A poorly conceived graph

Tufte, where are you?

A graph by Alan Kreuger, a top economist, who should know better than to draw misleading charts:

I have no quarrel with the data on the graph, of course. What I'm pointing out is how poorly conveyed the information content of the graph is.  The graph is so poorly conceived that it is misleading.

Looking at the graph, and without looking at the captions, do think the United States is improving or not? Our natural reaction on any numerical chart is that the top-right is where we want to be.  That is, however, not what the chart is saying.

The x-axis variable is inequality. The more left you are better, the more cohesive the society is.  So, for one thing, the x-axis should probably have been (1-inequality) or the axis should have gone from 0.40 on the left to the 0.15 on the right i.e. the x-axis ought to have been flipped.

The y-axis is similarly poorly conceived. The higher you are in the y-axis, the less the social mobility in the country, i.e. people born poor remain poor. Again, the y-axis ought to have been (1-elasticity) or its range flipped to go from 0.6 at the bottom to 0.1 on top.

Once you flip the two axes, then the take-away of the graph becomes extremely clear. The United States is unequal and getting more so.

Chart-drawing software (Mr. Kreuger probably used Excel here) has made it so easy to draw graphs that people forget to stop a bit and think about whether the chart actually conveys (visually: that's the point of a chart after all) what they want it to convey.