Yes, but budget it honestly. Five emails is an afternoon and the expected return is one useful reply. Spending a week in a public dataset's documentation reliably produces a research question. Do both, in that order, and do not let the emails become the project.
What if they say no but offer something smaller?
Take it immediately. A fifteen-minute call, a reading list, or an introduction to a PhD student is worth more than a vague yes to a placement that never materialises. Graduate students in particular say yes far more often than professors, and read methods sections more carefully.