Why did the pollsters get it wrong?

„Don’t be distracted by the polls, voters“ (via redjar, Flickr)

After first headlines of the “Hillary surprise”had vaporized, another question came to pundits’and voters’ minds: Why was it a surprise? Why did the pollsters get the whole New Hampshire thing wrong? Why did they predict a landslide victory for Barack Obama?
Now, that the dust has settled, theories from various blogs have come up and the pollsters themselves have come in to analyze and apologize (Pollster, John Zogby, Gallup’s Frank Newport and Rasmussen Reports have all had their say).

It looks like the Bradley effect does not apply here, as Mr Obama’s support was rightly estimated. Hillary Clinton’s surge seemingly had a lot to do with her showing an emotional side and, funnily, with her being more on top of the ballot than to Mr Obama.
In Germany, we had similar mispolling (ha, google that word): In 2005, pollsters predicted a landslide victory for the conservative CDU to oust the red-green government of Gerhard Schroeder. But the result showed it winning with a much smaller margin, leading to the grand coalition Germany now has. Two things come to mind:

1) In elections where parties or candidates stand for two different visions of a country (or at least suggest so), voters tend to make snap-decisions (not in a sense they did not think about it, but in a sense they decide very carefully and emotionally on election day). In the New Hampshire case, 34 percent of the Democratic voters had not made up their minds until three days before the election. As our media-driven and highly technological society becomes faster and faster, pollsters have to take into account that more and more decisions might be made shortly before going to an election.

This leads to 2): As the media (blogs included) like to speculate about the outcomes and what they mean, they tend to ignore high indecision numbers. Because – what should they write about them? Race is undecided? Give me a break, who would want to read it if another site has the headline “Will Hillary drop out”hours before the ballot (okay, we know Matt Drudge is extreme). Taking this into account, political future markets cannot claim to be independent – they are subject to spin and polls, too, as Slate has shown.

Will this be the last time the polls get it wrong? Probably not, but media and blogs will take a second looks before saying candidate is doomed (a reason why after New Hampshire people have been careful to cry “Romney’s finished”or “Giuliani is done”). But still – do still expect the campaigns to become less of a football game, where the wild cheering, the spin of tactics and meta-discussions about who will win might get more attention than discussions about subjects and ideas.