Chart junk illustrated

A word about chart junk

Charts that cheat by showing only part of the vertical axis will exaggerate any change in data compared to a chart with a complete vertical axis.

This example comes from two data points in a recent Pew study.

Chart junk in line charts

By truncating the vertical axis in a line chart, you exaggerate any changes reflected in the data.

The chart on the left shows the change from 2007 to 2014 using a complete (0-60) vertical axis.

The chart on the right does more than zoom in when it shows only data points 49-55; it changes the slope, making the change look more dramatic.

Technically, this is legitimate because of the clear “break” in the vertical axis. But the emotional/connotative impact of the two charts is very different, and that’s what gives me heartburn. And leads me to label charts like the one on the right as chart junk.

Here are some direct examples from the the study summary:

two pew charts

Two Pew charts illustrate this problem.

 

There’s not a lot of slope (change) evident in these charts. Try to imagine how little slope there would be if you saw the complete vertical axis.

I understand why charts are presented this way in print: for effect or because of space constraints. The effect is exaggerated change, so that’s an editorial choice. But on the web, where is the space constraint?

Line charts are traditionally used to show change over time. But in this case, there are only two points. A bar chart would be more visually neutral. And intellectually honest.

Bar chart comparison

In this example, two bar charts are more visually neutral.

 

One thought on “A word about chart junk

Leave a Reply