Nature of Science

Physics is the study of matter, energy and the interactions between them, and its ambition is unusual among the sciences: to explain the widest possible range of phenomena with the smallest possible number of rules. The same equation that describes a cricket ball falling in Lahore describes the Moon circling the Earth.

It is built on measurement. A claim that cannot be tested against a measurement is not a physics claim, however reasonable it sounds. This is why so much of an early physics course is spent on units, instruments and errors — they are not preliminaries to the subject, they are the subject's foundation.

The branches you will meet are mechanics, heat, waves and sound, light, electricity and magnetism, and atomic and nuclear physics. They are divisions of convenience rather than of nature; a single problem often needs several of them at once.

Textbooks often present the scientific method as a tidy sequence: observe, hypothesise, predict, experiment, conclude. Real science is messier than that — experiments fail, results surprise people, and good ideas arrive out of order. But the sequence is still worth knowing, because it describes what a finished piece of work has to look like in order to convince anyone.

The step that does the real work is prediction. A hypothesis that merely explains what has already been seen is cheap; anyone can invent one afterwards. A hypothesis that says in advance what a new experiment will show is taking a risk, and it is that risk which makes the result worth something.

A hypothesis must therefore be falsifiable — there must be some possible result that would show it to be wrong. "Heavier objects fall faster" is a good scientific claim precisely because it can be tested and found false. A claim that no experiment could ever contradict tells you nothing.

An experiment is only informative if you change one thing at a time. Three kinds of variable appear in every practical you will do, and being able to name them is worth marks in itself.

The independent variable is the one you deliberately change. The dependent variable is the one you measure to see the effect. The control variables are everything else you must keep the same.

If a control variable is allowed to drift, the experiment cannot tell you anything, because two things changed at once and you cannot say which caused the result. This is the single most common reason a school experiment fails to show what it should.

Consider testing how the length of a pendulum affects its period. Length is independent, period is dependent, and the mass of the bob, the size of the swing and the place you do it in are controls. Change the bob halfway through and the data is worthless.

No measurement is exact. Every reading carries an uncertainty, and a scientist's job is to know how large it is rather than to pretend it is zero.

Random errors scatter readings either side of the true value — a hand-operated stopwatch, a slightly different eye position each time. They are reduced by repeating the measurement and taking a mean, and they show up as scatter in a graph.

Systematic errors shift every reading the same way — a balance that reads 5 g high, a ruler with a worn end, a zero error on an ammeter. Repeating does not help at all, because every repeat is wrong by the same amount. They show up as a graph line that has the right gradient but does not pass through the origin.

An anomaly is a single reading that sits well away from the pattern of the others. It should be identified, and then repeated if possible — not silently deleted. An anomaly that survives repetition is data, and occasionally it is the most interesting data you have.

The two sliders separate the two ideas that beginners merge. Systematic error moves the whole cluster off the centre — the grouping can be tight and every reading still wrong. Random error spreads the cluster, so the average can be right even though no single shot is.

Physics works by building models — deliberately simplified pictures that capture what matters and ignore what does not. Treating a planet as a point mass, or air resistance as zero, is not carelessness. It is a decision that those details do not change the answer to the question being asked.

Every model has a range within which it works and outside which it fails. Newton's laws describe motion superbly at ordinary speeds and break down near the speed of light. That does not make them wrong; it makes them a model with known limits, which is the most any model ever is.

A theory in science is not a guess, despite the everyday use of the word. It is an explanation that has survived repeated attempts to disprove it and now organises a large body of evidence. But no amount of confirmation makes a theory final. A single reproducible result that contradicts it is enough to force a revision, and that willingness to be overturned is what separates science from other ways of holding beliefs.