Channeling TruthIsAll -- One of the Best Ever -- TruthIsAll

[autorank]

HOW TO MATCH AN EXIT POLL TO A  VOTE MISCOUNT:  USE BOGUS
WEIGHTS.

This is by far one of the best TruthIsAll posts ever.  He
makes two simple points that destroy any notion that 2004 was
a fairly counted election.  Either the uncontaminated National
Exit Poll on election day is correct and Kerry won OR the vote
count and the day-after “adjusted” National Exit Poll is
correct (1:25 p.m. 11/03/04).  We know that the voting process
was corrupted and that the count was as well, in all
probability (as evidenced by crazy results like an obscure
judge candidate for the Democrats out polling Kerry by large
margins in 28 Ohio counties).

We were never supposed to see the results from the Election
Day National Exit polls.  It was released by mistake and
captured by true heroes of democracy.  We were to see the
final NEP of 11/03/04.  Why?  Because it was ADJUSTED to
reflect the actual vote count.

That FINAL 11/03 adjusted NEP has two huge problems:  1) the
weight for Bush voters from 2000 was 108% of those who were
still alive and 2) the weight for Gore voters still living was
just 92% of those who voted for Gore in 2000.  They had to do
a lot of “adjusting” between the final election day NEP at
12:22 a.m. right after election day and the FINAL 11/03/04
official NEP released at 1:22 p.m. on the 3rd.  

Is there a reason that day-after NEP was adjusted to match the
“actual” vote count?  Who benefits?

They’re hoisted on the petard of their own frantic
inconsistencies.

THANK YOU TIA

==================================================

TruthIsAll:
=================================================+
 

HOW TO MATCH AN EXIT POLL TO A  VOTE MISCOUNT:  USE BOGUS
WEIGHTS.
http://www.geocities.com/electionmodel/BogusWeights.htm


Kerry won the 12:22am National Exit Poll (NEP) by 51-48%. In
the 2:05pm Final NEP, weights AND vote shares were adjusted to
MATCH the vote count. Bush won the Final, 51-48%. That the
Final NEP 43% Bush/37% Gore weights are MATHEMATICALLY
IMPOSSIBLE is beyond dispute.

Change to PLAUSIBLE weights in the Final and Kerry is the
51-48% winner - EVEN ASSUMING THE SAME VOTE SHARES. Therefore,
Kerry MUST HAVE DONE EVEN BETTER than 51-48%, since his
12:22AM vote shares were SUBSTANTIALLY REDUCED in the Final.

To believe the Final NEP, you must also believe:

1) That 52.57mm (43% of 122.3mm), or 108% of the 48.7mm Bush
2000 voters still living, turned out in 2004 - FOUR MILLION
more Bush 2000 voters than were ALIVE. FOUR MILLION PHANTON
BUSH 2000 VOTERS.

IF we assume a PLAUSIBLE 95% Bush 2000 voter turnout (46.25mm,
or 37.83% of 122.3mm), THEN the Final NEP OVERSTATED Bush
voter turnout by 6.3 MILLION.
 
2) On the other hand, the NEP apparently UNDERSTATED Gore 2000
voter turnout. The 37% weighting means that 45.24mm voted in
2004, or JUST 92% of the 49.2mm still living. So 8% stayed
home? Not likely.

IF we assume PLAUSIBLE 95% Gore 2000 voter turnout (46.75mm or
38.24% of 122.3mm), THEN the Final NEP UNDERSTATED Gore voter
turnout by 1.5 MILLION.

___________________________________________________________________


FINAL NATIONAL EXIT POLL 
2:05PM, 13660 RESPONDENTS

HOW VOTED IN 2000
(Impossible 43%/37% weights)

BUSH WINS: 62.5-59.3mm (51.1%-48.5%)

Voted..2004............Vote Share..........Votes (mm)
2000	Votes	Weight	Kerry	Bush	Other	Kerry	Bush	Other
No	20.79	17%	54%	45%	1%	11.22	9.35	0.21
Gore	45.24	37%	90%	10%	0%	40.72	4.52	0.00
Bush	52.57	43%	9%	91%	0%	4.73	47.84	0.00
Nader	3.67	3%	71%	21%	8%	2.60	0.77	0.29
								
Total122.27 100%      48.48% 51.11% 0.41% 59.28	62.49	0.50

_________________________________________________________________


FINAL NATIONAL EXIT POLL 
2:05PM, 13660 RESPONDENTS

HOW VOTED IN 2000
(Adjusted for plausible weights)

MAXIMUM WEIGHTS (100% turnout): 40.25% Gore/39.82% Bush
PLAUSIBLE WEIGHTS (95% turnout): 38.24% Gore/37.83% Bush
NO CHANGE in vote shares.

KERRY WINS: 62.6-59.2mm (51.2%-48.4%)

Voted..2004............Vote Share...........Votes (mm)
2000	Votes	Weight	Kerry	Bush	Other	Kerry	Bush	Other
No	26.22	21.44%	54%	45%	1%	14.16	11.80	0.26
Gore	46.75	38.24%	90%	10%	0%	42.08	4.68	0.00
Bush	46.25	37.83%	9%	91%	0%	4.16	42.09	0.00
Nader	3.04	2.49%	71%	21%	8%	2.16	0.64	0.24
								
Total 122.27 100%	     51.17% 48.42%0.41%	62.56	59.20	0.51

_________________________________________________________________
 
To analyze the NEP, start with the FACTS:
a) 2004 vote V2004 = 122.3mm 
b) 2000 vote V2000 = 104.8mm
c) U.S. annual death rate  = 0.87% (3.5% over four years)

Then we can proceed to:
1. Approximate the number of 2000 voters still alive in 2004.
   V2000A = V2000 * (1-.035)= 104.8*.965 = 101.1mm

2. Estimate the percentage of V2000A who voted in 2004.
   Assume PV2004 = 95%

3. Calculate the number of 2000 voters who turned out in 2004:

   V2K2004 = PV2004 * V2000A = 101.1 * .95 = 96.1mm

4. Calculate the number of 2004 voters who Did Not Vote in
2000. 
   DNV2000 = V2004 - V2K2004 = 122.3mm - 96.1mm = 26.2mm
   DNV2000 = NEW registered and OTHERS who did not vote in
2000.

5. Analyze the NEP time line:
   3:59pm: 8349 respondents 
   7:38pm: 11027 "
  12:22am: 13047 "

6. Calculate Kerry and Bush vote shares of DNV2000, Gore, Bush
and Nader. 

7. Calculate the national vote share: Kerry led 51-48% at each
time line.

8. The exit pollsters were left with a choice:
a) Tell the Truth: 
   Kerry won the National Exit Poll.
   The exit poll/vote discrepancy was BEYOND the 1% Margin of
Error.
   The numbers speak for themselves.

b) Change the weights and/or vote shares to MATCH THE VOTE
COUNT. 
    Even knowing that:
   -the weights were mathematically IMPOSSIBLE; 
   -the 3% vote discrepancy was TRIPLE the 1% MoE; 
   -the PROBABILITY of the discrepancy is 1 in 488 million.

9. The exit pollsters chose (b). 
They changed the weights AND the vote shares. 
The Final 2:05pm NEP of 13,660 respondents REVERSED the
12:22am NEP of 13,047 from Kerry 51-48% to Bush 51-48%.

10. The pollsters had an explanation for the discrepancy.
Well, sort of. It was due to Kerry voters responding at a
56/50 ratio to Bush voters. The Reluctant Bush Responder (rBr)

The weighted average exit poll response rate was 53%. Since
Bush won just 48% of exit poll responders, it implied that he
won more than 54% of the non-responding 47%. But there was a
problem. The rBr hypothesis was incompatible with the 43%
Bush/37% Gore weights. So the naysayers had to come up with
another explanation. 

11. The 43/37% weighting mix was due to "False
Recall".
FOUR MILLION Gore voters (8%) FORGOT THAT THEY VOTED FOR GORE.
THEY FORGOT JUST FOUR YEARS LATER.

So they told the exit pollsters they voted for Bush. 
And no, Bush 2000 voters DID NOT FORGET.
A ridiculous explanation, but it's a moot point. 

It doesn't MATTER who Gore or Bush voters SAID they voted for
or whether they forgot or lied.

Let's review the facts:
Fact 1: 43% of 122.3mm is 52.57mm. 
Fact 2: 48.7mm Bush 2000 voters were alive in 2004.
Fact 3: 48.7/122.3 = 39.8%. NOT 43%.
Fact 4: The Final Exit Poll inflated the Bush vote by 4
million. 

SO WHO THE EXIT POLL RESPONDENTS SAID THEY VOTED FOR IN 2000
IS IRRELEVANT. 
THE MAXIMUM NUMBER (MAX) OF 2000 VOTERS IN 2004 IS FIXED BY
THE FORMULA: 
MAX =  TOTAL VOTED IN 2000 - TOTAL DIED. 
NOTHING CAN CHANGE THAT FACT. 

THE 2000 VOTE COUNT IS INDEPENDENT OF 2004 EXIT POLL SAMPLING.

IT'S THE BASIS FOR DETERMINATION OF PLAUSIBLE EXIT POLL
"WEIGHTS".

PRELIMINARY "GROSS" WEIGHTS ARE BASED ON THE 2000
VOTE COUNT. THESE ARE "MAXIMUM" WEIGHTS WHICH CANNOT
BE EXCEEDED - BY DEFINITION.

TURNOUT PERCENTAGES OF GORE, BUSH AND NADER 2000 VOTERS MUST
BE ESTIMATED TO DETERMINE "NET" WEIGHTS FOR
CALCULATING NATIONAL VOTE SHARES. 



quere mas?

[BeFree]

1. Thanks TIA

Gawd knows how badly the Final Exit-poll was cooked to come up with numbers allowing * to remain squatting the white house. TIA proves that.

Thank you too, Autorank, for keeping this in front of the un-suspecting population. We must never give up. Our heritage, indeed, the soul of the country depends on citizens like yall to keep reminding us of where we've been so that we may oneday cure the cancer that will surely kill our American Democracy.

Kerry Won. The people lost.

[foo_bar]

2. doesn't this belong in the 9/11 forum?

Do not quote or link to "conspiracy theory" websites, except in our September 11 forum, which is the only forum on Democratic Underground where we permit members to debate highly speculative conspiracy theories. A reasonable person should be able to identify a conspiracy theory website without much difficulty.

http://www.democraticunderground.com/forums/rules_detailed.html

The website in question is operated by a tombstoned DUer:
http://www.geocities.com/electionmodel/TIAExitPollAnalysisPIThreads.htm

Re:

One can apply the Cum. Normal Distribution or Poission function (or both to confirm) to determine the probability of rare events occuring by chance only.. There are other models which may be applicable. I found these two perfect and easy to use. All you need is Excel; the functions are built-in.

I have also used this method to calculate the probability that
1- at least 15 JFK witnesses would meet unnatural deaths in the year following the assasination.
2- at least 16 world-class microbiologists would meet unnatural deaths in a 4 month period following 9/11.
3- The probability that at least a certain number of people would suffer from mad cows disease in a specific geographic area in a given year.

http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=104&topic_id=1777401#1777492

[file83]

7. That "link" wasn't linked - I think this OP found a loophole:

bury your source unlinked in the body of the post. Moderators probably won't see it. Hence....

[loudsue]

Go peddle your wares somewhere else.

Patriots in this country KNOW THE VOTE WAS STOLEN.

It's NOT a "conspiracy theory". To believe otherwise is a "coincidence theory".

Kerry won. Get over it.

:kick:

[autorank]

65. For you;)

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=105&topic_id=4835155&mesg_id=4835155

[loudsue]

76. OOO! Thanks, Autorank!

:loveya:

:kick:

[foo_bar]

74. "Patriotism is the last refuge..." -Samuel Johnson

[Peace Patriot]

25. It's not a "conspiracy theory" with numbers like this--and so many

corroborating numbers--for instance: that the Dems blew the Repubs away in new voter registration, nearly 60/40, in 2004; issue polls over the last two years showing a great majority of Americans disapprove of every major Bush policy, foreign and domestic, way up in the 60% to 70% range, including near 60% disapproval of the Iraq war way back before the invasion (Feb. '03), and 63% disapproval of torture "under any circumstances" (May '04); and Bush/Cheney's dismal approval ratings now, and, with a few short exceptions, throughout the entire junta.

Plus: Two of the biggest crooks in Washington DC, Tom Delay and Bob Ney, are responsible for the $4 billion "Help America Vote Act" (HAVA) electronic voting boondoggle, which permitted two rightwing Bushite corporations, Diebold and ES&S, to gain control of 80% of the nation's vote in 2004, using "TRADE SECRET," PROPRIETARY programming code with virtually no audit/recount controls, no paper trail requirement, no controls on private industry 'testing' of the machines, no controls on partisan vendors, no controls on lavish lobbying and "revolving door" employment, and underfunded regulation. They thus infused our election system with corruption and insecurity, and created a NON-TRANSPARENT election in 2004--the only kind of election that Bush could have won.

Non-transparent elections are not elections. They are tyranny. That is not a "conspiracy theory." That is a no-brainer.

Further, until recently, the CEO of Diebold was a Bush/Cheney campaign chair, major donor and fundraiser (a Bush "Pioneer"), who promised in writing to "deliver" Ohio's electoral votes to Bush/Cheney in 2004. ES&S, a spinoff of Diebold (similar computer architecture), was initially funded by rightwing billionaire Howard Ahmanson, who also gave one million dollars to the extremist 'christian' Chalcedon Foundation (who, among other things, tout the death penalty for homosexuals). Diebold and ES&S have an incestuous relationship--they are run by two brothers, Todd and Bob Urosevich.

These are the people who counted most of our votes behind a curtain of secrecy.

That is not a "conspiracy theory." That is a recipe for the overthrow of our democracy--and the imposition of a warmongering, criminal junta tied in with the establishment of a state religion, in defiance of the Constitution.

Open your eyes, Foo-Bar. The question is NOT whether the 2004 election was stolen. No proof of that is necessary. The conditions for an honest vote count were NOT PRESENT, while conditions for a dishonest, Bush-tilted count were abundant, at every point in the election process. To name a few...

--Diebold/ES&S touchscreens reported to be changing Kerry votes to Bush votes (and almost never the other way around)

--the ease with which any of these machines--voting machines and central tabulators--can be penetrated and altered (one hacker, a couple of minutes, leaving no trace...)

--the unreliability and hackability of these new machines, established time and again in independent tests

--the "trade secret" vote tabulation programming--code so secret that not even our secretaries of state are permitted to review it

--one third of the country voting with no paper trail of any kind whatsoever (no audit or recount even possible)

--the inadequacy of audit/recount controls overall (election laws not keeping up with the newly installed capacity for massive, speed-of-light fraud)

--evidence of corruption among election officials including lavish lobbying junkets and "revolving door" jobs (for instance, the former Repub Sec of State in Calif now working for Sequoia--the third big "trade secret" vote tabulation player)

--OTHER evidence of intent to commit fraud; massive violations of the Voting Rights Act, against Democratic voters, in Ohio and elsewhere

--the war profiteering corporate news monopolies' DOCTORING of their exit polls to FIT the results of Diebold's and ES&S's secret formulae.

No transparency. No checks and balances. Corruption, secrecy, lavish lobbying, and private, partisan corporate control all along the way. And, in ADDITION to all of this--to the conditions ripe for fraud, favoring a regime that is notorious in its contempt for the law--we have TIA's analysis, and the analysis of others (Dr. Steven Freeman, for instance, and the UScountvotes.org statisticians), that the numbers (exit polls/official count) don't add up.

With secret voting counting, you must PRESUME fraud--not the other way around. You are a fool if you DON'T presume fraud. And that is not a "conspiracy theory." It is common sense.

And if we don't face up this and change it, while we still have a window of opportunity at the state/local level, our democracy is over.

[OnTheOtherHand]

72. if you wonder why you have not won the argument

it has little to do with assessments of motives or opportunity. It has much to do with failure to cite corroborating numbers that are actually pertinent, and to defend them when challenged.

It doesn't matter how the war polled in early 2003, or how Bush polls now. The polls that arguably matter are the ones immediately before and after the election, and these make it eminently plausible that Bush won. He led in most of the horse-race polls, and his approval rating was good enough to win, as well.

If you are suspicious of touch screens in particular, then it would make sense to demonstrate that the touch screen results are especially suspicious.

Your shtuff about "DOCTORING of their exit polls" indicates that you either haven't read, or have chosen not to believe, explanations of how the exit polls actually work. (I still look forward to an explanation of how "Diebold's and ES&S's secret formulae" were implemented in New York.)

If you really, really want to discuss Freeman and the UScountVotes literature, let's go. Otherwise, it's an empty appeal to authority. (However, everyone agrees that the exit poll and official count don't agree, so we don't have to debate that.)

If you'd rather spend your time working for verifiable elections, I respect that. What I think you are missing about foo is that he has spent a lot of time actually investigating the 2004 election, by analyzing reports in the EIRS database. I betcha he knows things about the machines in 2004 that you don't (and I suppose probably vice versa). The fact that he disagrees with TIA has no bearing on all that.

[foo_bar]

73. strawmen aside,

I agree with most of your bullet points (although the reported DRE vote-switching was ES&S/Sequoia iirc), but a list of things-that-aren't-conspiracy-theories doesn't explain how TIA can prove JFK was assassinated by a cabal using only Excel. Similarly, Kerry's "99.8% probability of winning" and subsequent revisionism belongs with the "Flight 93 was a missile!" crowd (IMO), who also shift the burden of proof when pressed for a logical explanation (usually in the form of the strawman, "So you believe the official version?").

So no, I don't believe the official version of the election either, but that doesn't prove Bigfoot stole Florida, or that it was accomplished with mind control lasers or what have you. In fact, it can't hurt to take your advice about "presuming fraud" even where unsourced, non-peer reviewed urban legends are concerned.

(edit: bad link and missing prepositional phrase)

[Febble]

75. I'm delighted to find

that I completely agree with this:

The question is NOT whether the 2004 election was stolen. No proof of that is necessary. The conditions for an honest vote count were NOT PRESENT, while conditions for a dishonest, Bush-tilted count were abundant, at every point in the election process.

I would go so far as to say that it matters as much that Bush voters (though in a minority) were denied the right to vote by machine rationing in predominantly (but not solely) Kerry precincts in Franklin county.

But I have to take issue with:

--the war profiteering corporate news monopolies' DOCTORING of their exit polls to FIT the results of Diebold's and ES&S's secret formulae

After close of polls, projections for each state's results were made, increasingly, from the incoming vote-returns as well as from the responses from the exit polls. This is done every election, and was not done secretly. Details are available from the E-M website, and were available before the election, which was when I read them.

http://www.exit-poll.net/faq.html#a10

But you must know this?

[Usrename]

77. Yes, we know what you think about the exit polls.

You truly believe that the exit polls were contracted for and performed for all sorts of various reasons, every reason under the sun, except for the obvious reason, which is to find out how the people who went to the polls actually voted.

I know that it is a good argument that you make, that people can do exit polls for other reasons than to find out how people who went to the polls actually voted, but I'm just not buying it.

The reason that they do exit polling has always been to find out how the people who went to the polls actually voted.

That is until this election. This poll was either done in order to deceive, or else it is just like every other exit poll, and it was done to determine how the people who actually went to the polls voted.

Why is this such a difficult concept?

If you believe that the exit polls were done to deceive, then the OP makes perfect sense. If you believe that the exit polls were done to determine how the people who went to the polls actually voted, then the OP still makes perfect sense.

The only argument left is some tin-foil nut-case theory that E/M and the media conglomerates that contracted the polls had no interest in how anyone actually voted. It was some crazy make-work exercise in which all the numbers are actually supposed to be meaningless. Completely meaningless is the argument.

And completely meaningless by intent and design, because all of the reasons for them being completely meaningless are reasons that were well known before the money was spent. The mathematicians all knew it would be meaningless ahead of time, but they spent the money and took the polls anyhow. It is bizarre is all that I can say. But yes, you might be right. There may be idiots like that out there with the power to do such a thing. But why would they?

[Febble]

79. Well, the reason it is

a "difficult concept" is this:

If the pollsters state, in advance, that they are going to project the counted vote, and that the method they will use will involve incorporating vote-returns into the projections, and then that is exactly what they do...

why would anyone think that what they actually intended to do was to project the result from the polls alone, but that for some reason, instead, they did what they stated they were going to do, and and use the vote-returns in addition?

But I do appreciate the tone of your post.

I know that it is hard to believe that the exit polls were not designed as a check on the election process, and I rather agree that, American being something of a banana republic, perhaps international observers from the UN should do some independent monitoring, possibly using exit polls...but it is hard to see why anyone would think that the reason the 2004 exit polls were conducted was not for the reasons stated before they were conducted.

Except that I admit that probably not that many people read the E-M FAQ before the election. But I did.

It didn't, nonetheless, stop me considering, very seriously, that the reason the around-close-of-poll exit results were discrepant from the final results might have been because the count was wrong. Or from putting a vast amount of effort into trying to figure out whether it might be possible to disambiguate the two possibilities (error in the poll; error in the count); nor does it stop me remaining outraged by the broken state of your democracy. But it did, I admit, stop me considering that the projections were intended to be independent of the count, because it was stated in black-and-white that they wouldn't be, ergo, that their designed purpose was to predict the counted result, not act as a check on it.

But it doesn't mean that the pollsters were not interested in how anyone voted. They were (or their clients were) very interested. And while crosstabs performed on the National Exit Poll give you slightly different answers depending on whether you use weighted or unweighted data, the percentage point differences don't radically affect the information.

What the weights radically affect is whether what was a narrow margin between the two candidates put Bush in the White House or Kerry. Not why, to a reasonable approximation, particular groups voted for either.

Because even if you are right, and the poll was nearer to the true vote of the American public than the count, the fact is, tens of millions of people voted for Bush. The polls remain informative about who those people might have been, whether or not you are convinced that the total numbers are correct.

And I would add that I still believe that various forms of voter and vote suppression probably deprived Kerry of many thousands of votes, including votes in Ohio, which would have won him the election. But mostly those wouldn't show up in the exit polls.

In other words, while I don't think that the exit poll evidence provides good evidence that the election was stolen, I see plenty of evidence, as Peace Patriot outlines upthread, that the election was unjust, and that the injustice was disproportionately visited on Kerry's voters.

[Usrename]

88. I'll explain why I think this particular argument is wrong.

Let's take a closer look at the your alleged motive for performing an exit-poll in the first place. An argument is now put forward that the exit-poll information was not gathered in an attempt to project the actual vote. Instead a slightly different purpose is being claimed. The argument is being made that the exit-polling was merely done as a tool to allow the pollsters to make early calls on election night for the media outlets.

That viewpoint is not very sound, IMHO. The pollsters could just as easily have used an entirely different data source for that purpose. For example, they could have used the actual vote returns (by precinct even, publicly available at no cost, for free, in other words) from the previous election to accomplish that particular exercise. It would have yielded the same results as gathering exit-poll data and then assigning impossible weights to it.

Here is the simple example:

Team 1 takes actual vote returns from the 2000 election.

Team 2 uses M/E exit-poll data.

Both teams get the 2004 actual returns (the counts) at the same time and weight their data to forecast a winner.

Who calls the election first, assuming they both use the same threshold of probability to make the call?

I'm saying Team 2 has no advantage over Team 1. I think my point of view is proved out by the fact that Team 2 was required to wait until there were enough actual counts to allow them to assign weighting that was impossible. That would not be required by Team 1.

Team 1's weightings would have been realistic. They could have made the calls much earlier.

See what I mean? It's just math. Why not simplify the equation? If that was the question that they were trying to answer, why not do it the easy way?

The exact same process can be used to make up cross tabs based on the demographic data from some other readily available (free at no charge) source. Like the census data. I know you like to argue that they did a meaningless exit poll for some obscure reason, and that they knew ahead of time that the data would be meaningless, I get your argument loud and clear. I just think it's sort of weak, is all. Not convincing and it has nothing at all to do with the OP, either.

I know that you like to insist that the exit polls were meaningless, that this was intentional, and that they announced ahead of time that they would be meaningless. That whole purpose of performing them was so that they could be rendered meaningless. But if you are correct, if that was their true purpose, then doesn't that make it even a little likely that their purpose was to deceive the public. What other reason is there for performing a meaningless exit poll?

[Febble]

96. OK

Let's take a closer look at the your alleged motive for performing an exit-poll in the first place. An argument is now put forward that the exit-poll information was not gathered in an attempt to project the actual vote. Instead a slightly different purpose is being claimed. The argument is being made that the exit-polling was merely done as a tool to allow the pollsters to make early calls on election night for the media outlets.

Correction: the exit-poll information WAS gathered in an attempt to project the actual vote, precisely SO THAT early calls could be made. However, the other purpose, arguably more important, was to find out the characteristics of who voted for whom. The crosstabs.

But regarding the first, very conservative probability thresholds (t values) are set, and have to be exceeded before it is recommended that a state is called - the probability that a state has been won by either candidate is computed both from exit-polls and from vote returns. For a few states, this threshold may be reached on the basis of exit polls alone (Washington DC springs to mind, although it is not a state). For most, corroborative data is required from the vote-returns. The "safer" the state, the fewer vote-returns would be required for the t value to reach threshold.

That viewpoint is not very sound, IMHO. The pollsters could just as easily have used an entirely different data source for that purpose. For example, they could have used the actual vote returns (by precinct even, publicly available at no cost, for free, in other words) from the previous election to accomplish that particular exercise. It would have yielded the same results as gathering exit-poll data and then assigning impossible weights to it.

Well, if they'd done that:

1. They wouldn't have all the other data


2. They wouldn't be able to make anykind of estimate until the vote count was in. Read my stuff about t values above.

Recall: many elections aren't as close as 2000 and 2004, although many elections result in a change of government. Take Clinton:Bush in 1992. Bush was the incumbent; projections made from 1988 would have had Bush ahead, whereas Clinton won. The polls, correctly, had Clinton winning, even without reweighting, although the reweighting reduced his winning margin.

So you use the best data you can. And you certainly hope that your raw exit survey data is not biased. You try to ensure that it isn't. But, as I've said before, it isn't something you can assume - polls are prone to bias, especially when response rates are low. This is why, for most states, the t value does not reach threshold until a substantial number of vote-returns are in.

...

See what I mean? It's just math. Why not simplify the equation? If that was the question that they were trying to answer, why not do it the easy way?

Well, I think I addressed this very point in a previous thread, and I've also had a go in this post. Sure, if you think you have a biased poll, and the result isn't that different from last time, you'd be quicker using the previous result. But you don't know any of that until the thing is over (or at least half over).

The exact same process can be used to make up cross tabs based on the demographic data from some other readily available (free at no charge) source.

The cross-tabs aren't "made up". Take a look at them - I gave you the link to the data. Run some cross tabs yourself. They are interesting for at least two reasons (whether or not you use the weights given):

1. They are specific to the date of the election


2. They ask very specific and detailed questions, including who the respondent voted for, which enables detailed analysis to be made of the issues that may have influenced voter choice. Of course, if the poll is biased, the data will be less good (although, as I've said, this is the rationale behind reweighting - to try to restore the data, post hoc). But in fact, the data is not hugely different, for broad analytical purposes, whether you use it weighted or unweighted. The weights mostly aren't that great. Actually, the weights are worth looking at in themselves.

I know you like to argue that they did a meaningless exit poll for some obscure reason, and that they knew ahead of time that the data would be meaningless, I get your argument loud and clear.

No you don't because you seem to have it the wrong way round. I don't think the reason is obscure, I don't think the data is meaningless (weighted or unweighted) and I don't think they "knew ahead of time that the data would be meaningless" - but I also don't think they "knew ahead of time" that they would have such a large discrepancy (although they are certainly used to having a discrepancy, which is why they use the methodology they do).

I just think it's sort of weak, is all. Not convincing and it has nothing at all to do with the OP, either.

Well, fair enough. Although the OP is about the crosstabs, and my post was about the crosstabs.

I know that you like to insist that the exit polls were meaningless, that this was intentional, and that they announced ahead of time that they would be meaningless. That whole purpose of performing them was so that they could be rendered meaningless. But if you are correct, if that was their true purpose, then doesn't that make it even a little likely that their purpose was to deceive the public. What other reason is there for performing a meaningless exit poll?

Well, to re-iterate, this is not what I believe or insist. All I insist is that they were not designed to be a check on the count. They were designed to predict it (and, I should add, to provide data on the characteristics of who voted for whom).

What I will say, however, is that the attempt by the networks to inform the public as to what the numbers meant, or how they were derived, was wholly inadequate. I happened to read the FAQ before the election (actually, before the results were in - the reason I checked the FAQ on election night was because I didn't know how your exit polls projections, given your range of time zones, were made), so I knew. But CNN certainly didn't make it clear that the reason the projections had changed direction was not because a few more voters had been added to the dataset (mathematically impossible) but because they had started to incorporate vote-return data. But it doesn't seem to be the case that E-M was attempting to deceive anyone (because the info was on their FAQ) and I doubt CNN were deliberately trying to deceive anyone, I think it probably didn't occur to them that it needed explaining. After all, it's what happens every election.

I have to say, our election reporting is way better. Peter Snow explains everything as he goes, and his catchphrase, about the projections, is "it's all a bit of fun...."

[Usrename]

123. Do you want to rethink anything?

"the exit-poll information WAS gathered in an attempt to project the actual vote"

"All I insist is that they were not designed to be a check on the count. They were designed to predict it (and, I should add, to provide data on the characteristics of who voted for whom)."

I think we have made some progress. This might be a breakthough for you.

Now, if the prediction doesn't match the vote, what then? You are arguing, correctly, that E/M just changed the prediction to match the vote. And, since E/M announced ahead of time that they would do just that, then somehow this changes things.

Well, perhaps it does change things. Maybe they should call their exercise a deception instead of an exit poll. Yeah, that's the ticket! The E/M deception poll, performed to sway public opinion during stolen elections, not to be confused will the well respected and time honored practice of exit polls. Yes. Deception poll. It has the purpose spelled out right in the name.

Prove me wrong.

[Febble]

125. well, yes, I do rather think

that if you tell someone that what follows is a deception then it rather ceases to be a deception.

But what you seem to be forgetting, is that to any pollster (any social scientist) participation bias is expected. The reason they give you the recipe in advance, is that they are, in effect saying: "we'll ballpark this thing from the exit poll responses, and then fine tune it once we know how big our non-response problem has been".

Because they expect non-response error. That's why they ask their interviewers to record age, race and sex of non-respondents, so that at least they can do a crude correction for non-response bias by visible characteristics. But granted that we know - and they know, specifically - that non-response bias occurs, there is no earthly reason to suppose it wouldn't apply to non-visible characteristics as well. If men are more likely to refuse than women; if middle-aged people are more likely to refuse than young people (which can be verified and compensated for), why, intrinsically is it implausible that Bush voters would be more likely to refuse than Kerry voters? or, perhaps, specifically, rare voters who only vote when they are frightened (just a hypothesis)?

Well, it isn't implausible. It may not be true, but it's perfectly plausible, which is why it makes sense to weight to the count. The FLAW in the procedure is that you may not trust the count. But they did. Maybe they should have, but they did.

So not only did they not "deceive" because you can't really call it a deception when you do what you do every year, and you tell people in advance that it is what you are doing, and you subsequently release all the data so that people can see exactly how much of a reweighting for non-response bias you actually applied...

They didn't "deceive" because their purpose was not to disguise fraud but to compensate for non-response bias, in exactly the same way as they use their non-respondent data.

NOW we can start talking about whether they should have been more skeptical of the veracity of the count; or whether on this occasion, or even on previous occasions, the redshift (which there has been in every election since at least 1988 according to their own report) was not due to non-response bias but to fraud.

You can call them what you like, but if you want polls to monitor the accuracy of your election, you need a differently designed poll. And I can suggest many better ways of monitoring your elections.

[Peace Patriot]

127. You Edison-Mitofsky apologists keep forgetting that the 2004 election

was NON-TRANSPARENT.

Non-transparent elections are not elections. They are tyranny.

We are seeing that tyranny in every action of the Bush junta. And, as this house of democracy crumbles into dust, you are arguing what Edison-Mitofsky's intent was, in assigning a 108% turnout of Bush 2000 voters in 2004?

You also write as if Edison-Mitofsky was operating within some pristine scientific vacuum--rather than in a hotbed of unparalleled journalistic corruption and crime.

You are forgetting who employed Edison-Mitofsky--the same war profiteering corporate news monopolies who helped this junta manufacture a war in which tens of thousands of innocent Iraqis were slaughtered, and many more tortured, and that has resulted in chaos in their country, and in which thousands of U.S. soldiers have been killed or maimed for life; the same war profiteering corporate news monopolies who are helping this junta loot our country blind and destroy our Constitution.

A manufactured war, a manufactured election to endorse that war, and falsified exit poll numbers to confirm that result. That is the context that you keep leaving out of this discussion--the manufactured war and the manufactured, non-transparent, Bushite-controlled election, both brought to us by the SAME PEOPLE who brought us Edison-Mitofsky. It's like someone who would say, prove to me that Al Capone committed a single murder. Blood all over the pavement, Mr. Capone nowhere in evidence.

And by falsifying the exit polls, they denied the American people major evidence of election fraud, and helped to squelch protests and calls for investigation--unlike, say, in the Ukraine, where exit poll numbers and official count numbers were kept separate, and everyone could see that something was very wrong.

A few weeks ago, George Bush admitted that, yes, he had committed numerous felonies with warrantless domestic spying, and would continue to do so, because HE doesn't think it's illegal.

His SAYING this does not make what he is doing legal. His SAYING this makes his illegal actions an artifact of the "will to power." He now possesses a degree of tyranny whereby he can ANNOUNCE his intention to continue committing felonies with impunity. It is his will.

Similarly, with Edison-Mitofsky, their announcement that they were going to mix exit poll results and official results does NOT make it right or proper or transparent. They can do whatever the hell they please, because they are in the employ of the war profiteering corporate news monopolies, who are paying them big money and will cover for them, whatever they do, provided that it is in the interest of WAR PROFITEERS and the rich robbing the poor.

That argument--that they SAID they were going to mix the results--is like George Bush's statement that he intends to continue breaking the law. It is a position of supreme arrogance. They knew damn well WHO would be 'counting' our votes--two rightwing Bushite corporations--and HOW they would be 'counting' them--behind a veil of secrecy. They KNEW they would be the only check and balance against those numbers. And you cannot tell me that they did not shut down their reporting system, with a convenient "breakdown," when it was clear that Kerry won their exit polls, and deliberately doctored those numbers to please the thieves and crooks and liars who employ them, and the murderers and torturers and liars whom they wanted to keep in power.

That was their FAIL-SAFE. That obscure little "explanation" of their mixing the results was their fallback in case of a Kerry win (a candidate who was not especially anti-Iraq war, but who would have been beholden to a huge antiwar, grass roots majority; a candidate who would not likely have started that war in the first place, and who was known for his big-time criminal investigations--in short, a man whom the Bush junta could not afford to have in the White House).

E/M had to use that fallback. Their post-election secrecy points to it. The numbers point to it. And the Bushites' extraordinary, blatant vote suppression and other illegalities in Ohio and elsewhere--most likely needed because the electronic advantage to Bush had to be pre-programmed, and was not so easy to change on election day itself--also points to it--that Kerry won, and that any evidence of that win had to hidden from the American people.

E/M was right in the middle of this disgusting, corrupt, lying, war profiteering, traitorous, filthy murdering, torturing, fascist junta that is destroying our country and threatening to destroy the planet. They are being enriched by the bad guys--by the insane super-rich who care for nothing except for their own well-guarded enclaves of "safety" against us poor slaves and cannon fodder who are forced to suffer their wars and other disasters. I will give them no quarter. They get no benefit of the doubt. You don't have to look much farther than who signed their paycheck to know what they did with their 2004 exit poll monopoly.

[BeFree]

128. Bravo! A needed inspiration!

If you don't make a thread of this on your own, I will. I will steal it. It needs to be read far and wide. I have miseed your missives, Peace Patriot!

[Febble]

132. No

You Edison-Mitofsky apologists keep forgetting that the 2004 electionas NON-TRANSPARENT.

No I don't. It is right at the forefront of my concerns, together with the disenfranchisement of those who most needed President Kerry.

Non-transparent elections are not elections. They are tyranny.

We are seeing that tyranny in every action of the Bush junta.

Yes.

And, as this house of democracy crumbles into dust, you are arguing what Edison-Mitofsky's intent was, in assigning a 108% turnout of Bush 2000 voters in 2004?

You also write as if Edison-Mitofsky was operating within some pristine scientific vacuum--rather than in a hotbed of unparalleled journalistic corruption and crime.

No.

You are forgetting who employed Edison-Mitofsky--the same war profiteering corporate news monopolies who helped this junta manufacture a war in which tens of thousands of innocent Iraqis were slaughtered, and many more tortured, and that has resulted in chaos in their country, and in which thousands of U.S. soldiers have been killed or maimed for life; the same war profiteering corporate news monopolies who are helping this junta loot our country blind and destroy our Constitution.

A manufactured war, a manufactured election to endorse that war,

Maybe. I don't know if I hold CNN culpable for the war. Fox, maybe.

and falsified exit poll numbers to confirm that result.

No. You write great polemic, PP, but this doesn't follow. The fact that I detest the Bush regime as much as you do doesn't stop me seeing that the pollsters did what they said they would do in advance, what they do every year, which is to project the official result from a combination of exit poll responses and vote-returns. This is not "falsifying" in the way you appear to infer, although if the count is corrupt, the polls will indeed project a corrupt result. It would be "falsifying" if the exit polls were claimed to have any kind of monitoring function. No such function was ever claimed, and they were not designed to do that job. It is a myth that American exit polls are designed to monitor elections. Yes, exit polls are used, sometimes, to monitor foreign elections. Yes, I agree, American elections should be independently monitored. But you are accusing the pollsters of not doing a job they were not asked, or expected to do. FWIW, UK exit polls are not designed to monitor our elections either.

Your implication that there was a deliberate coverup of corrupt result is therefore false. What there was, instead, was a methodogical assumption that any discrepancy between poll and count would be due to bias in the poll. The assumption may have been unjustified in this instance; but there is no evidence of intent to falsify, and strong evidence that there wasn't. The methodology was clearly described in advance, and was not different in 2004 than in previous years.

That is the context that you keep leaving out of this discussion--the manufactured war and the manufactured, non-transparent, Bushite-controlled election, both brought to us by the SAME PEOPLE who brought us Edison-Mitofsky. It's like someone who would say, prove to me that Al Capone committed a single murder. Blood all over the pavement, Mr. Capone nowhere in evidence.

No. You are like someone saying that Al Capone committed a massacre. Everyone that Al Capone employed must have been complicit.

And by falsifying the exit polls, they denied the American people major evidence of election fraud, and helped to squelch protests and calls for investigation--unlike, say, in the Ukraine, where exit poll numbers and official count numbers were kept separate, and everyone could see that something was very wrong.

No. They published both the data, with weights, so you can analyse them, and they did an thorough investigation into their own methodology - they published a report that was short on analytical detail, but contained copious data as to exactly where the discrepancies had occurred and when.

A few weeks ago, George Bush admitted that, yes, he had committed numerous felonies with warrantless domestic spying, and would continue to do so, because HE doesn't think it's illegal.

His SAYING this does not make what he is doing legal. His SAYING this makes his illegal actions an artifact of the "will to power." He now possesses a degree of tyranny whereby he can ANNOUNCE his intention to continue committing felonies with impunity. It is his will.

Yes.

Similarly, with Edison-Mitofsky, their announcement that they were going to mix exit poll results and official results does NOT make it right or proper or transparent.

No, it is not similar. Announcing their methodology in advance DOES make it transparent. You may disagree that it was right or proper, but it was certainly usual and expected. The exit polls were not designed to monitor the election. Perhaps they should have been, but they weren't. They were designed to allow the networks to call the states, and to provide information as to who had voted fro whom.

They can do whatever the hell they please, because they are in the employ of the war profiteering corporate news monopolies, who are paying them big money and will cover for them, whatever they do, provided that it is in the interest of WAR PROFITEERS and the rich robbing the poor.

That argument--that they SAID they were going to mix the results--is like George Bush's statement that he intends to continue breaking the law. It is a position of supreme arrogance. They knew damn well WHO would be 'counting' our votes--two rightwing Bushite corporations--and HOW they would be 'counting' them--behind a veil of secrecy. They KNEW they would be the only check and balance against those numbers. And you cannot tell me that they did not shut down their reporting system, with a convenient "breakdown," when it was clear that Kerry won their exit polls, and deliberately doctored those numbers to please the thieves and crooks and liars who employ them, and the murderers and torturers and liars whom they wanted to keep in power ...etc..

I find this a complete misrepresentation. These are not arguments, they are assertions. As an argument, you seem to be saying: Bush is corrupt and evil and elections are non-transparent; therefore the count must have been false, therefore discrepancy between poll and count must have been due to fraud, therefore the pollsters must have been complicit in covering up the fraud.

I don't agree that these things follow from one another, and I don't agree that expressing them as passionately and as colourfully as you do makes them do so.

I am a scientist, not a poet. I will leave the poetry to you.

[FULL_METAL_HAT]

3. It will be an unbelievable 2 years come November ...the diligence of TIA's

It will be an unbelievable 2 years come November ...the diligence of TIA's

leadership could finally bear the kind of harvest we know is our due.

One word has become more and more prevalent in the examination of the not just the 2004 (and 2002 and 2000) election results, but in so many facets of the way the present administration has acted.

That word is certainly writ large in each and every of TIA's posts, as well as in everyone seeking to examine the events that govern our nation and govern our passion.

Without realizing it, the nature of the offense in this word has slipped out from our moral outrage and in a way lay dormant, under the cover of so many other thoughts, hopes, and rationales.

But it still lies there, an elephant grown so huge in the room that people can no longer do anything but consider "part of the furniture" as we try to do our things and live our lives AROUND its constancy.

D E C E P T I O N

Take a step back and feel your gut when you think that EVERYTHING we have experienced from our Government -- EVERYTHING -- has used DECEPTION ...

Now, for those who want a peek into WHY, I can tell you that, unequivacably, the following statement is 100% true:

"All warfare is based on deception"

There has been a war going on for nearly 6 years, and it turns out, I hate to say, we, the citizens, are the enemy now.

[Nederland]

4. FYI

Large bold fonts do not make up for poor arguments.

[Cessna Invesco Palin]

8. Damn straight.

Everybody knows EXCESSIVE CAPITALIZATION MAKES UP FOR POOR ARGUMENTS!!!!

Also, overuse of !s.

[loudsue]

22. And tiny fonts don't make up for keeping your head buried in

the sand.

[autorank]

5. The Deception regarding the last two elections is more than coincidence.

The invasion and occupation of Iraq was based entirely ON DECEPTION.

The mishandling of the Katrina disaster and aftermath was based entirely ON DECEPTION.

The illegal wiretapping by NSA was and is based entirely ON DECEPTION.

The * crew breaks every rule to realize their objective of unfettered and perpetual power.

Election fraud in 2000 and 2004 were and are the harbingers of each new phase of DECEPTION

IMPEACH NOW

[bleever]

9. Is anybody actually denying that deception is the hallmark of this

administration?

I think your point is pretty straightforward and relevant. If anyone has any real substantive argument against your point, I haven't seen it here.

:thumbsup:

[bleever]

10. Still waiting for someone to prove * won, in light of the mathematical

impossibility.

I guess we're supposed to be bullied into agreeing that 2+2=5.

Not gonna do it.

[Nederland]

11. You've got that backwards

The burden of proof is on you.

TIA never successfully proved anything because TIA always failed to take into account the clustering effect. Ignoring this basic statistical principle old speaks volumes about TIA's competence.

[EFerrari]

15. And yet, his basic assumptions about the Cabal

have been validated by every action they have taken, up to and including domestic spying.

Please. There are freeways just crying to be played on. :)

[Nederland]

20. Sad but true

I guess my point was that although his basic assumptions about the Cabal are true, his assumptions about the exit poll samples are hopelessly unproven.

[Name removed]

23. Deleted message

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

[Nederland]

30. As usual

A personal attack with no facts related to the discussion.

[Yupster]

143. And the exit polls proved

that Dukakis and Bush were neck and neck even though Bush won comfortably in 88.

And the exit polls proved that Gore barely won Alabama in 2000 even though he lost it by 15 %.

And the exit polls proved that Dukakis won Pennsylvania in 88 and that's why the networks called it for Dukakis and had to say oops the next morning when Bush actually won the Keystone State.

But the exit polls are never wrong. Except almost always.

[OnTheOtherHand]

145. I thought it was Illinois (in '88)

Gotta look that up. Darn you, Yupster! ;) Might've been different on different networks -- '88 was the Golden Era when networks still had separate exit polls.

Hey, look, I like exit polls. You don't have to give them that bad a rap. But they are not infallible truth machines, no indeed.

[Peace Patriot]

27. Nope, that is simply not true regarding elections: The burden of proof

is NOT on us, with Tom Delay and Bob Ney arranging for "trade secret," proprietary vote tabulation by rightwing Bushite election profiteers. Sorry. The burden is on THEM, and they cannot prove that Bush won, and all evidence, aside from their secret vote count, points the other way.

Non-transparent elections are not elections. They are tyranny--as Bush & Co. have underscored time and again since their fraudulent retention of power in 2004.

[Nederland]

35. So

...any person accused of the crime of election fraud is guilty until proven innocent?

An interesting jurisprudence philosophy you have there.

[Peace Patriot]

51. An election is not a court case, as I have explained elsewhere in this

thread. A court case requires an accused and evidence, and a presumption of innocence. An election is exactly the opposite. It requires everything on the table from beginning to end. There can be no exclusion of evidence. There can be no protecting the government's "rights" against the voters. The government HAS NO RIGHTS. Only WE, the people, have rights. Elections MUST be open. They MUST be transparent. And government assertions--about who won or who lost--must be provable. Not the other way around--citizens having to "PROVE" the government wrong.

Or you could have "elections" as they did in Stalinist Russia, where the government is always right, and its non-transparent vote counting is supposed to be accepted on faith.

It's basic democracy. Non-transparent elections ARE fraudulent elections--and it is the government's duty to prove otherwise, because, goddammit, they are our SERVANTS! We don't accept ANYTHING they say or do on faith! Governments and the powerful are not to be trusted. What the hell is democracy about, if not that?

Government is guilty of thievery and powermongering and unjust war and lying and corruption and all manner of fraud UNTIL PROVEN OTHERWISE by TRANSPARENT procedures, stringent ACCOUNTABILITY, and Constitutional CHECKS AND BALANCES.

And when those things are missing, you're damned right they are guilty until proven innocent.

[loudsue]

67. Peace Patriot, you're the best!

:applause: You are hereby nominated to lead the rebellion!! :applause:

You write so eloquently, and make the truth crystal clear. Thanks for all you do for Democracy!!

:yourock:


:kick::kick::kick::kick::kick:

[OnTheOtherHand]

39. well, it depends on the context

The OP says, "He (TIA) makes two simple points that destroy any notion that 2004 was a fairly counted election. Either the uncontaminated National Exit Poll on election day is correct and Kerry won OR the vote count and the day-after 'adjusted' National Exit Poll is correct (1:25 p.m. 11/03/04)...."

Doesn't seem like a big stretch to paraphrase that as "TIA proves that Kerry won." That is accepting a burden of proof, I would say.

So, if anyone here has the chops to defend that argument, bring it on.

Now, if you want to say that elections should be verifiable, and our elections are not, and you are not convinced that Bush won, and the election system itself should bear the burden of proof that he did, that is another topic.

[autorank]

41. PeacePatriot, once again I admire your skilled turn of a phrase...

"proprietary vote tabulation"...says it all doesn't it.

The Republican companies own the machines, they won't let anybody see inside, and VOILA we get
these ridiculous totals.

In another thread I made the undisputed assertion that when we say "the Republicans" it should
be taken to mean "the Republicans and the a MSM" because, as we all know, MSM=CM (corporate media)=Republican Party.

Through that gem of a marriage (from Hell) we have the vehicle to keep the public in doubt because nothing is EVER reported on this issue by "the Republicans."

Their time limited. Their day is soon to be done.

[Febble]

70. Actually, TIA is quite keen on clustering these days

He doesn't appear to know what it represents, but he bungs it in anyway, because of course it doesn't make any difference. Whether or not all states were within or outside their MoE or not is not really the issue. Even if they were all within it, the fact that significantly more states were redshifted rather than blueshifted means that that the discrepancy between poll and count is highly unlikely to be due to sampling error. If we go to precinct level analysis, the likelihood that the discrepancy was due to sampling error becomes infinitesimal.

So this raises the question: what did cause the discrepancy? Was it bias in the poll or fraud?

There are a number of approaches to answering the question, and one of them is examining the internal consistency of the crosstabs. Unfortunately the Gore/Bush crosstabs are poor ones to choose as a) the question was only answered by a subsample of respondents and b) it records recalled previous vote, not actual previous vote, and we know from panel studies that recalled vote is not a particularly reliable proxy for actual previous vote, and can overstate the previous vote for the incumbent.

I'd say, try another category, preferably one in which data is available for all respondents. Although even then, it is unlikely to be definitive, as what we want to know is who, if anyone, is missing from the poll. Postulating that the discrepancy is caused by polling bias implies that some demographic group is under-represented. If, for example, the under-represented group tended also to be rare voters, or swing voters, this might result in internal inconsistencies in the crosstabs, but would not necessarily mean that polling bias was inherently unlikely. But by now we are not doing statistics: we are testing various bias hypotheses for psychological plausibility, and that doesn't tend to come with associated p values. In other words, I don't think the crosstabs offer a particularly fruitful approach to disambiguating polling bias from fraud, although they are certainly interesting.

What seems a much more fruitful approach is to consider methods by which election theft might have been achieved, and from this, to construct testable operational hypotheses. So far, to my knowledge, these have not supported a theft conclusion. But I keep trying, and live in hope....

[Usrename]

78. I can't believe you are actually saying this.

Have you looked at the gender response? There is a major malfunction here. Get with the program. I'm not sure why you would say this. The 2000 vote is a historical fact.

The "MASSAGED" NEP results (or whatever term is acceptable to you, contract deliverable data items, the "OFFICIAL" results) violate reality, the physical real world.

No one is denies that, unless you do. Do you? Yes, people often get their opinion wrong, but this isn't opinion is it? How do you reconcile the numbers with reality? Give it a whirl. Feel free. Don't hold back. Tell us what you believe the numbers tell us. Inquiring minds would like to know.

[Febble]

81. Saying what?

Obviously gender, being a visible characteristic, will be held constant in the weighting.

Vote in 2000, being an invisible characteristic, will not be. In fact it won't be considered in the weighting at all, because the data is only obtained from a subset of the respondents.
So there are three reasons why the fact that the Gore:Bush percentages in the reweighted ("massaged" is fine by me; my own metaphor was "subjected to metamorphic pressure")sample don't jibe with the 2000 numbers:

1. The re-weights were crude (which they necessarily must have been)


2. Recalled vote is not a reliable indicator of actual past vote (which we know, from panel data)


3. The count was corrupt.

I DON'T believe the crosstabs tell us that the election wasn't stolen. But NOR do I believe that it tells us that it was. We are talking about survey data, and the alternative hypothesis to the fraud hypothesis is that the poll respondents were not a purely random sample - that those who "should" have been in the poll, but weren't were "drawn from a different population", as we say in stats, from those who were. And as those who "should" have been in the poll but weren't comprised at least 47%* of those selected, then that leaves a fair bit of room for a systematic difference. If the non-participants were more likely to be Gore-Bush voters; or DNV2000-Bush voters, then that would affect the crosstabs, as I said.

I maintain this is plausible, not that it is fact. But because it is perfectly plausible (to me) I don't buy the argument that it violates reality. Biased samples, especially with face to face selection, is all too real a phenomenon.

My case is simply that the exit poll evidence is very poor evidence for a stolen election.

That is NOT the same as saying the election wasn't stolen (although I am not convinced it was) and it is certainly not the same as saying the election was just.


*47% assumes all non-participation was due to non-response; it ignores the possiblity that selection itself was non-random.

[Usrename]

86. It's the weights that don't jibe with reality.

I didn't come up with them, someone else did. Why is this so difficult? If the data can't be weighted somehow to conform with reality (the results from both elections) and this is fact is made known, like in the OP, why is it so difficult for you to just admit, hey, the OP is correct. The facts don't fit the data. The weights they used are impossible. That just has no meaning I guess. You may be correct. There is no method here. The non-responders are what this is really all about.

So, I guess your argument now has become that as long as there are non-responders in any exit poll, the whole thing is meaningless. There is no point to it at all. And somehow, no one ever understood this, until you came along to point this out. I'm very skeptical here. It just doesn't make sense. No exit poll has ever had a 100% response rate. Are you saying that this makes all of them meaningless? And also, since randomness, in the sense you mean, is not possible to create, then that also must make exit polls meaningless.

[OnTheOtherHand]

121. to sketch some points from my PM

at the risk of repeating other posts --

A big problem with the OP is that while the 2004 final weights lead to "impossible" results for 2000 recalled vote, the 2000 exit polls -- which, we agree, were relatively accurate overall -- yield similarly "impossible" results for 1996 recalled vote, whether they are weighted or not. So, how can the "impossible" results in 2004 prove that something is wrong with the weights?

This isn't to say that the results are actually "impossible" or "meaningless." I think they pretty clearly mean that some folks say they voted for the incumbent when they actually didn't.

Surely Febble isn't arguing that "as long as there are non-responders in any exit poll, the whole thing is meaningless." As you say, every exit poll -- every poll, period -- has non-responders, and therefore the potential for non-response bias. Pollsters try to minimize the bias, detect the bias if it occurs, and compensate for it if so. Unfortunately, there is no reliable way to compensate for "bias" on the fly if it is actually fraud (in the count or, conceivably, in the poll).

Pollsters live with non-response error all the time. They don't seem prone to existential crises about it. Mitofsky was bummed that his survey was coming back 51-48 when the votes looked more like 48-51, but I doubt it ever crossed his mind that the data were meaningless.

[Nederland]

92. Interesting

Your comments are, as always Febble, insightful.

I lost touch with the whole TIA love fest even before he got banned. I guess my frustration with him was his refusal recognize that before you start calculating a MOE you have to demonstrate some sort of confidence that your sample is truly random. It would be one thing if he and I disagreed about whether or not a sample was random, but TIA acted like it wasn't even an issue. I still remember the time he calculated the MOE for a DU poll he put up, as if a self selected sample of DU members was somehow representative of, well, anything remotely random. As you mention, in the 2004 election the easiest way to do eliminate sampling error is to drop down to the precinct level, but at the time TIA never grasped even that. I suppose I should be glad he is making some progress on the self education front.

[OnTheOtherHand]

122. by the way, one thing about the MoE/clustering biz bugs me

As far as I can tell, the standard errors for the projection margins are estimated from the model itself -- not calculated by applying a pre-canned design effect multiplier to a calculated standard error a la Excel (binomial proportion).

E/M uses results from past races as a baseline. To get a crude idea of how this works, one might imagine a vector of 50 precinct-level deviations from past-race proportions: +3, -2, +4, +1, yadda yadda. The mean of those deviations would yield one estimate of the final outcome, and the standard deviation would yield a standard error of proportion. That's not how it actually works (inter alia, they use multiple races, consider ratios as well as differences, factor in absentee votes...), but it gives some of the flavor.

This sort of approach allows E/M to make use of data from past races; it also means that if precinct results are coming back with wild variations, the standard errors will be larger. (Of course, if precinct results are all biased to the same extent, the standard errors will not be larger.)

(EDIT: So, what bugs me is that I don't feel I've done enough to make clear that the whole "what is the MoE?" debate doesn't seem to reflect what the exit pollsters actually do. In fact, when it comes to an MoE for presidential vote in the national exit poll... AFAIK they don't compute one, because they don't make any attempt to "call" the popular vote winner.)

[Donailin]

71. Clustering effect?

So you adhere to the position that the Bush Administration -- The CLUSTERFUCK Of THE CENTURY -- who has started an illegal immoral war, profiteered on that war, committed treason, sold out the middle class, watch the poor drown in NOLA, loses and American city, fails to protect the borders, fires anyone who disagrees with bad policy if not twists their arms badly, looted our national treasure, spies on Americans especially their political opponents and in general just commits scandalous acts daily, would take their chances on fair and clean elections????

It seems you are the one who misses the "clustering effect"

[Nederland]

93. Perhaps you misuderstood

My position is that there is such a thing as the clustering effect.

Do you disagree? Do you believe that the clustering effect is an important issue to at least look at when looking at exit polls? If you do, then you and I have no difference of opinion about anything you or I have said so far.

[Amaryllis]

14. See this for graphic of exit poll timeline that even the math impaired can

understand:
http://www.oregonvrc.org/files/Timeline%20of%20Results_B&W.pdf

[Yupster]

142. Cool

That was like the beginning of Star Wars.

[ThomCat]

6. It doesn't seem to matter how much proof there is

of vote fixing and fraud. If the media doesn't care then nothing happens. Millions of sheep can go on blissfully pretending that the election was not rigged.

History might eventually look at all the evidence and state the truth, but that doesn't do us any good now.

[Peace Patriot]

49. ThomCat, most Americans are not sheep. They voted this junta out.

But they are suffering a lag time is getting the word on certain things--the fraudulent election system chief among them. And how can you blame them, when the war profiteering corporate news monopolies pull shit like DOCTORING their exit polls? We are reduced almost to a word-of-mouth culture--with the internet acting as our modern "Committees of Correspondence." In this circumstance, of war profiteering corporate news monopolies controlling news and opinion, it takes TIME for the truth to get around. And the failure of our election system--which included the failures of our Democratic Party leadership, of American journalism, of the courts, and of all other checks and balances--is rather mind-boggling, and hard and scary for people to absorb. We just have to keep hammering away, at, 1) the facts about the election SYSTEM, and 2) the obvious wrong results of the election system (a consistent 60% disapproves of Bush and everything he does--so what is he doing as president?). We also need to assure other Americans that THEY ARE NOT ALONE, because this is the most insidious effect of the corporate news monopoly illusion: that the rightwing is the majority. It most certainly is not--which is established by poll after poll after poll after poll, over a long period of time, on Bush/Cheney approval and on the issues.

MSM - mainstream media. Not.

Not, not, NOT the mainstream of this country. Not even close. Narrow fascist view falsely promulgated as "mainstream."

[zann725]

12. As always, TIA breaks down thenumbers, explains it better and with more

tireless enthusiasm than most. TIA primarily, and many other statisticians, and computer pro's convinced me beyond a doubt all along that Kerry won by AT LEAST 51-48%.

Interesting, I was recently talking with someone (clearly a Bush voter)...who upon noting my "Kerry" bumper sticker, said rather disgruntled/baffled, "Why is IT that people still keep those things on their cars? Why won't they let go...get on?" Chin-jutted, as always, I replied: "Because, FACT IS, Kerry DID win...by 5-7 million votes...according to most every well-respected statistician. We keep ALL these stickers on because more than you'd like to think, actually KNOW...and won't "let go" of the Truth...Kerry WON."

Deal with it Pugs...let that be your "hair-shirt".

TIA and his tireless, endless stats on the subject ARE right-on. Kerry won WITHOUT a doubt.

[autorank]

13. I had the same experience with a Kerry sticker...a family member.

Naturally, I was thrilled to see it and paused for a moment, even me, to realize the obvious point: 2004 was the second straight "crooked" Presidential election. Full disclosure of the raw data to all who wish to see isn't even necessary at this point. The summary data is so revealing in so many ways, the final 01/03/04 p.m. "adjusted" (to meet the actual "count") NEP results showing * winning simply crumble.

Reality based community conclusion: It was stolen, Kerry won.

Thanks to the brighter-than-me family member and you, I'm putting a Kerry sticker on my car. Nice protest movement until the cascade of gross negligence impels impeachment.

[bleever]

19. Even the Coincidence Theorists are welcome to add to the discussion,

but really would garner more support and have greater credibility if they could show how the highly improbably coincidence theories (via "within precinct error" or "clustering effect" or "reluctant Bush responders" or whatever) can make sense of the actual data.

Still waiting.

[OnTheOtherHand]

21. do you have a serious question?

Sir, I have contributed to dozens of substantive threads on the exit poll topic, and I don't think you have asked a question yet. So, I can only wonder what you are waiting for.

If you want to convince me that my arguments don't "make sense of the actual data," well, I am waiting too.

I personally don't think that clustering is very important in this debate, except to demonstrate that TIA isn't much of a polling guy. The basic problem is that lots of people are making claims about the uncanny accuracy of polling that actual pollsters never make. I mean, c'mon, folks, if thousands of polling experts were convinced that Kerry won the popular vote, wouldn't the public discourse be a teensy bit different?

Sigh. Hey, good luck out there.

EDIT: oh, looks like I am supposed to object to the "Coincidence Theorist" bit. Hey, whatever.

SECOND EDIT: did someone convince you that it is mathematically impossible that Bush won the popular vote? Just curious.

[kster]

24. I got a question

Alaska, Califonia, Hursti hack does any of this make you the least bit suspicious ? Their Failure to prop up their side or answer our charges in the media, should at very least throw up a red flag for you. If we had control of the media we would'nt be afraid to go public. Why are they?

[glitch]

26. Excellent question. K & R nt

[OnTheOtherHand]

28. why are you changing the subject?

kster, if the exit polls don't support fraud, they don't support fraud. As sequiturs go, yours is a non. Your completely unfounded rhetorical assumption that I am not "the least bit suspicious" is beside the point.

Now --

Alaska: the presidential result is close to the last poll result -- and contrary to some assertions, the Knowles/Murkowski result also seems to be close to the last poll result, although the provenance of the poll is uncertain. What specifically should I suspect?

California: what is your question?

Hursti hack: if I argued that the machines were hack-proof, the Hursti hack would be a big problem for my position. I don't, so it is irrelevant.

I'm not sure what the rest of your post means.

If you agree with the content of the OP, are you prepared to discuss it?

[kster]

29. Yes I agree with the op, Autorank you mind if I ask OTOH

one question in your thread?

OTOH.....Why does the Government Fail to prop up their side or answer our charges in the media? Exit polls, Election theft machines?

Autorank I will delete this post if you want.

[OnTheOtherHand]

36. what government?

Do you mean, why doesn't George W. Bush issue a press release explaining his interpretation of the NEP exit poll, or his opinion of DREs?

Umm, why would he bother?

AFAICS the problem with the exit poll argument is not that the government won't respond to it, it's that pollsters don't take it seriously. And that is a distraction from the arguments about the machines.

[kster]

50. Why can't we have this discussion

on TV, prime time so all Americans can decide, why is it always discussed online. But never on TV?

Why isn't the (Government) Democrats and Republicans the elected and the selected ones on TV to settle the debate.

I think Americans would like to know that in 22 states, the exit polls where off.

In Ohio they called a terror alert so no one could see the count.

And the fact that with the election theft machines everthing including the counting of our votes is done in secret.

Don't you think Americans are entitled to know all of this.

[OnTheOtherHand]

52. hey, you want to put me on TV? cool...

But let's get real. If they put this discussion on prime time, Americans are going to change stations.

If I had an hour to try to change Americans' minds about something, I would not choose to spend it talking about exit polls. If you are not interested in discussing the facts about exit polls, then I don't see much point in wrangling about whether the media is spending too little time doing it.

[Name removed]

53. Deleted message

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

[OnTheOtherHand]

54. sir, that is neither fair nor smart. n/t

[kster]

55. Why do you choose to do it at the DU?

You said "I would not choose to spend it talking about exit polls" just curious.

[OnTheOtherHand]

58. because I want progressives to care about facts

and the fact -- as opposed to the truthiness -- is that the exit polls don't provide useful evidence of massive fraud in 2004. The argument is a dog, and it makes progressives look like creationists. In fact, it is precisely an Intelligent Design argument, as is evident from the folks on this thread who refer to "Coincidence Theory."

[kster]

63. What about all the facts surrounding the exit polls?

Do you take them into consideration? McPherson certified illegal vote counting machines in california, this is a fact.

I know you like to stay targeted on exit polls, but there are so many other things around them exit polls that don't ADD UP, that it makes the "exit polls were off theory" real strong. To me anyway.

[OnTheOtherHand]

69. no, here is the thing you misunderstand about me, kster

I do not "like to stay targeted on exit polls." I think we should focus making sure that people get to vote, and to have their votes counted. And I think we should look for all available evidence of wrongdoing in 2004.

What you are saying is that you believe the exit poll arguments because they jibe with what you know from other sources. But that isn't the same thing as finding the exit poll arguments convincing.

You apparently think I am trying to argue you out of believing that Kerry won. I'm not. Heck, I hope you can prove it. But I will continue to insist that TIA's arguments are weak, because I think you have a right and an obligation to know the problems with them. If you want to be taken seriously by people who know about polling, you will be better off not offering TIA's arguments. If offering weak arguments isn't an issue for you, that is your choice.

Lots of people here want to believe TIA's arguments, and therefore aren't willing to look seriously at criticisms. I know this will not change.

[Usrename]

80. My turn for a question?

Do you think exit polls have any significance or value in determining whether an election is on the level? If so, what is the threshold of credulity? One-in-a-billion? One-in-a-gajillion? I'm just wondering what you think an exit poll that reeks would smell like. Many think this one reeks and yet you seem impervious to the odor.

I thought exit polls were like the canary in the coal mine. Are you arguing that our canary is alive and well and in perfect health? Or, should we get a new canary? Is this one too invincible, too impervious to harm to be useful? If that is your argument, you aren't coming through very clear to me.

[OnTheOtherHand]

83. let me try to take this point by point

Are exit polls of any significance in determining whether an election is on the level? Sure. I take all the facts I can get.

However, it isn't a matter of setting a statistical "threshold" that would tell you whether the election probably was or wasn't counted right. All a P value can tell you is whether a given discrepancy -- in this case the discrepancy between the exit polls and the count -- is likely to be due to chance or not. And we all agree that this one isn't.

That leaves us to sort out whether it might be due to fraud, to participation bias or some other problem with the exit poll, or to some combination of those. I've spent over a year working on that question. I don't know how you got the impression that I was "impervious to the odor."

In a nutshell, the 2004 exit poll errors don't seem to correlate with much of anything one might expect them to if they evinced fraud, and they correlate with lots of things one might expect them to if they evinced participation bias. For instance, they don't correlate (at the state level) with change from pre-election polls, and they do correlate (at the precinct level) with interviewer distance from the polling place. That's why I don't consider the exit polls useful as evidence of fraud in 2004.

Of course we should get a new canary -- no, better yet, we need a new mine. That is, we should build an election system that everybody trusts. Yes?

[BeFree]

106. Question

Quote: the 2004 exit poll errors don't seem to correlate with much of anything one might expect them to if they evinced fraud.

The question: What would be evidence of fraud? IOW, what type of errors would convince you fraud did occur? What would you expect to see in the exit-polls to make you even consider fraud might have been committed?

[Febble]

109. Me! Me! Pick me!

Bearing in mind that one has to make any inference of causality from a correlation with caution

I would expect the fraud explanation for the exit poll discrepancy to be supported by evidence that:

• Greater precinct-level redshift in precincts where we think fraud would be easiest (DRE precincts?)


• Greater precinct-level redshift disrepancy was correlated with swing to Bush since 2000.


• Greater precinct-level redshift in swing states.


• Greater precinct-level redshift where there was a Republican BoE


• State level discrepancy was correlated with pre-election poll discrepancy.


• Internal inconsistencies in the cross-tabs that are rather greater than those I have seen already (but I continue to look)

I would expect the participation bias explanation for the exit poll discrepancy to be supported by evidence that:

• Precinct level redshift was greater where the interviewing rate was low


• Precinct level redshift was greater where the interviewer had to stand far from the precinct


• Precinct level redshift was greater where the interviewer felt badly trained


• In short: precinct level redshift was greater where the opportunities for selection bias were greatest.

Of group one, the swing state correlation seems to be supported. However, there is an alternative explanation, which is that redshift in the past has been greatest where interest in the election is high. So there is also a plausible behavioural explanation.

Of group two, all are supported. However, there is an alternative explanation for the first two: fraud may have been greater in larger precincts, which also had lower interviewing rates; and interviewers may have been required to stand far from the precinct in precincts where there was fraud going on.

However, I do find the complete lack of a swing-shift correlation (no correlation between redshift and swing to Bush since 2000) fairly strong evidence against massive, vote-switching fraud, and I have yet to be convinced by any fraud mechanism that could account for a substantial proportion of the discrepancy without producing a swing-shift correlation. I keep searching, and welcome suggestions.

Bottom line: I don't think the exit polls are good evidence for vote-switching fraud on a large scale, and if anything, tend to suggest it that vote-switching fraud did not account for the discrepancy.

However, that is not the same thing as saying fraud did not occur, or that no fraud is compatible with the exit poll data. There are plenty of forms of disenfranchisement (including over and under votes; uncounted provisional ballots) that would be. But it is the same thing as saying that the exit polls are not slam-dunk evidence of fraud. Not close, IMO.

But this is EXACTLY the right question to ask, and I'd like to know if anyone else has answers.

[Peace Patriot]

133. Febble, you are making all kinds of hidden assumptions in this list--

and are leaving out the most critical facts that should preface any investigation of the exit polls: Are conditions for election fraud present? Is the election transparent and recountable? And if the answer is "yes" to the first question, and "no" to the second, WHO has the best opportunity to commit fraud, and WHO created the non-transparent, unrecountable conditions? In short, who is in control of the vote counting?

I repeat my point from above: You MUST consider the CONTEXT in which the exit polls occurred.

Considering that conditions for election fraud WERE present, in spades, and that the election was unquestionably non-transparent and non-recountable, and considering who played the most critical role in arranging that non-transparency--two of the biggest Bushite crooks in Congress--and considering that two pro-Bush electronic voting corporations consequently gained control of the vote tabulation with "trade secret," proprietary programming code, which has been proven to be easily hackable in a matter of minutes by one programmer leaving not a trace, and which can switch large numbers of votes over a wide grid of results, in impossible to detect patterns, at the speed of light, you SHOULD NOT make any of the following assumptions, in analyzing the exit polls:

--that the fraud would be easiest to detect in DRE precincts
--that the fraud would be easiest to detect in any comparisons with Bush 2000 (different voting system entirely)
--that the fraud would be easiest to detect in swing states(*)
--that the fraud would be easiest to detect in Republican-controlled precincts
--that small inconsistencies are not meaningful (i.e., cross tabs)

As for pre-election polls, there is such a significant difference between pre-election polls and exit polls--the one of possible voters, the other of actual voters just after they voted--that this should not be considered at all. (Remember Harry Truman! Pre-election polls can be very wrong; exit polls seldom are.)

(*)Re: your "alternative explanation" about swing states. There is no evidence of higher interest in swing states in THIS election--an election that was unique in many respects (if one can call it an election at all), and concerning which there was high interest and big turnout all over the country. The watchword among Democratic grass roots groups was "this is the most important election in our history"--and is evidenced by a nearly 60/40 blowout success by the Democrats in new voter registration in 2004.

You are not considering the most likely scenario, given the fraud capabilities in this election system and its non-transparency: that the pattern of electronic fraud was cleverly randomized to avoid statistical detection, and that small percentages were stolen here, there and everywhere, with some concentration in the swing states, and that these thefts occurred in the CENTRAL TABULATORS, not in the voting machines (the DRE switches might even have been a red herring). These small randomized shaves (of various kinds: switch from Kerry to Bush, switch from Kerry to 3rd parties, 'disappeared' Kerry votes, added Bush votes, etc.) over a wide grid of results would be very hard to distinguish from the little up and down blips of any exit polling, and could only be seen clearly in the aggregate--a 3% margin of victory to Kerry in the exit polls, overall. The most detectable would most likely be in the swing states--several of which were critical to a victory (because of the Electoral College)--and that is exactly what we find, looking at the numbers (and the redshift) in this election. Certain states had to be taken by Bush; that's where the most noticeable redshift bumps occur. Bush's popular vote edge could then be manufactured in a randomized way all over the country. (But I recall that TIA found an east coast redshift--which would make sense in terms of election dynamics--to a fraudster.)

I believe that the electronic fraud may have been limited by a need for pre-progamming** (not easy to alter on election day), and that this may have been the reason for the blatant, illegal, and risky vote suppression in Ohio. One of the reasons I believe this is the tenor of the country at the time, and the spirit among Democratic voters and grass roots get-out-the-vote groups (also the new voter registration stats, and the high turnout). I think Kerry's win was higher than the Bushites expected (more than the exit poll 3%) and that additional measures had to be taken, and that some of the old pols like Rove didn't trust the promises of the electronic voting wizards, and had plenty of old-fashioned Republican "dirty tricks"-type vote suppression planned and ready to go (too few precincts, too few voting machines in poor black areas, etc.). They were faced with a big Kerry win on election night (and a big rejection of Bush and his filthy war) and had to pull out all the stops to get even the modest margin they got (2.5%).

There is evidence of pre-election vote suppression--in Ohio, Florida and other places--and it's possible that the randomized global thefts were planned to be combined with blatant vote suppression, either to avert a Kerry win, or to divert attention from the main fraud. But I tend to believe they were faced with an unexpectedly high Kerry win, and had to take extra measures. (The highly visible suppression of black voters was the most likely to cause notice and post-election trouble--as it did--that's why I think it was a desperation measure.)

----------------------

A note about Republicans:

It's my opinion that Republican precincts were the most likely places that the Bushites would steal votes--both because high Bush vote counts in such precincts would be less likely to raise eyebrows, and because the Republicans in control would not likely raise questions about unusually big numbers for Bush (or unusually low numbers for Kerry*).

Republican precincts are ALSO places of high suppression of contrary views and tight social, and sometimes church-related, control over individuals. These precincts therefore may well be subject to a "non-response" or "wrong response" effect--not by Bush voters, but by Republicans who voted AGAINST Bush. I happen to think there were a lot of such votes--possibly even a big factor in the election. My evidence is anecdotal but it is compelling. And I have some social and church experience with which to gage what a Kerry vote among rightwing Republicans, especially Christian rightwing Republicans, would do to the social standing of an individual. The suppression can be severe, and may well have resulted in social Republicans, or religious Republicans, being fearful of admitting a Kerry vote to anyone in a public place. I frankly think there was a lot of private rebellion going on, but it would never be admitted in any circumstance where it could get back to others in their social or religious groups.

This kind of suppression is extremely rare in Democratic-dominated areas. Democrats are more tolerant, more respectful of others' views, and much less coercive in both social and religious settings. Further, there was quite a difference in the nature of Bush and Kerry supporters. I can remember elections when Democrats were possessed by the personality of the candidate (JFK and RFK are good examples). But that was not the case with the bulk of Kerry supporters, who were more possessed by the idea of ousting Bush--for which they were VERY enthusiastic--and were not especially fanatical about Kerry as a person (mainly because of his cool stance on getting out of Iraq). Contrast this with Bush supporters, many of whom do tend to be hard-line and fanatically attached to Bush. A Republican in a Democratic precinct who was voting for Bush would have no fear of social disrepute, or dirty looks, or slashed tires, or lost employment, or other punishments, as a result of telling a pollster of their vote of Bush. A Republican in a Republican precinct who was voting for Kerry would be inclined to keep their mouth shut about it, for fear of repercussions by the hardliners.

These are to some extent value judgments, I know--but they are also based on long experience.

If there is any bias in the exit polls, I therefore think that it is the opposite bias than the one E/M put forward--which, for one thing, assumed that Republican non-responders were Bush voters. That is not a safe assumption, if you take REAL political and social conditions into account. E/M also made the assumption that Bush voters--some of whom put Bush on a pedestal with Jesus Christ--and who had had four years of their heart's desire in anti-gay and anti-abortion proposals, and in war, mayhem and thievery by the rich, would at all be reluctant to admit to a Bush vote, as Republicans might have in the past. We were living in BushWorld by then, with Bush doing anything he liked, to a fawning news media. Why would any Bush voter be reluctant to admit a Bush vote? They would be more inclined to brag about it, and even be in your face about it--my experience of Bush supporters; not at all shy.

I think the exit polls had this hidden hole in them--the socially fearful Republican voter for Kerry--and they missed some of of the Kerry vote because of it.

I think the exit polls correlate closely to the real vote, with the exception of the above (which adds to Kerry's margin, making it more than 3%), and that, if all the votes had been counted, and all those who wanted to vote had been permitted to, Kerry won the election by a 4% to 5% margin of victory.

I consider the Edison/Mitofsky falsification of the exit polls to be the greatest disservice ever done to the American people by a polling organization and by the corporate news monopolies which commissioned the polls. The numbers were doctored in impossible and absurd ways to confirm a Bush win, in the knowledge that the election system was non-transparent. If ever we needed an honest pollster, it was 2004. E/M failed us completely, as did the five fatcat, war profiteering CEOs who run the news monopolies and all their toady, careerist, so-called journalists. I shall never trust E/M again on any matter. They are liars in the pay of people who helped lead us into unjust war--into murder and torture, into the loss of our very soul as a nation. What they did was the height of irresponsibility. It was unforgivable.

----------------------

*(California is an interesting case. Barbara Boxer (running as an incumbent for US Senate) won the state by a 20% margin in 2004. Kerry won the state by 10%. All of the difference between them is to be found in Republican precincts, which makes no political sense. They ran about even in Democratic areas. The difference does not correlate with method of voting, but rather with Republican political control. It's my opinion that this points strongly to the central tabulators (Diebold/GEMS) as the culprit in 'disappearing' Kerry votes, to reduce Kerry's national popular majority (a potential popular/Electoral vote difference haunted the Bushites, after 2000), and possibly also to pad Bush's popular majority, and that Republican precincts were chosen because an unusually low vote for Kerry would be less noticeable there, and the Republican election officials could be counted on to ignore anomalies in vote totals that favored Bush. The exit poll difference was small in California (a half a percent, as I recall--although in California that's a lot of votes). This means that the California exit poll could be off by as much as 9.5% (failure to detect Kerry votes). Some part of the Boxer margin (over Kerry) in Republican precincts might be due to incumbency or familiarity; some also might be Republican women voting for a woman (but not Kerry). Say half that 9.5% is accountable--which would mean that the exit polling missed 4% to 5% of the Kerry vote. I think the Republican Kerry voter/reluctant responder may have been a factor. My experience of California Republicans over the last several decades is that they tend to cluster in tight enclaves and exercise strong social control over each other--and some Republican areas can be quite rightwing extremist. Fear as to admitting a Kerry vote in public would be a factor there. At the same time, the OLDER Republicans--those who had joined the Republican Party in its more liberal heyday ('50s-'70s)--would be good candidates for Kerry voters (and my anecdotal information points that way). The Bushites may have had their reasons for not ALSO altering the Boxer vote (along with the Kerry vote), which I won't go into here. The Boxer/Kerry discrepancy remains one of the oddest things I've seen in a California election, or any election. Boxer is to the left of Kerry--and voted against the Iraq war--yet did better with Republicans???) (Are California Republicans secret leftists? Or are they crazier--more schizie--than other Republicans? A Bush/Boxer vote is very hard to figure.)

**(This is a rather important strategic issue--does the electronic fraud have to be preprogrammed, or, if it had to be preprogrammed in 2004, will that continue to be the case? I am hoping that close, refined analysis of the election numbers, including exit poll analysis--that takes the context of the fraudulent election SYSTEM into consideration--can help us figure this out.)

[OnTheOtherHand]

136. well, I wish Febble luck with all this, but some fact checks

Don't get me wrong -- it is a very interesting post. I'm not quite sure why you think you need to lecture Febble, of all people, on whether conditions for election fraud were present, but OK. Of course I agree with the principle that we should consider proposed fraud mechanisms in designing our tests.

Your translation of Febble's expectations as "fraud would be easiest to detect..." (my emphasis) is erroneous -- I don't know where to begin, beyond that. Maybe it's just a semantic quibble.

In 1992, most of the pre-election polls were more accurate than the exit polls were. Invoking Truman is a stretch, since the Gallup Poll then used quota samples and the last survey was in mid-October. It is probably a bigger stretch to argue that we should ignore the pre-election polls altogether. Tell the truth: would you still have said that if Bush had won Vermont? Pre-election polls could be wrong, sure, but would we expect them to be wrong in varying degrees from state to state in just the right proportions to disguise any evidence of fraud? Isn't tht a bit wishful? (Still, it might be worth it, if you can get TIA to agree to ignore pre-election polls as well!)

Nor is there any reason to ignore 2000 on the grounds that a different voting system was in place. The point is: if Bush steals a lot of votes in some states and few in others, why will those states not tend to stand out in comparison with 2000? One can come up with reasons, but I haven't encountered any parsimonious ones. (Kathy Dopp's reason is, roughly: "Hey, it's mathematically possible!" Yes indeed. But is it likely? It would be reasonable to question the premise that vote-stealing varies across states, except that exit poll discrepancies vary across states. What's up with that, if the discrepancies are the proof of vote-stealing?)

There certainly is evidence of especially high interest in swing states: turnout (% of voting-eligible population) was up a median 6.3 points overall across the 50 states and DC, but a median 8.2 points in the swing states -- 9.6 points in Ohio and 9.7 points in Florida.

The "small randomized shave" (SRS) argument doesn't really suit the exit poll evidence that I thought was supposed to support it. Some states experienced double-digit exit poll discrepancies, others not; the discrepancies vary wildly from precinct to precinct. If the exit polls support fraud at all, they do not seem to support small, randomized shaves! Actually, it is interesting that you should venture that. The latest NEDA report on Ohio purports to refute Mitofsky by showing that a constant degree of exit poll bias doesn't explain the data. This finding is not a problem for Mitofsky, since non-response bias is expected to vary across precincts, but is a problem for your SRS argument unless "small, randomized" means "somewhere between 0 and 40+ points in either direction." (OK, that is a bit polemical, but really, you have a problem here.)

(The small, randomized shave argument also doesn't really suit the election day evidence, which is why you offer a hybrid. But I actually agree that a hybrid model is plausible, at least in principle, so that is OK.)

"It's my opinion that Republican precincts were the most likely places that the Bushites would steal votes" -- OK, then, can we test that? (Trick question: we have spent months doing that!) Also, if we changed "Republican" to "Republican-controlled," we would be back to one of Febble's expectations that you derided; what exactly makes your expectation more compelling than hers? Hmmph.

I won't offer any general conclusion here. I look forward to seeing whether you can generate testable hypotheses that actually explain the available data nationwide better than alternatives.

You can assert indefinitely that E/M were guilty of "falsification," but I wish you wouldn't, because it tends to peg you as someone who doesn't know much about exit polls and doesn't want to -- not helpful in this context.

Cheers,
OTOH

[Febble]

137. Golly, you sound like my mother, PP

How I used to dread those fat envelopes....

And like my mother, you are making quite extraordinary assumptions about me.

But let me take it in bite sized chunks, and preface them by saying that no, I am not forgetting the things you allege I am forgetting. You are drawing a vivid picture by connecting a certain set of dots. I, to be honest, think that there are some important dots you yourself are leaving out, because they would change your picture (or at least make it more nuanced) and I think that some of your dots are spurious (yeah, I can do metaphor too....)

OK:

First off, you mistake an assumption for a hypothesis (remember I am a scientist.) I do not assume any of the following:

--that the fraud would be easiest to detect in DRE precincts
--that the fraud would be easiest to detect in any comparisons with Bush 2000 (different voting system entirely)
--that the fraud would be easiest to detect in swing states(*)
--that the fraud would be easiest to detect in Republican-controlled precincts
--that small inconsistencies are not meaningful (i.e., cross tabs)

What I proposed were testable hypotheses. To understand this, consider what your response were be if these were tested, and indicated that, for example, redshift was significantly (statistically) higher in DRE precincts than in precincts using optical scanners (or vice versa - there are good reasons for using two-tailed statistical tests). Failing to find such a correlation proves nothing. However, finding such a correlation would be extremely suggestive (although we have to be on guard for innocent collinear variables). This is why I am constantly on the lookout for testable hypotheses, because while, mostly, failure to that a particular effect is "significant" simply means that your precise hypothesis was not supported, finding a "significant" effect may well consititue important evidence.

So to consolidate so far: these are NOT assumptions; they are a list of hypotheses worth testing. At least one appears to be supported - redshift appears to have been greater in swing states. This is suggestive, although many interpretations are possible (including a rather boring mathematical one I won't go into here). And BECAUSE these are not assumptions, you need to back off a bit (as I would have said to my mother) because I am not up to what you seem to think I am up to, and in fact, am suggesting potentially fruitful lines of investigation - or at least soliciting more.

Second off:

You accuse me of not considering:

1. fraud "cleverly randomized to avoid statistical detection"

Oh yes I am. I don't expect you to read my posts particularly avidly, but if you had been you would have read the many occasions in which I have made the point that, perhaps contrary to lay intuition, random fraud would have been much easier to detect than non-random fraud. Random fraud is actually difficult to square with the data as we know it; however non-random fraud (fraud specifically targetted to hide the patterns that "randomness" would create" may well be consistent with the data.

2. "and that small percentages were stolen here, there and everywhere, with some concentration in the swing states"

Well, yes, and contrary to your implication, this is precisely what I have suggested, in several previous posts. Ask Land Shark.

3. "and that these thefts occurred in the CENTRAL TABULATORS, not in the voting machines (the DRE switches might even have been a red herring)"

Well, there is a bit of a problem here: I agree that the central tabulators may have been used for fraud, but that wouldn't account for much of the precinct level discrepancy observed in the exit polls, as the precinct level discrepancy was largely calcuated on precinct counts. But again, if you had followed my posts on this subject, you would know that this is a hypothesis I have been keen to explore.

4. "small randomized shaves (of various kinds: switch from Kerry to Bush, switch from Kerry to 3rd parties, 'disappeared' Kerry votes, added Bush votes, etc.) over a wide grid of results would be very hard to distinguish from the little up and down blips of any exit polling, and could only be seen clearly in the aggregate--a 3% margin of victory to Kerry in the exit polls, overall."

Actually about 6.5% but whatever. And yes, you would see it in the aggregate. I would also argue that it is difficult to envisage a scenario (and believe me I have tried to model several) where "these small randomized shaves" - if they were responsible for a substantial proportion of that 6.5% (insert: difference between margins) aggregate - would not also show up, aggregated, as a positive correlation between redshift and swing, which they don't. Either they weren't random, or they didn't add up to the redshift.

Third off:

You take me to task for my interpretation of the swing state finding, which a bit below the belt, as I actually pointed it out, and placed it in the "supports fraud" category, merely footnoting, as any good scientist should do, that a reasonably plausible alternative interpretation is also possible. This is QUITE DIFFERENT from arguing that the alternative is more likely, or that because there is an alternative, one shouldn't consider fraud. It is merely doing what good data analysts should do. I chalk the swing state finding in the column marked "consistent with a fraud explanation"). I would like to keep it there please, asterisked as necessary.

The rest of your post I won't address in detail, as it is a full-scale rant (and I love rants) not specific to me. I simply disagree with you about the pollsters, but I have addressed that point elsewhere. But I will end by saying:

I do not believe, in any circumstance, that "he who is not with us is against us". I may not be with TIA on his interpretation of the numbers. I may not be with you in your interpretation of more qualitative data. But this DOES NOT MEAN that I am against you, or what you want to achieve. I applauded an excellent point you made upthread. Frankly, ANY disenfranchisement is QUITE UNACCEPTABLE in a democracy, particular one that wields the power that America does over the rest of the world IN THE NAME OF DEMOCRACY.

But I happen to believe that bullet proof arguments are more convincing than flawed ones, especially when it comes to quantitative data. And, more importantly, I think that a well informed look at the quantitative data, can lead to real progress in finding out what went wrong, whereas I think that blanket assertions, while they may be great for inspiring activitist to action (and as I said, I like a good rant) can actually lead people in the wrong direction. I think (after extensive consideration) that the exit polls are the wrong direction. You are entitled to disagree.

But, like my mother, you seem to have got completely the wrong end of the stick about where I am coming from, or where I want this campaign to get to. That may well be my fault. But I have at least attempted to put the record straight here.


Edited to correct: Sorry PP, you are quite right with your 3 points. I am so used to thinking in terms of difference between margins at precinct level, I thought you had underestimated the discrepancy....)

[OnTheOtherHand]

110. OK, setting aside the rancor

(EDIT: Aw, shoot, Febble beat me again. OK, she appears to have all of mine plus about six others. And based on activity on other boards, she also appears to know all about paleontology. Sigh. Time for me to go mop the floor, I guess.)

First of all, your third question seems to be based on a faulty premise. I have always considered that fraud might have been committed. (Actually, I think we can be sure that fraud occurred somewhere -- but in this context, I assume that we are talking about fraud large enough to swing the popular vote. Only because that is what the exit polls are often supposed to demonstrate. I do realize that Kerry could have become president by winning Ohio regardless of the popular vote.) If you can register this fact as a fact, it may make things easier.

I suspect that your question is rhetorical -- that is, I think you are convinced that actually, I am setting up some impossibly stringent standard for Evidence of Fraud, and I will refuse to pay attention to any evidence whatsoever. But these questions are legitimate whether or not you are interested in my answers, so I will try again.

One answer to your first two questions appears right after the sentence you quoted. If the often large, widely varying state-level exit poll discrepancies evince varying degrees of fraud, then I would expect Bush to have done better (in general) relative to pre-election polls in the states where there was greater exit-poll "red shift." He didn't.

I also would expect Bush to have done better relative to his 2000 performance in the precincts with the greatest "red shift"; he didn't.

If the 2004 election was stolen on electronic machines, I would expect exit poll discrepancies to be larger in precincts with DREs -- or DREs and op scans -- or, hey, what the heck, DREs, op scans, and punch cards -- than in precincts with lever machines. They weren't. (Of course, an alternative fraud mechanism could lead to different expectations.)

I would also expect Bush to do better relative to his 2000 performance in precincts, counties, and states with suspect equipment -- especially those that recently adopted suspect equipment -- than in other precincts/counties/states. By and large, he didn't.

I would not especially expect exit poll discrepancies to be larger in precincts where the interviewer was far from the polling place, but they were. I would not especially expect them to be larger in precincts with young interviewers, but they were.

None of this is proof that Fraud Didn't Happen, or even that Kerry didn't win the popular vote. But in my opinion, if one takes all this evidence seriously, it is much harder to come up with a coherent scenario in which Kerry won the popular vote.

Now, on the other hand (TM), there is sound statistical evidence that misallocation of machines in Franklin County, Ohio reduced turnout and hurt Kerry. And there is sound statistical evidence that "caterpillar crawl" in Cuyahoga County cost Kerry votes (it is hard to tell how many). There are many unknowns about the extent of voter suppression in Ohio and elsewhere -- bear in mind that most of this would not show up in exit polls. And so on.

I hope that this benefits someone who reads it. I'm writing quickly, so I welcome requests for clarification, suggestions of other things to look for, and so forth.

[BeFree]

111. Good lord...

....do you guys get paid by the word? _{just kidding, right? a joke, ok?}

It was a real simple question: what signs of fraud would show up that would convince you that fraud had been committed? I just read a bunch of what makes yall think nothing happened. I've heard it all before, I don't agree with most of it. You have not convinced me. OK?

So, let's try it this way: In theory, what signs in a future election would convince you that fraud occured?

[OnTheOtherHand]

112. umm, humor is OK, sure

I am not trying to convince you of anything. I'm not sure that you have convinced anyone of anything, either, but that's OK -- no one has died yet, or anything.

But c'mon, you can answer the last question for yourself by reading my preceding post and changing any words that are specific to a particular election. Right?

There are others, of course. Even without cross-state variation, if the polls showed the Dem up five and pulling away, and then the Rep won, that would obviously be suspicious. Massive undervotes are and would continue to be suspicious. Discrepancies between the number of people who appear to have voted and the votes actually counted are suspicious. The reports of DREs flipping votes are suspicious, but would be much more suspicious if (1) there were more of them and (2) Bush did much better than expected in places where there were more such reports. I'm sure you can think of others.

[BeFree]

114. No one has died

Oh, I think a few people have died because Gore, and Kerry had their elections stolen. Yes, quite a few.

So, if a pre-election poll shows someone up by five points and increasing, and the opponent ends up winning, that would be evidence of fraud?

How about if someone wins a race by 5 more points than the exit-polls projected before 11pm? 6 points? How many points would it take before you could begin to think: Fraud?

[OnTheOtherHand]

116. I agree with your premise, yes

I meant, of course, that no one has died because we disagree. You could contest that, but things are confusing enough already, I think.

BeFree: How about if someone wins a race by 5 more points than the exit-polls projected before 11pm? 6 points? How many points would it take before you could begin to think: Fraud?

Please reread my post #110, which states:

OTOH: First of all, your third question seems to be based on a faulty premise. I have always considered that fraud might have been committed.... If you can register this fact as a fact, it may make things easier.

Conversely, if you refuse to register this fact, then meaningful discussion is impossible. If you did register the fact that I have never ruled out fraud, then how can you ask what level of exit poll error "red shift" it would take before I could begin to consider it?

Are you, or are you not, yanking my chain?

(EDIT to clarify meaning -- I meant "error" in the technical sense.)

[BeFree]

117. Of course

Everyone who knows anything about the elections has to consider there was fraud committed. That's a given.

But as a question to an expert such as yourself, I would expect an easy answer. I guess you think I am trying to capture you, or trick you into saying something that would make you look bad?

You have reason to be aware. But it really is such a simple question: What kind of shift would it take in an election to make you claim fraud? Actually, I think you already answered it earlier.

You did say a shift of five points, right?

[OnTheOtherHand]

119. why did you not answer my question?

Let me optimistically invent an answer on your behalf: "No, I'm not yanking your chain. I am asking what level of shift would intensify your suspicions." Would something like that be your answer?

(I can't tell why you are on this kick about my being an "expert." I try to know what I am talking about, and to ask when I don't, which is what I would hope of everyone else here. You seem to have a weird habit of asking questions, then verbally snickering at the answers as if you knew better, when as far as I can tell, you often don't. It's not a matter of who is an expert and who isn't.)

Now, without actually having answered my question, you pose your own: "What kind of shift would it take in an election to make you claim fraud?" As worded, that question is simple, but useless. The answer would depend on the nature of the evidence, as well as the type and magnitude of hypothesized "fraud" I was trying to assess. (The "shift" due to overvotes in Florida 2000 was probably less than 1 percentage point, but I find the evidence very persuasive.) Moreover, I would not rely on a single 'litmus test' if I did not have to.

I do not necessarily suppose that you are asking me useless questions in order to "trick" me, but I have to wonder what your purpose is.

Are you "really" asking whether the 5- or 6-point average exit poll discrepancy is suspicious? Yes, it is. That is one reason why I have spent hundreds of hours over the last year-plus studying it.

[BeFree]

126. You wonder what my purpose is?

You come hear with virtually nothing constructive. You tell us we are wrong, and that our evidence is wrong. I hate to say this but maybe everybody else is right: You are mere flaming.

My purpose: to uncover and help describe what I beleive was a stolen pair of presidential elections. You offer me nothing in that regard, and when asked simple questions evade answering. You don't want to discuss this issue. You want to control it. You have big problems with TIA, and others. I don't believe you are honest and fair in the discussions. I am but a barely litterate person, and you, a professor, right? You already have a great advantage over me yet you can't be totally open.

What is your purpose, I wonder. Is this how you get your kicks, by flaming activists? Sad.

[Febble]

130. I think the problem

is that we have two separate agendas here, that are not diametrically opposed, but end up looking as though they are.

In one corner we have activists like you, dedicated to uncovering a "stolen pair of elections", and all evidence, weak and strong, is grist to your mill. And I have some sympathy with that.

In the other corner we have people like OTOH, and me, and maybe Foo_bar and Nederland, who are similarly appalled (yes) at the broken state of American democracy, who also believe (probably - I do) that Gore was the rightful winner of 2000, and who also believe that 2004 was grossly unjust, but who do not think that the evidence points unambiguously to outright election theft.

So we end up actually fighting over how strong the exit poll evidence is for actual theft in 2004, instead of working together for ways of restoring transparency, trust and justice, which we all agree is imperative.

I think both groups have important roles to play in the future, and possibly in uncovering what happened in 2004. It might be difficult for both groups to see the role the other has to play at times. I will try and see yours more clearly. But I would submit that there is also value in critiquing the arguments used by your own side, which is what I see myself as doing, and I don't think the exit poll evidence is nearly as strong an argument for a stolen election as TIA appears to believe. I may be mistaken; and I may be mistaken that it matters, but I find it hard to believe that what I honestly find flawed arguments are the best way of making progress in restoring American Democracy. I'd like to be able to point the way to better arguments. I will try. But I will also try not to get in your way.

But I most sincerely wish you well.

Lizzie

[OnTheOtherHand]

131. OK, thanks for laying that out (really)

It's sort of a pain to fight with someone who doesn't lay on the line what he thinks the fight is about. This may help. And apparently you think there is something I'm not laying on the line, so feel free to keep after me (not abusing me, just pressing me) until it's clear.

I don't think my motives are that hard to understand. Like lots of people here, I try to find good information and give good information. Anything that went wrong in the 2004 election, I want to know about, so we can fix it. And this next part may be silly of me, but: I think we will probably do best if we have our facts straight.

Since polling is what I know most about, and since there happen to be (sometimes) a lot of bad polling arguments around the ER forum, it isn't surprising that you have the impression that I am mostly flaming. I have the same impression of you, but I am willing to set it aside at any time. Feuds bore the hell out of me.

"You tell us we are wrong" -- no, I don't think that is true or fair. I try to sort out what is right, what is wrong, and what is unknown. Same as you, I assume.

"and that our evidence is wrong" -- only when I think it is wrong. Do you think it is a bad thing to point out when evidence is wrong, even if it is wrong? Can you imagine why I think it is a good thing to point out when evidence is wrong?

"and when asked simple questions evade answering" -- not at all, at least not intentionally. I have answered dozens of questions from all sorts of people (and people have answered questions for me, too). If you honestly think I have evaded answering some of your questions, it may be because the questions aren't as simple as you think, or because I don't think you have paid much attention to my answers in the past. (Again, this isn't a matter of my being An Expert, it is a matter of simple courtesy.)

"You don't want to discuss this issue. You want to control it. You have big problems with TIA, and others." Whoa, are you suggesting that I am more of a control freak than TIA? Quite a few DUers across the spectrum can testify that it was often hard to say boo to TIA without getting clobbered. I respect him for saying what he thinks, but I don't respect him for making the same mistakes over and over again, and he said a lot of abusive things about people whom I respect. I'm not sure exactly what issue you think I want to control, but I plead innocent.

I am prepared to respect you if you get out of my face about six inches. There are lots of people at DU whom I see doing hard, good work, and although I may disagree with them about some issues, I give them respect and even defend them.

Look, if you disagree with something I write, just explain why. Or if you can't explain why, take that as a constructive challenge to learn something you don't know yet -- it will make you a more effective activist. Or don't take it anywhere; we can just disagree.

As far as I can tell, the feud started somewhere around here:
http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=203&topic_id=395592
I admit I got mad pretty fast -- your "yall" shtick set me off. Also, from my POV, you were defending the statement that "We know Kerry led the pre-election national polls" not because it was true, probably not even because you believed it, but because TIA had said it.

[BeFree]

147. It is a pain

Exit polls should be pure science without politics or business interfering. Mitofsky is a profit making business, so it really has no place looking at elections. As you have so well explained, mitofski is not interested in auditing elections, it merely wants paying customers, and delivers only what the customers want. So, as far as concerns elections, mitofski can go to hell. They need to get as far away from elections as is possible, and do something like find out who buys diapers, that would be more honest for them.

The 2004 election was stolen, and anyone who suggests otherwise without scientifically studying the election machines used to steal it, can go fly a kite. I have no interest in their words except to expose them as the anti-democrats they are.

[BeFree]

129. "No one has died yet"

Is that humor, or do you have a dark side? What the hell do you mean by:" no one has died YET".

Maybe you need help, if you think a discussion on this board might lead to death. That is the sickest thing I've ever read on DU! In fact, I'm calling an alert on you, right now.

[OnTheOtherHand]

135. now, this is strange

I just spotted this. I invite everyone to note the content and time stamps of posts 112, 114, 116, 117, 119, and 126, and then puzzle with me over the timing and rationale of post 129. I didn't think I was that scary. But, no, I don't think anyone is going to kill either of us because of anything we have posted, or whatever other dark meanings you might belatedly read into my post well upthread. My point -- I thought it was obvious, and it seems to have been -- is that disagreeing isn't dangerous.

Actually, I have been accused on DU of covering up for killers, or something like that, which is what I was alluding to in 116: someone might actually believe that if it weren't for 'people like me,' Bush would have been removed by now.

To put it differently, some people may believe that disagreeing is dangerous. And I suppose that is true. It seems like the human condition to me. I don't think this disagreement is especially dangerous.

(EDIT to add last two sentences, hoping to forestall further speculation about my dark side)

[BeFree]

146. Explain, if you can...YET

No one has died...YET. I take that as a threat. You can dance around it all you want, but I take it as a threat.

So explain the YET.

[OnTheOtherHand]

148. no, I don't think you take it as a threat at all

I think you are yanking my chain, again. I still don't know why you are bothering.

OK, you think exit polls should be done not-for-profit. They still wouldn't be infallible. But hey, I don't oppose non-profit exit polls.

[BeFree]

149. You don't think?

I tell you I do, and you don't think? That's a problem, I'd say.

Explain the YET. "No one has died YET."

I could see plain old, no one has died, or anything. But you threw in the YET.

You have connections to people who stand to lose millions of dollars and tons of prestige. It has happened to folks like me in the past.

Explain the YET.

[OnTheOtherHand]

150. reread #135, and stop making stuff up. bye. n/t

[Febble]

113. Well, as someone or other said

I apologise for the long post, but didn't have time to write a short one.

OK, I'll rephrase:

I would be convinced fraud had occurred in a future election if:

Redshift was correlated with:

swing to the Red candidate
precincts with suspect equipment
states with corrupt SoS's
counties with corrupt BoEs
anything that, a priori, would lead us to expect fraud.

Exactly the same principles apply to non-participation bias.
If redshift correlates with any stuff that I would expect to compromise random voter selection, then I would tend to suspect non-random voter selection.

And although I see more of the second than the first, FWIW, I'm not convinced either. That is rather my point.

[BeFree]

115. How could you discover...

... precincts with suspect equipment?
... counties with corrupt BOE's?

All with exit-polls? That's what I am asking.... how do you discover fraud with exit-polls? Oh wait, I know your answer to that... let me put it this way: What signs in exit-polls would convince you that fraud had occured? (in theory)

[Febble]

118. Well, I'm not saying it's easy

But one of the principles of statistical analysis is that you have much more statistical power if you start with an a priori hypothesis, and test that, rather than stare at the data and see what you can see. It's a bit like tealeaves. If you tell me that you've found the image of a giraffe in the tealeaves at the bottom of your cup, and that is amazing, and must be due to a giraffe conspiracy, I won't be very convinced. But if you say: I believe there is a giraffe conspiracy: if so, then I would expect to see a giraffe in the tealeaves at the bottom of my teacup - and you then you do, then I will start to become seriously alarmed about giraffes.

So if you think that the large exit-poll discrepancy was due to fraud, and that the fraud was on an unprecendented scale because of the huge increase in DREs - and if you then show that the discrepancy is significantly larger in DRE precincts, then I will be convinced. But if people just stare at the data, and then say - oh, look, there's a big discrepancy over there, and I've never liked the look of that county's BoE - then I will be harder to convince.

So yes, what would convince me that fraud had occurred is if any a priori hypothesis was supported by a test of that hypothesis in the data. I've thought of quite a few myself, but I am sure I have just scratched the surface. I did think that the swing-shift correlation was a good one, but that didn't show anything. If it had, I would have been VERY convinced, as it is not easy to think of a non-fraud explanation for the discrepancy being greater where Bush's improvement on 2000 was greater. But it wasn't.

Unfortunately, it won't work next time, as we already suspect fraud this time. The advantage in doing it this time was that the discrepancy in 2000 was unusually small.

[Febble]

120. From your responses

to OTOH, I gather that you might be asking what size" of discrepancy I might find suspicious. I would agree with OTOH - the size might be an indicator, but not necessarily the best one. I think 2000 was corrupt and that the wrong man became president, despite the fact that exit poll accuracy was fairly good. In 1992 is it was poor, but I don't particularly suspect fraud, although it is possible. In 2004 I definitely suspected fraud, because of the size of the discrepancy, because I suspected fraud anyway after 2000, and because of DREs. But it doesn't seem to be correlated with DREs or with advantage to Bush. So now I am less convinced.

What would convince me far more than exit poll evidence would be if swing from 2000 was correlated with DREs. Or with some other attribute that we might consider, a priori, to be associated with fraud. The shift could be quite small, but we would have far more statistical power, because the analysis would not be confined to a few handfuls of NEP precincts in each state.

[autorank]

45. I'm like blind Tiresias at some points on threads;) Enjoy. n/t

[bleever]

59. Hey, Tiresias:

Is it true that women are smarter?

Do girls just want to have fun?

Is there more land, or sea?

Inquiring minds want to know.

:hi:

[autorank]

66. Hey, Bleever...

Tiresias here…to answer all of your questions, bleeve it

1) Is it true that women are smarter?



Asked and answered.

2) Do girls just want to have fun?



Some times yes, sometimes no.

3) Is there more land than sea?

Well, yes there is more land than sea but more ocean than land.

Now, one for you…

What walks on four legs in the morning, two in the afternoon, and three in the evening?

:hi:

[bleever]

90. O wise Seer, thank you for favoring me with your wisdom.

By humble chance, I know the answer is Man, who crawls as a child, strides as a child and adult, and who hobbles with a cane in the later years.


Thanks for these threads, bro.

[autorank]

42. They won't answer the charges because 'THEY CAN'T STAND THE TRUTH"

...and we all know what that means.

Talking heads, with their blow dried coif, sitting behind desks or in lounges all day.

Why do you think they don't write stuff like this?
http://www.scoop.co.nz/stories/HL0603/S00016.htm


Because it rocks the boat, it tells the simple truth, and it makes *co look bad.

There are no answers to out sourcing our elections.

There are no answers to "proprietary" touch screens and tabulators.

There are no answers to a lack of public inspection of elections.

There are no answers to hiding key data from the public on the NEPs.

There are just lies and the lying liars who tell them.

[Nederland]

32. Agreed

I personally don't think that clustering is very important in this debate, except to demonstrate that TIA isn't much of a polling guy. The basic problem is that lots of people are making claims about the uncanny accuracy of polling that actual pollsters never make.

I agree and wish more people would realize that. The work done by USCV for instance, was very interesting and authored by people that actually understood statistics. TIA was obviously in over his head and could only manage to produced hundreds of rehashed all caps posts on how the odds of Bush winning were a gazillion to one.

[Peace Patriot]

37. "...wouldn't the public discourse be a teensy bit different?"

The "public discourse" in this country is entirely controlled by the war profiteering corporate news monopolies.

I am reminded of my telling an intelligent friend about who owns and controls the election system now--two rightwing Bushite corporations, using "trade secret," proprietary programming code. She replied, "But the Democrats wouldn't let that happen, would they?"

The issue is not, "what would the Democrats do?" The issue is, "what they DID do." They did let it happen--for whatever reasons of ignorance, corruption, fear or ties to the military-industrial nightmare we're living through.

The issue is not, "if the election was stolen, why is there no 'public discourse' about it?" The issue is, "given the grave non-transparency of the election system and the evidence of fraud, why is no 'public discourse' occurring?"

In the Ukraine, they got to see the exit poll numbers unpolluted by the official tally--and they proceeded to have quite a discourse--hundreds of thousands of people pouring into the streets, demanding a recount. Here, the exit polls were fiddled to match the official tally--an official tally with zero transparency, controlled by Bush partisans. That is one of the reasons that there is no 'public discourse.' The war profiteering corporate news monopolies were complicit in the stolen election. They failed to apprise the American people of the non-transparency, hackability, and partisan control of the election system that the Delay/Ney $4 billion HAVA boondoggle had created. They then doctored their exit polls to match the results of that non-transparent system, thus denying the American people major evidence of fraud. They are not war profiteers for nothing. They are PREVENTING discussion. They are black-holing the story. Because they, too, are guilty--just as they were on the pack of lies that the junta told about Iraq to justify the invasion. A complete and total failure of American journalism, under corporate monopoly control--which has included journalist participation in actual crimes as well as in government disinformation.

Same with much of the Democratic Party leadership. You cannot presume innocence. Not any more. Not given what is happening to this country. They had bad reasons for shutting the hell up, while Diebold and ES&S took over the election system. It's not normal. Among those reasons, fear is probably the most forgivable. Too many of us are thinking in the old paradigm--politics as usual; it will all balance out in the end; the two-party system is a check and balance. Those rules are no longer relevant. Those systems are no longer working. We are suffering a JUNTA, for godssakes! It's as plain and plain can be. And anybody, at this point, who persists in nickpicking about "cluster effects" in one piece of the OVERWHELMING evidentiary picture of a fascist junta is simply irrelevant.

Show me ANY aspect of this regime that NOT fraudulent. Can you point to ANYTHING that gives them legitimacy other than Diebold's and ES&S's secret vote tabulation formulas?

They are operating without the consent of the great majority of Americans--and, indeed, in defiance of the will of the majority--in everything they do, in everything they say, and in everything they fail to remedy.

I suppose you could say that they lied their way to re-election, with the help of the corporate news monopolies. But there is just too much evidence that the American people didn't buy their lies then, and don't now--with voter registration and voter turnout figures in 2004 pointing to a revolt against their lies and maladministration, and polling figures showing long-standing opposition to their disastrous war and other policies.

I think people should start there and work backward. WOULD these people commit fraud? Was the opportunity there? Who created the opportunity? Who controlled that opportunity? And THEN, what is the evidence for a fraudulent count?

Not the other way around--where some DUers begin: That it's not likely that the election was stolen, because the evidence for a fraudulent count can't "prove" a fraudulent count in a court of law, and can't name the specific perps. That evidence--if it exists--has been DELIBERATELY hidden from view, inside SECRET programming codes, written by anonymous programmers inside private, partisan corporations, who had the ability to fix the election with self-erasing, untraceable code, before the machines were even deployed.

But an election is not a legal case, where perps are innocent until proven guilty. An election is just the opposite, in fact. An election must be fully transparent, with all evidence on the table--AND with a presumption of guilt regarding any evidence that is NOT on the table. It should be presumed that people who have the means to steal an election, and are counting votes in secret, WILL use that power to their economic and political advantage, no matter who they are--let alone Bushites, with a history of fraudulent elections, deceptive practices, brutal powermongering, and contempt for the law.

The burden of proof is absolutely NOT on ordinary citizens and voters, in this or any case of non-transparent vote counting.

This is the thing that just drives me nuts! BUSHITE. CORPORATIONS. COUNTED. THE VOTES. WITH. SECRET. PROGRAMMING. CODE. AND. VIRTUALLY. NO. AUDIT/RECOUNT. CONTROLS.

How can anybody NOT get what that means? How can anybody with common sense presume innocence on the part of the Bushites, the war profiteering corporate news monopolies, and an opposition party half of whom voted to hand war powers over to these thieves and criminals?

How can anybody trust what THEY consider to be "public discourse"?

[OnTheOtherHand]

47. sigh

No, the public discourse is not "entirely controlled by the war profiteering corporate news monopolies." I guess we're just about done here.

I will make the offer again: if you can defend the OP, go ahead and do it. If you feel a certain urgent need to change the subject, hey, go wild.

[BeFree]

57. Defending the OP

.... two simple points that destroy any notion that 2004 was
a fairly counted election. Either the uncontaminated National
Exit Poll on election day is correct and Kerry won OR the vote
count and the day-after “adjusted” National Exit Poll is
correct (1:25 p.m. 11/03/04). We know that the voting
process was corrupted and that the count was as well, in all
probability (as evidenced by crazy results like an obscure
judge candidate for the Democrats out polling Kerry by large
margins in 28 Ohio counties).

We were never supposed to see the results from the Election
Day National Exit polls. It was released by mistake and
captured by true heroes of democracy. We were to see the
final NEP of 11/03/04. Why? Because it was ADJUSTED to
reflect the actual vote count.

That FINAL 11/03 adjusted NEP has two huge problems: 1) the
weight for Bush voters from 2000 was 108% of those who were
still alive and 2) the weight for Gore voters still living
was just 92% of those who voted for Gore in 2000. They had
to do a lot of “adjusting” between the final election day NEP
at 12:22 a.m. right after election day and the FINAL 11/03/04
official NEP released at 1:22 p.m. on the 3rd.

Is there a reason that day-after NEP was adjusted to match
the “actual” vote count? Who benefits?

The only reason for adjusting the exit-poll was to hide something. Indeed, why adjust the exit-polls at all? By the time the adjusted exit-polls were published, the official results were widely avaialable, making any further exit-poll numbers totally useless. But altered they were. Why? To hide something is the only logical answer. Hide what?

Well, after recieving $10 million for the work, the exit-pollers had to match the official numbers, otherwise their product would evidently be worthless. Well, they couldn't very well tell the payers of the $10 million "Yall wasted your money", now could they?

So, either they were hiding their incompetence, or they were hiding the true outcome. If someone feels the NEP was incompetent, we can understand. But if they were hiding the real outcome, we don't get it. In fact, they are still hiding the real raw numbers, so, one simply must conclude the reason for all the hiding is.....?

Now, OTOH, don't go into hiding, please, people would like to hear your expert opinions, I'm sure.

[OnTheOtherHand]

60. OK, I will answer it over here, too

Why did newspapers subscribe to the exit poll, and publish results from the exit poll, if the official results would "mak(e) any further exit-poll numbers totally useless"? What? are they just stoopid?

Nope, they publish the results because (as the NEP website puts it) the results "deliver rich details of who voted and why." IOW it's cool to have data about what percentage of young voters voted for Kerry, and stuff like that.

Why weight the results? Because it is kind of pointless to post or print results that show that Kerry won, if Bush actually won. And (if the official results are right), by weighting, they will get closer to the actual results.

Now, you write, "either they were hiding their incompetence, or they were hiding the true outcome." Nope, that's a false choice. They weren't hiding anything, since they were showing members and subscribers their results all along the line. The premise that polls only go wrong when the pollsters are "incompeten(t)" makes no sense.

EDIT: By the way, you didn't actually defend the OP, did you?

[BeFree]

61. So,

The $10 million spent on the exit-polls is justified because:...the results "deliver rich details of who voted and why." And this was worth the $10 million? That's a laugh.

Then you say:Why weight the results? Because it is kind of pointless to post or print results that show that Kerry won, if Bush actually won. LOL. It is pointless to post a result that contradicts the 'official' results? Yeah, pointless, if you want any respect, and another contract. Too, it really is so much easier to go along, to get along. ROFLMAO.

You seem to be saying the polls HAD to match the 'official' or they are pointless. That's rich.

But lets say there was value, and a point to the raw numbers, if you can open your mind that far. Because, as you say: ... since they were showing members and subscribers their results all along the line. Well, the poll results were showing raw numbers all along to the subscibers of the $10 million dollar project, and those numbers showed Kerry winning. But then, as the official results came in, the raw numbers were abandoned, and/or weighted, so that the numbers would all match up, and everyone would say: "Good job, Brownie"?

The OP doesn't need defending. Just reading it is adequate, if you can open yourself up to what it says.

[OnTheOtherHand]

62. what on earth?

"You seem to be saying the polls HAD to match the 'official' or they are pointless."

Funny, I thought that was your position. No, that is an incorrect paraphrase of my statement, and it evinces either careless reading, unwillingness to think, or intent to mislead, as far as I can tell.

(Why do you think newspapers subscribed to the exit polls?)

"the poll results were showing raw numbers all along" -- no, not raw numbers. See, you are apparently unable to open your mind far enough to learn anything, ever, from anything I try to explain to you. It is tedious.

I've been open to exit poll arguments all along. But TIA's are poor, for reasons I have explained, and that you refuse to listen to. OK, your choice. And if you want to believe that all the honest pollsters agree with TIA, hey, enjoy the bubble.

[BeFree]

64. Not raw numbers?

...no, not raw numbers. Then you stop saying anything. You don't explain what numbers were being shown, you just say NO. Then you make a personal attack. Sad.

What were they weighted with then? IOW, if they weren't raw they were weighted, right?

To sum up, just so you can get it straight.... TIA shows that the weights Missofsky placed upon their numbers, after 12am, were bad weights. Weights that he shows are impossible to make mathematical sense out of.

Then, to explain it all away, you say it would be pointless to issue numbers that would contradict the official results.

Then you issue statements that I am wrong but stop right there without any further explanation, just a big fat NO.

Then you personally attack me. That sums it up, for now. Thanks. I got your number.

[OnTheOtherHand]

68. no, what is sad is that we had this whole discussion months ago

and you either pretend not to remember it, or actually don't.

Why don't you review this thread
http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=203&topic_id=397259
and see if you have any serious questions?

Short: the projections and tabulations are always weighted; different weights are applied at different times.

TIA may find it impossible to make mathematical sense out of the weights, but if you are not prepared to explain why yourself, then you have no rational basis for agreeing with him, and we have nothing to argue about. Welcome to the real world. Some may find your certainty impressive in itself, but I do not. And you will have no trouble finding threads in which I explain why TIA's arguments are unconvincing.

[Usrename]

82. The reason it doesn't make mathematical sense is due to reality.

Bush only had a certain amount of voters in 2000. That's on the record and a historical fact. The NEP says that he had 4 million more than can possibly still be alive.

TIA doesn't make this claim - the NEP data from E/M makes this claim.

Get it? The NEP says this is true - TIA just points out the fact that it's a pantload. Right?

Do you think the TIA is mistaken for pointing out this fact? Is it not a fact?

Just what do you think the NEP says about the 2000 Bush voters? Obviously, according to you, it must be some other thing that they are saying, since you continue to insist that TIA's interpretaion of what they are saying is so completely and utterly wrong.

But it sure looks like they are saying that more people who voted for Bush in 2000 returned to vote for him 2004 than could possibly be alive. Simple enough.

But just WHY do you think, and keep insisting, that the NEP does not say what TIA says it does?

[Febble]

84. Listen, the weightings are fiction.

They are guesswork. They are an attempt to "correct" the responses for selection bias.

Now, on the one hand, as you, and TIA would claim, the reason the "correction" was required was that the count was corrupt.

On the other hand, as the pollsters no doubt assumed, rightly or wrongly, that the reason the "correction" was required was that the poll was biased.

If I conduct an experiment, and I find I have missing data (which happens very frequently), I have a number of alternatives. My statistical software actually lists several of them, and sometimes I choose one of their options, sometimes I do something else (provided I can justify it to a reviewer). One common method, for example, is to substitute the mean value for the whole sample for the missing data point. This does not mean that my data is now "correct"; it simply means that I can do sensible statistics on it, although at a cost of extra noise.

Similarly (though not identically), if, after the data is in, the pollsters discover it doesn't match the vote, and if, as they do, they assume that the count is correct (and they are, being pollsters, perfectly well aware that selection bias is highly possible) they attempt to "correct" their data for what they assume are missing datapoints.

But their correction doesn't meant they have now got every cell correct. It simply means that, crudely, the whole matrix has been warped into something that globally matches the count, although it may differ from actuality in detail. It is, if you like, a bodged retrofit to the data.

Now I do understand that you find this an outrageous proceeding - that you think that the exit polls should have been presented uncontaminated by the count as a check on it. But GIVEN that (see above) this is NOT what the exit polls were designed to do, and that the assumption of the pollsters was that the vote-count would be more accurate than their data, then this warping of the matrix to fit is, from their vantage, quite reasonable. They know that they had some non-response bias by visible characteristics (from the age/race/sex data on non-respondents); they now assume they also had some non-response bias by invisible characterestics (candidate preference). But all they know is the crude vote totals - They assume, broadly, that there are Bush voters missing from their sample, but they don't know who, precisely, those non-respondents were. If they were disproportionately Gore-Bush voters, then that would give implausible totals in the cross-tabs. And that is before we even consider the probability that recalled vote responses over-state the previous vote for the incumbent.

So again, all I am saying is: the thing isn't a slam dunk; it isn't an assault on "historical fact". We are talking about a survey, and surveys are prone to selection bias.

It is perfectly possible that the apparent selection bias in 2004 was not selection bias, but fraud. But there is no reason, in principle, to suppose that it was not selection bias, which is an extremely well-documented and researched polling phenomenon. Which is why I continually say that if you want good evidence for a stolen election you need to look elsewhere, particularly at forms of corruption that wouldn't even show up in the exit polls, like voter suppression.

[OnTheOtherHand]

85. OK, let me review

Data do not make claims.

The NEP results imply that if all voters had been asked, 43% would have said they voted for Bush in 2000, and 37% that they voted for Gore. 43% of all voters is indeed more voters than possibly could have voted for Bush in 2000.

I argue that some people say they voted for Bush when they actually voted for Gore. In fact, I argue that people often say they voted for the incumbent when they didn't. And I have offered several pretty strong (IMHO) lines of evidence.

- I've pointed to a bunch of surveys (mostly General Social Surveys) in which people exaggerate having voted for the incumbent.

- I pointed to a panel study where (depending on the weights applied) somewhere between 12% and 14% of the people who said in 2000 that they had voted for Gore, said in 2004 that they had voted for Bush.

- I pointed to the 2000 exit poll results, in which 46% of (weighted or unweighted) respondents said they had voted for Clinton, which would make over 48 million, although Clinton only got 47.4 million votes, of whom by TIA's method I think maybe 45.7 million would have survived to 2000. So, if we conclude that Bush stole millions of votes in 2004, how can we not conclude that Gore stole millions of votes in 2000?

- Then I started going back through other exit polls to show how they overstated the incumbent's vote share, too.

And TIA has no answer. AFAICS, to any of this. That is why I think his interpretation is completely and utterly wrong -- because he really doesn't seem to have one.

(EDIT to remove needless s. likely to be interpreted as snark)

(EDIT 2: Febble is also quite right to point out that there is no reason to take the weights literally -- one can't say, "If the weights don't fit, one can't acquit." But if the weighted results made less sense, I would be more worried. The 43/37 recalled vote result actually does make sense.)

[Usrename]

87. Please be more specific here OTOH

Are you saying that the same phenomena has been observed with regard to the 2000 exit poll results?

"- I pointed to the 2000 exit poll results, in which 46% of (weighted or unweighted) respondents said they had voted for Clinton, which would make over 48 million, although Clinton only got 47.4 million votes, of whom by TIA's method I think maybe 45.7 million would have survived to 2000. So, if we conclude that Bush stole millions of votes in 2004, how can we not conclude that Gore stole millions of votes in 2000? "

That is, that the unweighted results rest firmly in the real world, yet when weighted appropriately, those same data miraculously create an extra 5-6-7-million votes for one or the other candidates? I don't think so. At least I hope that is not what you are saying. The VNS data was highly accurate, I thought. It correctly predicted the missing votes in Broward, Duval and Palm Beach Counties in FL.

I hope that you are not saying that their weights "changed" the outcome in 2000. They called it rightly for Gore. So what's your point?

[OnTheOtherHand]

89. happy to be more specific

Here you can see rounded versions of the weighted 2000 exit polls, which show 46% of voters reporting that they had voted for Clinton in 1996 and 31% reporting that they had voted for Dole. Strange, since Clinton actually won by 8.5 points, not 15 (rather as Bush actually lost the popular vote, instead of winning it by 6 points). (Actually, all these discrepancies are somewhat wider, allowing for the people who report not having voted in the prior election.)

Inspection of the dataset archived with ICPSR shows that Clinton retrospectively won (so to speak) among 2000 voters by 45.6 to 31.3 if the weights are applied, or 45.8 to 30.5 in the raw data.

So, the 2000 data stand firmly outside the real world whether they are weighted or not -- that is, they indicate more people having voted for Clinton in 1996 than actually could have (although we could try to salvage the situation by making rather bizarre assumptions about differential turnout).

However, I argue that this result should actually not be surprising, as it jibes with all the other cases where votes for the incumbents are exaggerated.

Incidentally, your statement about the 2000 exit poll in Florida is dubious at best. What is the quantity of missing votes in Broward, Duval, and Palm Beach Counties that you think the exit poll "correctly predicted" -- and how could an exit poll predict missing votes in three particular counties, anyway? An exit poll of some 1800 voters in 43 precincts cannot reliably "predict" shifts of hundreds or thousands of votes out of over six million cast.

(EDIT to fix broken link)

[Usrename]

95. So are you saying that the weights used in 2000,

by VNS did or didn't change their results? Like what the OP is talking about? How is what you are talking about even relevant? Like you say, the data from 2000 shows the same thing, with or without weights. Are you arguing that this is also the case in 2004? Just what is your problem with the OP. It is accurate, no? You seem to be grasping at straws here. The VNS exit polls were just fine, IMHO.

And as to your question about VNS projecting the vote correctly, they did, I saw it. After FL was "uncalled" a VNS guy was on the air explaining why it was uncalled. It was well documented. There were illegal ballots that led to a lot of spoiled ballots. The spoiled ballots were real. The will of the people was thwarted. You must be aware of this. From memory it was something like 30-35k from Broward/Palm Beach and another 15-25k in Duval.

In total, I think it was 50k votes. Enough to determine the winner, which is why they called FL for Gore in the first place. And yes, no one denies the double-punched ballots really happened. It's an old trick. One that I first learned about back in the days of the Voting Rights Act. It's the reason butterfly ballots were illegal to begin with. If the ballots don't confuse the voters enough to make the proper amount of mistakes necessary for a win, then simply double-punch a bunch of them and blame it on voter confusion. It has always been a really popular in predominately black precincts. They used to show a great movie on it back when I was in grammar school. I've seen it two or three times, if not more than that. One of the many truths about the history of voting in the South.

In 2000, one of the black precincts in Duval had one out of every five voters voting twice for President. One in five. Black voters. Voted twice. I can't believe you don't know about this. Of course I think that VNS called it correctly. Why wouldn't I? And if you think that's dubious, well then at least I know who I'm talking to. Wow. You sound sincere, but if you don't understand anything about Bush v. Gore and what really happened in FL and why and how, then I don't know where to begin.

Maybe I don't understand why you are bringing up VNS. Is there also something similar that is wrong with their stuff. If so, I don't know what it is. It passes my smell test. Especially in light of everything that I know about the recount effort in FL. The OP refers to a different animal completely. Do you really not comprehend the difference, or are you practicing at being intentionally obtuse. Not trying to offend with this question. I just won't waste time responding if you are playing at the latter. It is just really difficult for me to believe that you don't understand the big picture here. The OP is correct. This reeks. Why is that so difficult to admit?

[OnTheOtherHand]

98. very perplexing

Yeah, I really am sincere -- I will say that up front. Let's try this a few more times and see if we can get on the same page.

Check my reasoning. The OP argues that we know the 2004 weights must be wrong because they lead to an impossible percentage of Bush-2000 voters. I point out that the 2000 exit poll, weighted or unweighted, offers an impossible percentage of Clinton-1996 voters. So if you think that the "impossible" in the 2004 exit poll proves fraud, how do you explain the "impossible" in the 2000 exit poll?

Please read that slowly, and then reread my last few posts again. We seem to be talking past each other pretty well.

I should've known that the Florida thing would get too picky. Umm, let's assume that Gore won among votes actually cast by some fraction of a percentage point (I think 50K is quite possible -- almost one point -- although I'm not sure what proportion of those votes even the most meticulous recount could have altered without statistical inference). But if the exit poll had initially shown Gore ahead by one point, VNS and the networks never would have called the state for Gore within the first hour or so. The sample simply isn't big enough to support that sort of confidence, even if everything else was perfect. Rather, the initial model estimates gave Gore a larger lead (over 6 points), and that helped bring about the initial call.

I think you are right that not much was wrong with the 2000 Florida exit poll itself. There were some problems with the model, and they way underestimated the absentee vote (and of course the exit poll doesn't shed any light on how corrupt the absentee vote might have been). The correct call from the exit poll would have been no call, or too close to call. That is statistical reasoning, and has nothing to do with whether Gore should have won Florida. You and I agree that he should have. The overvotes in Duval were huge in the election, but would have been less than a blip in the exit poll. The VNS results are compatible with a Bush margin of 500 votes, or a Gore margin of 50,000 votes, or lots of other outcomes. It is political, not statistical, to say that the exit polls rightly show that Gore won. However, other statistical analysis does support that inference.

Now, back to 2004. If the only important argument you find in the OP is that "this reeks," then we don't need TIA, do we? But if you think that 43/37 is strong evidence that Kerry won the popular vote, then a couple of posts up I offered at least four reasons why I don't think so.

I think there is a big difference between discussing the proposition that 43/37 proves fraud, and discussing the proposition that this reeks. When the OP claims that 43/37 proves fraud, and I explain why it doesn't, and I am accused of denying that this reeks, it seems like a bait-and-switch.

[BeFree]

99. Let's be clear

The 43/37 doesn't prove election fraud, the exit poll raw numbers prove the fraud.

What the 43/37 proves is that Mitofsky cooked their numbers. What the OP is saying is that the raw numbers prove fraud, and Mitofsky tried to hide the evidence by cooking the numbers.

Try to stick with the facts. The raw, unweighted numbers, before being cooked, show evidence of fraud. Mitofski is accused only of trying to hide the evidence.

But who can blame them? Why would they want to go out on a limb to save democracy? It ain't like they owe us anything... they are not soldiers, or unpaid activists fighting for America, they do it for the money, and sticking with the evidence would cause them to lose money. Who can blame them for taking the easy way out?

[OnTheOtherHand]

100. you ignored the content of my post n/t

[BeFree]

102. My we are testy, eh?

Your attempts at inciting me into personally attacking you will not work. It has worked on numerous others, and I have watched and learned.

You claim to be the owner of the field, but yet your personal attacks and one liner responses like the above show me that you have an ulterior motive when it comes to this issue. That's fine. This is an open board and all views are welcome.

What's really great about it is when dumbasses like me get to probe the upper eschelons of the field. Its all, believe me, in an attempt at becoming educated.

Thanks for your time, talent, and prowess. You have convinced me that we are on the right track.

[OnTheOtherHand]

107. you personally attack me in practically every post n/t

[Febble]

101. What proves that Mitofsky "cooked" the numbers

is this:

http://www.exit-poll.net/faq.html

How are projections made?
Projections are based on models that use votes from three (3) different sources -- exit poll interviews with voters, vote returns as reported by election officials from the sample precincts, and tabulations of votes by county. The models make estimates from all these vote reports. The models also indicate the likely error in the estimates. The best model estimate may be used to make a projection if it passes a series of tests.

(My bold)

and this:

ftp://ftp.icpsr.umich.edu/pub/FastTrack/General_Election_Exit_Polls2004/

Go to "datasets", then "National". You will need some kind of reader to read the files, but here are the columns you will find:

Respondent ID
State ID
Sample precinct number
State or National precinct
Questionnaire version number
Call number
Backside completion flag
Absentee poll/Election Day respondent
Respondent weight
In today's election for president, did you just vote for:
In today's election for U.S. House of Representatives, did you just vote for:
Are you:
Are you:
To which age group do you belong?
No matter how you voted today, do you usually think of yourself as a:
On most political matters, do you consider yourself:
Is this the first time you have ever voted?
2003 total family income:
Compared to four years ago, is your family's financial situation:
Are you:
How often do you attend religious services?
Would you describe yourself as a born-again or evangelical Christian?
What was the last grade of school you completed?
Do you or does someone in your household belong to a labor union?
Do you work full-time for pay?
Have you or has someone in your household lost a job in the last four years?
Have you ever served in the U.S. military?
Do you or does someone else in your household own a gun?
Are you currently married?
Do you have any children under 18 living in your household?
What type of telephone service is there in your home that you could use or be reached on?
Are you gay, lesbian or bisexual?
Are you of Hispanic or Latino descent?
Are you of Mexican or Hispanic descent?
Are you of Cuban or Hispanic descent?
When did you finally decide for whom to vote in the presidential election?
Was your vote for president mainly:
Which ONE issue mattered most in deciding how you voted for president?
Which ONE candidate quality mattered most in deciding how you voted for president?
Do you approve or disapprove of the way George W. Bush is handling his job as president?
How do you feel about the U.S. decision to go to war with Iraq?
Compared to four years ago, is the country:
Who would you trust to handle terrorism:
Who would you trust to handle the economy?
How do you think things are going for the U.S. in Iraq now?
How worried are you that there will be another major terrorist attack in the U.S.?
Compared to four years ago, is the job situation in your area:
Do you think the condition of the nation's economy is:
How confident are you that votes in your state will be counted accurately?
Did anyone call you or talk to you in person on behalf of either major presidential campaign about coming out to vote?
Did either of these candidates for president attack the other unfairly?
Which comes closest to your position? Abortion should be:
In general, does George W. Bush pay more attention to the interests of:
In your vote for president today, how would you rate the importance of the Osama bin Laden video?
Which comes closest to your feelings about the Bush administration:
Is your opinion of George W. Bush
Is your opinion of John Kerry
In general, does John Kerry mostly:
Which comes closest to your view of gay and lesbian couples:
Which comes closer to your view:
How concerned are you about the availability and cost of health care?
Do you think the war with Iraq has improved the long-term security of the United States?
Do you consider the war in Iraq:
Do you think the Bush administration's tax cuts have been:
Do you think things in this country today are:
If these were the only two presidential candidates on the ballot today, for whom would you have voted?
Did you vote in the presidential election in 2000?
Sex by race
Age 60 (four categories)
Age 65 (six categories)
Age 65 and over
No matter how you voted today, do you usually think of yourself as:
2003 total family income:
2003 total family income:
Are you:
Religion among whites
White protestant conservatives
White born-again or evangelical Christians
How often do you attend religious services?
Church attendance by religion
White/Protestant or other Christian/Evangelical/Conservative/Attend church at least weekly
What was the last grade of school you completed?
Do you approve or disapprove of the way George W. Bush is handling his job as president?
When did you finally decide for whom to vote in the presidential election?
When did you finally decide for whom to vote in the presidential election?
How do you feel about the U.S. decision to go to war with Iraq?
How do you think things are going for the U.S. in Iraq now?
How worried are you that there will be another major terrorist attack in the U.S.?
How confident are you that votes in your state will be counted accurately?
In your vote for president today, how would you rate the importance of the Osama bin Laden video?
Do you think the condition of the nation's economy is:
Who would you trust to handle the economy?
Who would you trust to handle the economy?
Who would you trust to handle terrorism:
Who would you trust to handle terrorism:
Did anyone call you or talk to you in person on behalf of either major presidential campaign about coming out to vote?
Did anyone call you or talk to you in person on behalf of either major presidential campaign about coming out to vote?
Did either of these candidates for president attack the other unfairly?
Did either of these candidates for president attack the other unfairly?
Do you or does someone in your household belong to a labor union?
Do you or does someone in your household belong to a labor union?
Married women
Mothers
Married with children
Working women
Have you or has someone in your household lost a job in the last four years?
Have you or has someone in your household lost a job in the last four years?
Unadjusted race
Unadjusted race (AZ, MT, OK)
Unadjusted race (AK)
Language of questionnaire
National region
Size of place code
Size of place code (3 categories)
Congressional district number

(My bold)

Note the column marked "respondent weight". This is the "cookery". If you want to analyse the numbers, you can either use the weights or leave them off. The "evidence" that "Mitofsky cooked the numbers" isn't hidden. It was announced before the election, and the actual weights, together with the raw data, were provided, last January, as they always are.

You may find, when you analyse the raw numbers, that they are evidence of fraud. But that is up to you. But it isn't hidden.

[BeFree]

103. You are missing the point

The point being that after the polls were closed, and well after the exit-polling numbers were in the can, the raw numbers were cooked to match the official numbers. The graphs, the arguments by TIA, all show the cooking.

Interesting that you say the numbers were established using official results. Are you saying the pollers were privy to official results BEFORE the polls were closed? Cause that sure seems to be what you are saying. Quote:


Projections are based on models that use votes from three (3) different sources -- exit poll interviews with voters, vote returns as reported by election officials from the sample precincts, and tabulations of votes by county. The models make estimates from all these vote reports. The models also indicate the likely error in the estimates. The best model estimate may be used to make a projection if it passes a series of tests.

From your quote, it seems to me that the projectors are privy to vote returns before the polls closed. Is that true?

Now, it seems projections are just that: meaning before. Therefore the projection period was OVER, therefore it need no longer be projected. Are you with me so far?

So if the projection period was over, say 12am when ALL the polls were closed, then the job was done? At that point in time did the projectors turn into analyzers? Still with me?

Ok. So the projectors had projected that Kerry won the vote. But then the analyzers came into play. And play they did. The graphs and almost all independent analyzations of the numbers show the projectors-turned-analyzers cooked the numbers. It is so simple. Most everyone can see it. Including a dumbass like me? Why can't you?

[Febble]

104. Well, I think you may be....

The point being that after the polls were closed, and well after the exit-polling numbers were in the can, the raw numbers were cooked to match the official numbers. The graphs, the arguments by TIA, all show the cooking.

Interesting that you say the numbers were established using official results. Are you saying the pollers were privy to official results BEFORE the polls were closed? Cause that sure seems to be what you are saying.

....

From your quote, it seems to me that the projectors are privy to vote returns before the polls closed. Is that true?

Yes: the cooking starts when the first vote returns start coming in (Eastern Zone) but while poll results are also still be coming in from more western time zones. As I've just demonstrated, not only do TIA's arguments show the cooking, the E-M FAQ announced the recipe. I read it on election day.

Now, it seems projections are just that: meaning before. Therefore the projection period was OVER, therefore it need no longer be projected. Are you with me so far?

Well, we need to sort out the time zone business. TIA is talking about the National Poll, which covers all states (it is a subset of the state poll sample of precincts), and no projection is made from that. It is used for the crosstabs. But because it covers the nation, some vote-counts will be incorporated from the east while poll responses are still coming in from the west. However, for the state "projections" these need to reach a critical value of probability of being "correct" (i.e. in this context, in line with the counted result) before a state is "called". The closer the state, the more vote-count data will need to be incorporated before the t value reaches threshold.

So if the projection period was over, say 12am when ALL the polls were closed, then the job was done? At that point in time did the projectors turn into analyzers? Still with me?

No, not with you, because we parted company a while back. Are we now talking about the National poll or the state polls? When the polls close for a state, all the responses should be in (although there may be some delays), and the vote-returns start coming in. So between the estimates made at close-of-poll (from the poll respondents) and the t value reaching criticality, more and more vote-count evidence will contribute to the t value. As I said, the closer the state, the more vote-count data will be required before the t value reaches criticality.

However, for the National poll, vote-counts from the Eastern time zone will contribute to the crosstabs before poll counts from the west have finished coming in. So TIA's timeline reflects not only the addition of the last group of voters respondents, but also reweighting from the incoming returns. That is why it made that big jump (not because late, Western poll responders were unanimously Bush supporters, or alleged to be).

Ok. So the projectors had projected that Kerry won the vote.

No. Again, I'm not sure whether you are talking about the national or the state polls. If you are talking about the national poll, when the first polls were closing, it indicated Kerry was ahead, as we know. That's when I opened my bottle of wine.

But then the analyzers came into play.

They were playing all the time. It's a dynamic process. The National Poll data showing Kerry ahead probably incorporated some eastern vote returns. The later screenshot incorporated more.

And play they did. The graphs and almost all independent analyzations of the numbers show the projectors-turned-analyzers cooked the numbers. It is so simple. Most everyone can see it. Including a dumbass like me? Why can't you?

Not only do I see it, I've given you the link to the recipe, and to the precise ingredients, the irony being that both are directly to the E-M website and data. Hidden in plain sight.



(Edited to add a closing bracket)
(Edited again to change a word for clarity)

[BeFree]

105. Ok

Here's what I see....

We are talking about the national polls, because that's what TIA's and the OP is all about.

The key here is that with just a few hundred added respondents from the time frame of 11pm to 12am, the numbers were cooked. That's pretty clear.

With all number assessments trend lines are to be respected, and the trend lines clearly show Kerry with a winning margin, until the 11pm to 12am time frame. And from TIA's analysis, we see when that trend line was ruined.

The contention is that the trend line was ruined/altered because of the corrupted machine counts - once having been looked at - made the pollers change their numbers to match the official count. To even things out. And this was done after over 90% of the numbers were in. The fantastic alteration of the trend line is evidence of nothing more than an attempt to bring the poll numbers in line with the official returns.

Try to stick to just that in your response, please.

[Febble]

108. I am more than happy

to stick to the National Poll, but in that case we have to consider the time zones. The eastern time zones would be generating vote-returns while the western time zones were still generating poll responses.

So there isn't a clear "before" and "after". The poll responses were phoned in throughout the day, the polls closed at different times in different states and time zones, and the vote count was not instantaneous, despite your crazy machine counting methods.

Now, it is true that that FAQ paragraph refers to state projections rather than the National Poll (although the National Poll is simply a sample of the state NEP precincts) but there is a specific methodology statement here:

http://www.exit-poll.net/election-night/MethodsStatementNationalFinal.pdf

for the National Poll, which states:

For the national tabulations used to analyze an election, respondents are weighted based upon
two factors. They are: (1) the probability of selection of the precinct and the respondent within
the precinct; (2) by the size and distribution of the best estimate of the vote within geographic
subregions of the nation. The second step produces consistent estimates at the time of the
tabulation whether from the tabulations or an estimating model used to make an estimate of the
national popular vote. At other times the estimated national popular vote may differ somewhat
from the national tabulations.

Number 2 is the cookery you are referring to.

The "trend line" you refer to does indeed reflect the fact that the early estimates reflected mostly poll responses, while the later estimates included vote returns. Quite clearly, the crossing of the trendlines does NOT reflect the addition of a few extra respondents, as I said in my previous posts. It is, in fact, mathematically impossible, and absolutely no-one is claiming, or has claimed, to my knowledge (because it would be daft) that it was. It reflects the incorporation of new data - the data from the vote-returns.

And the method used is this:

The initial poll responses are weighted according to data on non-respondents (those selected to be in the poll, but who refused, or were "missed" because the interviewer was busy with another voter). The data collected is necessarily on visible characteristics: age, race and sex. Invisible characteristics (like who the non-respondent voted for) cannot, of course, be collected. This data is applied as a "weight" to the responses in the dataset. So if men are under-represented in the poll (a larger proportion of men in the non-respondent category than in the respondent category), male respondents are given a "weight" greater than 1. Let us suppose the weight is 1.5. What that means is that every man in the poll is treated as though he were 1.5 men, to make up for the missing men. Same applies to age and race - a weight is applied for each characteristic. When the vote-returns start to come in, and it becomes apparent to the pollsters (not to you, because you don't trust the count, but bear with me) that not only were men under-represented, and maybe 50-60 year olds under-represented, but that Bush voters were under-represented, then each Bush voter in the poll is also given a weight greater than 1 (of course the weight is incorporated into the weight that voter already has from the age;race;sex weighting). So the initial weights give one answer (Kerry winning); when the Bush weights are applied, they show Bush winning. Because he is - winning the count, that is.

But my point is that this was not a post hoc, unanticipated cover-up, it was built into the poll-design, in order to compensate for non-response bias by invisible characteristics (presidential vote), and specified quite clearly in advance (although the networks seem to have made a poor job of conveying that this was what would be done). What I cannot accept that anyone was trying to "cover up" the "cookery" when the "cookery" was actually made explicit in advance on a public website, nor can I accept that the cookery is hidden when the actual weights are in the public domain, and can be downloaded by you, together with answers to all the the questionnaires from every respondent in the National Poll.

Now, all this raises the important question, the one asked by you, by TIA, and indeed by me, as to whether the polls were indeed more biased than usual, or whether, alternatively, it was the count that was wrong. And that is what we are trying to figure out, and what the OP is about. TIA takes the logical view that it must have been the count, as the reweights resulted in internal inconsistencies in the cross-tabs. I disagree with him on those particular crosstabs (Gore:Bush) for reasons I've given elsewhere in this thread. But in principle, if the re-weighted crosstabs look to be internally inconsistent, it would push the probability that the discrepancy was due to fraud rather higher.

But the fact that the reweighting took place at all is not, per se, evidence of fraud, nor of a cover up, because non-response is expected (and is why the polls are always reweighted) and because we were told in advance (if we looked in the right place) that this would happen, and because the actual weights are available for us to study, together with all the questionnaire responses.

They are really interesting. If you want to look at them, and can't open the files because you don't have appropriate software, let me know by PM, and I will send you a version something like an Excel file, if you like.

[Usrename]

124. Now you seem to be telling me not to believe my own eyes!

I told you I saw(!) the VNS guy on election night talking about the missing votes. This was an hour or so after FL was uncalled for Gore. According to what you are saying in your post, he had no way of knowing what he knew. Was he just a good guesser?

I do understand what you are saying about 2000 voters false recall. If you really believed that argument then I would think you would take it to it's logical conclusion. Assume that party has no effect on this false recall thing. Then fine, lets add another million, million-and-a-half false recal votes into the mix. Hell, lets say the Gore voters are twice as likely to have false recall. It still doesn't mean anything. So why work so hard, wriggling around to find a dubious explanation, when there is a far more acceptable one?

And the 6 points remark. I'm pretty sure who the winner would be with a 6 point lead. Why bother with an election in that case at all, not to mention exit polls? I'm not quite sure that you have a good handle on the precision thing. I don't think Gore ever had that big a lead in FL. As a matter of fact, I think he was trailing up until the last week or so. It was pretty much within the MOE for months, the way I remember it.

[OnTheOtherHand]

134. nope, not trying to question your eyes

"I told you I saw(!) the VNS guy on election night talking about the missing votes. This was an hour or so after FL was uncalled for Gore."

OK, if we are going to talk about Florida 2000 (which I don't think is central to the "big argument" here, but there is no reason not to talk about it), it might help to have this document as a common reference:
http://archives.cnn.com/2001/ALLPOLITICS/stories/02/02/cnn.report/cnn.pdf
Obviously I can't vouch that it is all accurate, although I haven't noticed any glaring errors.

I will guess that what you mean by "the missing votes" is alluded to on numbered page 13 (PDF page 15) of the doc at 9:07 and 9:20 -- a keypunch error that made it appear that Gore had 98% of the vote in the counted precincts in Duval County. I have no reason to assume that that error didn't happen just as described here, or that it had anything to do with the actual voting problems in Duval.

Later in the night, there was the 'opposite' problem in Volusia (2:08 am, ff.).

Does that help? I can't tell yet whether we are talking about the same things.

---

I do understand what you are saying about 2000 voters false recall. If you really believed that argument then I would think you would take it to it's logical conclusion. Assume that party has no effect on this false recall thing. Then fine, lets add another million, million-and-a-half false recal votes into the mix. Hell, lets say the Gore voters are twice as likely to have false recall. It still doesn't mean anything. So why work so hard, wriggling around to find a dubious explanation, when there is a far more acceptable one?

Sorry, I can't actually tell what you are trying to say here. False recall actually does matter, and ironically, it tends (rather weakly) to support the view that Bush won by a few points. Based on other exit polls (and other non-exit polls), I expect the incumbent to fare several points better in recalled vote than he actually did. In the results that TIA regards as the true exit poll results, Bush has only a two-point retrospective margin over Gore. I've looked at results from ten other exit polls in eight different elections, and that would be a smaller bump than any of them. The final-weighted 2004 results for 2000 vote actually make more sense (compared with those other exit poll results) than the earlier results. I didn't know it would turn out that way, I just looked. And TIA didn't. And, apparently, TIA won't. Is he afraid of something?

I say "rather weakly" because the amount of the bump varies widely, and the question wording keeps changing, and so on. But if Richard Nixon did four points better retrospectively in the 1976 exit poll than he actually did in 1972 (probably not due to massive Republican turnout!), then for Bush to only do two points better would definitely be weird.

---

I can't tell why you think I am "wriggling around to find a dubious explanation, when there is a far more acceptable one" -- explanation of what? If you are talking about the Bush and Gore percentages in the 2004 exit poll, I have a perfectly straightforward explanation: it jibes with every other exit poll, and every other National Election Study, and practically every other election I have ever looked at. That isn't wriggling around, it is connecting the dots.

If you mean that non-response bias is "wriggling around," I don't agree with that either; there is lots of evidence of non-response bias in past exit polls, and as I've pointed out, there are results from the 2004 exit polls that do not jibe with fraud and do jibe with non-response bias. I am not one of the people (I'm not talking about you!) looking at the scatterplots that show Bush not doing better in red-shift precincts and bellowing, 'That means NOthing, NOthing!' I mean, one can make it sound simple to explain the evidence if one insists on ignoring most of it; but if one tries to look at all of it, then things get more complicated. That isn't my fault, that is life.

---

The 6 point business gets pretty technical. The link above might help. I wrote that (in Florida) "the initial model estimates gave Gore a larger lead (over 6 points)...." That doesn't mean that Gore had a 6-point lead in the pre-election polls -- of course he didn't. It means that when the exit poll folks took their interview results, and compared them to various past races, and made some assumptions about the absentee vote, and so on, their initial best guess was that Gore was ahead by 6.

[Usrename]

138. hehe.. I can’t make it through the preamble.. that link you gave..

holy pantload, OTOH!!!

This thing had tears running down my face it’s so hilarious!

I’ve highlighted the only true phrase in bold for you to ponder!

”On Election Day 2000, television news organizations staged a collective drag race
on the crowded highway of democracy, recklessly endangering the electoral process, the
political life of the country, and their own credibility, all for reasons that may be
conceptually flawed and commercially questionable.

Their excessive speed, combined with an overconfidence in experts and a reliance
on increasingly dubious polls, produced a powerful collision between the public interest
and the private competitive interests of the television news operations and the
corporations that own them.

Their hyper-competition stemmed from a foolish attempt to beat their rivals to the
finish line in calling state-by-state winners in the presidential election, foolish because
few in the crowd knew then or know now which network got the checkered flag most
often. Foolish because each network funded its competitor’s work. Foolish, too, because
their haste led to two mistaken calls in the state that turned out to hold the key to the
outcome of the election. All, in turn, played an important role in creating the ensuing
climate of rancor and bitterness."

<snip>

That phrase in bold, when taken by itself, does speak volumes. Yeah, that part really did happen.

But this last sentence is just so over the top I almost soiled my pants when I read it.

"Those calls and their retractions constituted a news disaster that damaged
democracy and journalism.”

Yeah, sure.

Like the Supreme Court of the United States ruling that it is unconstitutional to count legally cast votes or the media’s failure to explain this fact in no way damaged democracy and journalism.

Thanks for the hoot. I’ll try to look at the rest of it if I can recover from the fits of hysteria. No joke. That’s a hoot.

[OnTheOtherHand]

139. glad to have brought joy

although I admit that I don't quite get it. You really don't think it was weird to call a state that was ultimately decided by less than a percentage point (indeed, officially, by about 0.01%) before the polls had even closed in the panhandle? and then to uncall it? and then to call it the other way, and then to uncall it again? Sounds like a mess to me. Heck, someone might even write a report about it.

It seems to me that, perfectly legitimately, you aren't really very interested in what the networks did on election night. You are mostly interested in who should have won. Thus, the initial call for Gore was Cosmically Right, and we can skip everything else and cut to how the media botched the SCOTUS story.

Well, thank goodness some folks wrote about that, too.

[Usrename]

140. My bad, I guess.

I thought the exit polls were supposed to capture the vote, among other things in the demographic data. Take a snapshot, so to speak, of the people who went to the polls to vote.

Wrong theory, the exit polls are supposed to take into account, somehow, how the media will respond and haw the courts will rule. Hence the 0.01% win. How can this possibly be relevant? They should have known how the official vote would be counted? Why is this of any concern to pollsters? Why all the spin? How did the tabulation stuff affect the snapshot? It didn't. I really do have quite a different grasp on reality.

It is such a simple concept to me. How, and when, the calls that were made, could have made an impact on the exit poll, is a real mystery to me. Or even that it should have made some kind of impact. Somehow, through some mechanism unknown to me, things in the exit poll, unknown things, change because the media reports on it. It's like a miracle or something.

How does that work exacly?

[OnTheOtherHand]

144. whoops, whoa, the exit polls have at least two purposes

Certainly they are supposed to take a snapshot of the people who voted. The tabulations provide rich details of who voted and why, or whatever that language is.

The interviews plus the quick counts plus the county totals -- but really mostly the latter two, now -- are used to make the projections. The report on Television Coverage on Election Night 2000 (or whatever it is called! I am writing fast) is about using the projections to call a winner. The projections of course take account of what votes are actually counted. If they didn't, then there would be no way to update the projections once the exit poll interviews were complete, if you see what I mean.

Maybe that is the problem -- I don't think I even explicitly explained that in my PM, and I don't know if you know it, but at any rate most people probably don't. The projection models replace the exit poll estimates with quick count figures as those become available. In fact (at least in 2004), the projection models used exit poll interviews from multiple precincts, compared those results with historical returns from just one precinct at the polling place, and this was deemed a reasonable short-term compromise precisely because soon the exit poll results would be replaced with the quick counts. That is how the projections work.

We could talk further about how the tabulations are weighted to official results, and whether they should be. If the official returns are more accurate than the exit polls (as they seem to have been in 1992, and in Florida in 2000, and probably have been in many other cases -- even if we allow for the fact that interviews measure voter intent and the official returns may not), then weighting to the official returns will generally yield better estimates. That is why the pollsters do it. DUers can choose to believe that they were trying to confuse us into believing that Bush was ahead in the exit polls all along, but that is a pretty poorly founded belief.

The point of the projections, of course, is to project who will win; how can that possibly not be relevant? That is what the 'exit pollsters' are doing between poll closing and the wee hours of the morning: updating their models with vote totals, hoping to call a winner as early as possible.

No, the calls do not affect the "exit poll" (by which I assume you mean the tabulations). Nor did the "tabulation stuff affect the snapshot," I guess -- I don't actually know what that means. The tabulations don't even include a projection of who will win the state or the popular vote (although the printed reports do have a topline). The point of the tabulations is show things like "how Catholics voted," not to reveal the winner.

The whole thing about court rulings is pretty much beside the point. Whatever the courts did about Florida 2000 was not going to appreciably affect the tabulations (unless, as I suggested in my PM, they had thrown out the entire Palm Beach County vote -- that might have an impact). Ultimately, there was no call in Florida in November 2000 (after the round of second calls was retracted, that is!), precisely because there was no way to predict how the courts would rule. But, again, the tabs show things like "how Catholics voted" and "how big the gender gap was," not "who won."

Unfortunately, part of what you are grasping seems not to be reality. Conceivably it will help if I post an example of why weighting to official returns should improve the tabulations if the official returns are more accurate than the exit poll. That will have to wait until tomorrow evening, probably. (Or maybe you are just committed to believing that E/M did something amazingly sneaky, apart from being incredibly obvious. If so, I probably will not change your mind.)

(EDIT to remove a cranky sentence, although the last graf remains somewhat cranky)

[OnTheOtherHand]

154. why weight?

(Some people have questioned whether OTOH has the cojones to
create long, non-proportional posts with lots of tables.
Certainly I opt for quality over quantity, but lest anyone
doubt my ability to tabulate....)

Here is an attempt to demonstrate succinctly why weighting to
the official results is generally expected to yield more
accurate tabulations. This analysis also provides hints to how
one might detect certain forms of vote miscount (accidental or
otherwise), but my main point is just that weighting to the
official results is a legitimate method, not a scam --
although, of course, it can yield _less_ accurate tabulations.

Assume two candidates, Able and Baker, and an electorate split
into two groups, athletes (40%) and bookworms (60%). 60% of
athletes vote for Able, 40% of bookworms vote for Able, and
thus Able gets 48% of the vote (Baker gets the other 52%). In
a hypothetical sample of 1000 respondents without bias or
sampling error, the results would look like this. Note that
the percentages are row percentages (% of athletes or of
bookworms).

 (unbiased)       Able             Baker         total 
athletes       240 (60.0%)      160 (40.0%)       400
bookworms      240 (40.0%)      360 (60.0%)       600
––––––––––––––––––––––––––––––––––––––––––––––––––-––
TOTAL          480              520              1000

Now, suppose that Baker voters tend to steer around the
interviewers, so they are underrepresented in the survey.
Specifically, say that 10% (i.e., 52 out of 520) of the Baker
voters represented above don't participate. For effect, I will
take away proportionately a few more athletes than bookworms:
so I am assuming (to make the example more complicated) that
athletes are more likely to avoid being interviewed.

 (biased)        Able             Baker          total 
athletes       240 (63.2%)      140 (36.8%)       380
bookworms      240 (42.2%)      328 (57.8%)       568
––––––––––––––––––––––––––––––––––––––––––––––––––––-
TOTAL          480              468               948

In this biased tabulation, Able appears to be ahead. (You can
compute the percentage if you want, but I deliberately don't
report it because actual demographic tabs wouldn't report it.)
For present purposes, focus on the Able demographic
percentages: 63.2% among athletes, 42.2% among bookworms. Not
bad, but definitely off.

Now, the official returns indicate that Baker has won with 52%
of the total vote. To weight to the official returns, we
multiply each Able respondent by a number somewhat less than 1
(about 0.948), and each Baker respondent by a number somewhat
greater than 1 (about 1.053), so that the Able respondents are
48% of the weighted sample.

 (weighted)      Able             Baker          total 
athletes      227.5 (60.7%)     147.5 (39.3%)    375.0
bookworms     227.5 (39.7%)     345.5 (60.3%)    573.0
–––––––––––––––––––––––––––––––––––––––––––––––––––––-
TOTAL         455.0             493.0            948.0

Notice that the percentage estimates are now better (but not
perfect). There is one subtle side effect: athletes now make
up 39.6% of the weighted sample instead of 40%. (If
athletes-for-Baker and bookworms-for-Baker had been equally
likely to avoid being surveyed, then the weighting would
actually retrieve _all_ the original percentages.)

Now, what if for some reason the official returns indicate
that Baker won with _53_% of the total vote? (For instance,
what if votes for Able are subject to higher spoilage rates?)
The weights will get a bit more extreme, and we will end up
with something like this:

(misweighted)    Able             Baker          total 
athletes      227.8 (59.7%)     150.3 (40.3%)    373.1
bookworms     227.8 (38.8%)     352.1 (61.2%)    574.9
–––––––––––––––––––––––––––––––––––––––––––––––––––––-
TOTAL         445.6             502.4            948.0

These percentages are still more accurate than in the
unweighted biased tab. Loosely speaking, _if_ the official
results are more accurate than the survey results (as they
were in this case), and _if_ the bias in the poll isn't
_heavily_ concentrated in some demographic groups, weighting
will generally produce more accurate percentages.

[Febble]

141. Usrename:

Neither I nor OTOH need ANY convincing that more votes were cast for Gore than for Bush in Florida in 2000. On every possible occasion, I link to Walter Mebane's The Wrong Man is President and to to Wand et al's The Butterfly Did It.

The point OTOH is making, however, is that even if all those votes had been counted, the state would still have been too close to call before all the votes were in.

This may seem like a geeky statistical point - but statistical inference is what we are talking about here. The exit poll is a sample. Any inference from a sample has a margin of error, as TIA has been explaining for as long as I've been reading his posts. And in 2000, the actual margin, even if the all those overvotes and butterfly ballot votes had been counted, was within the margin of error of the poll (I'm not talking about voter suppression here, as those voters wouldn't have been in the exit poll anyway.)

So yes - of course - the call for Gore was "correct" in the sense that Gore did, as we now know, receive more votes (whether or not they "would have" been counted had the SCOTUS not intervened). But it was not "correct" statistically. I sometimes do experiments that get the "right" answer but are still not "statistically significant". And if that "right" answer is not "statistically significant" I can't get it published, even if it replicates what other people have shown to be true. To put it another way, the Florida exit polls in 2000 did not have the "statistical power" to call the state, even had the vote count been fair. It should not have been called, ever, until every vote was counted.

The tragedy, I remember thinking at the time, is that because they muffed it, Gore lost the psychological advantage. If it had never been called for either, I somehow can't see even Bush insisting that he was the winner despite loss of the popular vote. So you'd probably have President Gore today.

So my point is: the networks were not "wrong" because they called the state for Bush after they called it for Gore. They were "wrong" because they should never have called it at all. Not because Gore didn't win, but because the race was far to close for the statistical power of the poll.

[Usrename]

152. not buying it, sorry

This is not reasonable

"The point OTOH is making, however, is that even if all those votes had been counted, the state would still have been too close to call before all the votes were in."

That's just making up facts that don't exist and never have existed. Sorry.

[Febble]

153. It certainly isn't making up facts

that don't exist, but it may be that I have not clearly conveyed what I meant. Or even that I have them wrong. Here are the facts as I know them:

What was the Bush's final margin in Florida before the SCOTUS stopped the count? A few hundred votes (Wikipedia tells me 537). Mebane estimates that tabulation problems deprived Gore of a 30,000 margin. Wand et al estimate 2000 lost votes for Gore (to Buchanan) due to the Palm Beach butterfly ballot, so we'll call that 32,000.

Now we will never know for sure how much of that 32,000 margin would have been counted had the SCOTUS not stepped in. The butterfly ballots were a lost cause, sadly, unless they had a re-vote, and although some of the overvotes for Gore might have been counted, depending on the standards applied, certainly not all of them would have been. So the margin Gore would have had, even if the votes had been counted according to reasonable Florida standards, would have been less than 30,000. But even if we give all 30,000 to Gore, that gives him a margin of one half of a percent, which, if we assume a similar sample size to the one we can ballpark from the UMich data for 2004 (someone might have better figures), is well within the MoE calculated on sampling error alone. And for reasons that have been well-rehearsed on DU (including the dreaded "cluster effect" - the enlargement of the MoE required to account for the fact that voters are not randomly sampled from all precincts but from a sample of precincts), and including the fact that all polls have to allow for the possibility of non-sampling error (eg non-response bias) there was no basis for a call.

NOW PLEASE DON'T GET ME WRONG. There were many other people who should have had a vote to cast for Gore, and who didn't because of voter roll purges. And there may have been other forms of electoral corruption. What I am saying does not concern these. Networks cannot call a state because they think that a candidate should have won votes that were never in fact cast, or were cast on equipment too badly designed or maintained to render them legally countable, or even which were stolen. What I am saying is that even taking a generous stab at Gore's likely margin had the Scotus not intervened, the margin was too small to be predicted from a sample of 5000 voters or fewer. The SCOTUS gave Bush a margin of .01%. Mebane gives Gore a margin of .5%

The exit poll didn't have the statistical power to tell the difference. And the vote counts up till Harris's intervention were a few hundred votes apart. So the state should never have been called until every vote was counted.

And that's the count that the SCOTUS stopped.

[Usrename]

158. You don't see the inconsistancies in you own views.

The votes were NEVER COUNTED.

How can I make this any clearer. The SCOTUS ruled. Don't count the votes or there could be irrepairable damage. What we do KNOW, even without a count, just by the number of spoiled ballots in 2 out of 64 counties, is that Gore lost a lot of votes by not having a lot of votes counted.

We only KNOW this BECAUSE of the EXIT POLLS.

"So the state should never have been called until every vote was counted."

But they will never be counted. So what is YOUR POINT??!!!!

http://archives.cnn.com/2000/LAW/11/30/gore.counting.pol/

[Febble]

163. Well what I'm saying

is that we DON'T know it because of the exit polls. We know it because of painstaking work by people in Florida like you, and researchers like Mebane.

All I'm saying, and really it is not a huge point, and actually is part of the tragedy IMO, is that the state should never have been called either for Bush of for Gore, not because Gore didn't win (he certainly had more votes cast for him), but because even if we estimate the size of his margin given the votes we know he should have had but weren't counted, the statistical power of the poll wasn't great enough.

No, they will never be counted. Gore should have been your president in 2000, and would probably be your president today. It is a tragedy that extends beyond America.

Look, I completely agree with you about Florida 2000. If it wasn't for Florida 2000 I wouldn't even have got involved in 2004. But that doesn't alter the fact that the state, even with Gore's true majority, was too close for the polls. They are too blunt an instrument to detect that kind of difference.

[Usrename]

164. This is a false assumption, not based in reality.

"but because even if we estimate the size of his margin given the votes we know he should have had but weren't counted, the statistical power of the poll wasn't great enough"

Your willingness to assume things like this is very troubling. There is much of your arguments and attitudes that flows from this false way of viewing reality. If I were so inclined, I could make some very strong arguments that the FL 2000 vote was not close at all, and well outside the level of closeness that you claim, so as to have some real statistical power to differentiate, that the counting process was so flawed that it is less reliable than the exit polls that you have no faith in. But what is the point of such an argument, anyway.

The exit polls remain what they were. This is totally separate from what happened when the tallys came in or the courts ruled. You are claiming some bizarre connection between the two. Like one affects the other somehow. They are distinctly different things. Both of them interpret the same reality, sure, the reality of people going to the polls and voting, but the exit poll is discrete from the courts. An accurate rephrasing of the statement above:

"since an accurate count of the vote was deliberately avoided, and because there is no method available to accurately estimate the size of his margin, given the votes we know he should have had but weren't counted, there is no possible way to gage the statistical power of the poll"

See what I mean?

[OnTheOtherHand]

166. the poll's statistical power does not depend on the vote count!!!

The estimated standard error on the Gore lead was on the order of 2 points depending on the race used, which means that Mitofsky and Edelman needed a 5- to 6-point lead given their call standard then, or an 8-point lead given their call standard now. They did not consider the exit poll interviews decisive, so they waited for fifty minutes. Repeat: the exit poll interviews did not support a call in Florida.

Fifty minutes later, Mitofsky and Edelman called the state for Gore based on a combination of the exit poll and early returns, in comparison with the 1998 governor's race, which seemed to be yielding the tightest estimates at the time. (They also assumed an absentee vote smaller than it actually was.) If they had used either of the other reference races, even then they would not have made a call.

The mean precinct within-precinct error in Florida in 2000 was -0.6, which tends to confirm that the race was very close even with respect to voter intention (at least among election-day voters) -- if it possible for the exit poll to confirm such a thing.

AFAICT it is incorrect to assert that Febble has "no faith in" the polls (that is a very unscientific way of approaching the question of their accuracy); and, again, the poll's statistical power does not depend on the vote count, official or actual.

Apparently we are confusing you by trying to meet you halfway. I believe you offered the estimate that Gore won (based on intended vote) by 50,000 votes; if so, an accurate exit poll and model would have yielded an estimate within the margin of error, and no call would have been made. Even if you believe that Gore won by, oh, 200,000 intended votes (I think that's about the total number of overvotes and undervotes), an accurate exit poll would have yielded an estimate within the margin of error, and no call would have been made. In point of fact, as far as one can tell, the initial estimate was within the margin of error; at any rate, no call was made based on the interviews alone.

If you have a reason to believe that the exit poll proves something in particular about the outcome in Florida 2000, it might be constructive actually to marshal some evidence.

[Usrename]

168. I offered what I know about two counties in FL 2000.

Two out of 64. The largest county, Orange County, Orlando, no one has a clue about. No one. Check it out yourself. They never even did the mandatory machine recount. Just a black hole. The entire I-4 corridor is a huge black hole. The facts about Palm Beach and Duval are just an illustration. Vote switching may also be a big problem, but I just don't know. But I won't make some wild claims about things I don't know.

I said a 50,000 vote discrepency in just two(2) counties. Jeesh. Buy a clue. Things are REALLY bad in FL. You have no idea. You think you have an idea, but you just don't know how bad that Harris woman was. Take whatever you think she may have been able to accomplish and multiply by several thousand. Then you might begin to understand why your arguments just don't wash with me.

How was the VNS guy able to claim, correctly, what he claimed on election night? See, there are things that are known and things that are not known. He knew about those two counties, but for everything else he had his head up his ass. OK, keep telling yourself that, but don't expect me to buy it.

[OnTheOtherHand]

171. you ought to back off

You are expressing opinions about exit polls, and you do not evince much knowledge of exit polls, nor much willingness to think about exit polls.

You made the rather strange claim that the VNS data "correctly predicted the missing votes in Broward, Duval and Palm Beach Counties in FL." You have never substantiated that claim, and I do not understand how you ever could.

It seems to rest on your recollection of what "the VNS guy on election night" said "an hour or so after FL was uncalled for Gore." Probably, what you heard someone explaining was (as in the report I linked to) that the vote count from Duval had been misentered -- as in the Konner et al. report, p. 13, which states, "(Later scrutiny determined that a VNS keypunch operator had entered incorrect vote-count data, which had the effect of making it appear that Gore had won 98 per cent of the Duval County vote tabulated up to that time.)" (My emphasis.) Obviously ballot spoilage in Duval would not have had that effect.

Regardless, the exit poll was too small reliably to detect spoilage of 50,000 votes in two or three individual counties. If someone from VNS really did go on TV around 11 or midnight to say any such thing, he was gibbering -- and, if so, I think more people would have noticed. So, do you have any other reason to believe your assertion?

(Hey, if you remember what network you were watching, maybe we can get a transcript and try to sort this out.)

Have you noticed, in fact, that you are pretty much ignoring the content of our arguments about exit polls? You have a lot to say about what you know (and believe) about the election, and what you think you heard some VNS guy say, and that is about it.

You are welcome to believe whatever you want about Gore's winning margin in Florida, and you may indeed be able to make the case. (NB: if I took your suggestion to "multiply by several thousand," I would end at least in the hundreds of millions, which perhaps we can agree to rule out.) But if you want to claim that the exit poll supports your view, you have to make a specific, substantiated argument to that effect. Without such an argument -- which you seem to have no clue how to formulate, much less defend -- you have only a false appeal to authority, plus an apparent propensity to cast unjustified aspersions on people who point out your mistakes. This is what I have tried to caution you against.

If you do not know how to argue intelligently about exit polls, and you are not willing to learn, why don't you just drop the exit polls and build your case about Gore's winning margin in Florida? That would be an intellectually justifiable approach. Statistical error 'works' in both directions; you are unlikely to choose a margin that is inconsistent with the exit poll interviews, although of course it may be inconsistent with other data.

False appeals to authority, coupled with unjustified aspersions, are in very poor taste in my opinion.

[Febble]

169. My "assumption"

is based on the research done after the election on how many votes were apparently lost for Gore. If you don't accept the research, and you think Mebane's estimate was way too, fair enough.

But the state would not have been called on the basis of the exit polls alone anyway. I don't know why it was called, because the early vote counts (miscounts and all) did not support a clear win for Gore, even if they should have done.

[BeFree]

94. The old: "We had this discussion months ago" ploy

That's what you resort too? And you assume that I think you are to be believed? Such vanity I have never encountered. Oh, yes I have... from the squatters in the white house. How does it feel to be here, all alone, supporting the squatters?

We all pretty much know the exit-polls this time around were cooked up hard. Microwaved, if you will, and no ampount of watering things down will make the hard facts disappear.

TIA has shown, time and again, how the pollers cooked their numbers well after the raw numbers and most official results were in and your pleadings to reverse the truths in all of his work is tasteless and un-appealing. I don't see anyone here agreeing with you, do you? Maybe you are out of your field?

[OnTheOtherHand]

97. not a ploy, a fact

I linked to the previous thread -- there is no doubt that it actually exists.

The exit pollers weighted the final results, yes.

In case you haven't actually read this thread, I am one of at least five posters here who has criticized TIA's arguments. It may suit you to believe that we are all Republicans, but we're not (granted that I can't personally vouch for every one).

Bye.

[EFerrari]

16. K&R for good work.

:)

:kick:

[autorank]

17. The REAL Exit Poll Results for OHIO and the "not so real"

From SoCalDemocrat, a historic post. Priceless and true.

[mom cat]

18. This work needs a much wider audience. Will TIA be publishinghis

work in a way that is more widely accessible?

[FogerRox]

31. Auto- "K n R" to get it on -RANK

[mom cat]

33. I have already Ranked his Auto about this and K&R'ed this too.

I am all in favor of piling on at this point. It would be a WOOT!

[FogerRox]

34. WOOT WOOT WOOT FOR TIA

[autorank]

44. I'm used to being Ranked on (or is upon);)

[autorank]

43. Stay tuned;) n/t

[mom cat]

48. I won't touch the dial! Thanks!

[texpatriot2004]

38. A million thanks Autorank and TIA nm

[bleever]

40. Kick.

:thumbsup:

[Peace Patriot]

46. For those who say, what difference does it make? --as someone upthread

seemed to say--a sort of despairing cry that nothing can be done, and that it's one for the history books, no more...

1. The American people are demoralized, disempowered, and, above all, DISENFRANCHISED. They need to know WHY this government is so out of control. The reason is that the Bushites are no longer beholden to the American people. They are beholden to Diebold and ES&S, and the fascist agendas of secretive rightwing billionaires, Saudi royals and other puppetmasters. People NEED to know this--if we are to have any chance of restoring democracy. They need to know that the election SYSTEM is itself a fraud, on its face--and that it's very likely that their fellow Americans did NOT re-elect these criminals. They need to be heartened. They need to be bucked up. They need to know that the very mechanism of our sovereignty--our right to vote--has been taken away, and that we can, and must, fight to get it back.

2. There is still a window of opportunity to restore transparent elections. Control over election systems still resides at the state/local level* where ordinary people still have some influence. We need to act as quickly as possible to restore our right to vote. We need to rally people. And this info on the stolen election, and on the mechanism of the theft, can and does rally them. It's like a lightbulb goes off in their heads when you tell them about it (--those who are able to absorb the failure of all of our democratic institutions, the failure of the press, the failure of our own party; not everyone can take that shock).

3. The fraud can potentially be overcome by huge turnouts. They have to protect the fraudulent election system from exposure, so they have to be somewhat careful and not too obvious. Also, it's likely that the electronic fraud has to be pre-programmed, giving Bushites and warmongers a pre-set advantage, that is not so easy to change on election day itself. (At least this seems to have been true in 2004.) (That is, Kerry could have overcome the fraud, if he'd won by, say, a 10% margin, rather than a 3% to 5% margin.) So there is hope in this, too--that we can start turning things around by overwhelming turnouts; and also, by overwhelming switches by voters to absentee ballot voting, a voter protest against the electronic voting machines that is already becoming quite noticeable. Absentee ballots are not the complete answer, by any means. We still have the electronic central tabulators to worry about. But it's helpful for recounts, and as a protest.)

----

*(The Bushites are maneuvering to take power over election systems away from the states in legal cases in Pennsylvania. See: www.votepa.us.)

----

Some resources:

www.votersunite.org (MythBreakers - easy primer on electronic voting--one of the myths is that HAVA requires electronic voting; it does not.)
www.verfiedvoting.org (great activist site)
www.UScountvotes.org (monitoring of '06 and '08 elections)
www.solarbus.org/election/index.shtml (fab compendium of all election info)
www.freepress.org (devoted to election reform)
www.TruthIsAll.net (analysis of the 2004 election)
Sign the petition (Russ Holt, HR 550, great bill-has 169 sponsors). http://www.rushholt.com/petition.html
www.debrabowen.com (Calif Senator running for Sec of State to reform election system)
www.votepa.us (well-organized local group of citizen activists in Pennsylvania, where important legal issues are at stake)

Also of interest:

Bob Koehler (-- four recent election reform initiatives in Ohio, predicted to win by 60/40 votes, flipped over, on election day, into 60/40 LOSSES!--the biggest flipover we've seen yet; the election theft machines and their masters are now dictating election policy!)
www.tmsfeatures.com/tmsfeatures/subcategory.jsp?file=20051124ctnbk-a.txt&catid=1824&code=ctnbk

Amaryllis (Diebold, ES&S, Sequoia lavish lobbying of election officials - Beverly Hilton, Aug. '05)
www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x380340

------------------------------------------------------------------

Throw Diebold and ES&S election theft machines into 'Boston Harbor' NOW!

:think: :patriot: :woohoo: :patriot: :think:

[Melissa G]

56. kick!

[annabanana]

91. If it's not admitted.... It can't be fixed!

It's a dirty old cesspool.... but it won't work if it's not cleaned out...

[Awsi Dooger]

151. Terrific thread

I haven't been here recently but posts by Febble and On The Other Hand are always extremely informative and educational. When I saw the topic of this bloated thread I knew I wouldn't be disappointed in the content.

I bet heavily on politics, and have since '96. My initial edge over my competitors in the 16-man betting pool was a mathematical understanding of voting trends, specifically statewide bias and how that figured to swing based on candidates, situations and issues. I saved and studied one preference poll and exit poll after another.

In '96 and '00 I distinctly remember wild statewide exit poll numbers on CNN late into the night. I printed them, but they were so bizarre I threw them away in a day or two once the weighted numbers were available. Imagine my high hilarity post-2004 when suddenly the early numbers are considered gospel. Come on guys, you're supposed to understand history before making tunnel vision conspiratorial assertions. It's like if a horse with lousy breeding leads the Derby until the final furlong before collapsing, then everyone screams fix since they have no background reference point.

On election night 2004 I posted one thread here, regarding the party identification number, which was a stunning 37-37 according to the national exit poll. Kerry was doomed at that point, regardless of state by state hysteria. The PEW study forewarning the party ID transfer had been dead on. I posted the link before the election and here it is again: http://people-press.org/reports/display.php3?PageID=750

We were the victim of slightly more advanced and widespread low tech fraud. IMO, it didn't alter the outcome. My fear is focusing on Diebold and exit polls allows the GOP to laugh like hell while quietly stealing or suppressing more votes in that fashion.

[OnTheOtherHand]

155. (blush) yes, I worry about priorities

The problem isn't the "conspiratorial," it's the "tunnel vision." We should try to do justice to all the facts, not just a few of them. (What troubles me most is not when people express the belief that Kerry won, but when people express the belief that "it's no use, 2006 is lost, etc." The evidence doesn't support that brand of fatalism IMHO.)

Thanks for the kind word. Heck, wonks have feelings too (sniff).

So, does your interest in betting politics mean that you won't be telling us what you think about Congress 2006 until after the bets are settled?

[Awsi Dooger]

156. Ha! I already screwed that up years ago

I won the election pool in both '96 and '98 and would have won in '00 if Gore had rightfully been elected. Instead, I came in 3rd, still very profitable among 16 guys, the vast majority heavy right wing and furious a lefty was besting them.

So I was flush and cocky and stupidly started divulging some of my tactics to my competitors, i.e. evaluating the individual states and districts in terms of long term trends instead of overreacting to supposed current form. I mentioned that I posted on DU which wasn't brilliant since I tipped my picks here on several races, specifically Colorado senate. Plus I have a PAN or Partisan Adjustment Number theory, basically an assertion that statewide political polls err in the same direction on a regular basis. It's kind of like a pointspread in football. For example, in Georgia I automatically adjust every statewide poll 4.5 points more toward the Republican. That is the extreme example. Most states have a PAN or 2 points or less, but that can be significant in tight races. Admittedly, it's still a work in progress and the sample size is so low, since there are extremely few statewide races, that the PAN is fluid and inexact. But still no doubt a benefit and far superior to merely accepting pre-election preference polls at face value, instead of evaluating how the state polls have fared in the past.

In almost every case, the result errs toward the partisanship of the state. For example, in 2004 you had Tony Knowles leading every poll for an entire year against Lisa Murkowski in the Alaska senate race. I wasn't buying it at all. The state has red PAN by several points and every indication is it is getting more significant, polling skewed toward the Democratic candidate. In 2002, Fran Ulmer polled very close to the elder Murkowski in the gov race but was buried by 20+ points. I bet on Lisa Murkowski in 2004 and she won fairly handily. In 2000, the late polls had Rick Lazio very close to Hillary on election eve. The odds on the race plummeted and I took advantage. That state had blue PAN, and the '98 senate race between Schumer and Domato indicated the PAN edge was increasing.

I simply have very little respect for statewide polls, in many cases. Certainly in contrast to national polls. So when you've got threads and theories embracing state polls, or state exit polls, I allow a private chuckle. Look at the consensus national polls from 2004, immediately prior to the election. The average and median was virtually identical to Bush's actual popular margin of roughly 2.5%.

Regarding 2006, I get criticized all the time for posting this theory here, but IMO the Democrats should all but ignore Bush and the Republicans leading up to the elections. I started posting that six months before the 2004 election and I'm convinced it would be the proper strategy.

Democrats may be well ahead in the generic congressional polls, but the favorability rating is very low, straddling 50%, so that's flimsy as hell. It's kind of like a football team that is a big favorite based not on its own strength, but the perceived inferiority of the other team. Well, once the kickoff takes place in a contest like that, suddenly the favorite has no inherent strength and the game is almost always more competitive than expected, if not an upset. Superiority is a rock but we don't have that in the minds of the American people. Right now it's more disappointment and bewilderment in Bush and Co.

Unless we emphasize enough forward thinking proposals to boost favorability, I'm worried the pickups will be lower than estimated. Negativity works for Republicans but Democratic fortune is tied to thoughtful but conflicted white women, ones who switched to national security concerns post 9/11. It was the only major voting block that switched dramatically from 2000 to 2004, from 48-47 to 51-43. It's really the only voting block that matters right now and I'm amazed that isn't emphasized and strategized daily. It's a federal race dynamic so my primary hopes are for big gains in governorships.

I would be astonished if we gained control of either chamber. Post 2004, the conventional wisdom was Democrats would be playing defense in 2006 in regard to the senate, given the states in play and considering we already picked up +5 net among this block in 2000. To project another similar gain is simply not realistic, since it would be a transfer of basically 1/3 of the seats over a six year period. My baseline belief is everything tends to drift back to the beginning. The GOP implosions need to be significant and sustained to overcome that early conventional wisdom, regarding the senate playing field.

I know much less about the House, other than redistricting has severely narrowed the number of floppable seats. I don't see the '94 comparisons as valid for that reason, and also that Democrats were not motivated in '94. We had finally regained control of the White House in '92 after 12 years, so that relief and satisfaction carried over to '94 and depressed our turnout. Even if the GOP base is annoyed at Bush for recent choices, they will be sufficiently motivated and Rove's GOTV strategy has improved immeasurably since 2000. Sorry to be less than giddy, but I've learned to look at things from a balanced to slightly pessimistic standpoint as as wagering guideline.

[OnTheOtherHand]

157. yeesh, I should have asked you more qq months ago

(The exit polls themselves tended to err toward the partisanship of the state -- not to change the subject back to that.)

I'm not surprised at all that Alaska has red PAN, but I never tried to figure it out. The 'debate' over possible fraud in Alaska in 2004 has been... oh, never mind. Anyway, it would be interesting to compare your PANs with E/M's priors and see whether you have independently come to similar conclusions about how to adjust pre-election polls. (Not that this would matter in AK.)

Yes, indeed, cheerleading and analysis don't go well together. I'm not at all giddy about 2006. A few points I would press you on, but we have months to sort them out. I agree that the generics don't mean much given that Democratic is still not thriving as a brand. (And what I've seen so far certainly supports your view that the Dems are predisposed to do best in governorships.)

I would qualify the "all but ignore Bush and the Republicans" bit -- call me old-fashioned, but I think challengers do need to run against incumbents to some extent! Nevertheless, 2004 should underscore that running up incumbent negatives doesn't guarantee victory if the incumbents can run up your negatives, too.

I'm reluctant to think of white women as a "voting block," much less "the only voting block that matters right now." I'm inclined to agree with anaxarchos that the "security mom" theme is overrated. Dunno, to some extent it's a cognitive bias: I rebel against anything that sounds like a monocausal theory. Security issues are huge for the Dems no matter what block(s) one ties them to. (I don't think they will benefit the Reps in 2006 in the way they did in 2002, but I am also thinking further down the road.)

[Awsi Dooger]

159. I think security moms are underrated

Chuck Todd of the Hotline stressed repeatedly during '04 that Kerry's fortunes were tied to his status with white women. It's obviously not an accepted theme because I posted it many times here before that election with little response other than dismissive. Part of it is local knowledge. I host debate watching parties among apolotical voters every presidential cycle. Mostly white women. They were unquestionably more security conscious in '04. When I quizzed on the top issues before debate #1, 9/11 and terrorism was a comfortable first place.

When I looked at the presidential and senate races from 2004, it was evident white women voted more conservative in the federal races than in gov races on the same ballot. There didn't seem to be a similar break among any other major voting block.

[OnTheOtherHand]

160. I will think about it some more

I'm content with the premise that Kerry was hurt by white women's concerns about security, and it seems very plausible that that was crucial in narrowing the gender gap.

I have real reservations about trying to target the security message to white women, partly because I don't know how to do that, partly because white men were worried about security, too. But there is a lot to think about there.

(EDIT: Also, so many Americans are so emotional about the security issue -- from various stances -- that it is probably hard for us to think lucidly about its impact on presidential politics. I've encountered many polarized views.)

[Awsi Dooger]

161. Here's info from one link

Not enough comparison to previous years or non-federal races, but interesting:http://www.washingtontimes.com/op-ed/20041218-100132-6503r.htm

"Reconsider 2004. In Ohio, a mere 119,000 votes (among the 5.6 million cast for Messrs. Bush and Kerry) separated the two candidates. A reversal of 60,000 votes would have given Mr. Kerry Ohio's 20 electoral votes and the presidency. White women, comprising 46 percent of the Ohio electorate, gave 55 percent of their vote to Mr. Bush and 45 percent to Mr. Kerry. Mr. Kerry would have won the presidency if he only could have increased his share of Ohio's white women voters by less than 2.5 percentage points (2.3 points, to be exact). He didn't even need a majority: only 47.3 percent. Similar relatively minor improvements in Mr. Kerry's white-women vote in the cumulative states of Colorado, New Mexico and Nevada would have been sufficient to catapult him into the White House.

"In Florida, which Mr. Kerry lost 52-47, he received 43 percent of the white-women vote, which comprised 38 percent of the state's total. To win the state's 27 electoral votes and the presidency, Mr. Kerry did not even need a majority; 49.6 percent of the white-women vote would have sufficed.

"Citing a post-election survey, Anna Greenberg reported in the December American Prospect that "Kerry lost white women without a college education this year by 23 points (38 percent to 61 percent)." The statistic offers plenty of food for thought for Mrs. Smeal and other disgruntled Democrats to consider over the next four years.

[OnTheOtherHand]

162. I will try to give you some more data to chew on soon n/t

[Awsi Dooger]

165. Looking forward to it

My job is sports stats so I'm bogged down there until college basketball ends in early April. I'd like to pick apart and understand political numbers better than I do now. That's why I love the scrutiny of exit polls by yourself, Febble and others, since I can apply some of the theories elsewhere.

My basic belief is 9/11 severely detoured the Emerging Democratic Majority trend by turning a segment of Hispanics toward the GOP, and creating the Security Mom brand. Whether it's merely a 5 year delay, or 10 years or not at all, I'm not sure.

As far as ignoring the GOP, yes every election is a referendum on the incumbent party. I've simply seen no evidence our piling on has any impact. Opinions on this guy have been entrenched for years, subject to self-inflicted downturns only. We've made dozens of extended attempts to rip Bush but I've never seen a sustained 3 or 6 month period to focus on Democratic ideas, to potentially boost our favorables. In hosting the debate parties for a decade, I'm convinced people want to vote for, not against, especially white women who I continue to believe are the most critical block.

[coffeenap]

167. Did you all see the post in GD? TIA is very ill:

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=364x621028#top

[Awsi Dooger]

170. Yes, I just saw it and replied

Just stunning news. I hadn't visited that website where he was posting for a month or more, so I had no idea anything was wrong.

Best wishes to TIA. The disagreements we've had take a severe back seat to health and well being. His passion and energy in the mathematical analysis have always been remarkable.