Feeds: 
symbol identifies content for paid subscribers only.
ROUND FOUR FOR THE THIRD WAY. Now that the folks at Third Way have weighed in to defend their recent report claiming that whites, males, and the affluent (and I suppose, by extension, affluent white men) catapulted the Democrats to power in the 2006 midterms, let me respond, and also add a few things.
First of all, I must repeat that the "normalization" technique they use is misleadingly wrong. If they don't like the baseball metaphor I employed last time, here's my response to their claim that they produced an apples-with-apples comparison of the two electorates by simply multiplying one to scale with the other: Take an apple, which is normally smaller than an orange, and multiply it by 125 percent and see if it looks and tastes like an orange. The normal turnout in an off-year is simply not the same in a presidential year, for a variety of reasons, of which Mark Schmitt's smart point about differential turnout in competitive v. non-competitive districts is just one. (Others are the lower number of political ads and contacting, less media coverage of politics that cycle, and so on.) This is why their results are artifactual, and it is also why I produced a very simplified example of two, back-to-back elections in which Democrats both do no better among whites and better among blacks in a congressional cycle and yet still draw a greater share of votes from whites during that cycle -- again, because of differential turnouts. Third Way offers no response to this demonstrated mathematical fact because, well, there is no counter-argument.
Notice, too, that in both their original report and their response here, Third Way makes a big fuss about Democrats improving their performance among "redder" groups, yet quite curiously ignores the performance among what they might call "bluer" subgroups. Indeed, given all the time they spent comparing the 2004 and 2006 cycles, it is curious that the Third Way folks failed to notice this damning fact: Democrats improved among their own, "bluer" Democratic-leaning subgroups! Based on CNN exit poll data, here is my quick math on some of those groups:
- Women: +3 in 2004 to +12 in 2006; difference = +9
- Non-whites: +45 to +50 = +5
- Income under $100K: +1 to +12 = +11
- Union households: +19 to +30 = +11
- Non-college educated: -6 to +13 = +19
- Non white/evangelicals: +13 to +22 = +9
- Church occasionally or never: +11 to +25 = +14
- Unmarried: +18 to +30 = +12
So how can Third Way state in its report that, "relative to the rest of the electorate, Democrats received fewer votes from among much of their traditional base -- African-Americans, unmarried people, and the poor"? Again, by conflating turnout with performance, that's how. Sure, there were fewer Democratically inclined voters, but Democrats didn't perform worse among such voters; in fact, they performed better.
This is why I predicted previously that, if performance levels among "bluer" groups were to be the same in the upcoming presidential election as it was in 2006, even if Democratic performance in 2008 among "redder" groups is like it was 2004, the Democrats could still win. In their reply they say my prediction is "in fact what happened in 2004. The Democratic base turned out in droves -- and Democrats lost." Yes, and thanks for making my point, folks: Democratic performance in 2004 among its base groups was not what it was in 2006, but if they get a normal demographic presidential turnout in 2008 and their base-group performance is like 2006, they will win. And they will also win even if -- here's the key point --their red-group performance turns back to 2004 rates. I'm not sure how much clearer or how many times I can state this prediction. And now that the folks at Third Way have done the higher performance data they so curiously overlooked in writing their report, are they prepared to concede the point?
Overall, as would be expected for any winning Democratic coalition, the Democrats' 2006 victories were primarily driven by a female, non-white, union household, unmarried, lower-income, non-regular-churchgoing coalition -- with an emphasis on "primarily." Sure, rich, rural, higher-income, white, and male voters (and combinations thereof) voted more Democratic in 2006 compared to 2004; I happily concede Third Way's point there. But heck, everybody in the country -- with the notable, within-the-margin-of-error exception of white evangelicals -- voted more Democratic in 2006 than two years earlier.
To therefore proclaim that affluent white voters were critical for Democrats in 2006 is very, very misleading.
--Tom Schaller
COMMENTS (2)
"Based on CNN exit poll data"
It appears that you compared the 2004 presidential exit polls with the 2006 house exit polls. Is this true?
If so why didn't you campare 2004 house exit polls with 2006 house exit polls?
Posted by: hadenough | May 17, 2007 7:35 PM
"Democratic performance in 2004 among its base groups was not what it was in 2006"
Well, then it seems that the thing for you all to do, if you want to win as a party, is to figure out what is you did to alientate your base in 2004.
"And now that the folks at Third Way have done the higher performance data they so curiously overlooked in writing their report, are they prepared to concede the point?"
No, because they're still gunning for John Kerry II.
Posted by: Anonymous | May 18, 2007 11:48 AM