Data Blogging: the Electoral College

Defenders of the Electoral College claim that it serves to amplify the popular vote margin of victory and therefore (theoretically) enhances the legitimacy of the elected president. I was curious what the data looked like so I plotted it up. (Data courtesy of wikipedia - god help me caveat emptor - here and here.)

This is a plot of percentage of electoral votes vs. the popular vote margin for presidential elections in the past 100 years. Sure enough, the trend shows that the electoral college 'saturates' at somewhere around a 20 percent margin. That is, a 60-40 split in the popular vote will more or less win you a landslide, although there is a fair amount of scatter in the data.

Which is all well and good if you're Reagan in '84 or FDR in '36 (the two largest electoral college landslides on the right end of the trendline), but some of the outliers are pretty interesting too. The two largest popular vote margins (Harding and Coolidge) didn't exactly run away with the electoral college. And for the really close elections (the ones where the winner might want some legitimacy amplification) the effect tends to be pretty small.

And then of course there's poor Al Gore - the only datapoint, in this century, to fall below the x-axis.

[Update: Barack Obama beat John McCain by 6.5% and received 365 electoral votes, which puts him right on the trendline (red dot), and quite close to Bill Clinton's first election.]