Wednesday, January 9, 2008

Explaining Away Polling Failure

Originally uploaded by

Lots of humility today from pundits and pollsters about how wrong the New Hampshire polls were on the Democratic side -- from Chris Matthews soulfully telling Clinton's communications director "I will never underestimate Hillary Clinton again" to John Zogby's instant analysis that "We seem to have missed the huge turnout of older women that apparently put Clinton over the top."

In an interesting little item on the Huffington Post a commenter observes:
"No one is talking about how the polls actually nailed Obama's number. Obama didn't lose this election. He stayed steady and Hillary surged ahead."
Many narratives will be challenged in the coming days and will be replaced by other convenient narratives. Among the most curious, and none-too-subtle, is that the bulk of spot reporting appears to assert that Clinton's victory was a "surprise." This, even though there is no evidence that Clinton was ever behind in New Hampshire -- except from now discredited polls.

Fiction can't support a news angle, so nobody should be reporting "surprise" as fact, or as anything other than something which confounded the pollsters, or some such construction.

Matthews blames the polled -- garbage in, garbage out, he says -- and is part of a chorus singing "Bradley Effect," a theory which holds that white people lie on the upside to pollsters about their support for black candidates.

But if Obama's poll numbers were correct -- if just his relative finish was wrong -- then this isn't what happened: it was Clinton's support that was incorrectly gauged. There is no fancy theory about anybody lying to pollsters on the downside about their support for women candidates. As Zogby suggests, chances are they just weren't counted.

Whatever the pollsters say about their New Hampshire failures, the better to increase confidence in their work for the rest of the campaign, I think there are a couple of safe conclusions we can already draw:
  • The only story coming out if Iowa was that Edwards' support dwindled
  • Democratic voters who peel away from their first choice won't necessarily break to "anyone but Hillary"

No comments: