Why were the polls so wrong in the New Hampshire primary (causing embarrassment for the press and pundits) -- and only on the Democratic side?
Many theories have been advanced in the media since Hillary Clinton's stunning upset over Barack Obama. One of them has been much contested: that white voters told pollsters they would vote for Obama but couldn't quite pull the trigger for the African-American candidate when the time came to cast their ballots. This allegedly counted for more than any sort of "late female surge" for Clinton.
Maybe when Bill Clinton referred to the "fairy tale" surrounding Obama he meant the fable that massive numbers of whites would actually vote for Obama when they had plausible alternatives. But this has been the elephant in the room almost totally ignored by the media until now.
Why did it show up (if it did) in New Hampshire and not in Iowa? The Iowa caucuses were quite public, this theory goes, while voters had a curtain to hide behind in New Hampshire.
An interesting new detail has now emerged seemingly bolstering that theory: not just advance polls, but some exit surveys apparently show that even coming out of the polls, voters in New Hampshire gave Obama about a 5% bulge -- if they were being honest. Where did all those votes go? Maybe he never really had them to start with.
Frank Newport, head of the Gallup organization, said that his numbers did not support the idea that huge numbers of older women turned out and were not fully accounted for in the projections. Democratic party membership is now strongly tilted toward women.
"I think it's very naive to dismiss the racial factors in this," said Larry Sabato, professor of politics at the University of Virginia.
The racial theory is far from proven, yet it was surprising to see lengthy probes of the poll debacle -- such as one by Ken Dilanian in USA Today -- that did not even mention the possibility of some sort of modest race effect.
"Anytime you've got white undecided voters pulling the lever choosing between a white and a black candidate, that is when the race issue is most important," Drew Westen of Emory University told Tom Edsall, the former Washington Post reporter now writing for Huffington Post. "Both campaigns' internal polls showed a 10 to 12 point Obama lead; to see that evaporate into a three-point loss, when he didn't have any gaffes, that has a ring to it."
Andrew Kohut of the Pew Research Center said, "The failure [of polling] on the Democratic side has to do with the fact that Clinton ran best among groups of voters who most often refuse polls -- poorer, less well-educated people. These are also the very people who are reluctant to vote for a black candidate."
And Kohut told the Associated Press: "You can't rule this out as an issue." He said the problem had not arisen in Iowa, where ''Obama was not the front-runner. He was not such a symbol, perhaps threateningly, to people who don't like blacks, that he might be president.'' He told NPR he would be drilling deeper into the results this week to see what shows up in this area.
In a New York Times op-ed on Thursday, he concludes: "In New Hampshire, the ballots are still warm, so it's hard to pinpoint the exact cause for the primary poll flop. …