Planning, evaluating and improving experiments1 h 30 min of study
This unit is ≈13% of the A-Level Physics, across 5 lessons. Full syllabus
Lesson 3 of 5 · Measurement, maths and practical skills
Graphs, linearisation and data analysis
8 min read · about 1 h 45 min with practice3 quick checks≈3% of the testCore: Core: tested on most papers
Reading is free. Sign in to tick off lessons, keep your place and track your mastery.
Graph work is where practical-skills marks are won and lost. In AQA Paper 3 Section A, Edexcel Paper 3, OCR A Paper 3 and Cambridge Papers 3 and 5, most of the analysis marks follow from one straight-line graph: choosing what to plot, finding a gradient and an intercept, linking them to the physics, and putting an uncertainty on the result. The rules are strict but learnable, and examiners apply them the same way every series.
By the end you’ll be able to
Choose axes and scales, plot points to within half a small square and draw a best-fit line or smooth curve
Calculate gradients using a large triangle and read intercepts, stating units that follow from the axes
Draw error bars and a worst acceptable line to find the uncertainty in a gradient or intercept
Linearise y = kxⁿ using a log–log plot and y = Ae^(kx) using ln y against x, and extract n, k and A
Draw valid conclusions from data, identify anomalous points and suggest how to extend the range or improve reliability
What the exam asks
“Plot a graph of … against …” Marks go for the axes (quantity/unit), the scale, accurate points and a best-fit line.
“Determine the gradient / intercept.” You need a large triangle and correct units. An intercept may need y=mx+c if the axis starts at a false origin.
“Show that the data support y∝x / ”, or
vii.Check your understanding
3 questions on graphs, linearisation and data analysis. Every option is explained once you answer.
Sign in to try the quick check
Answers are checked on our side, every option is explained, and your result feeds your mastery for this topic. It’s free.
The first 3 of 11 cards for this topic. Sign in and finish the lesson to review them with spaced repetition.
PromptCard 1 of 3
How should graph axes (and table headings) be labelled?
n
y=Aekx
“Explain why a graph of lnV against t is a straight line.”
“Use your graph to find g / r / C / n.” Relate the gradient or intercept to the constants.
“Draw error bars and a worst acceptable line; calculate the uncertainty in the gradient.”
“Identify the anomalous result”, “Suggest why the line does not pass through the origin”, “How could the range be extended?”
On Courselo you cannot draw on screen, so questions give you a table plus a plotted chart to analyse.
Core ideas
Plotting to examiner standard
Axis labels: quantity / unit, e.g. T2/s2 or ln(I/μA). A logarithm has no unit, so the unit goes inside the bracket.
Scale: the points must fill at least half the grid in both directions. Use 1, 2 or 5 units per large square, never 3s or 7s. A false origin is fine and often necessary.
Points: plot to within half a small square, as fine crosses or dots, not blobs.
Line: a single thin best-fit line or smooth curve with the points balanced either side along its whole length. Do not force it through the origin, and never join point to point. Circle an anomalous point and ignore it when you draw the line.
Gradient and intercept
Use a triangle whose hypotenuse covers more than half the drawn line. Read its corners from the line, not from data points, unless a data point sits exactly on the line.
Gradient units = (y-axis unit)/(x-axis unit). The gradient of s (m) against t2 (s²) is in m s−2.
False origin: if the x-axis does not start at zero, the line’s crossing of the left-hand edge is not the y-intercept. Substitute a point on the line into y=mx+c to find c.
Linearisation: make the physics a straight line
Relationship
Plot y
against x
Gradient
y-intercept
T=2πl/g
T2
l
4π2/g
0
V=ε−Ir
V
I
−r
s=21gt2
s
pV=nRT at constant T
p
1/V
y=kxn
lgy
lgx
y=Aekx
lny
x
V=V0e−t/RC
lnV
C=(d+e)2k (count rate, unknown offset )
The method is always the same:
Rearrange the equation into the form y=mx+c.
Name what goes on each axis.
Write what the gradient and the intercept equal.
Why logs work. Taking logs of y=kxn gives lgy=nlgx+lgk, a straight line of gradient n. Taking natural logs of y=Aekx gives lny=kx+lnA. If lgy against lgx curves, the relationship is not a power law. If lny against x curves, it is not exponential.
Logarithms in tables
Give lg or ln values to the same number of decimal places as the raw value has significant figures. For x=76.5 (3 s.f.), write lgx=1.884.
On a log graph, the error bar is Δ(lny)≈yΔy and Δ(lgy)≈0.434yΔy. Alternatively, calculate ln(y+Δy)−lny directly.
Error bars and the worst acceptable line
Draw error bars of ±Δy (and ±Δx if significant) on each point.
Draw the best-fit line.
Draw the worst acceptable line: the steepest or shallowest straight line that still passes through every error bar. Dash or label it.
Find the uncertainty:
%uncertainty in gradient=mbest∣mbest−mworst∣100%
The uncertainty in an intercept is ∣cbest−cworst∣.
A constant found from the gradient carries the gradient’s percentage uncertainty, multiplied by any power in the relationship. For example, g=4π2/m has the same percentage uncertainty as m.
Curves, tangents and areas
The instantaneous rate of change is the gradient of the tangent at that point. Draw it touching the curve, then use a long triangle.
The area under a curve can be found by counting squares (value of one square × number of squares) or by splitting it into trapezia.
Conclusions and anomalies
Proportional means a straight line through the origin (within the error bars). A straight line alone shows only a linear relationship.
A non-zero intercept on a line that should pass through the origin suggests a systematic error. Examples: a length measured to the wrong point, or a light gate that starts timing late.
Anomalies: repeat the reading if you can. Otherwise exclude it from the mean or the line and say so. Never discard data just because it disagrees with a prediction.
Improving the data: take more readings over a wider range, and more closely spaced readings where a curve bends most. Repeat and average each point.
Worked examples
Exam technique
Write the equation in y=mx+c form above the working. It earns the “gradient equals…” mark and stops you inverting the relationship (for example writing g=gradient/4π2).
Label the triangle’s corners with coordinates, e.g. (0.050, 0.25) and (0.550, 2.70). Examiners check that they lie on your line.
Check the sign. Decay and discharge graphs have negative gradients, and the physical constant is the magnitude: RC=−1/gradient.
Units of constants. Work them out from the axes. The gradient of ln(I/μA) against t/s is in s−1, while the gradient of a lg–lg graph has no unit.
“Suggest why the line misses the origin.” Name a specific systematic error that fits the sign of the intercept, such as measuring pendulum length to the top of the bob instead of its centre.
Never skip the analysis. Do the gradient and constant parts even if your plot is imperfect: error carried forward protects you.
Common mistakes
Quick recap
Axes: quantity / unit. Points fill at least half the grid. Plot to half a small square. One thin best-fit line.
Gradient: triangle over more than half the line, corners read from the line, units = y-unit/x-unit.
False origin: find the intercept with c=y1−mx1.
Power law: plot lgy against lgx (gradient n, intercept lgk). Exponential: plot lny against x (gradient k, intercept lnA).
Error bars on log graphs: Δ(lny)≈Δy/y.
Worst acceptable line through all error bars. Percentage uncertainty =∣mbest−mworst∣/m.
Proportional needs a straight line through the origin. An unexpected intercept points to a systematic error.