Crunching 2006 Numbers
An extended and rather heated exchange has broken out over at TAPPED regarding Third Way's recent analysis of electoral trends between 2004 and 2006, which, to make a long story short, suggests that Democrats main vote gains last year were in "red" elements of the electorate, especially white men and high earners. The report drew criticism from Tom Schaller, Mark Schmitt and Ezra Klein. Then TAPPED let the Third Way folks respond in a guest blog, and Schaller came back at them once again.
For all the fire in these posts, I have to say both sides of the argument have important, legitimate points to make. In particular, Schmitt is right, generally, about the different nature of the electorate in midterm versus general elections (though I don't know that there's much to gain from staring at comparisons of 2006 with 2002, given the anomylous nature of the latter). But Third Way's right that there's something significant about the ability of Democrats to do so well in a less congenial electorate. Schaller's right that looking at percentage performances among different subelements of the electorate shows a different picture than Third Way's, and avoids some of the pitfalls of the "normalization" methodology Third Way used to create its raw vote comparisons. But Third Way's right that comparing percentages is misleading as well, since small gains in large segments of the electorate often produce more votes than large gains in small segments.
I do have a couple of observations to add based on my own unpublished, unscientific analysis of 2004 and 2006 House exit polls a few months back. First of all, trends in some of the subgroups of the electorate partially undermine the assumption that Democratic gains among whites, men, marrieds, upscale voters and self-identified independents (all of which definitely occurred) can be interpreted as gains in "red" or "red-leaning" voters. In particular, when you break the electorate down into self-identified liberals, moderates and conservatives, Democrats gained roughly the same percentages across the board, without any significant change in the ideological composition of the electorate.
Second of all, and more importantly, the national exit poll trends disguised some very striking regional variations. In the Northeast, Democratic gains strongly reflected the trends Third Way talks about, concentrated among white upscale suburbanites. But ideologically, Democrats gained an amazing 10 points among self-identified liberals, more than twice the gain among moderates. The West, Democrats' second-best region, was like a different country, with gains heavily concentrated among less-educated white men, and in rural areas. In the Midwest, Democrats made no gains among suburbanites, and made surprisingly strong gains among African-Americans. And in the South, Democrats actually lost ground with suburbanites and gained nothing from moderates, while the African-American percentage of the electorate dropped significantly.
Topping off all these confusing variations is the fact that the 2006 exit polls showed double-digit Democratic gains among Latinos. But virtually everyone thinks the 2004 exit polls significantly understated the Democratic Latino vote, so it's hard to know how seriously to take that "trend."
All in all, probably the safest thing to say is that Democrats' fine year in 2006 owed itself to a variety of national, regional and local factors; that Dems did pretty well in categories of the electorate where they've been struggling recently; and that the single most important trend was the strong showing Democrats made among self-identified independents, who may be "swing" voters but aren't necessarily "moderates." It was neither the base-mobilization election so many people predicted; nor the classic Clintonian seize-the-center election others suggested after the fact. --