Here we go again with Public Policy Polling. They did a poll for the League of Conservation Voters on the 2013 race for Governor in Virginia, and the electorate predicted by PPP is a strange one.
You see, it doesn't look like Virginia.
OK, sure. I had it wrong in 2012 in that I did not expect the base turnout advantage Barack Obama got, so what I thought would be a 7-10 point Romney win, was actually a four point Obama win. I was sure 2012 would be an economy-driven, swing-voter election. It wasn't. Instead of talking about jobs, Virginia voters were talking about abortion. Republicans didn't trust their candidate on abortion. Democrats trusted theirs. Democrats showed up (I live deep inside Arlington and let me tell you, the lines were long here in the heart of Jim Moran's district).
Let's compare this poll with other polling. According to CNN, in 2012 Virginia's electorate was 53% female/47% male, 70% white/20% black, and a two party split of 55 D/45 R. But something very strange is happening according to PPP. Many black men will sit out in 2013. Just the men, and somehow an almost equal number of Republicans as Democrats. Or something like that.
That's right. PPP has the electorate becoming much whiter, but only very slightly more Republican. 56% female/44% male, 74% white/14% black, and get this, a two party split of 54/46. So somehow, the black electorate which went 93-7 for Obama is dropping by 7%, the white electorate that went for Romney 61-31 is going up by 4%, but the net result is only +1 for Team Republican.
And on top of that, with all of those electoral shifts, McAuliffe manages to take a state that went 51-47, and turn it into a 44-37 advantage for him.
Is it within the realm of possibility that Virginia's shifted left overnight, such that a huge demographic shift toward Republican-leaning blocs will only result in a negligible gain for Republican party identification, and will result in an even bigger win for the Democrat? Sure, I suppose.
Is it more likely that Public Policy Polling is just fudging the numbers again? I believe so.