A couple of months ago, I noted (based on a Research 2000/Daily Kos poll) that although Republicans were more likely to turn out in the midterms, that was balanced by the fact that there were fewer Republicans than Democrats. Multiply likelihood of voting by number of partisans, and the GOP advantage was only one point.
That’s still the case. According to the Gallup/USA Today poll—not to be compared to the other one, which had different questions and a different sample; I’m just making a general point—69 percent of Republican voters are enthusiastic about voting, only 57 percent of Democrats. (I get those numbers from Nate Silver, who I assume subscribes to a fuller version of the Gallup poll than USA Today is showing.) But only 28 percent of the sample self-identify as Republicans; 32 percent are Democrats. If we assume that only enthusiasts will vote and multiply it out, the poll predicts that Republicans will send 19 percent of registered voters to the polls in November—and Democrats, 18 percent.
Nate argues that what matters is the ratio of turnout levels, not the cardinal gap. True. But in using for his example a hypothetical electorate of “50 Republicans and 50 Democrats,” he obscures a crucial point. Democrats have more trouble motivating our base because it’s less homogeneous, and its less homogeneous because the Democratic party is larger. We don’t need to have the same intensity that they have in order to win. Conversely, the Democrats’ Party ID has gone down as our intensity has gone up. For better or for worse, we’re becoming somewhat more like the other guys.
Again, a one-point disadvantage is nothing to crow about. Independents are still very wobbly. And I wish the party ID gap were even larger (though it has been in some other recent polls). This will be a very, very tough election. But let’s not panic for the wrong reasons. The Republicans won’t win by bringing to the picnic more crabapples than we have apples.
I will continue my internet jihad on this point: as everyone knows, we don't vote nationally. If my Republican congressman, Rep. Rehberg, wins by 10 points — and he just might — that doesn't give him any more authority in Congress than a 1 point win would for our Democratic neighbor Walt Minnick. Thus, a nine point Republican advantage translates into a tie. And this is what intensity means, especially in larger (and well gerrymandered) states.
Can Rep. Minnick win? He voted against both the stimulus and the Affordable Care Act; not that this would actually matter to the people claiming that these issues motivate them . . .
(Rehberg won in 2008 with 64% of the vote — 307,000 votes. Minnick won with 175,000 votes.)
@CharleyCarp: theoretically, you're of course right. Empirically, national swings tend to track swings in Congressional seats pretty well. (I'm sure there's a fancy statistical name for why that is.)
This election could be different. In particular, the President's race could mean that the South will trend differently from everywhere else, and there's some evidence that he's uniquely and savagely unpopular there. And that could mean that a lot of Southern Democrats will lose their seats but not a whole lot of others. (And there aren't many Southern Democrats left, except safe ones from heavily African-American districts.) But that would be a very unusual and striking outcome, and I'm reluctant to bet on those.
I'm skeptical, because the "Republican" label is toxic and tarnished even among people who will absolutely and exclusively vote for the Republican, making the party ID question problematic.
"But only 28 percent of the sample self-identify as Republicans; 32 percent are Democrats."
Leaving 40% who self-identify as neither; Maybe how they're planning on voting has some relevance here?