Confirming the Baiman/Simon National Exit Poll Probabilities

[TruthIsAll]

You may have seen this post before....

This is the original Baiman paper:
http://www.freepress.org/images/departments/997.pdf

This is the latest Baiman/Simon paper (from yesterday):
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x211821

I see that they are  using Excel's normal distribution
function. Wonder where they got that from?

The odds are overwhelming that Kerry won the popular vote.
My numbers agree with Baiman. For Kerry's -2,35% deviation
from the exit poll, I calculate odds 1 of 23 million; Baiman
gets 1 of 25 million for Kerry.

For Bush's 3.01% deviation, it could be anywhere from:
 1 in 355,000 (1.30% MOE) to 1 in 280 billion (0.86% MOE).
Take your pick. 


National Exit Poll Sample

Total sample size: 13,047
Margin of error: 0.86%			
					
Weighted:
Female:54%	
Male: 46%			
					
Female	Male		
Kerry:54%	47%		
Bush: 45%	52%		
					
	Exit	Actual	Diff	 Predicted Odds
Kerry 50.8%	48.43%	-2.35%    50.78% 1 out of 23 million
Bush	48.4%	51.23%	3.01%	     48.22% 1 out of...aw forget it.
Diff	2.4%	2.8%	 -5.36%	    2.56%
					
KERRY			TIA	TIA	Baiman
MOE	StdDev	Kerry Prob	Odds	       Odds
1.00%	0.0051	2.03E-06	491,047	 783,054
1.30%	0.0066	1.96E-04	5,083	         NA
0.86%	0.0043	4.33E-08	23,078,528	 25,507,308
1.12%	0.0057	1.87E-05	53,406	 55,499

					
BUSH					
MOE	StdDev	Bush Prob	Bush Odds	
1.00%	0.0051	1.80E-09	554,926,377	
1.30%	0.0066	2.81E-06	355,135	
0.86%	0.0043	3.55E-12	281,703,861,098	
1.12%	0.0057	6.44E-08	15,509,423	

[GregD]

1. you are back!

Oh thank goodness.

I can't believe the crap going on these days. You have worked so hard. I'm so glad you are back among us TIA.

[nashville_brook]

3. no doubt! so glad you back!

and this new info is great!

[TruthIsAll]

2. The Final Exit Potes:Pick an MOE, any MOE and Kerry ALWAYS wins.

The final National Exit Poll sample (13047) gives a calculated 0.86% MOE.

The 11,000 sample-size from the CNN site are cleaner than the final numbers which were apparently Bushwacked to match the actual "votes".

Can you believe it?

Did you EVER hear of this kind polling practice? How can NPR justify it with a straight face? Do they think we are that stupid? Why the hell take an exit poll if you're just going to sol it with actual votes?

Is it an EXIT POLL or an EXIT POTE?

What is the statistical justification for this?
They "massaged" exit poll data with "votes" late in the exit poll count (the last 2000).

Even with contaminated exit polls, arbitrarily inflated for Bush, Kerry STILL comes out ahead.



tia

[davidgmills]

4. Could you comment on my post

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x215658


If I have spotted something, it may help us even more.