Case 02: Charts That Lie
Same numbers, three different stories. One of the charts is cheating. Catch it.
A made-up phone survey asked: which brand would you buy next? Four brands, 151 votes. A salesperson shows you three charts of the same numbers, and each one tells a different story. Your job is to work out how each trick works and how to spot it.
All data here is made up for the video. No real brand, no real votes.
Your mission
- Rebuild the honest chart yourself from the Brands data. Start the axis at 0.
- Break it on purpose. First change where the axis starts, then squeeze the shape of the chart, then pick only a slice of the months.
- Find a real chart in the wild (a post, an article, an ad) and catch one trick. Log it in the Your Turn tab.
- Never share a chart that contains someone’s personal details.
The evidence
Chart 1: honest. Axis starts at 0. Votes per brand:
Chart 2: the cut axis. Same two numbers, but the bars start at 48. Brand A now looks twice as tall as Brand B:
Chart 3: the squeeze. Same numbers again, but the chart is drawn tall and thin. A small difference looks like a cliff. Nothing in the data changed, only the shape of the frame.
Chart 4: the slice. Brand A’s monthly votes. A salesperson shows only March to June:
And here is the full year the slice hides:
The twist. The numbers never changed. Only the axis, the shape and the slice did.
Brand A has about a third of the votes (34%), so it is not “destroying” anyone. The honest gap with Brand B is 2 votes, or 4% of Brand B’s total. The sliced chart shows Brand A up 44% from March to June, but after the June peak of 52 votes it falls 65% by December.
Spot the trick
Ask four questions of any chart you meet:
- Where does the axis start?
- Is the chart stretched or squeezed?
- Which time period or group was left out?
- Are the exact numbers shown?
Last updated: [date]. Next case: Case 03: Does Money Buy Happiness? (Opens soon)