Practical Skills: Experiments, Apparatus and Graphs
Physics has a separate practical paper, and it rewards a distinct set of skills. This page covers what the mark schemes actually look for — planning, measuring, graph work and error analysis.
1. Planning an experiment
Plan answers are marked against specific points, so structure matters more than length.
Include, in this order:
- The variables — independent, dependent and control
- Apparatus — named specifically
- Method — numbered, repeatable steps
- Measurements — what you record and with what
- How you’ll process the results — a table, a graph, a calculation
- A conclusion — what the result would show
Marks are distributed across specific points in the plan. Tutors flagged this, and also the corollary: read the bullet points in the question and cover each one. The bullets are the mark scheme.
Don’t pad your answer. Writing about safety precautions when they weren’t asked for was flagged directly — irrelevant material earns nothing and wastes time.
The three variables:
Independent — what you change Dependent — what you measure Control — what you keep the same so the test is fair
The independent variable cannot also be a control variable. Listing the thing you’re changing as something to keep constant was a recorded error.
Only list control variables that genuinely matter. Padding the list with irrelevant quantities was recorded — name the ones that would actually affect the result.
2. Reading instruments
| Instrument | Read to |
|---|---|
| Ruler | nearest mm |
| Measuring cylinder | nearest scale division, at the bottom of the meniscus |
| Ammeter / voltmeter | check the scale divisions first |
| Thermometer | nearest division, eye level |
| Stopwatch | as displayed |
Work out what one division is worth before reading any scale. Miscalculating ammeter divisions and struggling with a measuring cylinder scale were both recorded errors. Count the divisions between two labelled values and divide.
Read at eye level to avoid parallax error.
Estimate sensibly. A recorded error underestimated a beaker’s volume badly — sanity-check readings against something familiar.
Include units on every recorded value, in the table heading and in your answers. Omitting units was recorded and is an easy loss.
3. Tables
- Headings with units, written as quantity / unit — e.g. length / cm
- Independent variable in the first column
- Consistent significant figures down each column
- Include columns for repeats and a mean
Use a suitable number of significant figures — match the precision of your instrument, and keep it consistent.
On the practical paper you generally don’t need to show working for simple divisions — but do show it for anything multi-step.
4. Graphs
Graph work carries a lot of marks and each element is marked separately.
Axes: independent variable on the x-axis, dependent on the y-axis, both labelled with units. Scale: must be sensible (1, 2, 5 or 10 per square — never 3 or 7) and should make the points fill at least half the grid. Points: small, neat crosses or dots in circles. Line: a best-fit line or smooth curve — not dot-to-dot.
Choosing a scale that fits is itself a mark. Struggling to fit the graph scale was a recorded difficulty — check your largest value against the number of squares before you start plotting.
Draw a best-fit line, not a zigzag through every point. Roughly equal numbers of points should lie either side.
Finding a gradient:
- Draw a large triangle on the best-fit line, using as much of it as possible
- Read the coordinates of two widely separated points on the LINE
- gradient = Δy / Δx
- Include the units — the gradient usually has a physical meaning
Take gradient readings from the LINE, not from plotted data points. And use a large triangle — a small one magnifies reading errors. Incorrect gradient calculation was a recorded error.
Not every experiment needs a graph. A recorded error drew one where the question didn’t call for it — check what the question actually asks for.
5. Accuracy, precision and reliability
| Term | Meaning |
|---|---|
| Accurate | close to the true value |
| Precise | repeat readings close to each other |
| Reliable | consistent when repeated |
| Anomaly | a result that doesn’t fit the pattern |
Standard improvements — these earn marks again and again:
Repeat and take a mean — the most-cited improvement in every topic. Measure a large quantity and divide — 20 oscillations, 100 sheets of paper. Use an instrument with smaller divisions. Use light gates or an oscilloscope instead of human timing, to remove reaction time. Maximise the distance measured to reduce percentage error. Insulate and use a lid in thermal experiments. Read at eye level to avoid parallax.
6. Sources of error
Be SPECIFIC. “Human error” earns nothing.
Tutors flagged this directly. Name the actual mechanism:
| Instead of… | Write |
|---|---|
| ”human error” | reaction time when starting the stopwatch |
| ”the reading was wrong” | parallax error in reading the scale |
| ”heat was lost” | thermal energy lost to the surroundings |
| ”the equipment was bad” | zero error on the ammeter |
Types:
Systematic errors shift every reading the same way — e.g. a zero error. Fixed by checking and subtracting the zero reading. Random errors scatter readings — reduced by repeating and averaging.
Match the error to the experiment. A recorded error suggested “parallax” where the real difficulty was lining up the pins in a ray-tracing experiment. Think about what actually limited this measurement.
7. Experiment-specific notes
Ray tracing (light): use sharp pencil lines, place pins far apart for accuracy, view pins at eye level, and mark the outline of the block before removing it.
Draw ray diagrams accurately to locate the image — and check rays are drawn on the correct side of the lens or block. Connecting rays wrongly on the far side of a lens was recorded.
Springs (Hooke’s law): clamp the ruler vertically beside the spring, measure the original length first, and remove loads to check the spring returns.
Thermal experiments: stir liquids, insulate the container, and allow the thermometer to settle.
Electrical experiments: ammeter in series, voltmeter in parallel, and check for a zero error before starting.
Density: ensure the object is fully submerged, and dry it before weighing.
Label everything on a diagram, including any height or distance you measured. Omitting the height from a diagram was a recorded error.
8. Mistakes that cost marks
Listing the independent variable as a control.
Padding with irrelevant control variables or precautions.
Not covering every bullet point in the question.
Omitting units in tables or answers.
Awkward scales (3s or 7s per square), or a graph filling less than half the grid.
Joining points dot-to-dot.
Taking gradient readings from data points, or using a small triangle.
Drawing a graph when none was asked for.
Writing “human error” as a source of error.
Suggesting an error that doesn’t apply to that experiment.
Inconsistent significant figures.
Frequently asked questions
What are the three types of variable? Independent (changed), dependent (measured) and control (kept the same).
How should I label graph axes? Quantity / unit, with the independent variable on the x-axis.
What makes a good scale? Steps of 1, 2, 5 or 10 per square, filling at least half the grid.
How do I find a gradient? Use a large triangle on the best-fit line: Δy / Δx, with units.
What’s the difference between accuracy and precision? Accurate = close to the true value. Precise = repeat readings close together.
How do I improve reliability? Repeat and take a mean.
How do I reduce the effect of reaction time? Time many oscillations and divide, or use light gates.
What is a systematic error? One that shifts every reading the same way, such as a zero error.
Why is “human error” not accepted? It is too vague — name the specific mechanism, such as reaction time or parallax.
Do I always need a graph? No — only when the question asks for one.
Quick revision checklist
- I structure plans around the question’s bullet points
- I state independent, dependent and control variables correctly
- I don’t list the independent variable as a control
- I name apparatus specifically and give numbered steps
- I include a conclusion
- I work out scale divisions before reading an instrument
- I read at eye level and at the bottom of the meniscus
- I head table columns with quantity / unit
- I choose sensible scales filling the grid
- I plot neatly and draw a best-fit line
- I take gradients from the line with a large triangle, with units
- I can name specific sources of error
- I can distinguish systematic from random errors
- I know the standard improvements: repeat, average, measure multiples, better instrument
- I include units everywhere
This page covers practical skills for the Cambridge IGCSE Physics (0625) practical paper and is written for Grade 9–11 / Year 10–11 students. It is based on teaching patterns observed across a large set of one-to-one IGCSE Physics lessons, with particular attention to the errors students make most often and the wording examiners reward. Always check the current syllabus and the practical paper requirements for your own exam series.
Finished this topic?
Saved on this device — no account needed.