The Third Way's Flawed Look at 2006

In an attempt to prove that their centrist ways are the winning ticket for Democrats, the Third Way studied the 2006 vote and issued a report, "Looking Red, Voting Blue." I haven't seen the report, but this summary from Cillizza at WaPo sums it up.

Let's start with the overall shape of the electorate as laid out in the Third Way survey. "Compared to 2004, the 2006 electorate was wealthier and whiter, and more religious, male, married and rural," the authors write.

The average 2006 voter had a median income 13.4 percent higher than that same voter in 2004; the percent of white voters went from 77 percent in 2004 exit polls to 79 percent in those same polls in 2006. Men made up 48.3 percent of the 2006 voting pool, a two percent increase over 2004; married people accounted for nearly 70 percent of the 2006 vote as compared to 63 percent in 2004.

"Not only did Democrats win, they picked up nearly all of their new votes among those who fall into the typical Republican profile of voters," says the study. "Millions of voters from constituencies that had given up on Democrats in the past -- whites, men, couples, the well off, rural Americans, and yes, even the middle class -- switched sides in 2006."

That first sentence reveals the major problem with this survey, one Cillizza should have picked up on. You cannot compare the composition of voters in a midterm election to a presidential election and expect valid results. Though they seem to try to address this in their methodology by equalizing for the numbers of voters ("we used the National Exit Poll surveys from the 2004 and 2006 elections and then employed a standard statistical technique to 'normalize' the results so that the overall turnout was the same in 2004 as it was in 2006") they messed up when considering the composition of those voters. The electorate in any off year is going to be "wealthier and whiter, and more religious, male, married and rural" because these are the people who vote more reliably. Any serious analysis should have compared 2006 to 2002 and 1998.

The survey does reiterate something we pretty much knew all along. The 2006 vote was predominantly about Iraq:

The prime issue that drove these atypical Democratic gains was -- you guessed it! -- the war in Iraq. Democrats gained 7.6 million voters in 2006 who said they "strongly disapproved" of the war in Iraq while simultaneously losing 2.9 million voters from 2004 to 2006 who said they approved of the war -- for a net gain of 4.7 million voters. Similarly, Democrats gained 6.4 million voters between 2004 and 2006 who said that the Iraq war did not enhance national security while receiving 1.7 million less votes among those who did.

Other than the Iraq finding, there can be little in this report that shows Democrats the way to go in 2008. Iraq is likely to be the key issue in 2008 again, and once more, the key for Dems will be differentiating themselves from the GOP--and particularly Bush--on the war. That means doing everything in their power to bring an end to the Iraq debacle. Even if they don't succeed in doing so, that has to be seen as a result of the intransigence on the part of the GOP and Bush rather than lack of effort on the part of the Dems.