The assumption with the study you mentioned only works when those included in the study are a random sample of the population, expected to be representative. I'

The assumption with the study you mentioned only works when those included in the study are a random sample of the population, expected to be representative. I'd argue that people who vote aren't a random sample and not representative of the population. In other words, when people don't like either candidate, they are more likely to stay home. Therefore, a percentage of the population that voted for a candidate could potentially be an interesting measure. Just pulled together some numbers by dividing the winner's popular vote by the total US population in that year, and was actually surprised at Trump's showing. He's at the median of all presidents since 1960. Not sure what meaning could be extracted from it, though, as there are many factors that would affect turnout. Trump 16 - 19.5% Obama 12 - 20.9% Obama 08 - 22.8% Bush 04 - 21.2% Bush 00 - 17.9% Clinton 96 - 16.9% Clinton 92 - 17.5% Bush 88 - 19.6% Reagan 84 - 23.1% Reagan 80 - 19.3% Carter 76 - 18.7% Nixon 72 - 22.3% Nixon 68 - 15.8% Johnson 64 - 22.3% Kennedy 60 - 18.9% Sources: http://www.multpl.com/united-states-population/table http://www.multpl.com/united-states-population/table http://www.270towin.com/historical-presidential-elections/ http://www.270towin.com/historical-presidential-elections/