National exit poll shows Kerry lead over Bush by 3%

[Time for change]

This information is obtained from a key link from a post in the Election 2004 Forum:
http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=214701&mesg_id=214701

I thought that it would be important to post this information here, because the above post has a massive amount of information on it, and this is an attempt to sum up some of the key points. I also think that this is very cool and interesting information, and deserves wide circultation.

One of the key links is a paper by Jonathon Simons:
http://www.scoop.co.nz/stories/pdfs/PopularVotePaper181...
This paper points out that Kerry's lead in the final national exit poll, with 13,047 respondents, was 2.6%, and that the probability of the discrepancy between this and the official vote tally (2.8% in favor of Bush) happening by chance was close to a million to one. It also provides some of the explanations (but certainly not all of them) why the exit polls are probably a more likely reflection of voter intent than the official vote tally.

Here is the site of the full national exit poll data, updated to 7:33 on Tuesday evening:
http://www.scoop.co.nz/stories/pdfs/Mitofsky4zonedata/U...
It shows Kerry with a 3% lead nationally (rounded to nearest percent), and contains 11,027 respondents.

Here is a summary of the demongraphic groups that helped put Kerry over the top. I have four columns in this table. The first lists the Demographic group. The second lists the percent of the poll sample that contained that demographic group. The third lists the spread for Kerry contained in that Demographic group (for example, if 60% voted for Kerry and 40% voted for Bush, the margin would be 20%). The fourth column lists the vote margin provided by that demographic group. That percent is obtained by multiplying columns 2 and 3 (for example, if females provided 50% of the sample, and their vote spread for Kerry was 10%, then the margin provided for Kerry by females would be 50% X 10% = 5%)

Group-------- % in sample----- Spread for K----- Margin for K
Female------- 54% ------------- 9% -------------- 4.9%
Black-------- 11% ------------ 80% -------------- 8.8%
<$15,000 inc.- 9% ------------ 33% -------------- 3.0%
Union house-- 24% ------------ 23% -------------- 5.5%
Postgrad deg. 17% ------------ 18% -------------- 3.1%
Democrat----- 38% ------------ 81% ------------- 30.8%
Liberal------ 22% ------------ 74% ------------- 16.3%
Moderate----- 45% ------------ 16% -------------- 7.2%
Jewish-------- 4% ------------ 54% -------------- 2.2%
Non-Evangel.- 78% ------------ 20% ------------- 15.6%
Occas relig.- 40% ------------ 13% -------------- 5.2%
No relig at.- 15% ------------ 32% -------------- 4.8%
Not military- 82% ------------- 7% -------------- 5.7%
Not married-- 38% ------------ 22% -------------- 8.4%
Gays---------- 4% ------------ 53% -------------- 2.1%
No gun in h.- 59% ------------ 20% --------------11.8%
Dec. last mo. 21% ------------ 15% -------------- 3.1%