Five charts that mislead without a single false number

A chart is an argument. The data supplies the evidence; the axis, the window and the shapes decide how loud each piece of evidence sounds.

1. The axis that skips zero

Two plans, two satisfaction scores: 48% and 52%. Start the bar chart's axis at 46 and plan B's bar is three times as tall as plan A's. Start it at zero and you see what the numbers say: a four-point gap.

2. The convenient window

Show a share price from its low in March to its high in June and you have a rocket. Show the five years around it and you have a wobble. Neither window is false; one was chosen for the story.

3. Two axes, one convenient crossing

Put ice cream sales on a left axis and sunburn cases on a right one, then scale each axis until the lines overlap. They'll look like one cause and one effect. Rescale either axis and the crossing moves wherever you like. If two series need comparing, index both to 100 at the same starting point and share one axis.

4. Circles sized by radius

Readers judge a circle by its area, and area grows with the square of the radius: A = πr². Double a value by doubling the radius and the circle looks four times bigger. Size circles by area, or better, use bars.

5. Totals without a denominator

"The north branch had twice as many late returns as the south." It also has three times as many loans. Per thousand loans, the north branch is better. A total compares sizes; a rate compares behavior.

A two-minute checklist

1. Does a bar chart's axis start at zero?
2. What period does it show, and what happened just outside it?
3. Are there two vertical axes? Could you rescale one to tell the opposite story?
4. Is size shown by area or by length?
5. Are these totals or rates? Out of how many?

Lesson 2.6 in the course turns this list into practice questions with real charts from public data sets. If you'd like the next note in your inbox, sign up below.