Choosing a Research Programme: What to Ask First

Every research programme aimed at school students sells roughly the same two things: a publication, and a path to a major science fair. Both sit outside the provider's control. Here is what to ask instead.

Question one: show me a finished project

Not a testimonial, not a list of topics — an actual completed piece of work from a past student.

Then read its limitations section. If it has none, the programme is not teaching research; it is supervising a presentation. Every honest piece of work states what it does not establish, because every piece of work has a boundary.

Read the question too. If it is broad enough to be a conference theme — "applications of AI in medicine" — the project was never going to produce a finding. A real question is narrow enough that the answer could be no.

Question two: what happens if the result is negative?

This single question separates serious programmes from packaged ones faster than any other.

If the answer involves changing the hypothesis, adjusting the data, or "finding a different angle" so the write-up looks positive, walk away. That is the opposite of research, and it teaches a student something actively harmful.

The right answer is that a negative result is a result. A student who tests a plausible idea, finds it does not hold, and explains convincingly why has done genuine work. An inflated claim built on a weak experiment is the real failure — and anyone who knows the field spots it in a minute.

Question three: who is the mentor, and what do they do?

"Mentorship" covers an enormous range, from weekly working sessions to a monthly email.

Ask concretely. How often do they meet? Does the mentor read the code or the data, or only the summary? Who writes the first draft of the report?

That last one matters more than it sounds. If a mentor drafts and the student edits, the student cannot defend the work in conversation — and the conversation is what the project is ultimately for.

Question four: what does the student own at the end?

A good programme ends with artefacts in the student's hands: the data, the code or the apparatus notes, the report, and a version history showing how it developed.

A weaker one ends with a certificate and a slide deck.

The test is simple. Six months later, could the student reproduce their own result? If the answer is no, whatever happened was not research.

The truth about ISEF

Families ask about the International Science and Engineering Fair constantly, and the pathway is routinely implied rather than stated.

Students cannot enter ISEF directly. They must first compete at a Society-affiliated fair and earn a place there. There is no application route that bypasses this step.

The scale explains the competition: the 2026 fair in Phoenix drew more than 1,700 finalists from over 65 countries, across 22 categories.

So a programme can help identify affiliated fairs and prepare a submission. It cannot promise selection, and language suggesting otherwise is describing something it does not control.

On publication

Some programmes advertise that students publish a peer-reviewed paper as first author in named journals.

Those outcomes are self-reported promotional claims, and the venues vary enormously. Publication in a journal that accepts nearly everything submitted does not signal what families assume — and an admissions reader familiar with the field knows which venues are which.

What holds up under scrutiny is simpler and harder to fake: a project the student can discuss in detail, including its weaknesses, for ten minutes.

Narrow questions finish

The most common reason school research fails is scope, not ability.

A question narrow enough to answer in the available weeks produces something complete. An ambitious one produces an abandoned draft, and an abandoned draft is worth nothing to anybody.

Students resist narrowing because the small version sounds less impressive. Every reader who matters prefers it — and the narrow question is also the one where the student retains enough understanding to defend what they did.

When is a project the right use of the time?

Honestly, not always. The decision depends on where the rest of the application stands.

If a student is below the published score range for their target schools, raising the score is the higher-value work — a score below a range is a filter, and no project removes it. Our guide to what an SAT score does and where it stops sets out that threshold logic.

Once a student is comfortably inside that range, the calculation flips. Additional points add very little, while a defensible project is the only part of the application nobody else could have submitted.

Two warning signs worth naming

Beyond the four questions, two patterns recur often enough to flag.

A fixed topic list. If a programme assigns projects from a catalogue, several students are producing variations of the same work each cycle. That is a course with a project attached, which is a reasonable thing to buy — but it is not research, and an admissions reader who has seen the template before will recognise it.

A guaranteed deliverable on a fixed date. Research does not run to schedule, because you cannot promise a finding before you have one. A programme guaranteeing a completed paper by a set date is guaranteeing the document, not the result, and the difference shows in the writing.

What we run, and what we do not claim

Our programmes are built around a defensible project and the ability to discuss it. We make no promise of publication or admission, because neither is ours to give.

Students choose by where their curiosity sits rather than by prestige: the physics research programme for measurement and uncertainty, the informatics research programme for building and benchmarking a real-time system, or the mathematics research programme for proof-based work.

Whichever direction, the test at the end is the same one worth applying to any programme you are considering: can the student explain what they did, and say clearly what it does not prove?