Pollsters Right On for GOP, Dead Wrong for Democrats
Aside from a select few (including NPR's Senior Washington Editor Ron Elving, who last week said it would be difficult for Obama to repeat his Iowa success), not many political pundits and media-types had predicted that Sen. Hillary Clinton would win last night's Democratic New Hampshire primary.
But it was pollsters in particular who looked bad. Not a single poll taken in the last two days before the primary had Clinton winning. Yet the pollsters were much more accurate about the Republican contest - most showed McCain with a three- to five-point lead. (His current total is 37 to 32 percent.) And they even got totals for Mike Huckabee, Rudy Giuliani, Ron Paul and Fred Thompson correct almost to the decimal point.
The Washington Post's The Trail offers some explanations for the discrepancy in the two sets of results: the "likely voter" modeling, with pollsters perhaps over-counting the boost of enthusiasm among Obama supporters following his victory in Iowa; and independents looked to have opted at the last minute to participate in the Republican primary, depriving Obama of crucial voters.
Maybe it was the role that being the first name on the ballot plays. In previous years, the state rotated candidates names on the ballot from precinct to precinct - in one voting place Obama might be first, in another it would be Bill Richardson. But that was not the model that was used Monday night. For the first time, the names were in alphabetical order in every precinct - so Clinton's name was always first.
"Stanford Professor Jon Krosnick, a survey specialist and expert witness in a lawsuit about ballot order in New Hampshire, has estimated a three percentage point or greater bounce for a big name candidate appearing high on the ballot," write Jon Cohen and Jennifer Agiesta at The Trail. "Therefore, if pre-election polls randomized candidate names, as most do, they would have underestimated Clinton's support by at least three points."
It's an interesting theory, but most polls had Obama winning by far more than three points - although that ended up being the difference between him and Sen. Clinton. Whatever the reason, you can bet there will be lots of discussion about the pollsters' failure in the next few days.
Update: Karl Rove had some interesting observations to make about polling in primaries on Tell Me More.
"... maybe we'd better not look at the recent polls. Maybe we ought to start not by killing the lawyers but by killing the exit pollsters. ... Its' very tough to poll a primary. We've endowed these polls running up to the primaries with a false scientific precision they simply don't have."
9:17 AM ET | 01- 9-2008 | permalink
