ThinkCritical.eu · Manipulation atlas
Numbers and chartsMisleading chart
A truncated axis, uneven scale or selective window magnifies a difference that barely exists in the data.
A chart looks like objective proof, so we read it faster and less critically than text. Shift where the axis starts, use uneven steps or pick a narrow window, and a small difference looks like a chasm.
A company leaflet shows two bars for customer satisfaction: “Us” and “Competitors”. Our bar is more than three times taller. But the vertical axis starts at 82%, not zero – the real figures are 92% and 85%.
How to spot it
- The vertical axis of a bar chart doesn't start at zero, so bar lengths don't match the numbers.
- The axis has no numbers, or the steps are uneven (1, 2, 5, 20…).
- Two charts side by side each use a different scale.
- The time window is suspiciously short or ends exactly where it suits the message.
How to check and what to do
- Check where the vertical axis starts and what steps it uses.
- Read the actual numbers and work out the change yourself: (new − old) / old.
- For a bar chart, imagine the axis starting at zero – does the dramatic difference disappear?
- Find the same data over a longer period, ideally from the original source.
Tip: Bar charts should start at zero. A line chart's axis can start elsewhere if that is clearly marked – but even then, read the numbers, not the slope.
Full explanation and exercises:
thinkcritical.eu/en/manipulation-atlas/misleading-chart