Initial Exit Poll Analysis
Print Friendly and PDF
The CNN website has posted preliminary national exit poll numbers for House races. They show a big victory for Democrats in House races, in the 55-45 range, in contrast to the GOP's 51-45 win in the last midterm election in 2002.

(My detailed analysis of the long lost exit poll data from 2002 is here.)

The exit polls show a general shift toward the Democrats among most demographic groups except blacks.

The GOP did disastrously among whites this year, winning only 50-48, compared to 2002 when they won 58% of the white vote. Relatively speaking, they performed better among minorities, where their share fell only from 23% to 20%-21%. As I've been saying for years, however, the white vote is the 800 pound gorilla of ethnic blocs, and if the GOP doesn't win at least 55% of whites, they are in big trouble.

The GOP won by only a 52%-46% margin among white men, a group in which they should win 60%, and lost 50%-48% among white women. The famous gender gap was in the 5 or 6 point range, which is not all that large historically.

The Republicans' share of the black vote increased from 9% to 11%, perhaps because they nominated black candidates in some statewide races, such as Michael Steele who put up a good fight in the Maryland Senate race. Not surprisingly, it didn't do them much good overall.

Historically, the GOP's fraction of the black vote bounces around the 10% mark, and it doesn't pay much to try to overanalyze why it goes up or down a few points from election to election.

Among Asians, Republicans drew 32%, down from 34% in 2002. Asians have been a consistently strong force for the Democrats since the mid 1990s.

The GOP's share of the Hispanic vote dropped from 38% to 26%. No doubt this figure will be given vastly more publicity than the 8 point drop among the hugely more important white vote. Prepare to read that House Republicans shot themselves in the foot by not voting to open the borders.

In the conventional wisdom, Hispanics are labeled a "critical swing vote," but in reality, they are more of a flow vote, going up and down with the white vote. For example, the best year the GOP ever enjoyed in the House vote among Latinos was Newt Gingrich's big year of 1994, which was also the GOP's best among white House voters in recent decades. In contrast, the Republicans' (small) share of the black vote goes up and down according to its own patterns, out of sync with larger trends.

So, if the GOP was down 12 points among Hispanics and down 8 points among whites, then the net loss among Hispanics relative to four years ago was only four points.

The exit poll claims that Hispanics cast 8 percent of the vote, but the exit polls normally exaggerate the Latino share. (They need to "oversample" Hispanics to get statistically significant results, but they seem to fall for their own numbers). The gold standard Census Bureau survey of 50,000 households done right after the election found Hispanics accounted for 5.4% in 2000 and 6.0% in 2004, and only 5.3% in 2002. (That's because Latinos don't turn out as much in the less exciting midterm elections.)

Assuming that the Latino fraction of the electorate in this midterm was about six percent, that would suggest that taking a stand on immigration cost the GOP about 6% times 4 percentage points lost, or about one quarter of one percent overall.

In contrast, if the Republicans did just 1 percentage point better among whites than they would have if they had rolled over and played dead on immigration, they would have gained from white voters more than triple what they lost among Hispanic voters.

On the illegal immigration issue, 30% of voters said it was extremely important, and 32% said it was very important, compared to 29% saying it was somewhat important and 8% claiming it was not at all important. Not surprisingly, the Democrats won about 65% of the vote among the 37 percent saying illegal immigration was only somewhat or not important.

Among the five-eights of the electorate rating illegal immigration as extremely or very important, however, the GOP could only win a bare majority of roughly 50-49. The GOP sent a confusing message on this big issue, with President backing amnesty and "guest" workers, as did the GOP's 2008 frontrunner John McCain, who co-sponsored an amnesty bill with Ted Kennedy. And the Senate's Chuck Hagel and Mel Martinez, both Republicans, sponsored the bill that passed the Senate that would have let in 66 million more. This kind of mixed message led to a mixed result that would prove fatal to the GOP hopes in 2006.

Print Friendly and PDF