Problems with "New, Improved Analysis of NH Democratic Primary Results."

I'd post this as a reply, but I think it's important enough to be in its own thread, so we don't get caught in a hypestorm. That thread claims:

Percentage of Machine Counted Votes - Percentage of Hand Counted Votes

05.46% Clinton
00.00% Gravel
00.00% Other
-0.60% Kucinich
-0.72% Edwards
-1.24% Richardson
-2.89% Obama

That's an 8.35% difference between hand counted & machine counted votes when comparing Clinton to Obama.

That looks pretty damning...except we're looking at completely different voting districts. Rural areas are more likely to have hand-counted votes, and urban areas are more likely to have machine-counted votes. And, according to the CNN exit poll (which, it might be stated, was "hand-counted") city-dwellers broke for Clinton 43/35, suburbites for Clinton 42/35, and rural voters went for Obama 39/34. It isn't that the machines led to the difference in voting, it's that rural areas were more likely to vote Obama and were more likely to have hand-counted ballots.

http://www.cnn.com/ELECTION/2008/primaries/results/epolls/index.html#NHDEM

And failure to look at that kind of information is why we shouldn't draw conclusions based on half-baked analysis.

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=132&topic_id=4009174&mesg_id=4009174

People crying these figures as 'proof' of machine-based tampering are, to put it kindly, displaying an extraordinary lack of political sophistication.

What I want to know is
1) How in the hell Obama has the most educated people voting for him, and they all live in rural areas, make more money than Hillary people, and are young?
2) How in the hell did Obama get weekly church goers and secular voters voting for him, but not people who go to church occasionally?

Both those things are counter-intuitive.

Say, in the cities, Obama has the educated/rich/young, but everyone else goes for Clinton. That way, Clinton wins the cities, and Obama wins those groups. In the rural areas, you're more likely to have an even level of education (and when you don't, there still isn't the socioeconomic divide that exists in the cities), so that's largely a wash.

As for #2? The storyline I see from that and other figures is that Obama appeals to two groups in New Hampshire: the leftists who are sick of Clinton's centrist-liberalism, and the centrists/right-wingers who are sick of Clinton's Democratic ultra-partisanship.

Or, think about it this (highly spun) way: the people who truly care one way or another--the people who have made their decision on their principles and stick with it--are attracted to Obama. The people who tepidly shuffle along go for Clinton.