Jack Rabbit's response to Nederland and TruthIsAll

[Jack Rabbit]

Refences:
THIS EXIT POLL SIMULATION TEST BLOWS AWAY THE NEDERLAND "PROOF"
Mathematical Proof: TIA is Wrong (Part 1)
Mathematical Proof: TIA is Wrong (Part 2)
The Unexplained Exit Poll Discrepancy: Explained

Mr. Nederland and Mr. TruthIsAll have provided the friendly readers of Democratic Underground with a lively discussion concerning the 2004 election and the science of statistics. TruthIsAll, who has a background in statistics, says that the election was stolen and he can prove it. Nederland, who has a degree in philosophy, claims that TruthIsAll's arguments are flawed and don't prove anything, although he remains open to the idea that something was amiss.

TruthIsAll is showing that exit polling falls outside the margin of error (MOE) and he trusts the exit polling more than he trusts the vote counters. He has presented numerous bits of data and trial runs to support his hypothesis.

Nederland has advanced two propositions in an attempt to show that TruthIsAll's case is a house of cards:

• Regardless of your sample size, if your sample is not representative of the total population, exit poll results can be outside the MOE; and
• When fraud occurs, it is impossible to determine whether or not an exit poll sample was representative.

The first of these is almost a self-evident truth; in fact, even if the sample size is sufficient and the sample representative, the resulting data still may be outside the margin of error, although it would be less likely.

The second of these is true, but only to a point. If fraud has occurred, it would be impossible to tell if an exit poll has occurred is true, but only in the location where the fraud has occurred. That is an important distinction to make.

Nevertheless, I agree with Nederland's point. TruthIsAll can run tests and trials until doomsday and all it will show is that there is some discrepancy between the exit polls and the actual count. In and of itself, it won't give me or anybody else a reason to suppose the exit poll is right and the vote count wrong. Unless there is some control group introduced to gives us more confidence in one or the other, it would be impossible to show fraud.

Let's get down to brass tacks. What we want to know here is whether or not the election of 2004 was stolen. How can we prove that it was, or at least raise some red flags?

The hypothesis is that fraud occurred in certain types of precincts, namely, those which used certain types of voting technology. We could even narrow it down to precincts with certain types of voting technology and a history of favoring for certain types of candidates. It may not be possible to show fraud occurred by looking only at those areas where it is suspected to have occured, or even less by looking at the aggregate data, but by looking beyond those areas where fraud is suspected and comparing them to places where it is not, one may show that something is quite probably very wrong.

What can be shown, and what TruthIsAll's data does not show, is that in areas where a certain voting technology was used (electronic voting with no paper trail or optical scanners) that there was a significantly wider discrepancy between the exit polls and the actual count than in areas that other types of voting technology; that this discrepancy consistently favored one candidate (Bush) or group of candidates (Republicans). We could also narrow the search and confine it to types of voter technology in precincts with a history of favoring certain types of candidates (Democrats). Furthermore, we would also have to examine the historical reliability of exit polls: how well have exit polls in the past reflected the actual vote count?

There is no reason to trust Mr. Bush or his people on these matters. There are more ways to rig an election than miscounting votes. In 2000, the state of Florida illegally purged voter rolls those it claimed were convicted felons, although this was part of a process that was so flawed that it is difficult to attribute it to an honest mistake. Furthermore, most of those purged were African Americans, who vote Democratic about 90% of the time (as anyone who has ever taken Political Science 101 knows). There is some investigation into the matter of whether, in 2004, the state of Ohio adequately supplied Republican precincts with voting materials while shorting Democratic precincts. For Bush and his people to arrange for a systematic miscounting of votes would not be beyond the realm of possibility.

Nevertheless, when it comes to determining whether votes were miscounted, we must begin with a scientific enquiry into the data; until that is done and unless that produces a red flag, we must accept the null hypothesis that states there was no fraud of this kind.

If TruthIsAll can show the kind of discrepancy that I have outlined above in the face of historical reliability of the exit polls, then I will agree that there is cause for further investigation. Perhaps Mr. Nederland will as well, although he no doubt will speak for himself on the matter.

[FogerRox]

1. Nice--good points--DO I have this right ? the question is where do we go

from here? What do we do now? What should our goal be?

Whats the next step?

Which I think would be far more productive than renewing for another season of the soap Opera: TIA VS the Naysayers.

[Jack Rabbit]

2. The first question is whether we have anything at all

I'm trying to lay down criteria that would show us that we have something or not.

To simply say that fraud occurred and cite an overall discrepancy between exit polls and the official vote tabulation doesn't show anything. It's just as likely the exit polls were wrong as that there was fraud.

However, if you can show that the discrepancy occurred in certain situations, such as in precincts with new voting technology or in those where Democrats do better, and didn't occur elsewhere, then you're on to something. There are tools that a statistician can use to show whether the discrepancy was significant. If it was, then it was probably not due to random chance.

For example (and I emphasize that I am making this up), if a particular state in all precincts with a heavy Afro American population we see that the difference between the exit polls and the tabulated count to be somewhat less votes (say 5%) for Kerry than for Bush in the tabulated vote over the exit polls, you might wonder why. If you look further and discover that in those same precincts that in areas with conventional vote tabulation technology the difference between the exit polls and the tabulated vote is maybe 1% and in those precincts where e-Voting was used it was about 30%, you might start smelling a very big rat.

I'm not saying that's what you'll find, because I really don't know. But that's the kind of thing that would raise flags.

[BeFree]

63. That's a stupid question

"...whether we have anything at all"

TIA has raised so many red flags, one must make a real effort not to see them.

Not only TIA, but numerous other posters, over the last 4 months, have been waving other red flags.

It amazes me that peope have not seen those flags, until the realization sets in that they may not want too.

Yet, we are inundated with posters who still profess some kind of seasoned knowledge here, without having read all there is too read.

If ya want me to whip out a red flag for you, let me know, I've a few stashed away.

The Election Was Stolen

[Chi]

3. I believe one would have to break down the possibilities....

Heya Jack Rabbit 8)
Thanks for the summary

I believe if one were to break down the possibilities of why exit polls could be so far off from the results of an election, there would be only a few possibilities.

One of these possibilities is 'random error'.
I believe that anyone with an open mind, and a good understanding of what Tia has done would agree, he has proven 'random error' was not the source.

Out of curiousity...would you agree with that one point?

[Trillo]

4. Paper Ballots as Control Group

My summary of Jack Rabbit's post: Because a control group is not there now, nothing can be proved; if nothing can be proved the null hypothesis of nothing is wrong must be accepted.

This reminds me of the following absurdity: if a bank doesn't track money going in and out of accounts, then it can't be proven money is ever stolen, therefore we can conclude money is never stolen.


To me it sounds like an argument for paper ballots as a control group.

[Jack Rabbit]

9. Not a very good summary

True, a statistician will accept the null hypothesis unless he has reason to reject it.

As things are now, you're stuck with nothing better than There is a discrepancy between vote tabulations and the exit polls; therefore, there was fraud. One could as easily say (as you know they will and do): There is a discrepancy between vote tabulations and the exit polls; therefore, the polling was inaccurate. Those hypotheses are not contradictory, but no one has any good reason to accept one and reject the other.

However, I also said that the data needs to be broken down in order to determine what happened. That may give us a case for saying there was fraud and exactly what the nature of that fraud was.

If, after breaking down the data, we see a pattern that correlates to the use voting technology and a significantly wider discrepancy between the exit polls and that the discrepancy consistently favors one candidate or is to the disadvantage of another, then we have good reason to assume fraud. Why, after all, would exit polling be any less accurate in precincts that use e-voting than in precincts that don't? The common sense explanation of that would be the electronic voting machines we're counting the votes accurately.

However, we don't really know if that's what happened. Yet.

To me it sounds like an argument for paper ballots as a control group.

It is nothing of the sort; however, I favor the use of paper ballots and hand counting them in a public forum; I also believe that e-Voting should not be used at all. As a professional computer programmer, I can tell you that it would be very easy to rig an election with e-Voting.

[eomer]

5. We already have a red flag.

Jack Rabbit wrote:

until that is done and unless that produces a red flag, we must accept the null hypothesis that states there was no fraud of this kind.

The unexplained 6.5 point disparity between the exit poll raw data and the official count (in other words, Mitofsky's "Within Precinct Error") is a red flag to me.

Can someone please tell me why it is not? Do you need a disparity of more than 6.5 points? Do you take the position that an exit poll can never be a red flag? What is your position for saying this is not a red flag?

[Jack Rabbit]

10. See the second paragraph in post number 9

A 6.5 point disparity arouses my interest. However, it doesn't tell me if I should be interested in the vote counters or the poll takers.

To make our case, we need to be able to say it was the vote counters.

Personally, I think they had more to do with it than the poll takers, but we can't take it to the bank yet.

[Nederland]

6. Nice Post

I think if we had access to precinct level exit poll data we would be in a position to pin point fraud. And yes, I'm aware of the effort here to deduce that information by analyzing precinct sizes etc. Hopefully we will eventually find out what happened...

[Mairead]

7. There's a trout in the milk

As Thoreau pointed out, some circumstantial evidence is strong enough to convince the open-minded, even if not the closed. We've found a trout in the milk, and Steve Freeman has held it up for all to see. That isn't proof that the milk was watered/election rigged, but the likelihood of anything else being the cause is vanishingly small. It's significant that Mitofsky and MediaInc refuse to release enough information, even protected by an NDA, for peer review.

[Chi]

8. Any chance....

Heya Mairead

"Mitofsky and MediaInc refuse to release enough information, even protected by an NDA, for peer review."

Any chance you have a handy link for a good article on this?
If not, don't sweat it. 8)

I'm not doubting, I just don't have any articles with the NDA and peer review stipulations.

Thanks in Advance

[Mairead]

16. There might be some, but I don't have them

I was extrapolating from the worldwide scientific standard: if you believe your results hold water, then you put them and your methodology out for peer review, even if only to a select group under NDA. With all the kerfuffle and questions around it, Mitofsky should have been practically shoving it into people's hands at gunpoint.

The fact that he's done the exact opposite is an enormous hint that there's something going on that can't stand the light of day.

My guess, since Mitofsky's not a dope, is that a peer review would show that his methodology was excellent and produced a very small amount of slop, the pre-'corrected' values (that Freeman used) were correct, and that therefore the only rational conclusion is that the election was stolen outright by a large number of people working independently toward a shared goal, each affecting a few votes.

I would also guess that, since his livelihood depends on him not having effed this up, that he's making it known to selected people who might employ him that his methodology was fine and he's accepting the blame 'for the good of the country' since revealing the truth 'would risk civil war'. Since most of those people are probably GOP, they have no reason to blow the whistle on him.

Just my guess.

[Chi]

47. I agree...

Well said, I agree 100%.

Thanks for the response.

[Jack Rabbit]

11. For the record, I am not satisfied with Mitofsky's explanation

Our investigation of the differences between the exit poll estimates and the actual vote count point to one primary reason: in a number of precincts a higher than average Within Precinct Error most likely due to Kerry voters participating in the exit polls at a higher rate than Bush voters. There have been partisan overstatements in previous elections, more often overstating the Democrat, but occasionally overstating the Republican. While the size of the average exit poll error has varied, it was higher in 2004 than in previous years for which we have data. This report measures the errors in the exit poll estimates and attempts to identify the factors that contributed to these errors . . . .

It is difficult to pinpoint precisely the reasons that, in general, Kerry voters were more likely to participate in the exit polls than Bush voters. There were certainly motivational factors that are impossible to quantify, but which led to Kerry voters being less likely than Bush voters to refuse to take the survey.

That Kerry voters were more like to participate in exit polls is sheer speculation on Matofsky's part.

In addition there are interactions between respondents and interviewers that can contribute to differential non-response rates. We can identify some factors that appear to have contributed, even in a small way, to the
discrepancy. These include:

• Distance restrictions imposed upon our interviewers by election officials at thestate and local level
• Weather conditions which lowered completion rates at certain polling locations
• Multiple precincts voting at the same location as the precinct in our sample
• Polling locations with a large number of total voters where a smaller portion of voters was selected to be asked to fill out questionnaires
• Interviewer characteristics such as age, which were more often related to precinct error this year than in past elections

Mitofsky is providing a whole bunch of maybes here, but no solid explanation.

While I'm not satisfied with Mitofsky's explanation, my dissatisfaction does not prove election fraud. Election fraud, however, remains a possible explanation for the discrepancy.

[eomer]

13. Is this a correct statement of your position?

Collecting and paraphrasing your statements:

There are two possible explanations:

1. the exit poll is wrong.


2. the official result is wrong.

The explanations so far are just a bunch of maybes, but no solid explanation.

You don't think there is cause for further investigation.

To me something clashes in that list of your thoughts.

[Jack Rabbit]

15. You've got it right until the very end

We can and should examine the data further. We need to examine it at a precinct level and then determine what those places where exit polls were most out of line with the tabulated vote had in common that was different from places where the exit polls more closely matched the vote count.

For example, if the fraud was committed with rigged e-Voting machines, we would expect to see exit polls more out of line with the tabulated vote in those areas than in places where conventional polling was used.

Therefore, we cannot find out any more than we already know looking at statewide data, but we may be able to discover something looking precinct-by-precinct.

[eomer]

17. Thanks for the clarification.

I agree that analysis at the precinct level to search for correlation of the disparity with other factors is a promising approach.

There is an obstacle, which is that the exit poll data at the precinct level does not disclose the identity of each precinct. So there is investigation required first just to determine which exit poll precinct matches which official count precinct. That obstacle is the reason no one has yet performed the analysis you're suggesting.

I also believe there are other promising approaches that can be pursued in parallel to this one.

[TruthIsAll]

12. One very important point must be made here.

You say:

"If TruthIsAll can show the kind of discrepancy that I have outlined above in the face of historical reliability of the exit polls, then I will agree that there is cause for further investigation".

Which polls are you talking about?

The preliminary exit polls, which are suppressed from public view?
Or the final poll, which is re-weighted to match the recorded votes?

If you are not familiar with the following jewel of a factoid gleaned from the preliminary and final national exit poll, then consider this:

In the 2004 PRELIMINARY exit poll of 13047, 41% of respondents said they voted for Bush in 2000. Since 41% of the 122 million voters is 50 million, then this is just .50 million (or 1%) below Bush's actual 2000 total of 505.5 million votes. This is entirely possible, because we know that there were a certain number of Bush 2000 voters who decided NOT to vote for him in 2004. Could that number be close to the .5 million difference?

In the FINAL exit poll of 13660, which was massaged to "match" the recorded vote, 43% of respondents said that they voted for Bush in 2000. THAT IS CLEARLY IMPOSSIBLE, since 43% of 122 million is 52.5 million (4% above the actual Bush vote).

THE ONLY WAY THEY COULD PAD THE 2004 BUSH VOTE WAS TO PAD THE 2000 BUSH VOTE. BUT THAT IS A LOGICAL ABSURDITY.
SO WHAT ARE WE LEFT WITH?

Why would anyone want to change an absolutely plausible weighting factor, which produces a result KNOWN TO BE ACCURATE TO WITHIN 1% AT MINIMUM AND PERHAPS EVEN LOWER, to a weighting factor which is KNOWN TO BE ABSOLUTELY IMPOSSIBLE, since it produces a value KNOWN TO BE 4% OVER THE ACTUAL RECORDED 2000 BUSH VOTE?

DOES THAT MAKE SENSE TO YOU?

SO IS THIS PROOF THAT THE FINAL EXIT POLL OF 13660 WAS WRONG
AND THE PRELIMINARY POLL OF 13047 WAS PROBABLY VERY CLOSE TO WHAT REALLY HAPPENED?

DO YOU AGREE THAT THIS IS PROOF?
IF NOT, WHY NOT?

THE ONLY WAY THEY COULD PAD THE 2004 BUSH VOTE WAS TO PAD THE 2000 BUSH VOTE. BUT THAT IS A LOGICAL ABSURDITY.

SO WHAT ARE WE LEFT WITH?

THAT IS NOT THE ONLY NUMERICAL ANOMALY. THERE ARE LITERALLY SCORES OF OTHERS. BUT LET'S START WITH THAT ONE.

NOW DO YOU AGREE THAT THE EARLY EXIT POLL OFFERS POWERFUL CIRCUMSTANTIAL EVIDENCE THAT FRAUD OCCURRED IN BOTH THE VOTE COUNT AND IN THE MANIPULATION OF THE PRELIMINARY EXIT POLL TO VALIDATE THE IMPOSSIBLE?

ANY COMMENTS?

[Jack Rabbit]

14. I'm not one to mess with success

I would want to see figures before they weighted them to match the "actual" outcome. Looking at figures that are weighted to match the outcome defeats the purpose of looking at the figures to determine if the outcome was fudged. That, after all, is what we want to know.

Nevertheless, that the pollsters changed a good method for an inferior one doesn't prove voter fraud. It isn't even circumstantial evidence of it.

What would be circumstantial evidence of it would be showing a correlation between polling data discrepancies and some other independent variable, such as the use of e-Voting.

The hypothesis I am advancing is that if exit polls were wrong, they were wrong across the board and that if there was no voter fraud they would be just as wrong in precincts where e-voting was used as in precincts where more traditional methods were used. If there was fraud, evidence of it should be seen by examining those precincts where the exit poll discrepancy with the tabulated vote was widest and then asking what those precincts had in common with each other but not precincts where exit polls and the tabulated vote more closely matched.

That is something that can only be determined by examining the polling data vis a vis the tabulated vote at a finer level than I've seen from you or anybody else up to now.

[TruthIsAll]

19. You are avoiding it, Jack. You are avoiding the FACTS.

There was a reason for the weighting change.
And you fail to acknowledge it.

You gloss over it.
You do not question it.

You just ask for more evidence.
I gave you evidence.
And you try to spin it away.

The data is at a very fine level here.
It can't get any finer.

The final 43% was incorrect.
It was incorrect by design.

The initial 41% was very plausible.
But not by design.
It represented the truth.

They took a chance, adjusted the number and hoped that no one would catch it..

Or maybe they didn't do the math:
But we did: .43*122 = 52.46


Maybe they were too rushed.
But now it's too late to do anything about it.
It's right there for the world to see.

The Final exit poll had to match the recorded 2004 vote in order to enable them to rewrite history.

But they could not at the same time match the recorded 2000 vote, which is history.

Have a nice evening, Jack.

[Jack Rabbit]

20. Were the pollsters part of a plot?

May I ask for evidence of that? I don't mean theories; I want something that will hold up in court.

Also, I want the evidence at the precinct level. Unless it's there, it's not fine enough.

[TruthIsAll]

21. .43 * 122=52.46 is NOT a theory, Jack. It's simple arithmetic.

And .41*122 = 50.02
That's not a theory, Jack.
It's simple aritmetic.

That's not fine enough for you?

[Jack Rabbit]

22. No, it isn't fine enough

From post 14:

(T)hat the pollsters changed a good method for an inferior one doesn't prove voter fraud. It isn't even circumstantial evidence of it.

[TruthIsAll]

23. I guess you have nothing left to offer. Just "no, it isn't fine enough"

Jack, after your obvious ducking of this very simple question, how can you expect anyone around here to try to present you with any facts?

You simply disregard them and ask for more facts.

This one factual incongruity gets to the heart of the issue.

And you are not being honest, Jack, by avoiding it.

[Jack Rabbit]

24. It gets to nothing that I can see

It shows, if it is indeed a verifiable fact (do you have a link, BTW?), that pollsters switched to a method that was not as good as the original method.

Does that show the election was rigged? No. If it does, you'll have to explain to me exactly how. I've got a little flu this evening and must not be thinking clearly.

[TruthIsAll]

26. It is indeed a verifiable fact. How did you miss it after reviewing

my recent posts?

The 3 stages of the National exit poll:

1- on CNN at 7:38pm, 11/02, 11027 sampled.
Kerry leading by 51-48.

2- On WP at 12:22 am, 11/03, 13047 sampled.
Kerry still leading by 51-48

3- On CNN at 2:04 PM, 11/03, 13660 sampled.
Bush wins by 51-48.

This is a recent detailed comparative analysis of all three.
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x343168

Here are the links to the exit poll data:

1- 7:38pm
http://www.exitpollz.org/CNN_national2.htm

2- 12:22 am (See below)
http:
//media.washingtonpost.com/wp-srv/politics/elections/2004/graphics/exitpolls_us_110204.gif

3- 2:04pm - the bogus FINAL
http://www.cnn.com/ELECTION/2004/pages/results/states/US/P/00/epolls.0.html


Here is the 13047 Exit Poll - the Final True result

[Jack Rabbit]

33. Thank you, Mr. TIA

I am looking at your analysis presented this forum last Sunday.

You argue that the CNN figures at 8 pm on election night are accurate because that poll states that 41% of the those in the sample voted for Bush in 2000; this computes to about 50.02 million votes, whereas Bush actually got about 50.4 million votes (pretty close!) whereas the figures for 2 pm the following day give the percentage of those polled as having voted for Bush in 2000 at 43%, which computes to 52.46 million votes, quite a few more than the 50.4 million he actually got. This, you say, impugns the later poll.

However, I see by your figures that 38% of the respondents at 8 pm voted for Vice President Gore in 2000, which computes to 46.36 million votes, and at 2 pm the nest day 37%, which computes to 45.14 million. Al Gore actually got just shy of 51 million votes in 2000.

I would expect if things were as proportional as you imply they are at 8 pm, then the Gore in 2000 figures should be not 38%, but about 41.8%; that would compute to 50.996 million votes, just about what he actually received.

Therefore, if you are going to argue that the discrepancy you cite on Bush's figures on the 2 pm/following day poll invalidate that poll, wouldn't you also have to say that the figures for Gore invalidate both the 8 pm and 2 pm polls?

You claim in post 19 that the discrepancies on the Bush figures in the 2 pm/next day poll were by design. Were the discrepancies on the Gore figures in both polls also by design? If so, what does that mean?

[TruthIsAll]

37. Thanks, Jack, you have just provided DU with another SMOKING GUN

Here's why we now have all the MORE reason NOT to believe that 41% in the PRELIMINARY EXIT POLL (and surely not 43% in the FINAL) voted in 2000 for Bush...and that 38% may very well have been accurate or even slightly lower than the percentage who voted for Gore.

Two points:

1. If 38% of poll respondents claimed they voted for Gore and 41% said they voted for Bush, why did Mitofsky and others say that Bush voters were more reluctant to speak to the exit pollsters?

Maybe Kerry actually did even better than the early exit polls which had him winning 51-48%.

2. How many of those who voted in 2000 DIED or became INCAPACITATED and never got to the polls? You would expect a certain amount of attrition, wouldn't you?

There HAD to be FEWER voters in 2004 who voted in 2000 then the ACTUAL number of those who did, not MORE.

SO THE FACT THAT ONLY 46.36 MM OF THE 51 MILLION WHO VOTED FOR GORE (if you believe 38%) RETURNED TO VOTE IN 2004 IS VERY PLAUSIBLE.

FURTHERMORE, IT IS ALSO POSSIBLE THAT FEWER THAN 4.64 MILLION DIED OR WERE INCAPACITATED OR JUST STAYED HOME AND THAT THE ACTUAL PERCENTAGE WAS A LITTLE HIGHER (39%?).

BUT WE DO KNOW FOR A FACT THAT THE NUMBER OF GORE VOTERS WHO TURNED OUT IN 2004 HAD TO BE FEWER THAN THE NUMBER WHO ACTUALLY VOTED IN 2000.

ANYONE CARE TO CHECK THE MORTALITY TABLES FOR DEMOCRATS?

BUT THERE IS NO WAY IN HELL THAT 2 MILLION BUSH VOTERS COULD HAVE MATERIALIZED WHO WERE NOT ALIVE IN 2000 TO BEGIN WITH. EVEN JESUS COULDN'T BRING THAT MANY BACK FROM THE DEAD.

In any case, Kerry won 59% of the 21 million new voters, Bush only 39%. And Kerry won the vast majority (65%) of Nader voters to Bush's 13%.

So maybe Kerry did even better than 50.90-47.90%.

If we assume that 40% of voters in 2004 voted for Gore and 40% for Bush, then we have:

Mix Bush Kerry Nader
17% 41% 57% 1%
40% 8% 91% 1%
40% 90% 9% 0%
3% 13% 65% 16%
100% 46.56% 51.64% 1.05%

Thanks, Jack, you have just added another smoking gun to the arsenal.

Have you converted yet?

[Jack Rabbit]

41. No, you haven't converted me

You know I'm a hard sell.

1. If 38% of poll respondents claimed they voted for Gore and 41% said they voted for Bush, why did Mitofsky and others say that Bush voters were more reluctant to speak to the exit pollsters?

First of all, Mitofsky speculates (like I care about his speculations!) that Bush voters were more reluctant to talk than Kerry voters. He gives no reason for such speculation. It's possible, but I don't find it convincing.

Maybe Kerry actually did even better than the early exit polls which had him winning 51-48%.

That comes perilously close to begging the question. We have to demonstrate logically there was fraud, not assert it.

2. How many of those who voted in 2000 DIED or became INCAPACITATED and never got to the polls? You would expect a certain amount of attrition, wouldn't you?

No doubt a number of those who cast votes for Gore in 2000 passed away before the 2004 election. Likewise, a number of those who cast votes for Bush in 2000 passed away before the 2004 election. Do you have any reason to think that so many more people who voted for Gore passed way before than the 2004 election than those who voted for Bush in 2000? I don't.

BUT THERE IS NO WAY IN HELL THAT 2 MILLION BUSH VOTERS COULD HAVE MATERIALIZED WHO WERE NOT ALIVE IN 2000 TO BEGIN WITH.

If 122 million people voted in 2004 and 50.4 million voted for Bush in 2004, then it is obvious something less than 43% of those who voted in 2004 voted for Bush in 2000; however, it is entirely plausible that 43% of NEP's sample did. I would agree that that could make the sample biased toward Bush.

Mix Bush Kerry Nader
17% 41% 57% 1%
40% 8% 91% 1%
40% 90% 9% 0%
3% 13% 65% 16%
100% 46.56% 51.64% 1.05%

Your argument assumes that while the percentage of the NEP sample who voted for the various candidates in 2000 were wrong, the proportion of Gore voters voting for Bush or Kerry and of Bush voters returning to vote for Bush is correct. If your going to fudge one column assuming its wrong, why assume the other columns are right?

What I'm getting at is rather simple. Kerry was leading in some exit polls and only after the polls were adjusted to fit the actual count did Bush appear to win the exit poll. I'm a little confused by that. Frankly, I'd rather just toss out that 2 pm/next day poll and ask the more obvious question: Why was there a discrepancy between the exit polls and the tabulated result?

The figures don't prove fraud. However, they don't disprove it, either.

To those here who are saying that I've missed the point that this didn't happen in previous elections, I say: Nonsense. I wouldn't consider this worth discussing at all if this weren't something new. There is something wrong either with the exit polls or the election tabulation. I want to know which it is.

I think the best way to determine that is to look at the data more closely and see if there is a pattern to the discrepancy. Was there a wider discrepancy in precincts that shared some common characteristic or were the polls just wrong across the board?

[TruthIsAll]

43. Time to call it a night. My response...

You know I'm a hard sell.

1. If 38% of poll respondents claimed they voted for Gore and 41% said they voted for Bush, why did Mitofsky and others say that Bush voters were more reluctant to speak to the exit pollsters?

First of all, Mitofsky speculates (like I care about his speculations!) that Bush voters were more reluctant to talk than Kerry voters. He gives no reason for such speculation. It's possible, but I don't find it convincing.

NEITHER DO I.
...........................................................

Maybe Kerry actually did even better than the early exit polls which had him winning 51-48%.

That comes perilously close to begging the question. We have to demonstrate logically there was fraud, not assert it.

I DID NOT ASSERT ANYTHING. I CONJECTURED THAT THERE APPEARS TO BE EVIDENCE THAT KERRY DID BETTER TTHEN 50.9%, BECAUSE HE'S A WINNER EVEN WITH WEIGHTINGS THAT FAVOR BUSH - AND WHICH ARE DUBIOUS AT BEST.
............................................................

2. How many of those who voted in 2000 DIED or became INCAPACITATED and never got to the polls? You would expect a certain amount of attrition, wouldn't you?

No doubt a number of those who cast votes for Gore in 2000 passed away before the 2004 election. Likewise, a number of those who cast votes for Bush in 2000 passed away before the 2004 election. Do you have any reason to think that so many more people who voted for Gore passed way before than the 2004 election than those who voted for Bush in 2000? I don't.

NEITHER DO I. THAT'S CALLED A RED FLAG.
.............................................................

BUT THERE IS NO WAY IN HELL THAT 2 MILLION BUSH VOTERS COULD HAVE MATERIALIZED WHO WERE NOT ALIVE IN 2000 TO BEGIN WITH.

If 122 million people voted in 2004 and 50.4 million voted for Bush in 2004, then it is obvious something less than 43% of those who voted in 2004 voted for Bush in 2000; however, it is entirely plausible that 43% of NEP's sample did. I would agree that that could make the sample biased toward Bush.

AH, IS THAT A SLIGHT CONCESSION? IF THE WEIGHTING IS IMPOSSIBLE, AND IT IS, THEN WHY NOT MAKE IT PLAUSIBLE BY CHANGING IT BACK TO WHAT IT WAS - EVEN IF IT MEANS THAT DOING SO INVALIDATES THE FINAL EXIT POLL AND VALIDATES THE PRELIMINARY POLL.
..................................................................

Mix Bush Kerry Nader
17% 41% 57% 1%
40% 8% 91% 1%
40% 90% 9% 0%
3% 13% 65% 16%
100% 46.56% 51.64% 1.05%

Your argument assumes that while the percentage of the NEP sample who voted for the various candidates in 2000 were wrong, the proportion of Gore voters voting for Bush or Kerry and of Bush voters returning to vote for Bush is correct. If your going to fudge one column assuming its wrong, why assume the other columns are right?

BECAUSE I HAVE NO REASON TO DISBELIEVE THEM. BESIDES,THE WEIGHTING WHICH IS UNDER STUDY HERE IS THE MIX. SO I CHANGED IT TO SEE THE EFFECTS ON THE KERRY MARGIN. QUITE LARGE - FROM 3% TO 5%.
...................................................................

What I'm getting at is rather simple. Kerry was leading in some exit polls and only after the polls were adjusted to fit the actual count did Bush appear to win the exit poll. I'm a little confused by that. Frankly, I'd rather just toss out that 2 pm/next day poll and ask the more obvious question: Why was there a discrepancy between the exit polls and the tabulated result?


YOU WOULD LIKE TO TOSS OUT THE 2PM POLL?
WHY? IT'S PROOF THAT SOMETHING IS TERRIBLY WRONG.

WHY NOT JUST SAY THE EXIT POLL WAS WRONG?
WHY RE-WEIGHT TO THE RECORDED VOTE AT ALL?
I AGREE, WHY IS THE FINAL EXIT POLL EVEN NECESSARY?
................................................................

The figures don't prove fraud. However, they don't disprove it, either.

OK, THEN, WE'LL JUST CALCULATE THE PROBABILITIES:
1) 1 IN 19 TRILLION BASED ON STATE EXIT POLL DISCREPANCIES (17 STATES BEYOND THE EXIT POLL MOE ALL IN FAVOR OF BUSH)
AND
2) 1 IN 2 TRILLION BASED ON THE VOTER 2000 CHARACTERISTIC (FOR BUSH TO GO FROM 47.90% IN THE NEP TO 50.73% IN THE VOTE, ASSUMING A 1.0% MOE)

SO DRAW YOUR OWN CONCLUSIONS. IF THE DISCREPANCIES ARE NOT DUE TO RANDOM ERROR, AND NOT TO THE RBR, THEN WHAT COULD IT BE?
...................................................................

To those here who are saying that I've missed the point that this didn't happen in previous elections, I say: Nonsense. I wouldn't consider this worth discussing at all if this weren't something new. There is something wrong either with the exit polls or the election tabulation. I want to know which it is.

I THINK YOU ALREADY KNOW. WE HAVEN'T EVEN DISCUSSED THE TANGIBLE CORROBORATING EVIDENCE OF 99% OF BUSH VOTES SWITICHING TO KERRY, ETC., ETC. ETC.
.....................................................................

I think the best way to determine that is to look at the data more closely and see if there is a pattern to the discrepancy. Was there a wider discrepancy in precincts that shared some common characteristic or were the polls just wrong across the board?

LOOK AT THE DATA MORE CLOSELY?
OF COURSE.

SEE IF THERE IS A PATTERN TO THE DISCREPANCY?
THAT WAS ESTABLISHED A LONG TIME AGO.

42 OUT OF 50 STATES MOVED IN FAVOR OF BUSH FROM THE EXIT POLLS - AND THAT INCLUDED EVERY BATTLEGROUND STATE EXCEPT OREGON, THE ONLY STATE WITH 100% VERIFIABLE PAPER BALLOTS.

[Jack Rabbit]

45. Response

The only legitimate question about the exit poll concerns whether we can reliably say that Kerry had a lead on election day. If we cannot say that, then we have no case. There may have been fraud, but we'd never be able to prove it.

I agree that weighting the poll to fit the "actual" result is dreadful methodology, especially when it is the integrity of the vote tabulation itself that is being called into question. That kind of weighting begs the question. The 2 pm poll is useless either to those impugning the integrity of the vote or those defending its integrity.

If examining the break down of respondents by 2000 vote they provide us with any ammunition at all, it is that they appear to make the exit poll biased toward Bush. A random sample of voters in 2000 should be expected to provide an almost even split between those who voted for Bush and those who voted for Gore. Instead, we get more Bush voters. Based on this sampling, if one were looking for conspiracy theories one might just as easily formulate one that states Bush actually won the popular vote in 2000. Of course, we know that is not the case.

Yes, that is a point in your favor. Use it wisely. However, I don't consider it a red flag. A yellow one, maybe, but not a red one.

There is no point in suggesting that Mitofsky has any sinister motives. If he did, there would be no yellow or red flags, only green ones pointing to a clear Bush victory.

42 OUT OF 50 STATES MOVED IN FAVOR OF BUSH FROM THE EXIT POLLS - AND THAT INCLUDED EVERY BATTLEGROUND STATE EXCEPT OREGON, THE ONLY STATE WITH 100% VERIFIABLE PAPER BALLOTS.

That is exactly the point I'm trying to get at. We have very good reason to be suspicious of the tabulated vote count; this suggests that there was fraud and that the mechanism used was new voting technology.

However, this is not proof to my satisfaction. Maybe I just have a tougher standard of proof than you on matters such as this. My suspicion of fraud is not proof of it. I'm looking for something that will hold up in court.

Let's consider what we're talking about here. We are saying that the 2004 election was rigged and that the results should be overturned. That is an extraordinary accusation. If you're going to make it stick, you're going to need more than you've presented. You will need statistics at a very low level of detail, much more detail than will fit on a post on an open Internet discussion board. State level statistics are simply too crude to be considered irrefutable proof.

I am not trying to refute your suspicions. I share them, although to what degree I'm not certain. If Bush won the popular vote in 2004, I doubt it was by 4 million votes as reported. However, if I were asked to overturn the election based on the statistical evidence that you have presented, I would not do it. And believe me: I would really, really want to overturn it.

So, what I have suggested here is that we get that level of detail. What I am suggesting is that to prove the hypothesis that there was election fraud and that the mechanism was new voting technology you will have to correlate discrepancies in polling data with the use of that technology. A study at a local level, if it shows what we suspect it would, would leave no doubt that the election was rigged and how.

If it were up to me and I were presented detailed evidence as I have outlined on this thread that where sophisticated voting technology was used that there was a wider discrepancy between the expected result and the actual result, that the discrepancy favored Bush in the actual tabulation, and that frequency of such occurrences places it beyond chance, I would consider that a red flag. Then its time to start seizing the records of voting machine manufacturers and election officials. That might lead to the kind of evidence that I would consider proof of fraud.

[kobeisguilty]

49. mmm....nice

I think everyone should listen to one of his main points:

...where sophisticated voting technology was used that there was a wider discrepancy between the expected result and the actual result, that the discrepancy favored Bush in the actual tabulation, and that frequency of such occurrences places it beyond chance, I would consider that a red flag.

Ask yourself, if you believe that fraud happened, how did it happen? Clearly, those in this board that point to fraud believe it due to voting machine influence by Republicans, or simply that these companies hired ex-felons....

Rather than debating methodological biases all day with Mitofsky or "proving" that the final exit poll data was wrong, we should focus our efforts to determine whether or not there is a significant correlation between precincts that use voting machines/type of voting machine/company owned and those that don't.

If such a case could be made in Ohio (which would not fall prey to the bullshit Dixiecrat theory), then this alone could be used as a smoking gun to convince people that something is awry. An argument for fraud simply due to the inaccuracy of the final exit polls is too susceptible to explanations of methodological bias. However, a voting machine/discrepancy study by either precinct or county would not be susceptible to such a defense.

[ihelpu2see]

38. the other fact that Mr. Rabbit misses is

Prior presidential election polls did not fall out of the MOE.... Why 2004?

[Jack Rabbit]

42. No, I haven't missed that

The fact that there is a discrepancy between the exit polls and the tabulated results shows something is wrong. It doesn't show what.

Nevertheless, we need to keep asking what.

[eomer]

27. It is time for more data to be opened up.

The preliminary exit poll shows that Kerry won.

The final adjusted exit poll showing that Bush won cannot possibly be correct because it says that there were more repeat voters for Bush than there were voters for Bush in 2000.

All the data available either shows that Kerry won or shows some internal inconsistency that makes it clearly fallacious.

The data needed to do the precinct level analysis that you are advocating is in the possession of NEP but has not been made public.

It is time for those holding the data to bring it out into the open so that these glaring discrepancies can be resolved.

[kiwi_expat]

53. NEP precinct data at www.icpsr.umich.edu/org/announce.html#nep

The data is from the Univ. of Michigan. Minvis has downloaded and deciphered it. Blue22 has put it in a table that he is willing to e-mail to anyone who PMs their e-mail address to him.

Please visit the "Ohio Exit Poll Raw Data" thread. Hamilton County NEP precincts have now been identified and can be compared to the recorded vote. More precincts will be identified very shortly.

Mytofsky, himself, admits that something went very wrong in the SW Ohio NEP precinct data. We are going to find out what it was. Citizen canvassing of selected NEP precincts is the goal.

Why do I always feel I am "off topic" in these other threads, when I state that full manual recounts of selected precincts is the best way to PROVE fraud occurred in 2004?

[kobeisguilty]

25. for one thing.

chill out, TIA. your conduct of discussion is highly combative to say the least.

What do you think the motivation was behind the pollster's moving to utilize an inferior method? Do you belive the pollsters, assuming fraud, are in on it?

Your find does indeed mean that the final exit polls are inconsistent. This fact is also interesting is that it is an anathema to those that believe the polls are simply skewed towards the democrats, as I believe if more Kerry supporters were polled than Bush supporters, we would have a smaller amount in the national sample of those who voted for Bush in 2000, instead we have more people claiming to vote for Bush in 2000 than who actually did.

But didn't we already know before that the exit polls were fixed to the final vote count?

[eomer]

28. Yes we knew that the exit polls were fixed to the final vote count.

But the problem is that the final "fixed" exit poll as published cannot possibly be correct since it implies more Bush voters in 2000 than there actually were.

So that's red flag number two.

Red flag number one was the 6.5 point discrepancy between the exit poll raw data and the official count.

All the data that is available does not add up.

Jack, the data needed to do the precinct-level analysis you are calling for is in the possession of NEP but has not been made public. This data, along with many other instances of evidence and data that have been requested, should be made public.

[TruthIsAll]

31. You have just disproved the Reluctant Bush Responder Theory

"Your find does indeed mean that the final exit polls are inconsistent. This fact is also interesting is that it is an anathema to those that believe the polls are simply skewed towards the democrats, as I believe if more Kerry supporters were polled than Bush supporters, we would have a smaller amount in the national sample of those who voted for Bush in 2000, instead we have more people claiming to vote for Bush in 2000 than who actually did."

Thank you for confirming the impossibility of the RBR theory, which uscountvotes.org debunked a month ago, after it was put out there by Mitofsky as a possible explanation for the exit poll discrepancies.

And don't expect me to chill out. I've been fighting the good fight much longer than you've been here, to stop now.

After you work hard to create hundreds of analytical posts and find the same naysayers conspiring to attack them every time, as if on que with faulty logic, fluffy straw men and WH talking points, perhaps than you will understand why you won't just "chill out".

If I'm combative, it's because getting some people to focus on interpreting the facts already out there, instead of fogging them up, often requires issuing a challenge directly to them: either take the dump or get off the pot.

[kobeisguilty]

48. ok

I just think you'd get a lot more people to take your viewpoint more seriously if your posts focused primarily on the logical content rather than the personalities behind those that disagree with you. Keep fighting the good fight, definetely don't stop now.

[anaxarchos]

18. Lemme get this straight...

After months of discussion and debate inside and outside this board, you want to restart with this "fair" position:

"Nevertheless, when it comes to determining whether votes were miscounted, we must begin with a scientific enquiry into the data; until that is done and unless that produces a red flag, we must accept the null hypothesis that states there was no fraud of this kind.

If TruthIsAll can show the kind of discrepancy that I have outlined above in the face of historical reliability of the exit polls, then I will agree that there is cause for further investigation. Perhaps Mr. Nederland will as well, although he no doubt will speak for himself on the matter."

Good. It's my turn to be a naysayer. Nah, I don't think I'll take it from the top again.

As far as I can tell, that's just Degenerative Democrats Disease...

Have a blast.

[bleever]

32. Hey, thanks. I was going to find the right words, and you already

had them.

[msongs]

29. name some elected dems who give a darn about any of this??

meaning what is said specifically here on DU and are actually using any of the theories put forth in these threads to actively do something to change the efraud and prevent it. I am curious.


Msongs
www.msongs.com/political-shirts.htm

[TruthIsAll]

30. Whether any give a darn is irrelevant. The truth is all that matters. nt.

,

[anaxarchos]

35. You're right....

No conspiracies for cops to fix...
No lawsuits for lawyers to win...
No stories for newspapers to print...
No sound bites for celebrities to promote...
No speeches for politicians to give...
No timbers for carpenters to frame...
No laws for Congress to pass...

Just us...

Fuck it.

[dzika]

34. 2000 Pres. Vote Characteristic in the 2004 Exit Poll

I have been making the assumption that not enough of the NEP data has been released to compare the 2004 exit poll to precincts/voting machines. Based on that assumption, I believe the following to be true.


FACT: Either the un-weighted exit polls are wrong, the weighted exit polls are wrong or both are wrong.

FACT: The most verifiable characteristic in the exit poll (with the exception of the vote count) is 2000 vote response. Based on that characteristic, one of the following must be true (based on TIA's analysis here):

1. The un-weighted exit poll is more correct than the weighted exit poll

2. The 2000 vote total was incorrect. George W. Bush actually won more popular votes that Al Gore.

3. The 2000 vote response characteristic in the 2004 exit poll can't be used to make assumptions about the vote count in the 2004 exit poll.

If you selected #3 then you then you will probably not believe that there is sufficient evidence to suspect the 2004 vote count based on exit polls until additional data is released or discovered.


Can anyone explain where my understanding is flawed? Thanks!

[Jack Rabbit]

36. I choose number 3

And you may see post 33 to see why I take issue with TIA's analysis.

However, the fact is that early exit polls had Kerry defeating Bush; that doesn't mean that the actual results were wrong, but it raises some questions that I would like to see answered. I'm not interested in polls weighted to match the "actual" result -- that's begging the question.

My contention is that only looking at the data in more detail can we determine whether or not there was fraud. So far, that detail has not been made available. That doesn't mean it won't be or that we shouldn't request it.

[dzika]

39. Where do we go from here?

My contention is that only looking at the data in more detail can we determine whether or not there was fraud. So far, that detail has not been made available. That doesn't mean it won't be or that we shouldn't request it.

Do you believe that there is no point in the continued analysis of the 2004 exit poll until more data is available?

If not, how would you suggest directing the efforts of analysis? Do you have any other recommendations for moving forward in our understanding the 2004 exit poll discrepancies?

[TruthIsAll]

40. TO POST # 37. JACK HAS JUST CONFIRMED THE SMOKING GUN.

His post thanking me was very gracious.

We should now thank him for questioning the validity of my use of the 2000 voter characteristic to prove that the final exit poll was bogus. That's because it gave me a chance to ponder his argument.

And then I realized that JFK was right: We are all mortal. We all die. Especially Gore voters. But apparently Bush voters come back from the dead.

Any DUers who may have thought that Jack put out the smoking gun will find that it is smoking even more now.

[davidgmills]

54. I am the resurrection and the life

He that voteth for me, though he were dead, yet shall he vote again.

GWB

[bleever]

56. Whosever believeth in me

need think no further. --GWB

(No offense to my Christian friends; you've probably gotten treated more shabbily than most by the blivet.)

[autorank]

44. In defense of "Truth" ,,,

...is all I intend. Here are the two points by Nederland endorsed by Jack Rabbit

* Regardless of your sample size, if your sample is not representative of the total population, exit poll results can be outside the MOE.

How is the representativeness of the sample size determined? By statistical rules and laws that are fairly well established. By historical comparisons of sampling and population results (1996, 2000 Presidential Elections). These requirements DO NOT include knowing the total population in detail and then determining if the sample matches that. This would make polling and just about any other type of statistical sampling unworkable. What were the procedures for the exit polls? Were there any reason to doubt the representative nature of the polls? Even the US Census, the Mack Daddy of data gathering and statistical analysis infers the total population from what is arguably a huge statistical sample. They cannot measure everyone. We had an approximation of previous exit polling in 2004 and it ties directly back to the methodology and accuracy of 2000 and 1996.

* When fraud occurs, it is impossible to determine whether or not an exit poll sample was representative.

This is simply wrong. When fraud occurs, the only way to determine the result of the ballot counting is through the Exit Polls. The caveat here is that the Exit Polls cannot also be fraudulent. Fraud in ballot counting (in whatever form) means you can't know what the vote really was. To say that absent this knowledge, you can't separately take a valid statistical sample makes no sense at all. This second point is a pure set up to negate polling altogether. It's a cute rhetorical device but not much in terms of real world analysis. Fraud took place in the Ukraine count. Exit polls pointed this out. The election was redone and the results changed when the fraud was corrected. Is Nederland arguing that the Exit Polls in the Ukraine lacked the "representativeness" to catch fraud.

TruthIsAll prevails in this argument. As do the other statisticians. Let me remind all of us that the basis for applied scientific inquiry, measurement, and inference is entirely reliant on the very same statistical laws and procedures cited and used by Truth. Let's stop standing Truth on it's head, accept that there are multiple paths to prove fraud in 2004, and move forward on all paths united to make sure it doesn't happen in 2006. To do otherwise is paralysis by analysis.

Thank you TruthIsAll for your great efforts.

[Jack Rabbit]

46. Please reread the root post

I have a small but important issue with Nederland's second point, namely, that the only place you can't really tell is in the area where the fraud occurred. This is because one really can't tell if there was fraud or if the polling was inaccurate. If one looks at polling outside the area where fraud occurred, then one could uncover fraud.

The caveat here is that the Exit Polls cannot also be fraudulent.

They aren't so much fraudulent as they can just be wrong for any number of reasons. An extreme example is the Literary Digest poll in 1936 that showed Alf Landon was going to trounce FDR. The poll was taken by telephone; however, in 1936, a telephone in one's home was a luxury. Most people who had a phone were upper income and, in 1936 as now, those in the upper income brackets favor the GOP. Obviously, the Literary Digest began with a biased sample and got a biased result. No one believes the editors of Literary Digest were out to deliberately deceive anybody; they just didn't know what they were doing.

A very readable book I recommend (no doubt TruthIsAll is familiar with it) is Darrell Huff, How to Lie with Statistics. Although written in 1954, it's still in print and its critique of statistical methods and fallacies is still valid.

Fraud can be uncovered as long as it is not perpetrated equally everywhere. In this case, the hypothesis is that the agent of fraud was new voting technology; if that is the case, then it will be uncovered by examining the discrepancy between exit polls and vote tabulation in places where new voting technology was uses as opposed to where it was not. Even if the exit polls are flawed, as long as they are consistently wrong in the same way, one would be able to uncover evidence of fraud this way. If the polling remains consistent and the voting patterns do not, then one is justified to wave a red flag.

[eomer]

50. The hypothesis is not that the agent of fraud was new voting technology.

The hypothesis is that a cacophony of frauds was committed, sordid details to be determined. I won't try to list them in this post.

This election is seemingly f***ed up in every way you can imagine. If fraud associated with one particular type of technology were overlaid on top of a noisy, fraudulent pile of ****, maybe you could spot it, maybe you couldn't. If you are a fraudster, chaos is your friend.

I should also point out that not all of the alleged frauds are high-tech. There are numerous reports of chain of custody problems and other breaches of procedure that mean even a stack of punch cards cannot be trusted as fraud free.

Also, in those precincts that use old technology, they really use new technology in the sense that even punch card ballots are counted by computer. So you can't really look for a correlation with new technology since they all use new technology.

Another problem is that different voting technologies have different spoilage rates even if there is no fraud. So how do you tell that a correlation is fraud and not a propensity for spoilage? In an otherwise orderly population maybe you could tell but, well, you know, pile of ****, yada, yada.

I think the essence of my disagreement with your proposition is that you are looking for too tidy of an answer. Don't get me wrong, if we get our hands on the precinct-level data, do the analysis and see a clear, strong correlation between voting technology and the disparity, I will add that correlation onto the large pile of red flags we already have.

But I don't share your fetish for this one possibility of a red flag when there are already so many other red flags lying around all over the place that it's getting hard to walk in here.

[anaxarchos]

51. One way to miss a smoking gun....

...is to set the house on fire.

[Carolab]

57. Right. And let us not forget the suppression and the tampering.

Two very important parts of the charges levied in the suits in Ohio.

We CAN prove suppression, through testimony as well as intimidation, false information, and insufficient machine-to-voter ratios.

We CAN prove tampering, through affidavits and (hopefully) through further discovery.

[TruthIsAll]

52. No, I've studied the classics, like Feller's "Theory of Probability"

I did not get my three degrees in applied mathematics in order to learn "How to Lie with Statistics".

Although I studied Probability Theory and Mathematical Statistics, they were only two out of 24 advanced mathematics and operations research courses.

I am not a statistician, which others wrongly have said I claimed. I never did. I am a quantitative software developer who has worked many years on engineering and financial problem-solving.

And due to my extensive experience in developing models which work, I am quite able to see through "proofs" and "theories" which have been presented here to camouflage real-world analysis with faith-based, pseudo science.

[autorank]

55. My response

My dear Mr. Rabbit: After working a long day dealing with all sorts of complicated issues and dynamics, I came home to what?...an intelligent response to my message. Now I have to think again.

You state: This is because one really can't tell if there was fraud or if the polling was inaccurate. Not correct. Exit polling and the voting-tabulation process are totally separate. Exit polling, properly conducted, will show within a certain percentage point, what the vote was. This process is based on generally accepted statistical procedures, methods, and laws. It doesn't matter if the ballot boxes were stuffed or the machines were hacked. You state: If one looks at polling outside the area where fraud occurred, then one could uncover fraud. If the intended purpose of exit polls is to catch fraud then the location where fraud takes place is based on a binary process. If the vote tallies match up to the exit polls then the vote tallies are likely (at some very high percent) to be accurate. If the tallies vary from the exit polls, then fraud is likely. Polling "outside the area where fraud occurs" is, therefore, based on the exit polls, not the other way around.

You state: They aren't so much fraudulent as they can just be wrong for any number of reasons. An extreme example is the Literary Digest poll... I'm familiar with that poll. There were quite a few intelligent people around then and I'm sure several of them pointed out the absurdity of this methodology. In the case of today's exit polls, or those since 1996 that Truth uses, there is intense scrutiny of methodology. The 1996 and 2000 Exit Polls were very close. The 2004 Exits were not that different in application. Therefore, the validity of the polls carries over from '96 and '00 to today and today's (2004) polls are accurate.


You state: Fraud can be uncovered as long as it is not perpetrated equally everywhere. Not so. They can steal the entire election in the same way with the same methodology everywhere, and we can still, through the Exit Polls, determine that the vote was tampered with. The Exit Polls are transparent, largely, while voting and tabulation is not. You state: In this case, the hypothesis is that the agent of fraud was new voting technology; No again. The agents were those who conspired to commit the fraud. It could have been a variety of technologies including hacking data lines, mainframes or other tabulation machines; or, think about it, simply changing a total and relying on fast foot work in cases where an unbiased audit is called.

I place my money on Exit Polls over actual vote tallies and believe that our discovery of fraud can start or be supported by Exit Polls.

:toast:

[TruthIsAll]

58. Thank you, Autorank, for your perfect analysis

Along with Peace Patriot, your postings are true classics.

[Sancho]

59. Forest for the Trees....

"My guess, since Mitofsky's not a dope, is that a peer review would show that his methodology was excellent and produced a very small amount of slop, the pre-'corrected' values (that Freeman used) were correct, and that therefore the only rational conclusion is that the election was stolen outright by a large number of people working independently toward a shared goal, each affecting a few votes."

I happen to be a social science statistician...there's too much focus here on the MOE and sampling error. All researchers know that hypotheses and the power to detect effects are somewhat biased by the "non-ignorable non-respondents" - and that's why Mitofsky is paid the big bucks to sample correctly. Mitofsky is VERY aware of the problems with his conclusion from a previously planned stratified sample - and he didn't pick a biased sample - even if that is the current report! Mitofsky could easily look at other characteristics (like who in Pinellas Florida voted for Castor) along with similar local things and see if there was fraud. He doesn't report that because he has already concluded there is a red flag. His report is hiding what he knows and I would conclude that from reading it. The explanations are improbable guesses without empirical evidence which should be available to the poll from the data.

Remember - the simple solution is usually the best - and PROOF is a legal requirement, not scientific evidence for a conclusion. I'm convinced there was a hacked election in some states and counties. My experience is that exit poll problems could be explained or investigated if Mitofsky and others in the court houses of Ohio were open with the raw data. It's simply not enough to say that some didn't want to answer the exit poll. There needs to be evidence that is the case. Not having open data must have reason....hmmmm.....wonder what the reason is? Mitofshy has a big paycheck and a monopoly - why cause a controversy?

Why not do an appropriate sample of key voters from the voting record and ask post facto (now) how they voted? Picking questionable counties and results would verify or deny the results - just as good as the exit - but still could be done. I'd bet that some Ohio and Florida counties would reflect the pre-election and unadjusted exit polls.

[autorank]

60. I am soooo saving this thread...and I'm taking it back to my

PhD program stat professor (of ahem... several years ago); do a little dance and tell him that I could have gotten an "A" without giving him that envelope full of Franklins. He used to live for the type of analysis TIA does and would regale us with tales of odds ratios and monte carlo analysis.

TruthIsAll :yourock: Thank you for your ongoing analysis. It is priceless!

[byronius]

61. This thread is gripping reading... but the emperor will decide.

At the core of it all is a general horror that any American would consider such an act permissible. As in any negotiation, as soon as one side chooses to step outside of the process, and lie, or dissemble for advantage -- the negotiations are over, whether both participants are aware of it or not. The delicate nature of democracy is utterly dependent on fair elections. Without them, all reverts to tribal struggles for raw power. And the other side seems to understand this perfectly.

I have to say, you guys are damned amazing thinkers and problem-solvers. Using my raw power, I hereby make you all Senators and Presidents. And what a damned great future it would be , too.

[libertypirate]

62. Circumstantially

I think it is even more statistically improvable that the glitch meter favored the resident...