How to spot a misleading graph
Launch library · evergreen read
A graph can distort a story through choices that seem technical but carry real persuasive weight, such as starting an axis somewhere other than zero, which can make a modest change look dramatic, or stretching one axis to exaggerate a trend that is actually fairly gentle underneath, once the scale is corrected back to something honest.
Checking the axis labels, the scale, and the time period covered is a quick way to test whether a graph is presenting information fairly. A trend that only appears when a very specific and narrow window is chosen deserves particular scepticism before being accepted at face value by a reader.
It also helps to ask what a graph is not showing. Leaving out earlier or later data, or comparing categories that are not truly alike, can make a chart technically accurate while still leaving a misleading overall impression in the mind of anyone reading it quickly and without pausing.