Misleading charts rarely contain false numbers; they mislead through choices about how true numbers are shown.
Some are deliberate, but many are accidents by someone who never checked the design.
Knowing the common tricks helps both in making honest charts and in reading other people's.
A truncated axis on a bar chart exaggerates differences, since the bars no longer show their full lengths.
Cherry-picked time ranges choose a start or end date that tells the preferred story.
Dual axes put 2 series on different scales in one chart, so their lines can be made to cross wherever the designer likes.
Pictures scaled in both width and height grow 4 times in area when the value only doubles.
Uneven intervals, such as years 1, 2 and 5 apart drawn at equal distances, distort the shape of change.
Running totals always rise, so they can hide a period when new sales actually fell.
Showing counts instead of rates can mislead, since bigger places have more of almost everything.
3D effects distort sizes through perspective, making the front slices of a pie look bigger.
Leaving out any sense of uncertainty makes small differences between groups look like firm findings.
In practice, checking the axis start, the time range, the units and whether figures are rates catches most problems.
Drawing the honest version beside a suspect chart is often the quickest way to show what it hides.
The common misconception is that a chart of real data cannot lie, when the design decides what the reader believes.