How to Answer Planning, Analysis and Evaluation Questions in A Level Physics (9702) Paper 5

Cambridge International AS & A Level Physics (9702) Paper 5 is a practical-skills paper completed entirely in writing. Under the current 2025-2027 syllabus, it lasts 1 hour 15 minutes, is worth 30 marks and contributes 11.5% of the full A Level. Candidates answer two compulsory questions. Question 1 is a 15-mark planning task; Question 2 is a 15-mark analysis, conclusion and evaluation task. Cambridge also warns that the physical context can sit outside the taught syllabus because the assessment is testing practical reasoning rather than recall of one memorised experiment.

Treat Question 1 as an experiment you could actually perform

The most useful planning habit is to imagine the apparatus in front of you. Identify the independent variable, the dependent variable and the quantities that must be kept constant. Then ask how each variable will physically be changed or measured. Cambridge's June 2024 examiner report says candidates should explain experimental procedures in detail, including how variables are controlled, how measurements are taken and how the data will be analysed. Avoid phrases such as 'measure the temperature' or 'control the length' when the method matters. State the instrument, where it is placed, which quantity is measured and what is held constant. In the 2024 report, candidates lost credit for naming a ruler for a cylinder diameter where a more appropriate measuring instrument was needed, and for describing a temperature sensor in a position that could not actually measure the required surface temperature.

Click to enlarge

Define the variables before choosing the apparatus

A plan becomes confused when apparatus is chosen before the variables are clear. Start with what changes and what responds. If the question investigates how wire length affects resistance, length is the independent variable and resistance, or a quantity used to calculate it, is the dependent variable. If temperature must stay constant because resistance also depends on temperature, say how heating will be limited or monitored. Use the quantity symbols and names given in the question. Cambridge's examiner report specifically notes that referring to the actual quantity, rather than writing only a vague word such as 'time', makes the procedure clearer and more likely to earn credit.

Control variables by describing the control method

A controlled variable is not fully dealt with when you merely name it. State how it will be kept constant. If the mass must be the same, use the same object or measure and adjust it. If temperature must be fixed, use a controlled bath or wait until thermal equilibrium is reached. If geometry matters, state which dimensions are unchanged. Choose controls that can realistically affect the measured relationship. A long list of irrelevant constants is less useful than two or three variables controlled properly.

Draw a labelled diagram when it makes the setup clearer

Cambridge's 2024 examiner report says the strongest planning responses often included a workable labelled diagram. The diagram should show the actual arrangement, not a collection of apparatus symbols. Position sensors, rulers, circuits, masses or supports where they would physically work. Use labels for the quantities that matter. A clear diagram can communicate geometry, alignment and apparatus placement more efficiently than a long paragraph, but it should support the written procedure rather than replace it.

Choose equipment that matches the measurement

The best apparatus is not always the most precise instrument you can name. It must suit the quantity, range and experimental arrangement. A micrometer may be suitable for a small wire diameter; a metre rule may be better for a long distance; a light gate may reduce timing uncertainty in a motion experiment; a digital multimeter may measure current or potential difference directly. Explain enough that the examiner can see why the quantity can be obtained. Cambridge repeatedly warns against naming an instrument without describing what is measured or how the reading is taken.

Plan a useful range and repeat strategy

Choose a range of independent-variable values broad enough to reveal a trend and enough values to support a meaningful graph. The exact number depends on the experiment, so avoid treating a memorised number of readings as a universal rule. Spread the values over the useful range instead of clustering them. Where random variation matters, repeat measurements and calculate a mean. If a dimension is difficult to measure accurately, several measurements at different positions may reveal variation and allow a mean. Do not use repeats as a substitute for fixing a systematic problem.

Linearise the relationship before deciding the graph

Planning questions often give a proposed mathematical relationship and ask how it could be tested or how constants could be determined. Rearrange the equation explicitly into the form y = mx + c. Then state the actual quantity to plot on each axis, what the gradient represents and, where relevant, what the intercept represents. Cambridge's June 2024 report highlights a recurring problem: some candidates wrote only y = mx + c without identifying the physical quantities on the axes. Others proposed the wrong logarithmic transformation or claimed that a straight line had to pass through the origin when the rearranged equation contained a non-zero intercept.

Click to enlarge

Say exactly how the graph confirms the relationship

Do not stop at 'plot a graph and see if it works'. State the condition. If the rearranged theory predicts a linear relationship, say that the proposed relationship is supported if the plotted points are consistent with a straight line within their uncertainties. Only claim the line should pass through the origin if the rearranged equation actually predicts zero intercept. Then state how the constant is calculated from the gradient or intercept. Rearrange the expression so the required constant is the subject. This is where the planning section becomes quantitative rather than descriptive.

Make safety precautions specific to the experiment

Generic safety statements rarely show much practical understanding. Link the hazard to the actual setup. If the experiment uses hot water or a heated metal object, identify the burn risk and a suitable precaution. If high current could heat a wire, reduce the current or switch off between readings. If a falling mass is used, consider where it could land. Cambridge's examiner report explicitly says safety precautions should be relevant to the planned experiment rather than general laboratory rules.

Question 2 starts with the table, not the graph

Question 2 supplies experimental data that you need to process. Complete the table carefully, preserving units and suitable significant figures. If logarithms are required, Cambridge's syllabus gives specific conventions: write the original quantity with its unit inside the logarithm, such as ln(d/cm), and remember that the logarithm itself has no unit. The number of decimal places in a logarithmic value should reflect the significant figures of the original measurement. Do not round transformed values so aggressively that the graph, gradient or final constant changes.

Plot points and error bars precisely

Use a sensible scale, label both axes with quantity and unit, and plot points accurately. Error bars should be symmetrical where appropriate and must reflect the uncertainty stated or calculated for the quantity. Cambridge's June 2024 report notes that inaccurate error bars and oversized plotted points cost candidates marks. A point should be small enough for its coordinates to be clear. The graph is part of the measurement process, so drawing quality matters.

Draw a balanced line of best fit

A line of best fit should represent the overall pattern, with points reasonably balanced around it. Do not force the line through the highest and lowest points. Do not join points individually. If the graph should be linear, draw one straight best-fit line. Cambridge's report notes that some lines lost credit because the plotted points were not balanced around the line. A transparent ruler and a sharp pencil make this easier.

The worst acceptable line is an uncertainty tool

Paper 5 does more than ask for a best-fit gradient. The syllabus requires a worst acceptable straight line that is either the steepest or shallowest line that can still pass through all the error bars. Distinguish it from the best-fit line by labelling it or drawing it as a broken line. The purpose is to estimate uncertainty in the gradient and intercept. The absolute uncertainty in the gradient is the difference between the gradient of the line of best fit and the gradient of the worst acceptable line. The intercept uncertainty is obtained in the same way using the two intercepts.

Click to enlarge

Use a large triangle for the gradient

Choose two easy-to-read coordinates that lie on the line of best fit and are far apart. Cambridge's 2024 examiner report says candidates do not need to use plotted data points; coordinates on grid lines are often better. The report also warns against triangles that cover too little of the line. Calculate gradient as Δy/Δx, not the reverse. Include powers of ten from the axis labels. A correct-looking numerical gradient can still be physically wrong if the scale factor is lost.

Find the intercept from the actual equation of the line

If the graph has a false origin or the y-axis does not visibly include x = 0, do not read the intercept from an arbitrary point on the visible axis. Use y = mx + c with a point on the best-fit line and the calculated gradient where necessary. Cambridge's 2024 report highlights candidates who read a y-value at a non-zero x-coordinate and treated it as the intercept. The intercept belongs to x = 0, regardless of where the plotted graph begins.

Treat uncertainties as part of the final result

Paper 5 expects you to move between absolute, fractional and percentage uncertainties and to calculate uncertainties in derived quantities. Show the working. If the final constant depends on both gradient and intercept, include the uncertainty contribution from both when the required relationship demands it. Do not assume the fractional uncertainty in the intercept is the same as the fractional uncertainty in the gradient. Cambridge's examiner report identifies this as a recurring error. Work from the best-fit and worst-line quantities that actually belong to each parameter.

Logarithms need careful interpretation

Some Paper 5 relationships become linear only after taking logarithms. Be clear whether the question uses lg, meaning base 10, or ln, meaning natural logarithm. The June 2024 report specifically notes confusion between using 10 and e when recovering constants from logarithmic intercepts. Write transformed axes explicitly. If the y-axis is lg(T/s), include the original unit inside the logarithm notation and remember that the plotted log value itself is dimensionless.

Show the equation before the numerical substitution

Towards the end of Question 2, the analysis often returns from the graph to physical constants. Cambridge's examiner report says the strongest candidates clearly stated the equation, substituted the graph-derived values, calculated the result and included an appropriate unit. Do not substitute raw table values when the question requires a constant to be obtained from the gradient or intercept. Use the result that the graph was designed to provide.

Original example: testing a power law

Suppose an original investigation proposes T = C L^n, where T is a measured time and L is a length. Taking base-10 logarithms gives lg(T/s) = n lg(L/cm) + lg C, provided consistent reference units are used. A graph of lg(T/s) against lg(L/cm) should therefore be straight if the relationship is valid. The gradient gives n, while the intercept gives lg C, so C is found using 10 raised to the intercept. This example shows why the algebra comes before the graph. If you plotted T against L without transforming the relationship, the gradient would not directly equal n. If you confused lg with ln, the recovered value of C would be wrong even if the graph itself looked excellent.

Common Paper 5 mistakes

1. Naming variables without explaining how they are changed, measured or controlled.
2. Drawing apparatus in positions that could not physically make the required measurement.
3. Choosing an instrument that is unsuitable for the required range or precision.
4. Writing y = mx + c without naming the actual graph axes.
5. Claiming a straight line should pass through the origin when the rearranged equation has an intercept.
6. Giving a generic safety precaution that is unrelated to the experiment.
7. Rounding transformed or calculated table values too early.
8. Drawing error bars inaccurately or using large plotted points.
9. Forcing a best-fit line through the first and last points.
10. Using raw data points instead of coordinates on the fitted line to calculate the gradient.
11. Calculating Δx/Δy instead of Δy/Δx.
12. Reading a y-intercept at a non-zero x-value because the graph has a false origin.
13. Ignoring the uncertainty in the intercept when the final derived quantity depends on both gradient and intercept.
14. Confusing lg with ln when recovering constants from a logarithmic relationship.

A reliable Paper 5 routine

1. For Question 1, identify the independent, dependent and important controlled variables before writing the method.
2. Describe the apparatus and exactly how every important quantity will be measured.
3. Choose a useful range, repeats and controls that make the test realistic.
4. Rearrange the proposed relationship into y = mx + c and name the actual quantities on both axes.
5. Explain how the gradient and intercept will be used to obtain the required constants.
6. For Question 2, complete the table with correct units, significant figures and transformed values.
7. Plot accurate points and error bars, then draw a balanced best-fit line and a valid worst acceptable line.
8. Use widely separated coordinates from the fitted line to calculate the gradient.
9. Show how gradient and intercept uncertainties are obtained from the two lines.
10. Finish numerical work with clear equations, substitution, units and appropriate significant figures.

Put it into practice

Take one recent Physics (9702) Paper 5 and review Question 1 and Question 2 separately. In Question 1, label every planning mark as variables, measurement, control, analysis or additional practical detail. In Question 2, label every lost mark as table, graph, best-fit line, gradient, intercept, uncertainty or final calculation. Then practise the weakest category on its own before attempting another full paper. Inside NeuraGeek, you can use Physics (9702) past-paper practice alongside the existing Paper 4 calculation guide, treating Paper 4 as theory-driven problem solving and Paper 5 as the design and analysis of real measurements.