Another National Exit Poll Mystery: The Gender Vote. Can you solve it?

[TruthIsAll]

Consider these facts:
In 2000, according to the Final Exit Poll:
1) The Gender split was 52% female/48% male.
2) Gore won the women's vote: 54-43%!
3) Gore won the popular vote.

Flash forward to 2004:
4) The Gender split widens to 54% Female/46% Male.
5) Kerry wins the Preliminary Exit Poll women's vote: 54%/45%

What you might expect, since women are a solid Democratic
block.
But no... 

6) In the Final Exit Poll (13660)Kerry's numbers are
"adjusted"  to match the recorded vote, 51-48%.
Sound familiar? 

To add insult to injury, the Gender split stays the same:
54%/46%.

What happened between 12:22 am (13047 respondents) and 2:05 pm
(13660 respondents)? Did women decide to abandon Kerry? That's
possible. But John Edwards? 

 
		NEP: 2000 Final vs. 2004 Prelim vs. 2004 Final 

2000 Final					Prelim: 12:22am		Final: 2:05pm			
						13047 respondents		13660 respondents			
														
	2000  				     2004				2004			
Gender	Mix	GORE	BUSH	NADER		Mix	Bush	Kerry	Nader		Mix	Bush	Kerry	Nader
Male	48	42	53	3		46%	52%	47%	1%		46%	55%	44%	0
Female	52	54	43	2		54%	45%	54%	1%		54%	48%	51%	0
	100					100%	48.22%	50.78%	1.00%		100%	51.22%	47.78%	0.00%
														
														
														
AGE														
18-29	17	48	46	5		17%	43%	56%	1%		17%	45%	54%	0
30-44	33	48	49	2		27%	50%	49%	1%		29%	53%	46%	1%
45-59	28	48	49	2		30%	47%	51%	1%		30%	51%	48%	0
60+	22	51	47	2		26%	51%	48%	1%		24%	54%	46%	0
	100					100%	48.17%	50.53%	1.00%		100%	51.30%	47.96%	0.27%
														
Education														
No H.S.	5	59	39	1		4%	47%	52%	1%		4%	49%	50%	0
H.S.	21	48	49	1		22%	48%	51%	1%		22%	52%	47%	0
College	32	45	51	3		31%	51%	47%	1%		32%	54%	46%	0
CollGrad24	45	51	3		26%	50%	48%	1%		26%	52%	46%	1%
PostG	18	52	44	3		17%	40%	58%	2%		16%	44%	55%	1%
	100					100%	48.05%	50.21%	1.17%		100%	51.14%	47.91%	0.43%
														
Income														
<15	7	57	37	4		9%	33%	66%	1%		8%	36%	63%	0
15-30	16	54	41	3		15%	39%	59%	1%		15%	42%	57%	0
30-50	24	49	48	2		22%	47%	52%	1%		22%	49%	50%	0
50-75	25	46	51	2		23%	53%	45%	1%		23%	56%	43%	0
75-100	13	45	52	2		13%	50%	49%	0%		14%	55%	45%	0
100+	15	43	54	2		11%	53%	45%	2%		11%	57%	42%	1%
	100					4%	53%	47%	0%		4%	58%	42%	0
						3%	58%	41%	1%		3%	63%	35%	0
PartyID														
Dem	39	86	11	2		38%	9%	90%	1%		37%	11%	89%	0%
Rep	35	8	91	1		35%	92%	7%	1%		37%	93%	6%	1%
Ind	26	45	47	6		27%	45%	52%	2%		26%	48%	49%	3%
	100					100%	47.77%	50.69%	1.27%		100%	49.69%	49.15%	1.16%
														
IDEOLOGY						Mix	Bush	Kerry	Nader		Mix	Bush	Kerry	Nader
Lib	21	80	13	6		22%	12%	86%	1%		21%	13%	85%	1%
Mod	50	52	44	2		45%	41%	57%	1%		45%	45%	54%	0
Con	29	17	81	1		33%	82%	16%	1%		34%	84%	15%	0
	100					100%	48.15%	49.85%	1.00%		100%	50.83%	47.95%	0.22%
														
FirstTimeVote														
Yes	9	52	43	4										
No	91	48	48	2										
	100													
														
Voted for														
Clinton		82	15	2	Gore	39%	8%	91%	1%		37%	10%	90%	0%
Dole		7	91	1	Bush	41%	90%	9%	0%		43%	91%	9%	0%
Perot		27	64	7	Other	3%	13%	65%	16%		3%	21%	71%	8%
Other		26	52	15										
None		44	52	3	No	17%	41%	57%	1%		17%	45%	54%	1%
					Total
Vote	100%	47.38%	50.82%	1.04%		100%	49.49%	50.10%	0.41%
											
WHEN DECIDED		
Today						6%	40%	53%	5%		5%	45%	52%	1%
3days		48	46	5		3%	41%	53%	4%		4%	42%	55%	1%
week		48	44	5		2%	51%	48%	1%		2%	51%	48%	0
month		49	45	5		10%	38%	60%	1%		10%	44%	54%	1%
Before		48	50	1		79%	50%	50%	0%		79%	53%	46%	0%
						100%	47.95%	51.23%	0.54%		100%	51.25%	47.47%	0.19%
														
RELIGION	
Prot		42	56	2		53%	56%	43%	1%		53%	59%	40%	0
Cath		50	47	2		27%	49%	50%	1%		27%	52%	47%	0
Jewish		79	19	1		3%	23%	77%	0%		3%	25%	74%	0
Other		62	28	7		7%	20%	75%	4%		7%	23%	74%	1%
None		61	30	7		10%	29%	70%	1%		10%	31%	67%	1%
						100%	47.90%	50.85%	1.18%		100%	50.77%	47.99%	0.17%


 
 

[LatePeriduct]

1. Let's see I'll take a spin at it...

Well umm a freeper would simply say.

Isn't it obvious? The NEP over-sampled the wrong gender in every single precinct! It was just a simple polling problem.

To which my uhh response would be...

Well if it was a polling problem, how come the over-sampling seems to be the exact same as the original sampling taken only now its favoring the republic side exclusively?

I'm guessing they would say "It's magic, these things just happen!"

To which I would remind him/her that it happened in basically every precinct. If its magic, it has to be jesus. Jesus must have done this himself.

Or to be more accurate, someone purged a whole ton of voters out of the rolls who might have represented a gender/affiliation/whatever you call it. On top of that it added more numbers to the republic side along with all the miscounted votes. That makes alot more sense. A switcheroo.