Jan. 10

Senate map with polls Downloadable polling data Previous report Next report News   Primaries/caucuses Already completed Before February 5 February 5 After February 5

Welcome	FAQs	Donations	Icons for bloggers	Previous elections
Electoral coll. graph	Poll graphs	Cartogram	Animated maps	Data galore
All Senate races	Hot House races	Political websites	Articles and books	Political humor

News from the Votemaster

How could the polls have gotten it so wrong in New Hampshire? Pollster John Zogby has a couple of explanations over at Huffington Post. First, 18% of the voters didn't decide how to vote until primary day. This is an extremely large number. Second, The voters don't like coronations. When Hillary Clinton was the favorite going into Iowa, the voters said Whoa!. Then when Obama looked like the favorite going into New Hampshire, they again said Whoa! Third, The period between the two elections was too short for the usual bounce to take effect. Fourth, Clinton began gaining Sunday evening, but the rolling average polls weighted the earlier days too heavily. Fifth, there was a large turnout of older women, who are very pro-Clinton. Sixth, many independents thought that the Democratic race was a done deal for Obama, so they chose to vote in the Republican primary for McCain. He added that her debate remark that electing a woman represented a huge change for the country may have motivated many women to support her.

Frank Newport of the Gallup poll has a different story. He says that there are two reasons polls can be wrong. First, the pollsters screwed up on methodology--they didn't interview enough high-income black women or interviewed too many young left-wing evangelicals or otherwise didn't have a good random sample. It's sloppy, but it happens some time. He doesn't think that was the case because all the polls said the same thing. Maybe Gallup fouled up, but how would that explain ARG, Zogby, Rasmussen, and the University of Hampshire having virtually identical numbers? Everybody forgot to call elderly women? Unlikely. Besides, everybody got the Republican primary right on the nose. If some major demographic group was missed, surely that would have wreaked havoc with the Republican polls too, no? The second possible explanation is that many voters did not decide until Monday or Tuesday and were possibly influenced by the debate Saturday and Hillary's almost crying Monday. Some women may have thought "Maybe she's human after all."

Charles Franklin at Pollster.com has some nice graphs showing how much the pollsters missed by and in which direction. What is interesting is that the scores for Obama were right on. What all the polls missed was a sudden last-minute spurt by Hillary Clinton.

We tend to agree with Newport and Franklin. The Iowa polls were right on the dot and the Republican primary in New Hampshire was accurately predicted and the Obama scores were on target. The only plausible explanation is that a lot of voters didn't decide which ballot to take (D or R) and who they were going to vote for until Monday or Tuesday and these late deciders were women who went strongly for Hillary Clinton.

So what happens now? The next primary is in Michigan next week, but it is likely that the Democrats will boycott it since its date violates party rules. On the Republican side, probably Romney, McCain and Huckabee will participate.

Gov. Bill RIchardson is expected to withdraw from the race today after dismal results in Iowa and New Hampshire.

We have two new polls for Florida. On the average, Clinton is leading Obama by 10% there, and that was before her victory in New Hampshire. On the Republican side, one puts Huckabee ahead by 8% and the other puts behind by 5%. Clearly it is unsettled there. A South Carolina poll has Obama ahead by 7% and Huckabee ahead by 12%, but that will change shortly as the New Hampshire results sink in.

The polling results for all states are available as a Web page and in .csv format.

WWW www.electoral-vote.com