Why weren't the exist polls brought up?

[catgirl]

Aren't they some of the strongest evidence we have?

[dogindia]

1. extremely good question

[myschkin]

2. Good question!

As I said before, they don't see the whole picture... :-(

[carolinalady]

3. I wondered the same thing I hope the GAO investigation

brings out some info on that subject.

[wiggs]

4. good question!

Because we didn't really want to rock the boat, just give the appearance that we are doing so. A fine line was walked today, for some reason.

Although a lot was said, a lot wasn't said as well.

[slackmaster]

5. Exit poll "evidence" is too easy to brush off as irrelevant

All they have to say is "There's only one poll that counts - the election itself."

[Lerkfish]

13. but they were relevant in the Ukraine.

[nursebear]

23. This isn't the Ukraine

Different rules, different systems, different laws....

[Kansas Wyatt]

26. Correct, this is....

The Banana Republic of America, formerly known as the United States of America.

[spooked]

6. Someone please get through to C-SPAN about the exit polls

n/t

[justsomegirl]

7. Because that's not the point of today's exercise.

The point of this whole thing was not to stand in congress and scream, 'they stole the vote' (contrary to the wishes of many people on this forum) -- the point was to spotlight the fact that many, many americans were denied their right to equal protection under the law, because they were not given the right to vote.

The point is that every citizen over the age of 18 has a right to vote and that in Ohio, many citizens were not allowed their Constitutional right to participate in the democratic process.

The exit polls are not germane to the argument.

[catgirl]

8. But it shows problems with our voting system

[nursebear]

28. Perhaps the exit polls were wrong?

I don't mean to fan the flames of controversy, but has it occurred to anybody that perhaps the election results were right and the exit polls were wrong?

Follow me on this for a second:

If I have a bag filled with 10 socks, and I pull out two yellow socks and one red sock, what's my assumption?

I would (logically) conclude that 2/3 of the socks are yellow and 1/3 are red.

Okay... that's reasonably sound logic... and that's kind of what we saw in the exit polling.

But then suppose I pulled out all 10 socks, and found out there were 5 red socks, 4 yellow socks, and 1 green sock.

Would my "exit polling" still be correct? Of course not. I'd have to go with what the final results were when I pulled all 10 socks out.

Likewise, assuming the exit polling data is correct and the final vote tallies aren't is not logical.

And in this election, more than most elections in the past, one has to look at where the exit polling was done. If there were more exit polls in urban areas (which is likely), they're going to skew to the Democrat side.

My own state of Washington voted overwhelmingly for Kerry, but there's only half dozen counties that went "blue." If most of the exit polls set up where in Seattle/King County, it's going to be over represented for Kerry.

Y'all need to get over the exit polling irregularities. You can't prove it, it's easily dismissed (as it should be), and holds no scientific or statistical basis. IF you want to prove voter fraud, you'll need to find actual people who were disenfranchised. You need to find actual evidence that votes were tampered with. And above all you need to realize that there are probably a lot of politicians over the years who have squeaked by because of voter fraud (on both sides of the aisle)... they won't necessarily want to open up that can of worms and see their own elections sullied by questionable voting practices.

I doubt anybody will listen to this, but it's time to move on...at least from the exit polls...

[davidgmills]

50. Please see my post at #48

Counting ten items is one thing, counting 120 million, with umpteen different counting mechanisms is something else.

[Yupster]

66. The New Hampshire recount

ended the exit polls as relevant in my opinion.

[davidgmills]

31. Wrong.

Exit polls are very germaine to equal protection. Whether one doesn't get to vote to begin with or whether one's vote is lost electronically, the result is the same.

Exit polls are the only check on the system. They are the only way to "check your work." Like the math teacher used to tell you, the only way of accurately estimating whether your answer is clse to being correct.

Here is the best anology I have heard, from a freeper no less. The actual vote is like an atomic clock and the exit poll is like an extremely expensive Rolex watch. They both are capable of keeping excellent time. But sometimes even the atomic clock can get off and if it does, you need that Rolex as a backup.

I really don't buy the comparison between our election process as it now stands and the accuracy of an atomic clock. It's more like an atomic clock that is seriously broken. But the point is that we need both actual tabulation and exit polling and we need both to work and be checks on each other.

[iconoclastNYC]

9. One (R) Repren said : "the only thing less credible than exit polls..."

Are godzilla movies. They dont want exit polls, it makes it too hard to cheat.

[Doctor O]

10. Exit polls can not be evidence because

1. NEP warned the networks that they were not to be used for reliable indication of calling races.
2. In 2002 exit polls were inaccurate
3. In 2000 the exit polling was so inaccurate, the networks fired the exit pollsters and hired NEP (Mitofsky)
4. In 1992 and 1996 the exit polls called the correct winner, but they were up to 8 points off in favor of the Democratic candidate.
5. In NH they recounted 11 precincts that showed the biggest disparity in votes and after analyzing those found that the exit polls showing Kerry ahead in the state by 16 point in the exit polls to be inaccurate.

In 2004 Carter worked as an election inspector in Venezuela he found that it was a fair election, but yet the exit polls showed Chavez being beat by his challenger by a large amount over 16 points, and it was run by a leading US pollster Doug Schoen who has done exit polling.

Therefore, the exit polls were never meant to be evidence only an in dictation of potential problems.

[davidgmills]

48. Your analysis is right only if

You can prove that the actual tabulation was right. Give me proof that the actual tabulations for 1960, 64, 68, 72 ...96, 00, and 04 were spot on before you can tell me the exit polls were "wrong."

Freeman makes the excellent point that the myth of the count being accurate is just that, a myth.

He points to the census takers who believe the actual count of the 2000 census was off by as many as six million people and that experts agree a statistical poll would have been much more accurate. The census takers had months to get it right and they couldn't. There was much less at stake and it was not nearly as partisan. So if the census is a guidepost as to how well we count big numbers, we should have no faith whatsover in our ability to count an election.

The best analogy I have heard is that the system ought to work like an atomic clock (actual tabulations) with a fine Rolex (exit polls) as a check on the system.

Right now the atomic clock looks broken and may well have been broken for some time. In that case, I'll go with the Rolex.

P.S. I don't own a Rolex. Pick your favorite timepiece for the analogy if you wish.

[Blue22]

63. Forget the polls, let's get our hands on some tabulators

We've been going in circles with the polls.  Without the raw
interviews, the conflicting exit poll results are open to
interpretation by anyone.  Let's face it, most of us wouldn't
know MoE from Curly or Larry.  Let's put that discussion on
the back burner until the raw interviews are released in March
according to Mitovsky.  Let's focus on the tabulators.  Here's
a good start: 
http://www.scoop.co.nz/mason/stories/HL0307/S00065.htm  How
about if everyone in a state that used any kind of electronic
tabulator submitted a freedom of information request to their
state and county BOE.  You can use Bev Harris' FOIL request as
a good starging point.  Whatever we receive we post here for
everyone to examine.  You can never tell when someone will
spill some proprietary code.

B-)

[Pacifist Patriot]

11. I heard a couple of references to exit polls.

They weren't hammered on, but they were mentioned. One Republican, sorry can't remember who...they all look alike, said the challenge was bogus because democrats wanted us to base our results on exit polls rather than certified votes. On the other side of the aisle, I believe I heard at least two senators mention the discrepancy with exit polls but nobody pressed the issue. Clinton may have been one of the two. Probably as a segue to bringing up Ukraine.

[euler]

12. If all you read is DU, then...

...that is the impression you would have. The US exit polls aren't evidence for anything. Our exit polls aren't comparable to the ones in Ukraine, Germany and most of Europe. They design exit polls as true random samples. This one wasn't designed that way.

This information is readily accessible via Google. Here is my advice to all DU members. If someone you don't really know makes a statement of fact here on DU, and it is something that you can verify yourself by using Google, why not do so ? Especially if the topic is important to you. There is a lot of misleading stuff written here. And the claims made here about exit polls are the worst.

Hopefully people are more receptive to this now. Please note: I'm not telling you to do or believe anything. I'm asking you do do yourself a favor and verify what you're told by DU posters you don't know.

[TruthIsAll]

14. Mr. Mitofsky begs to differ with you.

[Mistwell]

16. What Bunk!

He's ON THE RECORD saying his poll, which you just posted a copy of, can not and should not be used to prove or disprove election results. That it is just a rough guide of demographic data.

Why does TruthIsAll feel the need to exagerate the Truth to make his point?

[TruthIsAll]

38. Rough Guide? Sure, with a 1% MOE. n/t

.

[Mistwell]

43. Or not

He tells you it is a 3%, not a 1% (and, despite your picture of the newspaper, the newspaper got it wrong and the actual website FROM the folks who ran the poll get it right).

In addition, it's a 3% FOR THE PURPOSES STATED. In other words, for the purpose of demographics...not the actual vote.

[TruthIsAll]

44. Got it wrong? 13,047 RANDOMLY-SELECTED IS A 0.88% MOE.

I AGREE WITH MITOFSKY.

[foo_bar]

51. that's called cherrypicking

You agree with the parts of Mitofsky that support your preordained conclusion, then ignore his stated MoE.

Before the election, you cherrypicked every poll that had Kerry leading, claiming the other ones were "biased" (Newsweek, Time, CNN/Gallup, Zogby when he disagreed with you). This supported a preposterous 99.whatever% probability, as any dissenting datapoint was discarded. In science this is called fraud.

Once the veil was lifted, you spent two months trying to rationalize your initial false assumption with a crescendo of new false assumptions. When you were called on it, you invoked the "with us or against us" fallacy and refused to admit fault, not unlike a certain world leader.

Now there's nothing to show for any of it, except a conviction that God was proverbially on your side.

[davidgmills]

53. Since you have accused TIA of fraud

I'm just curious, what are your mathematical qualifications? He's posted his before.

[foo_bar]

54. a mail order diploma does not an argument make

Nor does expertise in one field (a subset of mathematics) apply to others (science, logic, metaphysics). My background is in science, and your rebuttal is known as a fallacious appeal to authority:
http://en.wikipedia.org/wiki/Appeal_to_authority

You don't have to be a mathematician to debunk Creationism, geocentrism, and other faithful substitutes for reason.

[davidgmills]

58. Re: Fallacious appeal to authority

I'm a lawyer. The first question we always ask anyone who is claiming expertise is to ask for their credentials.

No credentials -- no talkie.

That's the way it works in court.

[foo_bar]

60. Re: Re: Fallacious appeal to authority

And we aren't in a court, so your qualified expertise is another inapplicable appeal to authority (of all things!)

If your claim wasn't a logical fallacy, your lack of mathematical credentials would prohibit comment on this thread, as would TIA's admission that "I am not a statistician":
source

[davidgmills]

64. Last I checked claims of fraud

have legal conotations. You accused him of fraud. I just wanted to know the credentials of someone who would accuse someone else of scientific fraud.

You are absolutely correct that my background does not give me the abilty to assess TIA's judgments or yours. Sometimes the only means I have of assessing something technical is a person's credentials.

If you have never had a math course in your life, I would like to know that. For what it's worth, I had calculus in college and my father and brother are both PhD's in biochemistry.

I am a personal injury lawyer, have been for 27 years and know a considerable amount about medicine (initially I was a science major). My father taught biochemistry to medical students and graduate students for 40 years. I use M.D.'s, economists (PhD's), and rehabilitation experts (PhD's psychology), all the time. All of these people deal with a lot of statistics.

So I think I have a legitimate right to inquire about the background of someone who makes a claim of science fraud.

Moreover, it is not just TIA's claim that the MOE's of the exit poll were one per cent. Freeman and Baiman think so as well. Maybe they are not good enough scientists for you either.

But the real question is not the exactitude of the exit polls, it's the exactitude of the exit polls compared with the exactitude of the actual tabulation. If the MOE of the exit poll is 3% and the MOE of the actual tabulation is 6%, the exit poll is still more accurate.

Freeman raises the question of the other part of the equation -- just how accurate are the actual tabulations? If we are going to be criticising the MOE of exit polls, shouldn"t we be just as critcical about the MOE's of the actual tabulation?

[foo_bar]

65. to be "science fraud", it would first have to be science

I don't believe I made the latter claim, and by extension the former ("in science, this is called fraud"; note the prepositional phrase). Nor has TIA, to my knowledge. If it were passed off as science (note the subjunctive), it would be immediately discredited for reasons stated above: cherrypicking datapoints, preceding a hypothesis with an unfalsifiable conclusion, calling the other scientists (to borrow the example) a bunch of "trolls" and "freepers" (no link, as the posts in question no longer exist: consider it hearsay in the metaphorical DU court in the sky), and changing one's base assumptions from week to week in accordance with a predetermined conclusion.

So I think I have a legitimate right to inquire about the background of someone who makes a claim of science fraud.

You would have that right in a court of law, in a state in which you passed bar. On a message board, you have only your disembodied voice. You can factually rebut an argument, or you can change the subject.

Sometimes the only means I have of assessing something technical is a person's credentials.

Herein lies the difference between the truth (lower case t) and the law. Mathematical proof is subject to a far more rigorous standard than 9 out of 12 jurors, or "expert testimony" by paid experts.

But since you've taken an interest in me personally:
http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=237348&mesg_id=238713

I too indulged in Calculus in college, but my vocation has focused the application of cognitive science (pseudoscience, in truth) to computer science, much of it rooted in mathematics and philosophy (although differential/integral equations haven't reared their head since graduation). Or it did, before an early retirement on the heady stock options of the late 90s.

I've also had clients who were distinguished in their fields, but I wouldn't brag about it here, as though credentials could be absorbed by osmosis. If you think discarding data that disagrees with a foregone conclusion is defensible, defend it; character witnesses don't fly in the court of truth.

[davidgmills]

76. As a non-scientist, I can tell you what I am looking for

and that is a consensus of the best mathmatical, scientific, and statistical opinion of the accuracy of the exit polls.

Was it slightly less than the one-percent as TIA claims, or about what Baiman (came in at 1.1%) claims, or Freeman (I'm not sure he has come up with his national number yet but I believe he thinks its in that range), or was it the one-percent claimed by the Washington Post article, or could it be over 3.5% like Mystery Pollster (Mark Blumenthal) claims, or is it some other number, more or less?

I don't really care to get in the argument of whose analysis is right and why because this is not my field. (BTW, I read the entire thread you cited because I wanted to see the context -- but I see no need to comment on it). All I want to know what the consensus is.

What's yours? Got a number in mind? I'd like to know it. I don't care to see the analysis. Just tell me whether 1% is in the ballpark and if not give me your best opinion of what it is.

Here's why. Freeman has changed the debate. He has challenged the myth (by the vast members of the public) the the actual tabulation has an MOE of near zero. Well it is obvious that it doesn't when you think about it. He's got us thinking about it now. He's got us thinking about the MOE of exit polls vs. the MOE of the actual tabulation. As between the two, I'm inclined to believe that even if the MOE of exit polling is 3% the MOE of the actual tabulation is likely to be much greater.

And what my agenda is, if you care, is how best to sync both up and have both be as accurate as is reasonably possible. I am afraid that much of the upcoming debate will focus on the elimination of exit polling, which is the only apparent check (even if flawed) we have on the system. I believe that would be disasterous.

Rather than the elimination of exit polls, I believe that next election we will need a unquestionably accurate exit poll in place to deter what many of us suspect as fraud. It is my goal to see it happen. You can have all the election reform in the world, but if you can't check your work it is all for naught.

[foo_bar]

77. I agree in principle with most of what you said

What's yours? Got a number in mind? I'd like to know it. I don't care to see the analysis. Just tell me whether 1% is in the ballpark and if not give me your best opinion of what it is.

Here's where we differ: I consider this a variation on "how many angels dance on the head of a pin?"

You're right that the actual tabulation has a nonzero MOE, but polling MOEs themselves have their own nonzero MOE, barring a truly random sample. While it's possible to design an exit poll around the most stratified sample in conception, no quantity of fudge factors can reproduce a "true randomness" that may not even exist, even extraneously to statistics that (always over-)generalize populations from samples ("So what do we do now, foo_bar? Wait for death?") (Even the hypothesis of a random coin flip has been disproven (the distribution typically favors the side on which it was flipped, which typically favors heads). A random coin-flipping machine would need to be fed a truly random sequence of numbers, which is arguably a chimera; the closest thing we have is to measure atmospheric noise, which itself presents biases in a world of finitely regressing causalities (as anyone who's even tried to define randomness without a circular definition would attest, "The generation of random numbers is too important to be left to chance.")) Figuring out what's random with respect to human beings is utterly unattainable without a nascent science approaching Asimov's fictional "psychohistory", as we can barely define the parameters of human behavior, much less a consensus definition of randomness or the probability density of speaking truth to pollsters. All MOEs derived from inferential statistics (as in, sample sizes under 100%) are potentially underestimated for reasons of bias above and beyond "sampling error", which statisticians can only (over-)compensate for with recursive inferences about the road not taken.

But I digress. To assess the "true" (as in, mathematically provable beyond an estimation or a teleological inference) MOE of the Mitofsky poll, I'd have to be present for the sampling or have access to raw data, and then I'd have to be omniscient. So I wouldn't look to inferential statistics for a hard answer on anything. The "true MOE" can't be axiomatized until we can truly extrapolate human behavior from a truly representative sample, as (somewhat bitingly) postulated in Asimov's Foundation. But I wouldn't find much work as a statistician with this attitude. The minimum MOE is TIA's chi-square, the maximum is all of it but the part that was counted correctly. In other words, don't waste your time looking to prove the unprovable; either go out and find direct evidence of fraud, or run with the hypothesis and find ways to curtail fraud next time around, or meet random (:)) people to remind them of their collective importance to the greater scheme.

[foo_bar]

78. in less words

Rather than the elimination of exit polls, I believe that next election we will need a unquestionably accurate exit poll in place to deter what many of us suspect as fraud. It is my goal to see it happen.

An unquestionably accurate exit poll is out of the question, like an unquestionably accurate "actual tabulation", for mostly the same reasons. If you could conduct such a poll, you wouldn't need an election (but to dispute pesky charges of disenfrachisement and one man one vote*). If you could exit poll 100% of voters, you would be called a voter-verified receipt.


*edit: this is the premise of another Asimov short story, the name of which I can't recall. The powers that be conclude that voting is too messy and inaccurate to conduct in good faith, so they select a truly-random Average Joe every number of years and extrapolate his beliefs to the populace at large with some sort of hokey process involving robots and the talent portion of a beauty pageant.

[davidgmills]

79. My rationale for asking a question that may have no answer

I can appreciate your unwillingness to answer a question you feel has no answer because you believe the premise is wrong.

I do not believe in life after death, arguably the whole basis for most religions. I do not go to church, because I can't agree with the premise on which most people base their faith. Like Nietzche, I tend to believe that man created God because man couldn't face the concept of his immoratlity. Nevertheless, I have to deal with the consequences of people and their religious beliefs every day.

There are some scientists who say that there is no such thing as time. They may be right and that time is no more real than religion may be and that man has created time because he can't face the reality of there being no such thing as time. Real or not, we try to measure time anyway. I mention time because one physicist, who has posted on Mystery Pollster and Stones Cry Out, has made the analogy of comparing tabulated votes to an atomic clock and exit polls to a highly accurate Rolex. It can certainly be questioned how accurately even these devices can measure time when time may not be real.

But I happen to like the analogy because like it or not, I have to deal with the consequences of another man-made creation, elections.

Just as time needs a measurement to satisfy the needs of man, so does the concept of elections. I have to deal with the consequences of those who believe in elections just as I have to deal with the consequences of those who believe in time.

Elections bring us two measuring devices, actual tabulations and exit polls. They exist in the minds of man. I have got to deal with the consequences of their existence. So as long as I have got to deal with them, I would like these two measuring devices, as flawed as they may be, to sync up as best as possible.

[foo_bar]

80. I didn't mean to lapse into epistemology

My consensus-reality point is this: a sample of 120 million people has more potential to measure public sentiment than a sample of 10000 people. It's still a form of inference, but it's the most complete sample we have. To suggest otherwise is to assume that humans are redundant, even unworthy of suffrage, a sentiment that Asimov parodied in his speculative fiction.

In other words, the poll that needs fixin' is the one on election day. Audit trails, quis custodiat, even "smart cards" could go a long way in reducing the MOE where it counts. "Syncing" it with exit polls is irrelevant until science can demonstrate a random person, a technology that remains the realm of science fiction. Zogby can make a gambler's best guess, so he wins more often than an amateur, but he loses some too. The pollster's MOE is merely hedging a bet, and even the hedge needs hedging when the "random sample" allows for voluntary noncompliance. Barring a major breakthrough in chaos theory, the 100% sample of 120M+ (not really, but close as it gets*) is the place where we can learn the most about our own intentions.

*edit: If TIA's mistaken assumptions regarding 1/sqrt(n) were applied universally, the (not-so-random) sample of 120 million would imply an MOE of 0.0091%. To categorically assert that the smaller poll is more accurate (in spite of calling the pollster a co-conspirator in fraud, and thus no more reliable than the large sample) is no different from discarding pre-election polls on a mere hunch of "bias", which inadvertantly created more bias than it mitigated. That's where the snowball began rolling.

[davidgmills]

81. Certainly I would agree that a sample of 120 million people

Has the potential to measure public sentiment better than a sample of 10,000 people. But I think you hit the nail on the head when you used the word potential.

The problem is accurately counting 120 million anything much less people. And has Freeman has pointed out, there is all kinds of incentives to miscount. Imagine trying to have a group of people count 120 million single dollar bills. Not only would the human abilities of concentration be put way past their capacity, but the potential for skimming off a bill or two, or twenty or fifty, or 100, or whatever on the assumption they would never be missed is enormous. The financial stakes of tabulating an election are just as high as the financial stakes of counting money.

And I guess I have come to the conclusion that until we have a breakthrough, like chaos theory, a gambler's best guess may be more accurate.

[TruthIsAll]

83. David, did you read those 20 books from you-know-who?

How come he (won't mention the name) did not give you a direct answer and asked you to pour through them to find his rebuttal(s) to my stuff?

[davidgmills]

88. Sorry I don't remember who told me to read the 20 or so books

Was it the Stones Cry Out guy? In any case I didn't read them, and I doubt that whoever it was gave me a direct answer, because I haven't gotten very many direct ones.

If someone doesn't want to answer because they don't think the question can be answered with today's level of undertanding, well so be it.

[Name removed]

84. Deleted message

Message removed by moderator. Click here to review the message board rules.

[foo_bar]

86. it's not ad hominem, it's the TIA version of the scientific method

You did mention that the burden of proof is on the accused. While turnabout isn't fair play, you demanded my credentials without offering your own. From your evasiveness it's clear that you aren't particularly proud of your higher education, except to suggest that it took place*.

*on edit: over and over again, whenever someone criticized your assumptions. I didn't raise the subject until you used the number of degrees, if not the type, to ward off examination by your DU peers on many an occasion. When you were asked for specificity, the subject quickly changed to the accuser's authority.

[foo_bar]

90. re: version 2 of your post

You are a charlatan.

That's how you lose an argument. Edit for DU rules: "For example, if someone posts factually incorrect information, it is appropriate to say, "your facts are wrong," but it is not appropriate to say "you are a liar."

Did you ever coinsider that the sample is real?
And that the votes were rigged?

For about a second, before I considered Occam's Razor. That 3 million votes being stolen without a peep is an unfalsifiable doctrine instead of a single polling pool being off by the same margin on a sample 1/10000th the size is a testament to your unique approach to truth. Kind of like dropping Zogby when he had Bush ahead, then re-introducing him when he didn't.

Where is your analysis?

Up and down this page. If you read as much as you posted, you'd have some truthful assumptions under your belt by now.

You believe in faith-based science.

Projection, thy ironic handle is TruthIsAll. The faith required to accept your unfalsifiable hypotheses exceeds that of any science.

[TruthIsAll]

55. You are the one cherrypicking and confusing national (1% MOE)

with state (3% MOE) exit polls.

So Mitofsky was right. Both times.

He said 3% for state exit polls. That's true for a poll of 1,000.

He said 1% for the National Exit Poll. That's true for a poll 10,000.

So there is no conflict.

He was right.
You are wrong.

Put away your strawman.

[foo_bar]

59. Mitofsky seems to think otherwise

The margin of error for a 95% confidence interval is about +/- 3% for a typical characteristic from the national exit poll and +/-4% for a typical state exit poll. Characteristics that are more concentrated in a few polling places, such as race, have larger sampling errors. Other nonsampling factors may increase the total error.
source

In the words of Mitofsky: "The presidential exit polls released at poll closing time when they were completed had an average error of 1.9 percentage points. There were no mistaken projections by Edison/Mitofsky or any of the NEP members... Only the unauthorized leakers and bloggers were misled - a fate they richly deserved."
source

He said 1% for the National Exit Poll. That's true for a poll 10,000.

When did Mitofsky say this? Inference doesn't count.

Put away your strawman.

In the words of Inigo Montoyo: "I do not think you know what the word means."

[chorti]

91. You are confusing what Mitofsky says

The MOE of a characteristic of the national poll is about +/- 3%. The exit poll is not primarily designed to predict the characteristics of the electorate, it is designed to predict the voting of the electorate. The MOE of the national vote predicted by the exit poll is about +/- 1%. The WP is right.

[foo_bar]

92. so the WP is right, and Mitofsky is wrong?

"The presidential exit polls released at poll closing time when they were completed had an average error of 1.9 percentage points. There were no mistaken projections by Edison/Mitofsky or any of the NEP members... Only the unauthorized leakers and bloggers were misled - a fate they richly deserved."

This suggests that Mitofsky believes the margin of error exceeded 1.9 percentage points (on average). Is that a correct reading?

[chorti]

95. the polls released at poll closing time, yes

As Mitofsky etal then checked their polling with actual precinct results in the precincts where they polled, they were able to reduce the MOE to about 1% by midnight, when the WP published those results.

[foo_bar]

96. but the Mitofsky email came two weeks after the fact

Is he not saying that the 1.9% deviance was within the MOE?

[chorti]

98. doesn't matter when he wrote the email

He is saying that the exit polls released at poll closing time, such as at 7PM, had a MOE of 1.9%. He is NOT saying that the exit polls released at midnight had a MOE of 1.9. Read that closely. This is important. What happens at poll closing time? Mitofsky is immediately sent the results for all the precincts where they did polling by the local election officials. That is part of the agreement - they expedite delivering specific precinct results to Mitofsky. He then compares his polling of the sample precincts to the actual results in those precincts, repeatedly adjusting the weighting of the national and state samples, with each new precinct comparison. Each re-weighting of the sample reduces the margin of error. So that by midnight or 1 AM, the margin of error is down to about 1.0%. This is when the WP numbers came out.

[foo_bar]

99. did he actually say the MOE was 1.9%?

I took his statement to mean the 1.9% error was within the MOE (i.e., >=1.9%).

[TruthIsAll]

101. Chorti, don't you see? Facts don't matter. Now that you have

confirmed the validity of the 1.0% MOE, you have also implicitly proved the election was stolen. foo-bar will never stipulate to that simple truth.

That's why the naysays try so hard to fog the MOE issue.
Once you agree that the MOE is 1.0%, its all over.
I have computed the MOE as 0.88%, which is 1/sqrt(13047).

The MOE is exactly 1.0% for a 10,000 sample size (1/sqrt(10000).
But why quibble? I will not contest the 1.0% MOE supplied by Mitofsky to the WP, even though that seemingly small 0.12% difference makes the odds go much higher - into the BILLIONS!

They tried to obscure and manipulate the Ohio exit poll charactersistics, but...

******* THEY FORGOT ALL ABOUT THE 13,047 NATIONAL EXIT POLL! ******

They just didn't have the time to "manipulate" the results.
Thank you, Jonathan Simon, for the download.

THE WP NATIONAL EXIT POLL graphic will become a HISTORIC document.
IT IS THE BIGGEST SMOKING GUN AMONG MANY OTHER STATE EXIT POLL SMOKING GUNS.

Note that Bush's vote tally rose by 3.01% over his WEIGHTED exit poll calculation. That's 3 times the MOE, or 6 standard deviations from the mean.

The odds are 547 MILLION to ONE against that occurrence.

This confirms your INDEPENDENT calculation from the other day.
You concluded that the odds were 500 million to 1 against the final absolute vote tally, based on the impossibility of the actual voter turnout for Bush.

Our analyses, although done completely independent of each other, are in full confirmation.

And Baiman and Freeman also agree.

[Mistwell]

67. It's not a strawman

If something is a fallacy, it isn't necessarily a strawman. You seem to have that word confused with the word "fallacy". In this case, you seem to think his statement is false. However, it IS on topic. It is not an imaginary opposition set up only to be easily confuted. He is not mischaracterizing your position only to make it seem weaker than it is so he can refute it. He is DIRECTLY attempting to refute your statement. It might be a false refutation (though I personally agree with him and think you are wrong), but it is in no way a strawman.

I've seen you use the word strawman several times now in an improper way. I just figured you would want to know that you're misusing it. Since, you know, Truth Is All to you.

[Yupster]

68. Cherrypicking

TIA had a chart showing the pre-election polls favored Kerry. The Pew Poll had Kerry up by one point.

I thought the pre-election polls were kind of all over the place, with a slight lean to Bush, so I checked into it.

The Pew Poll actually had Bush up 51-48 in their final poll. They hit the final numbers exact.

So I asked TIA what was up with his Pew poll showing Kerry up 1 %.

He said he used registered voters not likely voters.

That was the last time I took his stuff seriously because in a close election, you can cherrypick information to show anything you want.

[chorti]

94. please check out report on pre-election polls

Yupster, please check out my report on how well the pre-election polls matched the final adjusted exit polls.
<http://www.indybay.org/news/2005/01/1713046.php>

I used an average of all the pre-election polls. The pre-election polls were accurate in the blue battleground states but not accurate in the red battleground states. Nationally the pre-election polls matched the exit poll (projected backwards) within 1.1 points on Oct. 21 and within 0.4 points on Oct. 28. The pre-election polls then showed a major surge by Kerry. He was down by 3.8% on Oct. 28 but closed to only 1.2% down on Nov. 1. A 60-40 split by voters deciding on election day would have brought the national vote to a dead-heat. I believe that happened. Yet, the reported margin was about 2.7%. The exit poll (projected backwards) showed only a minor Kerry surge the last 4 days. I believe this exit poll was manipulated.

[TruthIsAll]

100. And you cherry-picked LV polls. n/t

.

[TruthIsAll]

41. The truth is in that graphic. Read the notes at the bottom. n/t

.

[foo_bar]

71. in other news that fits in print:

If you believe a tertiary rendering over Mitofsky's own statements, you might be cherrypicking.

[TruthIsAll]

82. Nice picture, but once again you resort to theatrics.

1)The 1948 election was not exit polled.
2)The last pre-election polling was done several weeks before the election.
3)In 1948, polling was nowhere as accurate as it is today.

So much for your Truman analogy.

[foo_bar]

85. the clinical term for your subject header is "projection"

It wasn't a Truman analogy, it was a fallibility-of-journalism analogy. But if one jpeg supports your cherrypicked conclusions of November, I suppose that outweighs the pollster's thoughts on the subject, or the 1/sqrt(120,000,000) Margin of Error of 0.0091% on election day (using your fascinating theory of randomness) for a much smaller sample by a pollster you accused of conspiring in fraud.

[euler]

72. I can't see anything here that shows Mitofsky differs with me

Be less cryptic so I can answer you.

[TruthIsAll]

87. READ THE NOTES. READ THE NOTES. READ THE NOTES.

Is that being too cryptic?

[Freddie Stubbs]

73. Funny how no Representative or Senator thought that this was worthy of

talking about today. :shrug:

[euler]

74. Despite repeated requests, you obviously never...

...researched how exit polls are conducted or the different types of exit polls that exist.

The exit poll conducted for the 2004 presidential election was designed to allow the MSM subscribers to report who voted for which candidate and why. In other words MSM wanted to know the demographics of the election. So, Mitofsky designed a exit poll that would give them that. Therefore, all the demographic information in the exit poll (which you have been kind enough to post here), is accurate because the exit poll was designed for this purpose.

The exit poll was not designed for election verification or fraud detection. So, you can do all sorts of mathematical gymnastics on the data, but you will never be able to prove or disprove fraud.

On page 3 of Stephen Freemans's paper title "The Unexplained Exit Poll Discrepancy," Freeman acknowledges this: "The pollsters have taken great pains to argue that their polls were not designed to verify election results, but rather to provide election coverage support to subscribers"

A footnote on the same page states: "Warren Mitofsky, the founder of Mitofsky International, is credited with having invented the exit poll"

If the man who invented exit polling tells me that his exit poll was not designed to verify the election, then I believe it and so should you. But, you know all this already.

[catgirl]

15. What is the "misleading stuff written here"?

and "the claims made here are the worst"? I agree you
shouldn't believe everything you read, that's why I don't
read mainstream media very often. Who are you suspicious of here?
I don't know you, so I guess I should be suspicious of you.

And the exit polls WILL be used in the Arnebeck lawsuit, I'm sure.

[Mistwell]

18. It's in this very thread

Virtually every post TruthIsAll makes about the exit polls radically exagerates the extent of the evidence and the analysis. I think that is the bulk of what the poster was referencing.

[foo_bar]

70. the TIA of this week contradicts the TIA of two weeks ago

This thread:

Got it wrong? 13,047 RANDOMLY-SELECTED IS A 0.88% MOE.
I AGREE WITH MITOFSKY.

Two weeks ago:

NPR/ Mitofsky has not released the raw data. Why? Could it be because it would confirm a Kerry win? Occam's razor would seem to apply in this case.

source

This isn't cherrypicking or dishonesty by omission; it's an ever-changing assumption that can't be falsified. Once it is proven false, the finish line moves, so "Weapons of Mass Destruction" becomes "spreading freedom" becomes "deposing a cruel dictator" becomes "whoever said anything about WMD?", as it were.

[TruthIsAll]

89. Mitofsky has not OFFICIALLY released the data.

But Simon downloaded it for us.
You call the chart cherry-picking?
Mitofsky agrees: 1% MOE
Gore won the popular vote by 2-3%.

The Normal distribution applied to Mitofsky's own numbers say that the odds were 1 out of 547 million that Bush could exceed his polling numbers by 3.01%. That's SIX standard deviations from the mean.

Do you even know what the standard deviation is?

Oh, once again.

I HAVE PRESENTED MY ANALYSIS.
SIMON HAS SHOWN US HIS.
FREEMAN HAS SHOWN US HIS.
BAIMAN HAS SHOWN US HIS.
WE ALL AGREE.

WHO AGREES WITH YOURS?
OH, I FORGOT.

YOU HAVE NOT PRESENTED YOUR ANALYSIS.

WHERE IS YOUR ANALYSIS?

[foo_bar]

97. the final analysis is 286 EV Bush, 252 EV Kerry

The faith-based conspiracy theories are unworthy of analysis, as you can create an infinite number of alternate timelines with different baseline assumptions, as you demonstrated. Why not calculate the odds of Bush winning 2000 legitimately? Because it's a done deal, and the threads sink like rocks. Hopefully you'll discover the "next big thing" in time to suspend the illusion of Rumsfeldian infallibility.

[davidgmills]

52. Please see my post at #48

Where on Google does it say that we know for certain the tabulations of the last 20 elections have been spot on?

When exit polls are "wrong" they are only "wrong" because the tabulation is presumed to be "right." That requires a leap of faith.

[Freddie Stubbs]

17. Because they were not arguing that Bush didn't win

Did anyone here any Representative or Senator say that Kerry really won? :shrug:

[jayctravis]

19. Right!

Bush said because Ukraine's exit polls didn't match that how could they POSSIBLY be indicative of fair elections...

[McCamy Taylor]

20. I tried to get through CSPANs phone lines but never could.

Exit polls are what the Special Prosecutor should focus on. I guess Congress is focusing more on election reform, but the fact that they wont give them to Conyers indicates that there is a problem the MSM/private election companies and government accountability

[Machiavelli05]

21. The more people bitching about trying to overturn the election

the worse the Democrats will look.

[bones_7672]

22. Exit polls have no legal standing. n/t

[catgirl]

24. This discussion isn't about overturning the election

(although that would've been nice). It's questioning why
they weren't mentioned today. Why don't they have merit
in our country anymore?

If it was a reversed situation, you better believe the repugs
would be shoving them in our faces.

[Not a Sheep]

25. Evidence Kerry won? Dems said they weren't trying to prove that. /eom

[info being]

27. Because our system is more of a sham than Ukrainian democracy

[bemis12]

29. It's cheap and lazy to compare the US to the Ukraine.

They didn't get a new election because of exit polls. They got a new election because there was hard evidence of systemic fraud and voter intimidation. Comparing the US to the Ukraine shows how little you know about the Ukraine election.

http://www.telegraph.co.uk/news/main.jhtml?xml=/news/2004/11/28/wukra28.xml

"They started to beat voters and election officials"

"People's faces were cut from blows to the head. There was blood all over"


You tell me where in America voters were beaten, poll boxes destroyed, and ballot boxes were opnely stuffed. Then we can talk.

[catgirl]

34. How 'bout voter registration forms destroyed?

[info being]

42. You're right. The fraud is more sophistocated here.

[mattclearing]

30. I think they decided to score points rather than cast doubt.

They goaded the Republicans into supporting a weak unverifiable democracy, and screaming sour grapes when nobody contested the actual victor of the election.

This makes Republicans look stupid, which is only fitting.

[FreepFryer]

32. because that would immediately be associated with tinfoils, and Mitofsky

hasn't legally released their data.

(in my opinion).

[Machiavelli05]

39. totally agree! If Exit pollign were hammered in - it would have

illegitimized the entire issue in the minds of way too many.


Look how exit polls have created such difference of opinions here - imagine when the audience is the uninformed, and less Dem. partisan public?

[candice]

33. Exit polls were mentioned multiple times

..I was moving back and forth between the two chambers (two CSPANS) during the "debates."

[catgirl]

40. Did you notice how different the levels of passion were

between the chambers? What's up with that?

[brindis_desala]

45. It's the reason I didn't expect more than one or two senators to

support Senator Boxer. The Framers devised the Senate because they were suspicious of "we the people". The Senate is expected to be the "sedate" house (read sedated) Passion is a no no. It's antiquated and anti-democratic just like the Electoral College. It's why these elections are so tough to get right.

[catgirl]

47. Time to bring some young blood in

These unpassionate sedators are stepping on justice.

[brindis_desala]

49. Time to scrap it altogether. I don't know what state you live in

but how can it be "democratic" for Wyoming with 450,000 people to have the same representation as California that has 36 million and the 6th largest economy in the world?

[FreepFryer]

56. Well, they made a step - they're allowed to refer to each other by name.

No more "My friend from the great state of Missouri'. Or at least, no admonishment for 'slipping up' and using names, as had been the tradition since the founding.

[genieroze]

35. Because they only count in other countries and against candidates

that Republicans don't like.

[candice]

36. Zogby has called for a "blue-ribbon panel" to investigate what went wrong

...he wasn't expecting entire states to flip at the last minute.

[RaulVB]

37. Because they gave the "victory" to Kerry...

I mean are you crazy?

"We are Americans..."

That shit only counts in Ukraine!:argh:

[catgirl]

46. They'll be seen in a court of law

I'm savoring the moment "guilty" is read to Blackwell.

[Liberty Belle]

57. Clinton talked about the Ukraine as a great example,

though she didn't specifically mention exit polls, as I recall.

[Laura PourMeADrink]

61. You can use the Nov 1 polls with the same effect as exits. Like

TIA brought up one time - the closeness of exits to the pre-election polls. Bush was down an average of 2.3 of the last polls I saw. He "won" (barf) by 4. That means he gained 6.3 points in ONE NIGHT

edit: FLORIDA

[Melissa G]

69. Here is a link to the Freeman op ed

commenting on exit polls

In three national elections over the past 13 months, the official count was sharply at odds with an independent national exit poll. As in the former Soviet republics of Georgia and Ukraine, U.S. exit polls projected a clear victory for the challenger. John Kerry was projected to win the national popular vote by a 2 percent to 3 percent margin and was ahead in nearly every closely contested state. Of course, the official counts, as in the other nations, showed an almost mirror image victory for the incumbent party candidate.

http://www.sfgate.com/cgi-bin/article.cgi?file=/chronicle/archive/2005/01/06/EDGOQAL6VA1.DTL
Kick the DU thread if you like it.

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x259888

[IndyPriest]

62. Got a lot done in one day. There will be more days ahead. n/t

[KaliTracy]

75. the issue at hand was disenfranchisement. The other things

will bolster this -- but one CLEAR message needed to be spoken today.

[Peace Patriot]

93. "The Election System is broken"--but we can't challenge the result.

See my post at

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x262606

on Dem leaders lack of power and "denial." Basically, they CAN'T change the outcome--they've given away all their (our) power--so they HAVE TO deny it.

Gave it away to Diebold and ES&S, and before that, gave lots of other things away (f.i., their Constitutional power to declare war).

They can't even ASSERT the INVALIDITY of the election due to blatant non-transparency and illegal vote suppression--things for which there IS "absolute proof"--and get a fair hearing. They can't ASSERT it. Because they have no power to do anything about it.

The combination of invalidity and all the inferential evidence of a Kerry win cries out for a fair hearing, a remedy.

The bind that WE are all in is that we can't produce "absolute proof" that Kerry won BECAUSE OF the non-transparency. Non-transparency that the BushCons SET UP.

But that isn't my point here. And frankly, I don't think we could have overturned this election even WITH "absolute proof." No power.

My point: What we saw yesterday was our naked powerlessness. There should have been a fair hearing, an investigation. We couldn't even get that.

Some heroes stood up and said what they could--in the face of insults and lies. That's all they COULD do.

Conclusion: It's up to us to recover our right to vote, locally--while we still can. It's our last and only chance to save our democracy.