Yesterday, Hillary Clinton won the New Hampshire Democratic Presidential primary, beating Barak Obama despite the fact that the final opinion polls conducted before the election showed Obama with quite a large lead. I have to admit that I experience not a little bit of schadenfreude when pollsters experience discomfiture, largely because I think that polls have taken an outsized place in the media's coverage of events and in way that public officials shape policy. Overemphasis on poll data turns elections into horse races where legitimate differences in policy among the candidates are ignored, and instant opinion polls discourages politicians from actually attempting to lead.
But complaining about polls is like the Pope issuing a bull against the comet. They are here to stay, and we might as well learn what we can about them. If you wonder what in the heck happened with the polls in New Hampshire and you're not satisfied with the shoot-from-the-hip glib garbage that most pundits spit out, be sure to read The Mystery Pollster. He's a real live pollster named Mark Blumenthal (joined by others now) who began blogging immediately after the 2004 Presidential election in response to an offensive amount of ignorant bloviation from many quarters about the significance of exit polling, and he has been a reliable producer of quality analysis about the meaning, strengths, and weaknesses of various kinds of polls. He certainly does have a quantitative bent, but I think that he's perfectly comprehensible for even the most math-challenged reader. Check him out early and often in this election season.