Home / Critical Analysis of Data & Models / Misleading Graphs
3.4a · Critical Analysis of Data & Models · Sub-skill
Spotting the tricks used to make graphs tell a misleading story — truncated axes, inconsistent scales, and cherry-picked data.
Build it up, step by step
Click each step below to reveal it — work through them in order the first time round.
A truncated axis doesn't start at zero. This can make small differences look dramatic — a bar chart where the y-axis starts at 90 instead of 0 makes a change from 92 to 96 look huge, even though it's really only about a 4% increase.
Watch for axes where the intervals aren't equal (e.g. jumping 0, 10, 50, 100), or two different scales used on the same graph without being made clear. Pictograms are especially prone to this: scaling an image up in both height and width to represent a bigger number exaggerates the size difference, since area grows faster than the actual value.
A graph might only show a short time period, or specific categories, chosen specifically because they support a particular narrative, while excluding data that would tell a different story. Always ask: what date range/categories are shown, and what's been left out?
Systematically check: does the axis start at zero? Are the scale intervals equal? Is the labelling clear and not misleading? Is any data missing or selectively shown? State clearly and specifically what is misleading and why, rather than just saying ‘the graph is wrong.’
A company's bar chart shows sales rising from £48,000 to £52,000 over two years, with the y-axis running from £45,000 to £52,000. Explain why this graph could be considered misleading.
Because the y-axis is truncated (doesn't start at £0), the visual height of the second bar appears roughly 4 times taller than the first, even though sales only increased by about 8.3% (from 48,000 to 52,000). This exaggerates the apparent growth to someone glancing at the chart.
Test yourself
A politician presents a line graph showing crime rates ‘falling dramatically’ over the last 2 years, but the y-axis only runs from 40 to 45 (crimes per 1000 people), and the x-axis only covers the most recent 2 years, even though data is available going back 10 years.
(a) Explain two ways this graph could be considered misleading.
(b) Suggest one change that would make the graph a fairer representation of the data. [4 marks]
(a) First, the y-axis is truncated (starts at 40, not 0), which exaggerates the visual size of any fall in crime rate, making a small change look dramatic. Second, only showing the most recent 2 years (when 10 years of data exists) could be cherry-picking a period that particularly suits the argument being made, hiding a longer-term trend that might tell a different story.
(b) Extending the y-axis to start at 0, and/or showing all 10 years of available data, would give a fairer, more complete picture of the actual trend.
Practice
Five short questions on misleading graphs. Work through them, then reveal the mark scheme to check.