A good research question is only the start. Designing is where you turn that question into a practical, step-by-step plan — a methodology so clear that another physicist could pick it up and replicate your experiment exactly. The goal is a valid procedure that collects enough high-quality data to answer your question properly, with every variable, measurement, and safety point spelled out. This stage is where careful thinking earns real Internal Assessment marks.
📘 What you need to know
Clearly identify and justify your independent (IV), dependent (DV), and controlled (CV) variables
Controlling the CVs is what makes it a fair test
Justify the range of the IV — aim for at least five values to establish a clear trend
Justify the quantity of measurements — at least three trials per value, so you can take a mean
Repeating trials reduces random error and helps you spot anomalies
Write a replicable method: precise apparatus, a safety section, and numbered steps
A pilot study — a small trial run — checks the method works before you commit
Identify and justify your variables
Every design opens by listing and explaining its variables. There are three kinds, and being crisp about them is the backbone of a fair test.
Change one thing (IV), measure its effect (DV), and hold everything else constant (CVs) — that’s a fair test.
Independent variable (IV): the single factor you deliberately change to see its effect.
Dependent variable (DV): the quantity you measure, which responds to the change in the IV.
Controlled variables (CVs): all the other factors that could plausibly affect the outcome. You must explain how you’ll keep each one constant.
Justify the range and quantity of measurements
It isn’t enough to state your measurements — you have to justify them. That means explaining both how widely you spread your data and how many times you repeat it.
Range of the independent variable
Plan to collect data across a sensible range, with a minimum of five different values of the IV to reveal a clear trend. And say why that range: for a pendulum you might write that lengths from 0.20 m to 1.00 m were chosen because shorter lengths give a period too rapid to time accurately, while longer ones are impractical in a standard lab.
Quantity of measurements
Repeat the experiment at each value of the IV — a minimum of three trials is recommended. Repeating lets you calculate a mean, which reduces the effect of random error and helps you identify and discard anomalous results.
≥ 5 IV values
→ and →
≥ 3 trials each
→ gives →
reliable mean
A quick way to think about it: five values gives you the shape of the relationship (the trend), while three trials gives you confidence in each point (the reliability). You need both — lots of repeats at a single length tells you nothing about the trend, and single readings at many lengths can’t be trusted.
Design a valid, replicable methodology
The method is the detailed, step-by-step procedure. It must be a logical sequence clear enough for another physicist to replicate exactly, with precise apparatus details — “measure the length using a 1.0 m ruler with 1 mm divisions”, not “use a ruler”. Creativity counts too: using video analysis to track a falling object is more reliable than a stopwatch and the naked eye.
🧭 Structuring your methodology
Materials and apparatus — a full equipment list and, where useful, a clear labelled diagram of the set-up
Safety, ethical, and environmental — a brief risk assessment naming key hazards and specific precautions
Procedure — the numbered, step-by-step instructions
A labelled diagram often does more than a paragraph of text. For a specific-heat experiment, showing the insulated set-up makes your control of heat loss instantly clear.
A labelled set-up showing the insulating polystyrene cup and lid — the key to minimising heat loss in a thermal experiment.
Different approaches and pilot studies
Hands-on lab work is the most common route, but not the only one. Your investigation could draw on databases (e.g. NASA or CERN data for a property like star luminosity) or simulations (e.g. PhET, for processes too difficult or dangerous to run in a school lab).
Whatever the approach, a pilot study — a small-scale trial run — is invaluable. It checks your method actually works before you commit. A quick pilot of a resistance circuit, for instance, confirms the ammeter and voltmeter are wired correctly and that the power supply gives a suitable range of current and voltage.
Quick recap: name and justify your IV, DV and CVs; spread the IV over ≥ 5 values and repeat ≥ 3 times for a reliable mean; write a replicable method with apparatus, safety, and numbered steps; and run a pilot to check it works.
EXAMPLE 1
Designing an oscillation investigation
Research question
“What is the relationship between the length of a simple pendulum and its period of oscillation?”
Variables
Independent: the length L of the pendulum (m), from the suspension point to the centre of mass of the bob.
Dependent: the period T of oscillation (s), the time for one complete swing.
Controlled: the mass of the bob; the amplitude (kept below 10° for the small-angle approximation); air resistance (minimised, away from draughts).
Justifying the method
Time 20 complete oscillations and divide by 20 — this reduces the impact of human reaction-time error, giving a more precise period.
Place a fiducial marker at the equilibrium position; start and stop timing as the bob passes it, for a consistent reference point each swing.
EXAMPLE 2
Designing an electrical resistance investigation
Research question
“What is the relationship between the length of a constantan wire and its electrical resistance?”
Variables
Independent: the length L of the wire (m), between the voltmeter probes.
Dependent: the resistance R (Ω), calculated from R = V ÷ I using voltmeter and ammeter readings.
Controlled: the cross-sectional area (same piece of wire); the temperature (low current to avoid heating); the material (constant resistivity).
Justifying the method
Investigate lengths from 0.10 m to 0.80 m for a wide spread of data.
Measure the diameter with a micrometer at several points to confirm uniformity and calculate the cross-sectional area accurately.
Switch the supply on only briefly for each reading, to stop the wire heating and introducing a systematic error.
💡 Top tips
Detail is key. Instead of “measure the length”, write “measure from the bottom of the clamp to the centre of the bob using a metre ruler (±0.001 m)”
Safety is not an afterthought — be specific: “a crash mat under the bob in case it detaches”, not “be careful”
Justify every choice. Explaining why you picked an apparatus or range shows you’re thinking like a physicist
Pilot first. Ten minutes of trial run can save you from a whole flawed data set
⚠ Common mistakes
Writing a vague method that couldn’t be replicated — it won’t score well
Stating measurements without justifying the range or number of repeats
Treating safety as a token line rather than a real, specific risk assessment
Forgetting to control a variable that plausibly affects the result, breaking the fair test
Up next: Controlling Variables in Physics — we move from listing your controlled variables to the practical techniques that actually keep them constant: calibration, insulation, reducing friction and resistance, and accounting for background radiation.
Want this to actually click before the exam?
Book a free meeting and let’s work through the tricky bits together.
