Six ways a chart misleads without containing a false number.

This is the rule people half-remember, and the half they drop is the important one. Bars encode value as length, so a bar starting at forty encodes a length corresponding to nothing, and the ratio between two such bars is arbitrary.

A line chart is about change rather than magnitude, so a non-zero baseline is legitimate — but a heavily truncated one converts a two per cent move into a visual collapse, and the reader's impression is set by the shape rather than by the labels.

The honest test: if the point is that a change is small, does the chart let it look small?

Month-to-date plotted against complete months is the single most common way a chart shows a collapse that is not happening. The line falls off a cliff on the last point and the cliff is the calendar.

It is also the easiest to do by accident, because the data is correct and nobody has done anything dishonest. Look at the last point on every time series and ask whether that period has finished.

A rate of forty per cent means nothing without knowing forty per cent of what. Rates are where charts mislead most often, because the numerator is usually obvious and the denominator usually is not.

Alphabetical ordering of categories wastes the chart's most valuable dimension on information the reader already has. Sorting by value is what turns a lookup table into a finding, and its absence is often a sign nobody looked at the chart analytically.

Two series on two different axes can be made to cross wherever the author likes, simply by choosing the two ranges. There is no correct scaling, which is the argument against dual axes generally: any relationship the chart appears to show is a property of the axis choice.

Below about thirty observations, one member moves the rate by more than most real effects. A segment of twenty-four people reported as a percentage invites a comparison that the sample cannot support.

Counts are less satisfying and more honest, and they make the sample size impossible to overlook — which is usually why the percentage was chosen instead.