Note: if you’re reading this and images are not showing, visit the original post on my blog. The blog post contains interactive plots which help in understanding the point I’m making.
I’ve recently come across this graph (on Twitter) from the Economist:
You can read the article here (archived for posterity). There are many things wrong with this chart. First of all, the economist is fitting a linear regression to some data points, and does not provide anything else to the reader, namely the regression coefficients, their standard errors, and the R².