Physics Research Programme with a Real Project

A physics research programme for school students should produce one thing: a project the student can defend. Not a demonstration of something already known, and not a title that sounds impressive attached to work nobody can explain.

Uncertainty is the whole subject

If there is one thing that separates a physics project from a science demonstration, it is this: a result quoted without its uncertainty is not a physics result.

Students arrive expecting the hard part to be the apparatus or the theory. It is almost always the error analysis — systematic against random error, propagating uncertainty through a calculation, and knowing how many significant figures are actually justified.

This is where the programme spends disproportionate time, because it is also the part that makes everything else honest. A student who can say "the difference we measured is within our uncertainty, so we cannot claim an effect" has learned more physics than one who reports a confident number they cannot justify.

Three tracks, no laboratory required

A frequent worry is equipment. It need not be a barrier.

Experimental projects use accessible equipment — a phone's sensors, simple optics, pendulums and circuits all support genuine measurement with real error analysis.

Computational projects simulate a system and compare against theory. The discipline here is validation: run the model on a case with a known answer before trusting it on the case you care about.

Data analysis projects use the enormous volume of genuinely public physics data — astronomical surveys, particle physics releases, climate records.

None of these is a lesser option. The choice depends on what a student can actually access, and a well-executed computational project beats a poorly-executed experimental one every time.

Narrow beats ambitious

The most common failure in school research is scope. "Investigating quantum computing" is a topic, not a question, and no amount of work converts it into a finding.

"Does the period of this pendulum depend on amplitude at angles beyond the small-angle approximation, and does our measured deviation match theory within our uncertainty?" is a question. It can be answered in weeks, the answer means something, and — crucially — the answer could be no. That is what makes it research.

Narrowing takes the first two weeks and students often resist it, because the narrow version sounds less impressive. Anyone evaluating the work will think the opposite.

Not fooling yourself

Richard Feynman's line is the operating principle: the first principle is that you must not fool yourself, and you are the easiest person to fool.

Concretely, this means writing down what you expect before collecting data, recording the runs that did not work rather than quietly discarding them, and resisting the pull to stop collecting as soon as the data looks agreeable.

These habits are taught explicitly, because the incentive to produce a tidy result is strong and nobody is immune to it.

Negative results are results

A student who tests a plausible hypothesis, finds no effect, and explains convincingly why has done real research.

The failure mode is the opposite: an inflated claim resting on a weak measurement. Anyone who knows physics spots it in a minute, and for a university application that is considerably worse than an honest modest finding.

Saying this at the start changes how students work. It removes the incentive to force a positive result out of data that does not contain one.

Science fairs and the ISEF route

Families ask about ISEF, and the pathway deserves stating plainly: students cannot enter directly. They must first compete at a Society-affiliated fair and earn a place there.

We help identify affiliated fairs and prepare a submission, but no programme can promise selection, and you should be wary of any that implies otherwise. The 2026 fair drew more than 1,700 finalists from over 65 countries — the competition is genuine.

What a school project is, and is not

It is not published science and should never be presented as such. What it demonstrates is that a student can work the way a researcher works: narrow a question, measure carefully, quantify uncertainty, and state honestly what the result does not show.

For a university application that is a strong and truthful claim — and an applicant who can discuss their own project's weaknesses for ten minutes is immediately distinguishable from one who has memorised a description.

Practical details

About 20 weeks at five hours a week, one-to-one, from Class 10 upward. Physics and mathematics to Class 10 is enough; Python is taught where a computational track needs it.

Students interested in the same method applied elsewhere can look at our computer science research intensive or the mathematics research programme.