Interesting data from exit polls:

[AP]

http://www.msnbc.msn.com/id/3762821/

GENDER
- For all Dean's rolled-up sleeves and aggressive masculinity, he did better with women than men, and for all Edward's hair, he did better with men than women. Dean and Kerry were the only two who did better with women than men.

RACE/GENDER - RACE
- Clark and Lieberman over-performed with non-white males and Edwards underperformed with non-white males, according to one measure. Another table breaking down voters by race says that 2/3 or the non-white voters (5% of turnout) were Latino or "other" (non-AA), but shows that that there was no statistically significant distribution of that vote. Resolve those contradictions as you like.

AGE
- Lieberman over-performed with retirees (the smallest of the four age groups) and Kucinich outperformed with 18-29 year olds.

WEALTH/CLASS
- Clark underperformed among the highest income bracket, and Edwards over-performed. Edwards underperformed among the poorest (but over-performed among the second lowest income bracket). Dean, and Clark to a lesser degree over-performed among the poorest.

ECONOMY
- Only 1/5th of voters thought they were better off today, but Lieberman over-performed in that group. Clark underperformed in that group. Kucinich outperformed among the 36% who thought they were worse off.

EDUCATION
- all Dems underperformed (ie, rounded to 0%) among the 3% of voters who did not finish high school. Dean underperformed and Clark over-performed among people who only graduated from high school. Edwards underperformed among the 1/4 of voters with graduate degrees, while Dean over-performed with that demographic. Lieberman over-performed with college dropouts/people with associates degrees. The difference between having or not having a college degree for Dean was was a 40% swing.

UNION MEMBERSHIP
- Only for Clark was there a significant difference between support among union and non-union households, but it seemed to be with the non-union member of the household. For Dean and Kerry there was a difference in support depending on the voter's membership in the union (non-members were less likely to support them).

MILITARY SERVICE
- Only for Clark and Dean did military service result in significant differences in support (in the predicted directions). Lieberman over-performed and Edwards underperformed in households in which someone other than the voter served in the military. Clark over-performed in households where the voter served in the military.

GUNS
- There was a noticeable, but slight increase in support for Edwards in households with guns, and the difference was bigger than for any other candidate. It looks like Edwards's message about gun safety/protecting second amend. got a little bit of traction.

PARTY IDENTIFICATION
- Dean significantly (and Kerry to a dramatic, but lesser degree) underperformed among the 4% of people who thought of themselves as Republicans and Lieberman significantly over-performed. If Republicans voted for Dean, either they lied in the exit surveys about their politics, or they were significantly outnumbered by those who showed up to vote for Joe. (Those Joe voters almost definitely denied Clark and Edwards delegates, which makes Joe a total ass in my book.)

POLITICAL IDENTIFICATION
- Dean significantly over-performed and Kerry and Edwards underperformed among people who identify themselves as liberal (which, to me, suggests that people are really confused about what constitutes liberal). That considering yourself liberal resulted in a quadrupling of Kucinich's support indicates some voters knew which end was up. Dean's support, and Lieberman's in the other direction, were dramatically consistent with people's perception of their politics. No self-professed moderates or conservatives voted for Kucinich.

FIRST TIME VOTERS
- nobody over-performed among the 5% of first-time voters. This issue was sold as a big one for Dean, but he didn't seem to deliver on it.

RELIGION
- Clark over-performed among non-religious voters (interesting for a man of so many religions). Dean outperformed among "other - Religious" (as did Kucinich) to such a degree that it dragged down Clark and Edwards, but not Kerry. Nobody out-performed among the 4% of Jewish voters, not even Lieberman. However, Lieberman over-performed among regular (once a week) church goers, to the degree that it dragged down Clark and Edwards.

TIMING FOR CHOOSING THE CANDIDATE
- Edwards and Lieberman did best among voters who decided for whom to vote recently. (So there really was some Joementum -- or, the media brainwashed people into acting upon what they were selling.) Dean way outperformed among people who decided a month or more ago (which was slightly more than 1/3rd of the voters). Edwards would have benefited from more time and money, in my opinion -- which I hope isn't the obituary for the Democrats' chances of beating Bush in the fall -- in which case, McAulife's head might roll.

REASON FOR SUPPORTING CANDIDATE
- Dean and Lieberman voters tended to support their candidate because of how they felt about the issues, rather than electability. For Kerry it was, dramatically the reverse.

BUSH HATRED
- Needless to say, the more the voter hated Bush, the more they liked Dean (the reverse being true for Lieberman voters), but only 2% of voters were enthusiastic about Bush.

REPEAL TAX CUTS
- Candidates supporters definitely know where their candidates stand on the Bush tax cuts (repeal all or just the breaks for wealthy), with their support correlating to how they felt about that issue vs how their candidates feel.

CIVIL UNIONS
- Favoring vs opposing civil unions translated to dramatically different levels of support for Lieberman and Dean in the predictable direction (for the rest, it didn't result in a significant difference). So this issue isn't really on the minds of voters except for the less than one-third who voted for Dean and Lieberman.

POPULARITY
- Kerry and Kucinich supporters tend to like Clark and Edwards, the rest don't. Only Kucinich supporters tend to like Dean. No other supporters than Kerry's tend to like Kerry (but Kerry supporters thought very highly of him). Only Edwards and Kerry supporters tend to like Lieberman. Only Kucinich voters tend not to like McCain. Edwards had the best favorable/unfavorable. Dean had the worst.

WAR
- Oddly (since they're almost identical on this issue) Kerry over-performed and Edwards underperformed among strong disapprovers of War. Dean and Lieberman support was predictably polarized on this issue as well.

ECONOMY
- Oddly, Edwards under-performs and Dean over-performed among 1/5th of people who think the economy is poor -- weird since Edwards talked about the economy and Dean talks about war and Bush. It seems that this probably relates to the polarization thing. If you're mad at Bush, you like Dean. Stranger still, Edwards outperforms in the group who are not so worried about the direction the economy is heading (however, this was only 8% of the vote). I can only guess that he did much better with this demographic in IA and that part of the reason he only got 12% was because this is the message which wasn't reaching voters.

MOST IMPORTANT ISSUE
- Edwards outperformed among the 28% of voters (combined) who thought economy/jobs and taxes were the biggest issue. Kerry underperformed on taxes (6% of voters), but over-performed on jobs (22% of voters). The most popular single issue was health care in NH (28% -- almost 5 times more than people who cared most about taxes). Clark seriously underperformed in this group. Clark was very strong among people who cared about war/national security combined. Unlike Clark, Dean split these issues -- under-performing on national security, over-performing on war (predictably -- since his chat has made these issues incompatible). Edwards seriously underperformed on both those issues combined. Dean underperformed on taxes/economy. Oddly, Kerry slightly underperformed on nat'l security/terrorism and war (of course he was strong on these issues, but relative to his winning percentage, these were weak issues for him).

CHARACTERISTICS
- Edwards seriously (and Dean significantly) underperformed with voters looking for experience, but not many voters really cared about that. Edwards over-performed on compassion/empathy and having a good message -- which, combined, a quarter of voters cared about the most, however, individually, those characteristics ranked lower than standing up for what you believe and electability (which, combined, were ranked number one issues by twice as many voters). Voters cared least about ability to shake things up in DC (where Dean way over-performed). They cared most about standing for what you believe, where Dean and Lieberman did well and (curiously) Edwards, Kerry and Clark did poorly. 20% of voters cared about electability (a third less than cared about standing up for what you believe). Everyone underperformed on that one except Edwards, who performed, and Kerry who significantly over-performed. So, it looks like people look at these candidates and think that, if they're doing what they need to get elected, they're not standing up for what they believe. Obviously, there's a ton to be gained from having the message that you can do both, and there will be a ton to be gained if Democrats (ie DU'ers) stop criticizing the candidates in this regard, and if Democrats develop and a rhetoric which undermines the inevitable attempt by Republicans to exploit this dichotomy as a wedge issue. (So, will DU'ers be Republican enablers, or not?)

GEOGRAPHY
- Except for Dean and Lieberman there was almost no geographic differences in support, which is probably why the tally as the precincts were counted remained remarkably steady.

------
So, therein lies a road-map of where the candidates went right, where they went wrong, how their messages were and were not received, and what they need to do to help themselves and hurt their opponents.

Incidentally, whey I say over-perform and under-perform, I'm comparing the breakdown for the individual category to the candidate's performance with all voters. Kerry consistently outperformed relative to all the candidates. So, for example, even if he weren't as popular among Protestants as Catholics, and therefore underperformed among that group, he may still have gotten more protestant voters than all the other candidates combined.

Basically what I'm saying here is, for example, if you were a Lieberman supporter, you were probably old, white, rich, not neccessarily Jewish with very conservative politics, and you like your politicians to say what they believe and not what's popular. If you're an Edwards supporter, you might be a wealthy white guy who cares a lot about taxes but might not think that the economy is going horribly in the wrong direction. And, remember, this data is for NH only. I'm sure that comparing it to data for IA is very revealing.

[NewYorkerfromMass]

1. Dean got the liberals on his 'anti-war' stance

I predict he will win the rabidly Bush hating, highly post college educated, and extremely liberal Upper West Side of Manhattan.

[AP]

3. ...who aren't looking for someone who's strong on national security and...

...don't care about experience.

This data is fascinating.

[NewYorkerfromMass]

5. My pleas to my constituents went unheeded on that point

When the Broadway Democrats endorsed Dean I began jumping up and down in e-mails screaming: "Defense, Defense, Defense!!!" But I was ignored.
I'm still worried Dean will make an impressively strong showing in the NY primary.

See Nov 14: http://abcnews.go.com/sections/Politics/DailyNews/fieldsharpton.html

[Skip Intro]

2. What I don't get about all the stats being thrown around today

is that this data is based on the same exit polling that had, at 8pm last night, the media talking about a very tight race between Kerry and Dean, which turned out to be way off from the actual results. (Unless the exit polls were accurate, which is something else altogether.)

But if these exit polls were so wrong on who people voted for, then why should any other info. from them be treated as accurate?

[AP]

4. I don't think the data is wrong. I think we're learning something

about the story the media wants to sell. They want to make it sound like Dean is still alive.

This data reveals a few other lies in the media today...in stories referring to this same data.

I'm tired right now, but I was listening to Juan Williams while typing this and he was saying things that directly contradicted what I was typing at that exact moment while referring to the exact same thing I was reading. I wish I could remember what it was that he said...I'll have to wait until the next hour and listen again.

[Lydia Leftcoast]

6. I caught a possible contradition:

Oddly, Edwards under-performs and Dean over-performed among 1/5th of people who think the economy is poor -- weird since Edwards talked about the economy and Dean talks about war and Bush.

MOST IMPORTANT ISSUE
- Edwards outperformed among the 28% of voters (combined) who thought economy/jobs and taxes were the biggest issue.

But fascinating, nonetheless.

[AP]

9. It suggests a lot of people who are doing OK still are troubled by the

economy, and a lot of people who aren't doing well are not troubled by the economy -- or at least aren't looking to the government to solve their problems wiht a fairer tax code and program to create more, better-paying jobs.

There is a very good reason for this. Many people who have professional jobs do better year over year, no matter what. It's just a natural progression -- you know more, you do your job longer, you get raises, you get promotions.

Some people realize that they're on a trajectory, and even thought they're always gong up, the government actually can take them off that upward trajectory in either direction. For example, say you expect to make 4% more every year. A Democrat creates and economy that gives you 6%. You're really happy. Well, what if Bush takes 2% of your 4% and gives it to his cronies. Some people think they're just doing better becuase of themselves. Others realize they've been robbed.

Edwards is appealing to people who got 2%, feel like they should have gotten 4%, and think Edwards is going to create a society which gives more to the middle class, ie, 6%.

The other things that ties into this is that Republicans create a perception among the working class that they're cowboys -- if you do well, it was indivualism that got you there. If you didn't do well it's becuase blacks, women and immigrants were handed something that you think you were entitled to without having to work as hard as those other people to get. Those people might suffer, but don't look to the government to give them a better society and economy. Those people don't look like they voted for Edwards in NH. They think the economy isn't that great, but they don't think it's becuase of the tax policy or for any reason that relates to the government creating more jobs.

[CWebster]

7. "No other supporters than Kerry's tend to like Kerry "

could pose a problem. If they don't like him now, what happens when the shine starts to fade and the baggage starts to get aired? Kerry support pivoted on contrasting with Dean--who had taken so many hits as the front runner, as preferrable in electibility. What happens when his liabilities are blared all over the place and people second guess their reactionary decisions?

The only way you can entrap people to vote for Kerry is to vote against Bush, and that is a sad commentary on the best the Democratic party has to offer.

[THUNDER HANDS]

8. if you notice, Kerry supporters tend to like everyone but Dean

Kerry supporters have higher opinions of other candidates. Dean supporters don't like anyone but Dean, Clark and Edwards supporters don't like anyone but maybe Lieberman.

Kerry supporters like Kerry and everyone else....but Dean.

[molly]

15. Hmmmm - I think you are correct!

[spooky3]

18. As an Edwards supporter, what you say about them isn't true for me,

and since I haven't systematically studied everyone, I don't know if it's true of anyone. Let's please not make sweeping generalizations we can't support.

[PROGRESSIVE1]

20. I'm a Clark supporter and I like Kerry, Edwards, Kucinich, and

Sharpton. I do not like Lieberman and Dean.

[AP]

10. Since Kerry supporters tend to really like him, you can imagine the sort

of polarizing war that DU'ers (like me) had with Dean, except with Kerry the fights will be over whether he's personally likeable.

I think we've already seen the beginning of this.

[Gin]

12. Is it possible that these people didnt tell the truth....since your vote

is a private matter.

[AP]

13. I think this data is generally very accurate.

[spooky3]

17. it is, but why agree to participate if you don't want to tell the truth?

[AP]

11. kick

[AP]

14. I just realized where the Joementum came from:

Non-jewish conservative voters who attend church once a week?

That sounds like those right-wing religious fanatics. Probably the same ones who are looking foreward to the war to end all wars in Israel (what's that called? the rapture?).

I just bet they were all told to go out and vote for Joe by ministers taking their orders from DC Republicans.

I guess we can all thank Gore for giving us Joe.

[windansea]

16. very good info AP

thanks!

[cindyw]

19. This part makes no sense.

"No other supporters than Kerry's tend to like Kerry (but Kerry supporters thought very highly of him). "

Howe can this be true. 40% of the people voted for him. So given this comment, you would say that only that 40% would have favorable opinion of him. BUT they also said that among this total group they found a 75% favorable rating.

Anyone want to explain this to me?

[AP]

21. Kerry's favorables were 72:25; 51/72 of favs came from his own supporters

and, as voteres will act like morons, 1% of Kerry's voters have an unfavorable opinion of him (perhaps motivated to vote for him only because of electability).

Spread among the voters of other candidtes is the remaining 21% favorables and 24% unfavourables. And among voters for each other candidate, there were more people who thought of Kerry unfavorably then favourably.

One obvious explanation for this is that, becuase Kerry was in the lead, the supporters of all other candidates looked at him as the person standing in their way. Nonetheless, I don't think there was any other candidates who was more loathed than liked by the supporters of each and every one of the other canididates' voters.

[spooky3]

24. A hair split on the data interpretation, sorry, AP

The tables are quite confusing, but I think this is the way that they add up, or I may be misinterpreting what you said. (Basically I think you are right in the conclusion). Your analysis is extremely insightful. JE ought to have you on board as a paid consultant!!!

The numbers to the right of the 72% favorable add up to 100% (approx), not to the 72. So, I interpret the %s under the Kerry column to mean that 51% of the 72% who are favorable are Kerry voters. So only about 37 percentage points of the 72% are Kerry voters. There are 35 percentage points left rather than 21. So 35% of the other candidates' voters are favorable toward Kerry, but that means that 65% of them are unfavorable toward Kerry.

In contrast, people who did not vote for Edwards were much more favorable toward him than were people who did not vote for Kerry were toward him. Otherwise, we would have gotten a favorable vote for Edwards that was much lower overall than the 73% that he received, since fewer people voted for him yet his overall favorability rating was a touch higher than Kerry's.

My interpretation, which requires some speculation and info from the other questions, is that (a) Edwards is a strong second choice because many people who did not vote for him have a favorable view of him, much more so than Kerry and perhaps other candidates (I didn't look at their numbers as closely); (b) primarily because of the regional familiarity with and liking for Kerry (and to some extent Dean), many people made their minds up a long time ago to vote for one or the other of them, but would be happy with Edwards too.

What do you think?

[AP]

26. oops, yes.

Should have caught that. Thanks for keeping me on my toes. (Which you're doing alot!!!)

I wish I caught that earlier so I could go back and edit it. Oh well. It was the middle of the night when I wrote this!!!

Without looking at the fav/unfavs for candidates other than Kerry, what you say about second choices sounds right.

As for long term-commitment & regional preferences, 1/4 of voters picked their candidate more than a month ago, and, for 71% of them, that was Dean and Kerry. 40% from a week to a month, and most of them were for Kerry (with Dean having a steep drop off in that time, as the polls showed). But for people making up their minds in the last three days, although the edge goes to Kerry, Dean and Edwards picked up, Dean in proportion to his final results, and Edwards doing better than his final results. So, if you knew them and liked them months ago (an advantage for Kerry and Dean) you were ahead of the game with these voters.

[spooky3]

27. Actually, AP, you shouldn't have been so kind as to agree with me.

I looked at the poll results again and decided I was wrong, at least about the 35%/65% split. But I still think the conclusion was basically right.

[bigtree]

22. John Kerry looks good in the Gallup analysis

http://www.gallup.com/poll/releases/pr040127.asp

Gender

Kerry is the top candidate among both men and women likely voters, leading Dean by 11 points among men and by 13 among women. Kerry does slightly better among women (39%) than men (34%). Clark and Lieberman are both polling better among men than women.


Age

To date, the conventional wisdom (as well as the finding of many polls) is that Dean has the market cornered on the youth vote. Indeed, Dean cites the enthusiasm of his young supporters in Iowa as the reason for his now-famous scream in his concession speech there. In New Hampshire, however, Kerry (35%) gets a larger share of the vote among 18- to 34-year-olds than does Dean (32%), though the difference is within the margin of error. Dean's support is not as great among older voters in New Hampshire, while Kerry scores similarly among all age groups. As a result, Kerry has a fairly strong lead among middle-aged voters (he leads Dean by 10 percentage points) and older voters (he leads Dean by 17 points). Edwards does considerably worse among the youngest age group (getting just 6% support) than he does among voters aged 35 to 54 (12%). Clark and Lieberman, like Kerry, score about the same among all age groups.


Education

Dean's national lead in fundraising is often attributed to his appeal to highly educated, wealthier Democrats. In New Hampshire, Dean does considerably better among those with a postgraduate education (29%) than among those with a four-year college degree (19%) and somewhat better among those who do not have a college degree (24%). But again, Kerry does just as well as Dean does among postgrads, and much better among those with less education than that. Kerry leads Dean by 22 points among college graduates and 14 points among non-college graduates. Edwards receives a larger share of support from college graduates than postgraduates or non-college graduates.


Ideology

Some of the largest subgroup differences in primary voting this year are found among ideological groups. Even so, New Hampshire voters of all political persuasions support Kerry at high levels. Other candidates' support is more related to ideology. Specifically, 40% of self-identified liberal New Hampshire voters are supporting Dean, while only 3% support Lieberman and just 7% support Clark. Among self-described conservative primary voters (which amounts to only one out of six primary voters), 24% support Lieberman and 16% support Clark, while just 8% support Dean, placing him fifth in this subgroup. Moderate voters' candidate preferences generally mirror the results of the overall ballot, though Clark tends to do somewhat better and Dean somewhat worse among this group.


Party Registration

The New Hampshire primary is a "semi-open primary," meaning residents who are not currently registered to vote, or who are registered as independents, may vote in the primary (Republicans may not vote in this year's Democratic primary). As such, significant numbers of independents often vote in the presidential primaries in New Hampshire, and Gallup's polling suggests that a larger number of independents may vote in this year's primary than has been the case historically. Again, the data show Kerry leads comfortably among both groups. If New Hampshire were a closed primary state (in which only those currently registered as Democrats would be able to vote), Clark and Lieberman would probably not be faring as well as the polls indicate. If larger numbers of independent voters go to the polls as predicted, that would likely mean higher levels of support for Clark, Lieberman, and Edwards.

[AP]

23. I just noticed a weakness for Dean in rest of country:

MOST IMPORTANT ISSUE:

I've read before that health care isn't a high ranking issue nationally. Jobs and economy is (especially in the south -- perhaps the NH economy isn't doing so poorly?). Oddly, Health care was the most important issue for voters. Dean did well with these voters, but not so well on the jobs/economy issues.

I don't think Dean's going to see another state for a while where they rank his strong suit so high. Clark, however, looks like he'll get a bump in places where health care isn't as big a concern as in NH, but he might be in trouble if that concerns is replaced with concerns about the economy and jobs rather than with war/terror.

[The Magistrate]

25. Worth Keeping Toward The Top, This

[progressivebydesign]

28. Choosing for "electability" is nonsense!! Don't do it!!

There is no electable or non-electable. That's utter bullshit. No one can predict, truly, who will do well against Bush. Until Rove & Co. release their hounds on the eventual nominee, will we see who is "electable" is who is not. Polls that try to predict that are useless at this point. The economy, the security, the mood, of America in November cannot be judged now. Too much can happen. The average voter doesn't know jackshit about who would beat Bush..

Vote with you heart in the primaries. It's one thing to support a candidate for the primary that believe in, and writing in the name of someone besides the Democratic nominee in November. I can make one absolute guarantee today. The person winning the nomination will be a Democrat. We need one Democrat and one Republican to have the election in November. If our nominee wins in November, then I guess he was "electable".

I refuse to be pushed into supporting someone I would never have supported in the first place, in a primary, just because some corporate politician types and the freakin' media tell me he's "electable". That's nonsense.

[AP]

30. I disagree. I think there are way to play out the arguments

and come up with great arguments about electability. What's harder to predict would be unforeseen circumstances -- if dems nominate someon overly identified with an issue with Bush's control, can/will bush remove that issue (as Nixon did with Vietnam vis McGovern)?

I think the intereting thing about NH and electability is that a bunch of NE"ers thought a liberal NE buddy of Ted Kennedy who has been embraced warmly by every liberal single-issue special interest group will be electable.

I think you can play out that argument and see where it will fall apart.

[LiberalBushFan]

29. Wow another instance of media whoring revealed

CNN was using the percentage of votes from veterans to claim that Wes Clark was unpopular with them, by failing to compare these percentages with overall results (Clark in fact did better with veterans that the general population). Then they talked about Kerry, who did not overperform with veterans, explaining how he did a great job of appealing to them by bringing with him people whose lives he saved, etc (yes, lives, now it's multiple people???).

[AP]

31. I could be wrong, but I think what you can gather from that

is that if Kerry gets absolutely higher numbers of veteran voters than Clark, it's obvious that he's doing a better job in some respect in terms of getting these voters. But if Clark is getting a strong response from veterans, in that they make up a larger percentage of his own voters compared to all voters drawn to him, there is something about his persona which is working more than other things with these voters.

You can't really argue that, if all these candidates received the same number of votes, you would have seen more military people going to Clark, and that they're coming from Kery, because the point is that all things aren't equal -- Kerry's getting almost 250% more voters than Clark, and twice as many military votes, absolutely.

But, you can argue that being in the military is something that leads military voters to lean towards Clark more than other factors, and it might be something he should develop if he wants to take military votes from Kerry, becaue military voters seem LESS interested in Kerry than when compared to all voters combinded are interested in Kerry. And, in fact, the only reason Kerry isn't getting these voters in or out of proportion with his overall percentage is becuase of the presence of Clark

The other thing this breakdown tells you is what, if your candidates' problem is just that people don't know your candidate yet, what you might expect to find if they do. If everyone knows your candidate, there's no much you can do once you know these numbers.

These numbers also let you get a sense of how your candidate will perform in states with different demographics, and circumstances. For example.If you go to a state with a big military vote, Kerry is going to get a big percentage of that vote, and Clark will take much of the remainder, (but so will Dean because he got more military votes that Clark).

To summarize: I'm afraid that I might be misleading people with the way I looked at these numbers. I'm not trying to figure out so much, "if you're in the military, who are you more likely to vote for?" -- the answer to that is Kerry, Dean, and then Clark, with Edwards not being too far behind Clark.

What I wanted to think about is more like, "if you support {X candidate} what are you more likely to look like?" Ie, within the limited universe of people who support the candidate (without regard to how big or small that sample is) what kinds of people are represented. Clark had a significantly smaller universe than Kerry, but within that universe, it looks like relatively more people had, for example, military experience. When the universes are of very different sizes, these inter-universal comparisons might overstate matters.

Confusing?