Politics isn't rocket science, to the eternal detriment of the egos of pundits and consultants. It's just math: get more of the folks who will vote for your candidate to the polls, and you win. In Virginia on Tuesday, a race that was expected to be a blowout wasn't, for the simple reason that, one, incumbent Sen. Mark Warner's (D-Va.) political scientists didn't get his people to the polls and, two, that the people who did go were less likely to vote for the Democrat.
(Warner currently leads by 0.8 percent and has declared victory, but Republican Ed Gillespie isn't conceding yet and might request a recount.)
The map below shows the drop-off in turnout for each county and city in Virginia between 2012 and 2014. Every county saw a drop, but certain (bigger) counties saw a much bigger one.
The drop, particularly in the Northeast, overlaps with where much of the Democratic growth in the state has been over the last decade.
Stephen Hanna of the University of Mary Washington created the map below as part of a research project by Stephen Farnsworth, Stephen Hanna and Benjamin Hermerding. It contrasts the shift in the vote for Mark Warner between 2001, when he became governor, with his vote in 2014. The swollen blue areas were more likely to vote for Warner over the span of 13 years -- but also overlap with the areas where turnout dipped the most.
To put a fine point on it, those areas voted more Democratically in the 2012 election than the country on the whole. (The data here shows vote margin relative to the national split between Obama and Romney.)
If you compare the turnout drop with the 2012 vote directly, Warner's problem is obvious. The dashed line below shows the general trend, but the second tab shows the dramatic plunge.
There's a bit of a chicken-and-egg problem with turnout. When you compare how Warner did in Virginia counties with how Tim Kaine (D) did two years ago, 98 of Virginia's counties and cities voted more Republican than in 2012. Only 38 voted more Democratic. That's in part because Democrats stayed home, but might also be because Warner had softer support. If you look at counties and cities in which Warner beat Ed Gillespie (R), even counties Warner won had shifted to the right. As Farnsworth, Hanna, and Hemmerding put it: "The sea of rural blue that Warner enjoyed in 2001 has been replaced by deep red Republican sentiments. Rural voters who once backed him are now far less likely to split their tickets."
But that second tab! Turnout dropped in Warner counties almost twice as much as in Gillespie counties. And as any rocket scientist can tell you, contests based on simple arithmetic get a lot closer when one side's numbers start to drop.