IF THIS PRELIMINARY NEP WEIGHT WAS CORRECT, WHY WAS IT CHANGED?

[TruthIsAll]

A total of 122 million voted in 2004.

According to the FINAL exit poll of 13660, 43% voted for Bush
in 2000, or 52.46 million. Only one problem with that: Bush
only got 50.5 million votes in 2000.

The FINAL weighted exit poll was OFF by 2 million votes (4%).

According to the PRELIMINARY exit poll of 13047, 41% voted for
Bush in 2000, or 50.02 million. 

The PRELIMINARY exit poll was MORE ACCURATE than the FINAL. It
was only off by .50 million votes. That's exactly 1%, which
happens to be the exit poll margin of error.

The logical question becomes: If the PRELIMINARY was more
accurate, why was its weighting adjusted from 41% to 43% in
the FINAL?

In the PRELIMINARY:
20.74 million (17%) did not vote in 2000.
Bush won 39% or  8.08 mm of these voters.
Bush won 41% or 50.02 mm of those who voted in 2000.
Bush won 13% or  0.63 mm of the 4% who voted for Nader et al.

So the total Bush vote was 58.73 million, according to the
preliminary poll. His actual vote was 62.028 million.
Where did he get these 3.3 million votes?

In the FINAL: 
Bush won 45% or 9.33 mm of the new voters.
Bush won 43% or 52.46 mm of those who voted in 2000
(IMPOSSIBLE).
Bush won 21% or  0.76 mm of the (revised) 3% who voted for
Nader.

Therefore, the total Bush vote is 62.55 million, according to
the FINAL exit poll (which we have shown to be in error by
4%).

So which do you believe?
Did Bush win 58.73 million votes as determined by the
PRELIMINARY poll of 13047 respondents, and which has been
proven (at least for the voter history weights) to be accurate
to within 1%?

Or did Bush win 62.55 million votes, as determined by the
FINAL exit poll? The FINAL was adjusted to match the recorded
vote and is in error by 4%, according to the adjusted voter
history weights applied to the total recorded vote.

	CNN	7:58pm	11027	Sample		WP	12:22am	13047	Sample	CNN	2:04pm	13660	Sample
     
	Mix1	Bush	Kerry	Nader		Mix2	Bush	Kerry	Nader		Mix3	Bush	Kerry	Nader

		
VOTED IN
2000	Mix	Bush	Kerry	Nader		Mix	Bush	Kerry	Nader		Mix	Bush	Kerry	Nader
No	17.0%	39%	59%	1%		17%	39%	59%	1%		17%	45%	54%	1%
Gore	38.0%	8%	91%	1%		38%	8%	91%	1%		37%	10%	90%	0%
Bush	41.0%	90%	9%	0%		41%	90%	9%	0%		43%	91%	9%	0%
Other	4.0%	13%	65%	16%		4%	13%	65%	16%		3%	21%	71%	3%
	100%	47.09%	50.90%	1.19%		100%	47.09%	50.90%	1.19%		100%	51.11%	48.48%	0.26%

GENDER
Male 	46.0%	51%	47%	1%		46%	52%	47%	1%		46%	55%	44%	0
Female 	54.0%	45%	54%	1%		54%	45%	54%	1%		54%	48%	51%	0
	100%	47.76%	50.78%	1.00%		100%	48.22%	50.78%	1.00%		100%	51.22%	47.78%	0.00%
														
PARTY
ID	Mix	Bush	Kerry	Nader		Mix	Bush	Kerry	Nader		Mix	Bush	Kerry	Nader
Dem 	38.0%	9%	90%	1%		38%	9%	90%	1%		37%	11%	89%	0
Rep 	36.0%	92%	7%	1%		35%	92%	7%	0%		37%	93%	6%	0
Ind	26.0%	45%	52%	2%		27%	45%	52%	2%		26%	48%	49%	1%
	100%	48.24%	50.24%	1.26%		100%	47.77%	50.69%	0.92%		100%	49.69%	49.15%	0.27%
														
												

[tcfrogs]

1. Give it up, TIA

Seriously, I know you're trying.

[TruthIsAll]

2. Seriously, I know you're trying also. n/t

.

[tcfrogs]

3. Your intentions are good

but, jeez, I do statistical analysis for a living, and I just don't buy what you keep feeding us.

[TruthIsAll]

5. You don't "buy" it. That statement is meaningless.

Back it up with more than a curt one-liner, buddy. So far, no one has been able to refute the work of Freeman, Baiman, Simon or 10Phd's at uscountvotes.org - or of mine.

What makes you think you can?

You say that my intentions are good.
Very condescending.

I've already done the work , as have others at DU who have strived to get the facts out in their own area of focus.

Are there any posts of yours available for review? You will have to do a better job than to just toss a one-liner. Take it as a challenge. Are you sure you are up to it?

Are you ready to employ the statistics you supposedly do for a living? Because if you are, we are ready. And it's not just myself who you'll be dealing with, but also some very knowledgeable, no-nonsense DUers who will be more than happy to jump in and put your arguments to the test. Remember, we've heard it all from others like yourself over the past four months. And none have succeeeded in putting a dent in the analysis of myself or the people at uscountvotes.org. On the contrary, naysayers like yourself have only helped by giving us the opportunity to expose the fallacy of their arguments.

Before you enter the ring, I suggest you read the papers from Baiman/Simon, as well as and Freeman and the 10 math/stat PhDs at uscountvotes.org. I assume you are already familliar with my posts; if not, there are hundreds to choose from to commence your attack.

You should probably start by reading this article by a famous mathematics professor at Temple University:
http://www.independent-media.tv/itemprint.cfm?fmedia_id=9953&fcategory_desc=Election%202004

I know you'll need some time to train before entering the ring.

Good luck.

[anaxarchos]

11. There is a paragraph in the paper you cite (Paulos) that speaks...

...directly to the naysays and getoverites on DU:

"Absent any proof or compelling reasons for the differences between the final tallies and the exit polls in the swing states, I don't understand why these gross discrepancies are being so widely shrugged off. After all, the procuring of random samples is far more of a problem for ordinary telephone polls, where the minority of people who cooperate with pollsters presumably differs in some way from the majority who don't. Still, these polls are not dismissed with the same impatient nonchalance as this year's exit polls."

"impatient nonchalance"....

[dzika]

12. c'mon tcfrogs...

You must have seen something so obviously wrong with TIA'a conclusion that you think it doesn't even require a serious response.

There hasn't been much data released by NEP so given what we know, TIA's question seems fair. Why does the margin of error seem to increase in the 2000 voter response characteristic after the exit poll is weighted to reflect the final vote totals?

Or are you saying that the analysis of the exit poll is a waste of time? If so, I would like to know why.

Thanks.

[farmbo]

10. Don't you EVER give up, TIA!

The exit polls are Rove's fingerprints at the scene of the crime. That is why you have been targeted for a subtle framing operation by ne'er-do-wells, including some on this board. Your posts are being increasingly, and aggressively trivialized. And, most recently, your very motives are being questioned.

If a small town police department assigned a couple of detectives to this case, it would have already been blown wide open: tens of thousands of voters in SW Ohio counties who supposedly voted for Bush, then inexplicably voted for down- ballot democrats; fifteen thousand votes (3/4th of which were for Bush) which mysteriously showed up in Miami county AFTER all of their precincts had already reported; and of course...my personal favorite...the Warren County phony terror lock down.

So many coincidences, so little time.

Well, don't despair. Howard Dean has appointed a committee of professionals to examine the Ohio vote and their review will apparently be backed up by the US District Court in Columbus (Judge Sargas).

The GOP still has many tricks up its collective sleeves and in the end, the GOP controls the levers of power. This makes them all that much more fun to beat.

Never give up! Never surrender!

[slor]

4. I believe you...

and the truth is what it is. Thanks for working to show that POS stole yet another election, it will go a long way towards election reform, when the people grow weary of the catastrophe he is leading us into.

[1democracy]

6. How many votes were "changed"?

I like to read your statistics and perhaps you have gone into this before...if so, sorry...
if Bush's votes increased by 3.82 million (if I subtracted correctly), can you tell from your numbers if Kerry's were reduced correspondingly or if there were simply vapor votes added to Bush?

I am sure that there are multiple causes for the shift but I'd like to know the primary modus operandi for the vote fraud.

[TruthIsAll]

7. OK, let's see what happened to Kerry's votes.

In the preliminary poll, based on the 2000 voter breakdown:
Kerry won 50.90%.
Bush won 47.09%

In terms of total votes:
Kerry = 122* 50.90% = 62.1 million
Bush = 122* 47.09% = 57.4 million

Caveat: we are working with weights rounded to the nearest 1%.

[TruthIsAll]

9. Cont.

Final recorded vote:
Bush: 62.028 mm
Kerry:59.027 mm

In the preliminary poll, based on the 2000 voter breakdown:
Kerry won 50.90%
Bush won 47.09%

In terms of total votes:
Kerry = 122.26* 50.90% = 62.23 mm, 3.20mm over his recorded vote.
Bush = 122.26* 47.09% = 57.57 mm, 4.46mm below his recorded vote.

So Bush picked up 1.26mm votes in addition to the 3.20mm switched from Kerry.

Caveat: we are working with weights rounded to the nearest 1%

[wiggs]

8. Something's definitely rotten. Keep going.

The question is....who in authority is pursuing this also? Who is looking hard at the Ohio machines, the Florida polling tapes, the New Mexico provisional ballots?

Is anyone with the Justice Department, who has subpoena power, looking at all the math, testimony, lockdowns, internal memos, etc to put together some sensible tampering story?

TIA, if your suspicions and our hopes are correct -- that Kerry won and Bush is less popular than the MSM storyline would have us believe -- then there MUST BE some principled FBI agents, congressional investigators, journalists, pollster, or political operatives who are pursuing all the smoke to find the underlying fire????

[TruthIsAll]

13. Nederland, we are waiting for your response. n/t

.

[TruthIsAll]

14. Well, tcfrogs? n/t

.

[TruthIsAll]

15. Yes, Mistwell, you are right. There WAS fraud in this election.

As is made very clear by the obvious manipulation of the preliminary exit poll weights for the 2000 voter demographic as shown above.

[dzika]

16. It's pretty quiet in this thread.

Could this be the smoking silence of agreement?

[Nederland]

17. No

We've given up on TIA. Its an agreement we all made. We figured that our posts merely added undue attention.

Best of luck all.

[TruthIsAll]

18. Is it a coincidence or conspiracy that you chose not to respond ...

to this OBVIOUSLY IRREFUTABLE SMOKING GUN?

[kobeisguilty]

19. a lot of talk

for no substantial disagreement beyond a one-liner dissing TIA... if you disagree, then state where you disagree.

[Nederland]

20. Response

I've realized that trying to argue every single little difference I have with TIA is extremely time consuming. Perhaps I'm right, perhaps s/he is right. I suppose time will tell. However, given that I've written nearly a hundred posts in the last five days on the subject, I hardly think you can accuse me of not responding intelligently and stating where I disagree. Therefore I find it unnecessary to restate my positions. Go back and look at my other posts if you really want to know.

Basically I'm satisfied with where things ended up. I got TIA to admit that clustering had an effect on the MOE. That's a start. At this point however, I think that's the best I can hope for at this point. This debate will continue on in the blogosphere and eventually a form of agreement will be reached. That is the real test for TIA--to see if his/her ideas can be sold to the world outside DU.

Best of Luck.

[TruthIsAll]

21. I'm not trying to "sell" anything to anybody. Just doing the math.

Who am I to try to sell anyone on The Law of Large Numbers?
Or the Central Limit theorem.
Or the Normal Distribution.
Or the Binomial Distribution.
Or Monte Carlo Simulation.
Or simple algebra.

No one has to sell anyone on these mathematical truths.

Either you are familiar with them or you are not. Those who are ignorant of them will just argue the talking points fed to them by the MSM, never realizing that they are attempting to refute the irrefutable through misinformation and flat-out lies.

Those who are familiar with them but who choose to argue the same talking points, are just dishonest. Like the MSM.

Those who are familiar and choose to use them can live with themselves knowing that they are using the best tools at their disposal to get at the truth.

[Lucky Luciano]

24. Hey I know about all those things you mention!

I am reading Glasserman's "Monte Carlo Methods in financial Engineering" right now. I just started it.....I had to get through some other stochastic calculus including Ito integrals and Ito's lemma.....Martingale representation theorems, Girsanov theory for the prelims.

[TruthIsAll]

25. I never said that I believed this "cluster effect", no more than

I believe the Reluctant Bush Responder.

So the reluctant Bushies just clustered around and decided not to respond? I guess they had something better to do.

And the eager Dems clustered their way straight to the exit pollsters? I guess they had nothing better to do.

[anaxarchos]

22. "We"?

You mean "you" (collective you - an anti-conspiracy conspiracy?) start 5 threads in a week, throw everything but the kitchen sink, and then leave to avoid "undue attention"?

Hmmmm.... on second thought, wise choice...

[TruthIsAll]

23. Prelim Exit Poll 2000 Bush voter weight (41%) may have been exactly right

The 41% weighting produced a value of 50.0 million for the Bush 2000 vote that was lower from the actual by only .50 million. Yet it may very well have been accurate if in fact 1/2 million fewer 2000 Bush voters voted for him in 2004.

On the other hand, we KNOW that the Final Exit Poll 43% weight is definitely incorrect, because Bush only got 50.5 million votes - and the weighting factor applied to the 122 million votes in 2004 produced a 52.5 million number.

[TruthIsAll]

26. UPDATE: 41% for Bush was MUCH TOO HIGH, due to death and incapacitation.

We must assume that MORE than 0.50% of Bush 2000 voters DID NOT VOTE IN 2004.

HOW MANY PASSED ON OR COULD NOT OR WOULD NOT VOTE FOR BUSH?
SURELY MORE THAN 500,000.

THE NUMBER HAD TO BE CLOSER TO KERRY'S 38%
YET IN THE FINAL EXIT POLL THEY RAISED IT TO 43%!!!!
AND THEY LOWERED KERRY'S TOTAL TO 37%.

WTF?

THE FLAG IS DRIPPING RED.
THE GUN IS SMOKING MORE THAN EVER.

[andym]

27. This is your most important result

This thread represents your most important result.

Pointing out the impossibility of 43+-1% for Bush is important because it clearly shows the final weighted/corrected exit poll is inconsistent/flawed.

And it is consistent with the earlier exit poll being more accurate and the idea that Bush got fewer votes than the official results reported.

But when you raise the possibility that even the earlier (unweighted?) exit poll is also potentially inconsistent, you have provided ammunition for the skeptics who claim that there was error in the exit poll methodology. Because once you show that one aspect of the supposedly good data is impossible, it makes the entire set suspect. I think you should investigate this latter point further before emphasizing it. Perhaps you should include the 1% MOE which adds another possible 1,220,000 to the pool of possible voters who have passed on. Remember its 41+-1%, so at the lower end of the range it's 40% * 122 million, or 48.8 million, which is 1.7 million less than Bush received in 2000.

[TruthIsAll]

28. andym, thanks. It is SO important it deserves its own thread.

.

[TruthIsAll]

29. Everything you say is true, but the key is the change from 41 to 43%

It's the 2% change in the obviously wrong direction to 43% which raises the major flag. The 41% is probably too high as well - but by how much?

The 1% drop in Gore voters from 38 to 37% is another flag, in that it makes the Kerry numbers lower as well in the weighted average.

So we have a few smoking guns here:
1- The nitial 41% is itself suspect and may be 1+ lower.
2-The initial 41%-38% spread is suspect, since Gore won more votes.
3-The move from 41 to 43% is totally without justification.

It would be interesting to calculate the percentage of voters who have passed on since 2000. An initial approximation would be the total percentage of deaths for individuals of voting age. But even this would not account for all those who failed to return to vote.

Its getting curiouser and curiouser.

[kster]

30. Kick...........n/t