Planning an Experiment

A good experiment answers one question cleanly. Most of that work happens before any apparatus is touched.

Variable — any quantity that could change during an experiment.

Independent variable — the one you deliberately change.

Dependent variable — the one you measure in response.

Control variables — everything else, held constant. These are what make the conclusion valid: if a second quantity changed at the same time as your independent variable, you cannot say which one caused the effect you saw.

Example: investigating how the current through a filament lamp depends on the potential difference across it. You change the p.d., you measure the current, and you keep the same lamp, the same leads and — as far as possible — the same temperature. Swapping the lamp halfway through would ruin the data, because any change might be the new lamp rather than the p.d.

Choosing a range and interval — take readings across as wide a range as the apparatus safely allows, with enough points to show the shape of the relationship. Six to ten well-spaced readings usually beat twenty crowded into one corner, because a trend only reveals itself over a range.

Pilot runs — set the apparatus up and try a few readings before committing. A pilot tells you whether your range is sensible, whether the effect is even measurable, and whether anything is about to melt.

Choosing instruments by resolution — pick the instrument whose resolution is small compared with the quantity you are measuring. A millimetre ruler is hopeless for the diameter of a wire, because the reading and its uncertainty are the same order of magnitude; a micrometer resolves to 0.01 mm and makes the measurement meaningful.

Example: timing a trolley over 20 cm by hand is dominated by reaction time — perhaps 0.2 s of uncertainty on a 0.4 s measurement. Light gates remove the human entirely and cut the uncertainty by orders of magnitude. Choose the instrument to fit the size of the quantity, not out of habit.

Risk assessment — identify what could cause harm, how likely it is, and what you will do about it. Lasers need goggles and a controlled beam path; heavy masses need a clear space beneath them; hot apparatus needs time and tongs.

Writing the method — the test of a method is whether a competent stranger could follow it and get your results. That means stating what you changed, what you measured, how you measured it, what you held constant, and how many repeats you took.