« Strategy Notes:
John Belisarius

How Big A Role Did Fraud, Ballot Theft and Suppression of the Vote Play in The Election? | Main | God, Guns and Gays: Testing the Conventional Wisdom About the 2004 Election »

Do the Exit Polls Indicate Voter Fraud?

There are two lines of analysis that are typically used to justify the claim that the 2004 election result was somehow stolen by the GOP. The first is various bits and pieces of "evidence"--the precincts in Cuyahoga County, Ohio with more votes than registered voters, the counties in Florida (Baker, Holmes) with huge Bush margins but big Democratic registration advantages, etc.--that supposedly indicate vote tampering. I find this evidence profoundly unconvincing and think Farhad Manjoo and others have it basically right: there's not a lot of there there. Vote tampering does not appear to have happened on the scale necessary to affect this election.

The second line of analysis invokes the now-infamous early releases of the NEP exit poll data, which showed Kerry with a 3 point national lead, solidly ahead in Ohio and also leading in Iowa, Nevada and New Mexico. The reasoning, laid out most clearly in a paper, "The Unexplained Exit Poll Discrepancy", by Steven Freeman of the University of Pennsylvania, is that exit polls are very accurate surveys and highly unlikely to produce the results referred to above by chance if the real world results truly were +3 Bush, etc. Therefore, the reasoning goes, our measurement of the real world (the actual vote counts) must be wrong and the original exit poll results right. Conclusion: there's something very funny going on with this election.

But there is a huge problem with this line of reasoning. The exit polls have always drawn samples that are off the real world results and have always had to be corrected (weighted) to eliminate bias, reflect new turnout patterns and, in the end, just flat-out conform to the election results. This year is no different (though it is possible that the magnitude of these corrections has been greater than normal).

Here is my understanding of how the exit poll samples are weighted, based on what I have been able to ferret out so far. (No doubt, I'm not getting it entirely right, but it's damnably difficult to track down good information about this--exit pollsters have never made much effort to publicly explain and document their methods.)

1. Samples are weighted to correct for oversampling of precincts (for example, exit polls have historically selected minority precincts in some states at higher rates than other precincts) and for non-response bias (exit poll interviewers try to keep track of refusers by sex, race and age).

2. Samples are weighted to correct for changing turnout patterns in the current election, since the sample design is based on past turnout behavior.

3. Samples are, in end, simply weighted to correspond to the actual election results. This is done by first weighting exit poll results in sample precincts to the true precinct results, as they are known, and then weighting the overall sample to the overall election result, once it is known.

At what point are these various weighting procedures performed? That's difficult to say because of the lack of public documentation of exit pollsters' methods. But it appears to be the case that weighting of flavors one and two takes place at least partially during the day (and continuously through the day), while the third flavor naturally has to wait until actual election results start to become available.

So where were we in this extremely complicated weighting process when those first +3 Kerry exit polls hit the CNN website? Who knows? (And exit pollsters have not exactly clarified the issue since).

But it's certainly clear that those data had not yet been weighted (or at least very much) to reflect the actual election outcome (again, part of standard exit poll procedure, not anything peculiar to this year). But how much had they been weighted to reflect the other factors (1. and 2.) mentioned above?

Possibly much of this weighting had already been done. If so, then the rest of the sample correction--that took their data from +3 Kerry to +3 Bush--was done by good old-fashioned weighting to the election outcome. Or perhaps it was some combination of additional weighting for factors 1. and 2. plus weighting to the election outcome.

Who knows? Again, exit pollsters don't seem to be particularly eager to share this information. Nor do they seem particularly eager to clarify how common it has historically been for exit poll samples at that time on election day to be that far off from the actual election result.

The issue of historical comparisons is an important one. Part of what has led to the brouhaha over this year's exit poll is people's lack of knowledge about how exit polls have been conducted in the past.

Consider this. The unweighted--completely unweighted--data from the last four presidential elections before this year are as follows:

1988: Dukakis, 50.3; Bush, 49.7

1992: Clinton, 46; Bush, 33.2

1996: Clinton, 52.2; Dole, 37.5

2000: Gore, 48.5; Bush, 46.2

President Dukakis? Obviously, the unweighted data have always been highly problematic and--interestingly--have always shown a strong Democratic bias. Now these unweighted data from past years do not, admittedly, correspond to where we were in the weighting process on election night this year when the +3 Kerry poll hit the 'net--those data had presumably already been weighted to some extent to correct for factors 1. and 2.--but it is still food for thought.

Of course, it's entirely possible that exit poll samples this year, controlling for similar points in the weighting process, were more "off" than in past years. I can't say at this point and I urge the NEP to make the appropriate historical comparisons available to answer the question. But, even if so, this is hardly evidence of skulduggery in the real world; much more likely it reflects the enormous--and perhaps increasing--difficulties of conducting surveys of this complexity in a rapidly changing country.

Of course, additional inaccuracy in the exit poll samples this year (if true) is not a development completely devoid of implications. It could mean that some of the specific results from the survey are less reliable than in the past. (I personally have my doubts about some of the numbers, like those for Hispanics.) But that's a far cry from assuming an election has been somehow stolen or tainted.

My advice: calm down and concentrate on what's really important--beating them next time.

Comments

Ruy, with all due respect, and I hold you in the highest regard, you are missing the point. You can write off some of the flakey bloggers, but you can't write off Greg Palast. Millions of "spoiled" ballots in every election are not counted. 54% of them are African American. That is scandalous. As Mr. Palast put it, "You want to know who won this election? Count the votes at the back of the bus." It's more than just vote tampering, Tamany Hall style. It is a long-standing, systematic disenfranchisement that we keep letting them get away with. Your always-brilliant statistical analyses do not address the issue, for instance, of why Ohio refused to replace there punchcard ballot system, particularly in largely black and Democratic areas, knowing full-well that they were legally skewing the vote. You do us all a grave disservice when you tell us that fraudulent numbers just don't add up. Jim Crow has nothing to do with statistics. It had to do with a good ol' boy netword getting away with treachery until someone finally put a spotlight on it. Please don't be guilty of helping the Republicans turn off the light.

Ruy, the history is compelling, your logic is sound and you're probably right. However, there is some smoke, and while there may be no fraud, peeling back the covers to see what's been going on in there can't hurt; if only to let everyone know that election fraud will be sniffed out and prosecuted, no matter what imperial majesty is in power.

I am confident, simply due to the law of large numbers, that fraud occured. Whether slight or pervasive, I believe it's in the interest of protecting the process that every hint of election misbehavior be investigated.

Do you think that ballot spoilage might have some role in this? If Democratic voters have a higher rate of spoilage (I've heard that they do), maybe that's why their opinions count more in exit polls than in the actual results.

Thank you, I have been looking for something like this. Very good analysis of the raw numbers.

For an analysis of the finished numbers, this from the Caltech/MIT Voting Technology Project, seems very solid (and I haven't seen it get as much attention as it deserves). They also have a good analysis of the silly Dixiecrats=fraud claims from commondreams.org.

Ruy,
Thanks for the info on the exit polls. I understand that the polls noted by Freeman were very accurate when it came to the senate races. Why would that be....and inaccurate for the presidential?

Thank you, Ruy, for the rational explanation. Much of the media coverage of this issue is dripping with condescension for the "conspiracy theorists," but doesn't attempt to explain what actually happened. That only fuels the fire.

It seems that if the media applied some pressure on polling organizations instead of wasting effort beating bloggers over the head, much of this could be quickly cleared up.

You say that vote tampering did not happen enough to affect the election. That's not the point. Did vote tampering happen at all? If it did, it's totally unacceptable to let this go. We do need to focus on how to win the next election, but if we promise folks that every vote counts, then we need to stand behind that promise.

I don't think the problems with exit polling can be so easily dismissed as mostly unworthy of our concern.

To begin with, it seems pretty clear that the numbers people have been quoting, which suggest great discrepancies, are NOT completely "raw", and so pretty much expected to be far off, as with the numbers you produce for, say, Dukakis vs. Bush.

In fact, the numbers people have been quoting are almost certainly weighted, otherwise they would not have been reported in major media such as CNN. Whether, say, Freeman in his paper used the final weighted numbers (before adjustment to the actual reported tallies, so creating automatic "agreement" in outcome) is certainly an open question (a question the pollsters could easily answer but refuse to). Yet obviously most of the weighting had taken place in those numbers, simply given the late time at which they were reported (1 AM on Wednesday following election day, as I recall).

However Freeman's argument, based on natural but not confirmed assumptions about the numbers he uses, may hold up, there's a larger, very important reason to be concerned about these exit poll discrepancies.

The point has been made a number of times that exit polls are used in many countries as an independent check of the accuracy of the tallied vote, and discrepancies too large are regarded as signs of election fraud. Are we to believe that exit polls in the US, surely as well understood an electoral environment as any in the world, are, in contrast, so riddled with potential biases that we can hardly take their results seriously?

Clearly, exit polls are an essential component in our continued good faith in our elections, and so in our democracy. They are the only independent check we really have on the official tallies, a check that is only more important, not less, as new technologies supplant older, harder-to-hack methods of voting.

For the sake of democracy, the exit pollsters, whatever their proprietary feelings may be about their methods and numbers, are obliged to divulge fully and honestly all information that pertains to our continued trust in their methods and reported results.

Clearly, we need to understand if the errors in this election are unusual by historical standards. We need to understand under what conditions they would ordinarily conclude that there was a problem with the tallied vote, as opposed to the exit polling itself.

The exit polling is, I gather, paid for largely by a consortium mainly comprising news organizations. These sponsors, if they genuinely care about truth and democracy, should do all they can to pressure the exit pollsters to divulge what's needed so that we can have continued faith both in exit polls and in our democratic elections.

I wholeheartedly agree. The important thing here is not to prove that Republicans some how "stole" this election. People have been claiming that for the last four years about 2000, but it didn't make much of a difference, because nobody really listened except for those who were voting Democratic anyway.

We need to focus on convincing people that voting Democrat is a good, worthwhile cause, not convincing them that Republicans are going to eat their babies and steal their democracy from them. That's not a fight we can win.

At the end of the day, professionals are either justifiers or truly objective analyzers. The former is the black hole of our civilization -- indeed there is this OVERWHELMING draw, approaching infinite gravity, towards the former -- including not only the obvious temptations but the fact that the availability of data and even the data themselves are skewed by the process of justification. I could use my own life here as an example of the latter issues but that would take me too far afield.

In the case of EXIT POLLS, we have a problem. If you skew or taint exit polls to incorporate the "actual" (ie reported) polling results, you are assuming what you are trying to prove -- namely that the reported results were in fact accurate. It is amazing to me that, where questions about the DIVERGENCE between exit polls and reported results are raised, with the possibility of vote tampering, one would use data BASED on the reported results to alter the exit polls, thus creating a GIGO feedback loop.

One thing that should be demanded is access to the raw data. And this access is needed now before the election is certified Dec 12. I disagree that it is merely a case of a few minor glitches, including exit polls that were mysteriously accurate in NONSWING states like Illinois, but showed a Kerry win in Ohio, Florida and New Mexico. There were also wide divergences in swing and marginal swing states that still went to the predicted winner, albeit by margins far more in the Republicans' favor than the exit polls showed: WI, PA, NC, and NH among them. This pattern is ESPECIALLY fishy.

We also know that the Republicans have been engaged in a full court press to distort the voting outcome, including massive systematic challenges, mainly of voters of color. It is often emphasized that the 160,000 or so provisional ballots were in all likelihood insufficient to tip the scales in Ohio -- which was won by Bush by 136,000 +or- votes. Well that one factor could cut the gap down to less than a quarter of its current size quite easily, leaving the results vulnerable to patterns of OTHER kinds of vote obstruction following a similar script. An excess of votes (I saw a list of counties, not just one) could cut the lead still further. FULL investigation should be urged by all REAL Democrats, with the idea that if there were enough factors to tip the election, they could be found and held accountale BEFORE Dec 12.

The thing that is most striking about the fraud issue is not just that there are claims like Hartmann that, though easily dismissed byDavid Corn of the Nation (I spoke with the two on WNYC recently)
but that there are a pattern of MANY reported frauds, many though not all cogent claims, pointing in the same direction and concentrated in key swing states. The analysis of the exit polls seems VERY weak -- as these same polls were taken as quite strong indicators when they were the evidence at hand. It appears that only an unwillingness to take seriously the possibility that fraud etc determined the outcome of the election, and to pursue this UNRESPECTABLE agenda that is clearly disfavored is governing the same strategists who never even went as far as Jonathan Chait at the New Republic on the flipflop spin or than the stuff I have repeatedly pointed out about the Matt Bai distortion of Oct 10 and its three weeks of campaign and press progeny.

I really feel that there is a cravenness factor here, an issue DIFFERENT from where the DLC or Kerry stand on a political spectrum. Especially in the campaign but also now to a lesser extent this is the MOST distressing feature of the Democratic Party leadership and strategy. But now, those who were silent about the flipflop spin being bogus for five months (although David Corn claimed he had been covering it "from Day 1" -- I haven't found that coverage and anyone can email me who has such info either from David Corn or any other source of mainstream press debunking of the inaccuracy of the flipflop spin PRIOR to the Republican Convention on ANY substantial scale (not just a line or two and not just pointing out the hypocrisy or defending the $87 billion vote issue). Chait reported its absence and then tried to explain it away.

Now with the fraud issue raised, the hounds that didn't bark at the flipflop spin or the Bai distortion are all of a sudden barking in opposition to the claims. My what a surprise! The only DLC Democrats or other Democrats worth a bucket of spit are those willing to BUCK these trends in the Party and the press. That's much more important than key issues like NAFTA (which I have always opposed).

It is extremely odd that there is such reluctance here to take the issues of fraud combined with decisiveness of election challenges of valid voters etc. more seriously. Instead of trying to explain the issues away, and with a 16 page article on the subject at the Wikipedia website disappearing without a trace, REAL Democrats should be going to the mat to insist on a FULL investigation RIGHT away, when it could matter, instead of trying to explain why the exit polls, which may very well have been right, were "wrong". That was what was wrong with the whole campaign, and is still wrong. YES it is disfavored in the media, with mere claims and counterarguments on SOME issues taken as "proof" in the (Matt Bai source NB, NO op ed columns on the flipflop spin throughout the campaign along the lines of Chait etc) New York Times -- VERY unjournalistically, from the standpoint of objectivity rather than of justifying the lying.

Here I follow the guidelines of C Wright Mills on objectivity -- "I have tried to be objective; I do not claim to be detached." But the air of 'detachment' often substitutes for objectivity when justifying the lying -- like sweetness in candy fools the body into thinking you are giving it something with vitamin C.

Bottom line, the issue of whether these issues MIGHT have tipped the election should be a MAJOR focus. Expertise in polling should not be mobilized to explain AWAY the exit polling issues that dovetail with concrete observations on the ground -- but the input elements should be demanded NOW. And what is it about exit polls that made Ruy (and Dick Morris) so sure that they were right until the line was (now) that they were wrong and the reported results were right? These polls are said to be more reliable than most polls -- so why were they off so much, and so concentratedly in key swing states but not so much nonswing states llike Illinois?
More on this issue later. A web site recommended by a friend who is FAR more skeptical than I (although I am skeptical and not insisting that we have enough information to say that the election WAS stolen, only that I SUSPECT it was and, in a perfectly objective cosmos, would say the odds are better than even):
http://www.cs.indiana.edu/~shockema/pfc/election.html

Here's a good representative source, one that leads to others, that states the issue nicely on exit polls. (Note especially the stats that support my contention of much greater discrepancies as a pattern in the crucial states than in nonswing states): AND THERE IS MUCH ELSE THAT IS FISHY

http://www.opednews.com/lestrange_110604_concede.htm

Jonathan Simon of Alliance for Democracy notes, "Statistical discrepancies were identified in key battleground states that exceeded the margin of error of the exit polls. In Ohio, for instance, a shift of 3.1% toward Bush converted a 52% - 48% exit poll "victory" for Kerry into a 51% - 49% electoral "victory" for Bush. Other critical states showed similar anomalous results, each of which should occur less than 5% of the time. In the group of 12 critical states selected for analysis, exit poll vs. tabulated vote shifts exceeded the polls' margin of error in 4 cases, which according to statistical analysis would occur only 0.2% (or one five-hundredth) of the time in the absence of significant mistabulation of votes." Simon also notes that exit polling appears inexplicably to have been significantly more accurate in nonbattleground states, than in the states that were crucial to a Bush victory.   Rather than objectively exploring reasons for these identified discrepancies, the networks now glibly claim exit polling based on scientific methodology is completely unreliable, and have all but forgotten that there was a deep and widespread concern about the reliability and security of the vote tabulating apparatus leading up to this election.

[ ...]

Americans concerned about voter fraud in Election 2004 are urged not to concede their vote to George Bush, and to prepare for further pro-active measures designed to empower millions of legal voters who continue to be disenfranchised by party politics. To join the WE DO NOT CONCEDE Coalition, send an email to the organizer, Kat L'Estrange, at signup@donotconcede.com with I DO NOT CONCEDE MY VOTE in the subject line and contact information in the body. Contact Jonathan Simon for questions regarding statistical data at VerifiedVote2004@aol.com.  

STATS: Critical States (12) Exit Poll Data Bush% then Kerry%, # of respondents, then time of poll ET, and "Red" Shift% Colorado 49.9 48.1 2515 12:24 AM 2.6% Florida 49.8 49.7 2846 12:21 2.5 (Florida 51.4 47.6 2862 1:01 0.6 ) Michigan 46.5 51.5 2452 12:21 1.0 Minnesota 44.5 53.5 2178 12:23 3.0 Nebraska 62.5 36.0 785 12:22 4.3 (Nebraska "critical" because of ES&S dominance and history) Nevada 47.9 49.2 2116 12:23 2.2 New Hamp. 44.1 54.9 1849 12:24 4.9 New Mex. 47.5 50.1 1951 12:24 1.8 Ohio 47.9 52.1 1963 7:32 PM 3.1 (Ohio 50.9 48.6 2020 1:41 AM 0.3) Penn 45.4 54.1 1930 12:21 3.4 Wisconsin 48.8 49.2 2223 12:21 (-)0.3 Iowa 48.4 49.7 2502 12:23 2.0 12 (Critical) State Average Red Shift +2.5% Note that because of rolling updates, some states may have been relatively pure at c. 12:20-12:25, while others may already have been slightly corrupted. My guess is that most of these were still OK at these times.

Non-Critical States (35) Exit Poll Data Bush% then Kerry%, # of respondents, then time of poll ET, and "Red" Shift% Alabama 58.1 40.5 730 12:17 AM 4.2 Alaska 57.8 38.8 910 01:00 AM 4.0 (keep in mind AK time zone, this is still early) Arizona 52.8 46.7 1859 12:19 2.5 Arkansas 52.9 46.1 1402 12:22 1.1 Calif 46.6 54.6 1919 12:23 (-)1.5 CT 40.9 57.7 872 12:22 3.4 (CT 44.4 54.7 872 12:53) 0.2 DC 8.2 89.8 795 12:22 0.3 Delaware 40.7 57.3 770 12:22 4.8 Georgia 56.6 42.9 1536 12:22 2.2 Hawaii 46.7 53.3 499 12:22 (-)1.2 Idaho 65.7 32.9 559 12:22 2.6 Illinois 42.4 56.6 1392 12:23 1.6 Indiana 58.4 40.6 926 12:22 1.6 Kansas 64.5 34.1 654 12:22 (-)2.7 Kentucky 58.4 40.2 1034 12:22 0.9 Louisiana 54.7 43.9 1669 12:21 2.1 Maine 44.3 53.8 1968 12:22 0.8 Maryland 42.3 56.2 1000 12:22 0.5 Mass 32.9 65.2 889 12:22 3.7 Miss 56.5 43.0 798 12:22 3.3 Missouri 52.0 47.0 2158 12:21 1.5 Montana 58.0 37.5 640 12:22 (-)0.3 ND 64.4 32.6 649 12:22 (-)2.4 OK 65.0 34.6 1539 12:23 0.8 Oregon 47.9 50.3 1064 12:22 (-)1.3 RI 34.9 62.7 809 12:22 3.4 SC 53.4 45.1 1735 12:24 4.4 SD 61.0 36.5 1495 12:24 (-)1.8 Tenn 58.0 40.6 1774 12:23 (-)1.7 Texas 62.2 36.3 1671 12:22 (-)2.0 Utah 68.1 29.1 798 12:22 2.5 Vermont 33.3 63.7 685 12:22 5.2 Wash 44.0 54.1 2123 12:38 1.6 WV 54.0 44.5 1722 12:24 1.8 Wyoming 65.5 30.9 684 12:22 2.7 35 (NonCritical) State Average Red Shift +1.4 The following state data was obtained too late and highly contaminated with actual tabulation results: NJ 46.2 52.8 1520 12:50 (-)0.2 NY 40.9 58.2 1452 12:52 (-)0.4 NC 56.5 42.7 2167 12:48 (-)0.4 Virginia 54.1 45.4 1431 12:56 (-)0.4

Important Points To Note: The average sample size for the critical states is 2109, for the noncritical states 1192, roughly half the size. All else being equal, higher sample size correlates with smaller margin of error, greater accuracy. And yet: the average Red Shift in the noncritical states is 1.4%, but in the critical states it is 2.5%. Why? -Jonathan Simon

I actually think it is important to keep the focus on the fraud, ballot theft and suppression of the vote issues NOW, while we still can, so perhaps something can be done about it. Otherwise, not only will nothing change next time around, but it's likely to be worse. It's human nature to try to get away with things, and if that succeeds, to try to get away with yet more the next time. Who is to say the margin of voter fraud would not be greater in 2006? And even greater still in 2008? The media and of course, the Bush administration and Republicans, have largely ignored this issue until now. For one, the e-voting solution of an unverifiable black box is completely unacceptable. If we don't talk about these issues now, when people are still angry about not being able to cast votes that count (or even cast them at all) in their own country, then when is the right time to talk about it? We risk people disengaging from the political process even more than they already have - and voter turnout is key to any success we may enjoy in 2006. I believe John Belisarius stated the case quite well.

Seems to me that across the board we have too low a standard for what amounts to a call for an audit. All this 'we know Bush won' business is just unnecessary.

I agree that Bush probably won, but it would be very useful (including helping us win next time) to do a full audit, ie, recount, of the Ohio vote, for example. As with modern arrogant clueless corporations, many a bug will crawl out from under if the rocks are turned over. It's doubtful that much of the squirreliness will favor Democrats.

Why so timid in calling them out on this?

In essence what you are saying Mr. Tuxiera is that, the exit poll data which Dr. Freeman got his hands on was not sound. Not sound in the sense that it was not properly corrected for all bias. i.e. The data was not sufficiently random and the reported margin of error should actually have been higher!

Assuming sound data, I did not find any problems with Dr. Freeman's paper. All the calculations were straightforward. Mr. Bush had a 5% swing in his favour in PA and OH and FL compared to the exit polls. The odds of this happening by chance is almost zero given the confidence intervals and sample sizes in Dr. Freeman's paper.

Of course once the exit poll results are adjusted to conform to the actual count, it no longer serves as a check against the final results. At that point the exit poll becomes useless in this capacity.

If we actually get a recount in New Hampshire we will get some clarity in this controversy. The swing in Mr. Bush's favour there was 9.5 percent.

For all our sakes, Mr. Tuxiera, I hope you are right and that we just have bad exit poll data.

Ruy,

I am astounded regarding your (lack of) analysis concerning this obvious fraud of an election. I have always admired your analysis of polling, but I am totally non-plussed in your lack of same regarding the exit polling in this farce of an election.

Rather than dwell on the myriad of detail which I could write pages on, I will just pose several questions to you:

1- Do you believe Bush stole the election in 2000?
2- Do you believe that Bush/Rove would steal the election in 2004, if they knew they were sure to lose?
3- Do you realize that the Republicans fought tooth and nail to keep a paper trail from being used on the touchscreens?
4- Of course, you must be aware of the relentless effortss of Jeb Bush in Florida and the Sectretary of State in Ohio to disenfranchise the black vote?

4- Do you realize that the exit polls were correct within 1% in virtually all the states, except for a chosen few: Ohio, Florida, New Hampshire, Nevada, New Mexico, North Carolina. Here are the 4pm exit polls vs. the actuals:

These exit polls matched the actuals within 1%. Nothing wrong with these polls:
Exit-4pm Actual Exit-4pm Actual
Kerry Kerry Bush Bush
AZ 45 45 55 55
LA 43 42 57 57
MI 51 51 48 48
IA 49 49 49 50
NM 50 50 48 50
ME 55 53 44 45
NV 48 48 49 51
AR 45 45 54 54
MO 46 46 54 53
IL 55 55 44 44
CO 46 46 53 52
Avg 48.45 48.18 50.45 50.82

A 27% deviation for Kerry, .37% for Bush. Pretty good exit polling, no?

Now look at these states: The exit polls diverged from the actual vote by over 3% for Kerry and Bush.

But let's focus on OH and FL:
WI 52 50 47 49
PA 53 51 46 49
OH 51 49 49 51
FL 50 47 49 52
MN 54 51 44 48
NH 58 50 41 49
NC 48 43 52 56
Avg 52.29 48.71 46.86 50.57

It just so happened that Ohio and Florida, the two states Bush needed to win, were the ones which magically flipped for him.

Can you explain that? Is it not obvious that the Exit Polls were correct and the Votes were rigged? And if it's not obvious to you, at least you should have entertained the distinct possibility.

Yes, the Professor's 250 million to one are true odds. Just review the binmoial distribution function and do the calculation in Excel.
The inputs are:
1) 18, the total number of states.
2) 7, the number of states which moved beyond the MOE to Bush.
3) .975, the probability that just ONE state would move beyond the MOE to Bush.

That's it. Just these three input intothe Excel Binomial Function.

And the probability that 7 out of 18 states would all deviated beyond the MOE, all to Bush?
ONE out of 187 million.

I guess Bush is a very liucky guy.

Say the word, Ruy: It's F_R_A_U_D

I see nothing to lose by looking into allegations of vote fraud and suppression in Ohio. It seems that two independent candidates are already asking for a recount in Ohio, the recount rules in Ohio are more specific than those in Florida so the results may prove interesting. Having said that, I doubt that a recount in Ohio will deliver the 137,000 votes Kerry would need to become President, nor do I think it's likely that an investigation into e-vote fraud will prove that 137,000 votes were stolen. I do believe that some vote fraud will be uncovered by an investigation, as well as widespread attempts to suppress the Democratic vote in Ohio. Any vote fraud deserves to be publicised and doing so may prevent such fraud in the future. The repubs may not have affected the outcome of this election by stealing and suppressing Democratic votes, but what about 2008, or 2012? This anti-democratic activity has to be stopped and it won't be stopped by saying "Oh well, we'll get them next time". The only way to stop this activity is to expose it and to see that charges are brought against those involved. Seems like a worthy fight to me, and besides, what else do Democrats have to do right now?

I would submit that a large part of the brouhaha over the exit polls is the lack of transparency with the election process. Whether massive fraud occurred or not, it is not tolerable or excusable that a system exists where such fraud _could_ happen, not when it would be just as easy to have a transparent election system. Confidence in the democratic process is an essential part of a functioning democracy.

American citizens MUST demand full publication about WHO the exit pollsters are, and thier methods for weighting the exit polls in the 2004 election. A full explanation in, layman's terms (as Thomas Paine would say), must be published in the local newspapers, on T.V. and in as many magazines as possible. Then, if Kerry actually won, Bush must be immediately removed from office. Immediately.

Ruy:

I have no idea what's going on, and, with due respect, you seem to not have much more insight than me.

That said, the biggest problem I have with your explanation, and historical comparisons is the behavior of the networks on the election night of each of those previous elections.

In 1988, the netwroks acted like Bush won form the moment they went on air.

In 1992, they acted like Clinton won, but not by a landslide.

1996, they acted like a landslide but treated the night as a drama of whether Clinton would get 50 percent.

In 2000, they acted like it was impossibly close the whole way - except for the Florida fiasco.

In 2004, the early reactions were that Kerry had won.

2004 sticks out like a sore thumb.

My specualtion is that the exit polling was faulty - but sometihng different happened this year, that seems undeniable to me.

The problem for a lot of us isn't that there's evidence the race was stolen; it's that in many cases there's *no* evidence at all. Electronic voting without a paper trail (and we've had something like that where I vote in Tennessee for years) means subtle but meaningful manipulation can occur with no way to discover that it ever happened. It doesn't do us any good to focus on "beating them next time" if the goal posts for beating them keep being moved.

Hi, Ruy,

You make a good point, but if the exit polls have really been so skewed toward Democrats all these years, then why were all the veteren newscasters and pols so convinced by them this year? Remember Florida in 2000? All the exit polls predicted Gore winning the state, and when the votes were finally counted, that proved to be the case. It's all right to say you want to move on, but first let's investigate the polls first. By the way, Moveon.org has a petition going to investigate the polls...
Blue Jean

Has anybody looked at the possibility that the '88 election was stolen too?

"My advice: calm down and concentrate on what's really important--beating them next time."

Of course we need to beat them next time, but that's no reason not to make sure that this election was run fairly. There are discrepancies in the polling and voting data. We don't know how significant they are; the first priority should be to find out before it's too late.

I'm not talking about assuming that the election was stolen, I'm talking about proving to eveyone's satisfaction that it wasn't.

The goal isn't necessarily to prove that Kerry really won (though if he did, unlikely though that may seem, we certainly should do everything we can to prove it). The goal should be, first, to make sure all the votes from the 2004 election are counted properly, and then to fix our broken electoral system so we can have some confidence in the results of future elections.

I'm convinced the 2000 election was stolen. If we don't fight to make sure that the 2004 election (which isn't over yet) is conducted fairly, it may not matter whether we win in 2008.

I am surprised that so many "respectable" Democrats are quick to dismiss the fraud allegations.

It seems that these allegations could be settled quickly with recounts in several states that are at the center of the controversy (obviously OH and some or all counties in FL, but also states that would not change the election outcome, such as NC, NH, and NM). What's the downside in conducting such recounts to make sure that the machines tabulating the votes were working properly?

That it would cost too much money? Is it better to leave a large part of the population suspecting that their votes will never be counted?

That it would raise doubts about the fairness of the process? A huge number of Democrats believe that the machines were rigged. A recount in several states showing that no such rigging occurred would go a long way toward cementing Bush's legitimacy.

Beating them next time is important, but I want to make sure all the votes were counted correctly this time first. We need to give people confidence in the system.

I would be curious to hear Teixeira's view on the other story floating around the net, about North Carolina.

The gist is that early votes (pre-election day) lined up almost exactly with election day results for a dozen down-ticket races, but not for the Presidential and Senate race where there was a huge shift toward Republicans on the day of balloting.

Democrates are not by and large ready to submit that the election wasn't stolen. This site like others, has been infiltrated by Republicans that diguise themselves as liberals.

This election was stolen, and until we expose the problem, really expose it we can kiss any hope of winning anything in the future. All the senate seats were also stolen. The religious thing is just a ruse. The statistics behind the statement dosen't stack up. Historically, exit polls were not incorrect as the media and Republicans would like to have you believe. Further, they were only incorrect in the senate races and the battle ground states. All of which had Kerry ahead at 4pm on Nov.2!!!!

There is one primary reason Dems should require that these election returns be thoroughly checked, and all indicators of fraud be exposed. It helps to ensure the integrity of the elections next time.

There is a second reason: those of us who ARE convinced there is a problem aren't going to stop talking about it based upon dismissive edicts by riders who haven't been in as many rodeos as we have.

LBJ once said "if you're gonna be a politician, when you walk into a room, you better be able to tell who's with you, and who's agin you." Some things are known first at the gut level, and the pieces all fit manipulation in Florida and Ohio.

What possible reason would Bush NOT contrive to rig the election in those two states?

In my opinion (call it a wild guess if you like), it's possible that Bush would have won legitimately in 2004, but that the Republicans cheated anyway, just to make sure. They knew as well as anyone that the election was close, and that stealing a few thousand votes in swing states could make the difference.

It looks like Bush won Ohio by about 140,000 votes. Suppose it turns out that, say, 40,000 of those votes were stolen. Does that mean we shouldn't worry about it? No -- it still matters, and we still need to find out what happened (and punish the guilty if it comes to that).

And once we've done that, we need to make sure nobody can steal another election, and *then* we can worry about winning the next one. (Which isn't to say that we can't do all three at once.)

As someone who worked for MoveOn's GOTV campaign in Cleveland, aptly called "Leave No Voter Behind," the larger issue seems to be the perceived credibility and integrity of the voting process. This has major consequences for our ability to get people to the polls: how do you convince someone that *every* vote counts when there's widespread doubt about voter fraud? If people are convinced the election will be stolen, whether justified or not, it's much more difficult to get them to the polls; it's certainly more difficult to stand on their doorstep and tell them with conviction that their voice will be heard. Every vote should count, period. We shouldn't have to aim for a "fraud proof" margin of victory.

Like many of you all, I have been coming to this site for over a year, and I have been treated to expert and clear-headed analysis. On this issue though, I think Ruy misses the point.

The exit polling discrepancy is just one issue here. The lack of a national standard, IT security and fault tolerance, audit trail, ability to to do recounts AND the partisan hacks in charge of election certification all casts serious doubt of all election outcomes. This shadow won't go away until it is investitgated and corrected.

I don't care if counting the votes tokes a week to accomplish, thereby depriving the networks of a big election night viewership, but if that's what it take to make sure all the votes count, and take the (s)election of our public officials out of the hands of vote counters, statisticians and the Katherine Harrises and Ken Blackwells of this world, that's what we should do.

P-

I agree with Paul's comments.

The complaints being addressed by some are simply part of a greater picture, which must be addressed. Our system of counting votes is too important to leave to party hacks in Ohio and Florida, who have opportunity and motive to cheat.

We must have a national ballot for president, and it must be counted by standards that assure transparency in the counting process.

We cannot get people to the polls if there is widespread lack of faith in the counting.

Would you buy a lottery ticket if you thought your stub wasn't really going to be in those from which the winner was picked? Of course not, even if it was.

I appreciate analyses of the voting, and I appreciate rationales that say Kerry would have lost even with some votes moving over. But they beg the question.

If a company discovers someone is stealing, they find it, no matter how little the money is. Should we expect less of elections?

For anyone who tells me and those like me that we should stop worrying about the election process and focus on making Democrats more electable, I say "save your breath, I can hold two thoughts in my head at the same time."

Yes, it is quite possible that the exit polls were flawed. But...

The explanations put forth to date have not been particularly persuasive.

Explanations have been:
--Too many women. (And pollsters don't know how to weight for that? Hard to believe.)
--Too many Westerners (but there was a similar trend in all battleground states, regardless of location.)
--Too great a reliance on early returns (but the same trends held later in the day.)
--The latest: Republican voters don't want to talk to exit pollsters to the same degree that Democratic voters do. (Jury still out.)

At the same time, the U.S. is disputing the November 21 Ukrainian election results on the basis of exit polls. So, Ukrainian exit polls are valid, but U.S. exit polls are not? Or maybe the U.S. media consortium should have hired the firm that polled in the Ukraine, rather than Mitofsky/Edison?

The exit poll results from all the battleground states favored Kerry relative to Bush. Yes, that might mean that the polls are flawed. And it also might mean that tabulating machines were hacked or that other irregularities took place that produced a final result inconsistent with the exit polls. The American public deserves a careful, nonpartisan investigation of what happened, and an accurate tally of the vote. Just what the U.S. is demanding for the Ukraine, come to think of it!

Ruy Teixeira rightly cautions us about the reliance on unweighted exit poll data, but throws the baby out with the bathwater.

The various reasons advanced to date for the discrepancies between the exit poll data and the U.S. presidential results have been unpersuasive:
--Too many women sampled. (Pollsters suddenly don't know how to weight for gender?)
--Early results too heavily weighted. (But the discrepancies persisted until late in the evening of November 2.)
--Too many Westerners sampled. (But discrepancies occurred on the state level throughout the U.S.)

The jury is still out on the latest theory: Bush voters were less likely to participate in exit polling than were Kerry voters.

While fraud might not have taken place, it is appropriate to remember that fraud could produce exactly the same results we saw on November 2: exit polls widely favoring the challenger, followed by an unexpected surge for a relatively unpopular incumbent.

It is, of course, ironic, that the United States and others have been using exit poll data to question the results of the November 21, 2004 presidential election in the Ukraine. The apparent willingness to accept the Ukrainian exit polls raises a number of questions:

--Why are Ukrainian exit polls considered more accurate than their U.S. counterparts?

--Or should the American media consortium that orders up the U.S. exit polls have dumped Mitofsky/Edison for the firm that did the work in the Ukraine?

--And, most importantly, why are American concerns about the integrity of the U.S. election system being minimized? Doesn't the integrity of the U.S. election process deserve as much concern as the integrity of the Ukrainian process?

An independent, non-partisan investigation of the discrepancy between the 2004 exit polls and presential election results is urgently needed. The integrity-- and perceived integrity-- of the American electoral system is too important to let the questions persist.

Roy,

As long as you're trying to address the issue, you really need to get the problem straight. It's not that we expected the exit polls to be exactly right - it's that they were within their margin of error in most states and most contests, but wildly off for the presidential race in a few states - well outside their margin of error, and always in favor of Bush. Why weren't the polls consistently wrong, if there was a uniform shift towards Bush? Why was the margin of error apparently so far off?

You have not explained why the recent results are outside the margin of error. In referring to past races, if you want to show that exit polls are normally this far off, you need to show not just the votes and the exit polls, but the margin of error for those past exit polls.

It's comforting to be able to chalk up the claims that have been made about election fraud to "conspiracy theorists." Americans — me included — don't want to have to think that we no longer live under democratic rule. And I don't mean "Democratic Party rule"; I mean that we don't want to think we have no say in who constitutes our government.

But one thing your apologistic article didn't give me was an understanding of two things:


1) Why have exit polls historically been within a few tenths of a percentage point from being accurate, as explained in the Freeman paper you mentioned?

2) Why when exit polls are off in foreign countries — such as the Ukraine or Venezuela — does the U.S. insist this is conclusive evidence of fraud?

And, as to that last point, why do we, along with numerous other countries, insist that elections where exit polling has revealed such fraud be done over?

It seems that exit polling is good enough to determine whether or not the regimes of Third World countries have "cheated" at the polls, but when it comes to the U.S., things are different.

Your comments don't explain why.

And, if the claims are true, your suggestion — "calm down and concentrate...on beating them next time" — is a proposal to do nothing. Period. You cannot "beat them" if your vote has no meaning.

I would want for ALL of the votes to be counted. By All of the votes, I should think that the people who are out of state whether it be vacation or work, whether it be because a person is in Iraq as a marine, navy, army air force, national guard, or a doctor, a nurse, a missionary, or whatever they might be doing.... How come we have never heard about their popular vote!

I recommend that in the US, that those of us that are proud of our vote, sign it or put our thumb print on the ballot!

It is something that I am going to throw out there!