(Bumped -- kos)
[Map of New Hampshire]
In the past I have spent a bit of time looking at the demographics and voting trends of New Hampshire. Though I know a bit about the state, I haven't even picked up a calculator yet, so consider what follows a very rough first draft of an analysis of the town-by-town results of yesterday's Democratic primary in New Hampshire.
Clinton appears to have won yesterday's primary in the field. She also appears to have pulled voters sought by the Edwards campaign. Obama did well in the areas where in 2004 Dean did well, while Clinton's performance appears to have roughly mirrored Kerry's.
Obama did extremely well throughout the Connecticut River Valley (essentially the Vermont border) and the bucolic Monadnock region (east and north of Keene). This region is the most Democratic part of NH-02, which is the slightly more Democratic of the state's two Congressional districts.
The only area of the Connecticut River Valley (the Vermont border) where Clinton appears to have done better is the Charlestown Hinsdale/Winchester area in the extreme SW corner of the state, which is an indication that the Clinton field effort was better, as that area essentially gets no NH television. In fact, they get almost no television, because of the topography and the distances from the towers, so most people have cable or, more likely, satellite. And their radio is more likely to be from Albany NY than from NH. So the voters in those towns were least likely to see broadcast advertising, and the communications they received from the campaigns would have been heavily weighted to mail, phones (which are of much less value in New Hampshire than in most places because of state law prohibiting pre-recorded messages being sent to people on the Do Not Call registry, which is about 90% of the state) and personal contact from campaign field workers.
The college towns and the state capitol all went for Obama, but Durham (NH-01, home of UNH) didn't give him much of a margin. Students were back in town at Dartmouth college in Hanover, and Obama won Hanover with 58%. But his performance in the other college towns—Durham (UNH), Keene (Keene St) and Plymouth (Plymouth state) was good but not astounding.
There appears to have been an education/income gap, with Obama getting the more upscale communities and Clinton getting more downscale communities and Boston commuter towns. In addition to the college towns and the state capital of Concord, Obama also did well in upscale communities like Amherst, Bedford and Bow, and in the wealthy areas with lots of resorts and second homes, like Sunapee, the Lakes Region, and in the ski towns. But Clinton did well in Manchester and the Boston bedroom towns like Salem and Atkinson, as well as the I-89 I-93 corridor towns like Merrimack and Londonderry. In fact, in the Republican-leaning areas that are made up heavily of Massachusetts tax evaders, she beat him solidly.
But the most stunning examples of the socio-economic gap in the electorates was in the most depressed communities of the state, like Claremont and especially Berlin. These are areas where one would think Edwards' economic populism would fare best, but both Clinton (43%) and Obama (35%) swamped him in Claremont, where he drew only 17%. In Berlin, Clinton pulled a whopping 50%, while Edwards only bested Obama by a single point, 23% to 22%.
Outside the college towns and Concord (and the outlier of the SW corner of the state), the denser the population, generally the better Clinton performed, while the sparser the population, the better Obama performed. This suggests that Obama's media and messaging may have been better, but Clinton's field was better. As DavidNYC discussed earlier, there appears to have been a Michael Whouley effect, which may explain why the polling was so off. Looking at the areas where Clinton performed well—predominately downscale areas or the Boston commuter communities—and remembering that the exit polls showed a huge gender gap in favor of Clinton, it appears that the Clinton campaign must have very deliberately and successfully boosted the turnout of middle-class and working-class women. This has always been one of her strongest demographics, and seeing how strongly she performed in the middle class and more economically depressed parts of NH, it appears that was her margin of victory.
The press will be analyzing the debate, or the "cry" and the response to it from John Edwards and Barack Obama, or Bill Clinton's finger waving, and try to figure out what changed in the polling from Saturday to Tuesday. That’s what the press will naturally do, because they can't imagine something they never saw, and that was independent of what was on broadcast and in the papers on Sunday and Monday, could explain why the results were so different than the polls predicted. It's likely that some men who may have voted for Obama shifted to Republican primary to vote for McCain, which would have drawn off some support for Obama. But this win by Clinton doesn't appear to have been brought about by anything else that changed in the last two days. No, it appears at this early stage of analysis that the pieces were in place for this win all along, and that the "secret weapon" of the Clinton campaign was their field program to significantly boost turnout with their strongest demographic, single women and women with less than a college degree.
In the end, one might conclude that were it not for Obama's win in Iowa, Clinton may have won by a much larger margin. But despite Obama's win in Iowa and Edwards' second place finish, Clinton had the votes of many more women already in the bag, and that's what saved her in New Hampshire. Thanks to her field program, she lives to fight another day, and we now know that the campaign will go on at least through February 5th and possibly well beyond.