Well, well, what do you know...

[texpatriot2004]

http://whatreallyhappened.com/2004votefraud.html

[Name removed]

1. Deleted message

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

[applegrove]

2. Remember how many Dems stayed home because they thought Kerry

had won with leaks of 'exit polls'. Remember how many votes were suppressed with long line ups. Remember how only 60 something % actually show up at the polls. Remember that the religious right all showed up.

There was voter suppression. But the exit polls lied along with the people who said they could not get enough machines for some heavily democratic ridings.

Kerry did concede.

This teaches us how hard we have to work in 2006. And that as each Dem shows up to vote or wakes up on that day - they have to repeat to themselves... no matter what happens...what i hear on the news. No matter who tries to boss me around. No matter how hard it is raining or how much I cannot afford that babysitter. No matter how much I feel the diebold will take my vote away.. I have to make it to the voting booth today. No matter what.

[Griffy]

3. Kerry won.. get over it.. we were robbed.. Diebolded... had..

the machines and counting was rigged.. if you cant see that you need to do more research! Could Kerry have run a better campaign, definatly... did more peple vote for him that day, yes... did bushco own the voting machines and hold the offices needed to challenge anything.. yup ... would only a fool think bushco would play fair?...

[applegrove]

4. Very few - perhaps only Bush supporters - feel Bush played fair.

How exactly votes were suppressed remains up for discussion. But everyone agrees that we have to be vigilant the next time around.

And we need voter transparency laws.

[Griffy]

5. Kerry won.. you still talking about supression... thats not the point

the supression is an old technique, and it was used no doubt... but it wasnt enough, Kerry STILL had the numbers, so they counted the votes Enron style! It was rigged... can you say it.. Kerry won, and I was robbed!

Of course we need transparency... and what was clear in '04 was the rigged system!

You dont get it I think, or choose not to express it... Kerry Won!

We were vigilant.. we watched them steal it.. we yelled THIEF THIEF and still he was inagurated... so stop with the mealy mouthed "how votes were suppressed" crap and read! We know the basics, we know were the odd numbers are... we know about vote flipping machines and locked down courthouse for secret tabulating! WE KNOW...

[Old and In the Way]

9. Applegrove wants really, really hard to believe that Republicans

would never, ever use compiler code to manipulate votes. It wouldn't be playing by the rules to do that. As we can see, they are honest as the day is long and there is no criminal reason that compels them to rig the vote to win. It really was that close.

Remember all those banks of laptops in Karl Rove's boiler room? Were they all watching CNN?

[applegrove]

10. Studies have already pointed out that "exit polls" were wrong. You

don't think the 'leak' of exit polls favourable to Kerry convinced many to stay home & not vote?

Come on. How hard would it be to play with an unscientific sample like exit polls. Repukes could have just taken all the parking spaces in front of the polls for themselves. And passed them on to another repuke. They could have been told not to answer polls.

So many ways for exit polls to be wrong. I am not saying that Repukes are honest - I'm just saying they play the games they cannot get nailed on.

Why exit polls have been off in places like Bush Texas elections ... before diebold was invented.

The exit polls guys. Keep your minds open.

[Old and In the Way]

15. No I don't.

I think that's crap as a matter of fact. People were absolutely motivated to vote for Kerry. People were pro-active in their desire to cast a vote. Did you see the crowds Kerry was drawing? 20-30,000 people. How about the "crowds" Bush was drawing? I never saw a crowd turning out for Bush. What I saw were tightly controled close up shots that never showed the lack of people turning out.

The Republican Corporate Media conspired with the Republican Party to frame the race as close so the Republican owners of the voting machines could steal it.

I don't buy this "people heard about the exit polls and stayed home". Sounds like something a reasonable Republican would say to cover election fraud.

[Griffy]

31. OK.. NO one knows the true results.. why do you default to bush...??

we cant recount millions of votes (paperless Evoting), so ultimatly its about who you believe... I dont believe the MSM about much, and I dont care what the exit polls said, what I see is a criminal regime that uses propaganda to mislead and obfiscate the truth. They know intellectuals will work hard to evaluate the merits and give them a benefit of the doubt, and you prove them right.

The thing you are missing is trust in your gut... the feeling that liars will lie and cheaters will cheat.

I see you twisting any way you can to keep some door open that means that leads to bush won, and I cant see why.. are you trying to be fair... open minded?... if so, shall we discuss intellegent design on equal footing with evolution... after all we need to keep an open mind... and perhaps the world is flat... we need to hear both side! Balance between lie and truth cant be met in the middle!

Try explaining this.... at 11:00pm or so on election eve the lead to bush was about 140,000 with like 60% precints reporting. Now we KNOW the rural voting precints were done, closed and had reported.... but we KNOW that people in the cities were STILL IN LINE and waiting to vote. As the hours went by bushes lead barly moved.... hour after hour.. as these precints finally closed and reported the numbers didnt change... SO, that means that those lines were close to 50/50 bush/Kerry voters. SO, you tell me, go look at the video of the people waiting to vote and you tell me! You are lost in trying to make sense of that which is hard to accept, WE WERE ROBBED... our democracy is on hold since the coup of 2000. THIS IS FACT.

and how about Clint Curtis... how does that figure into your exit poll BS, it doesnt... the exit poll talk is a red herring and you are chin deep in it!

this whole thing angers me, that DUers would be bush apologists... that after all the lies and death and greed you would try to give the benefit of the doubt, using bushco talking points no less. Believe it or not thats what your doing, and you are slowing the truth...

We cant get voting fixed till we get Americans to KNOW its broken.. as in the wrong man was inagurated.. PERIOD!

[applegrove]

36. I don't default to Bush. I default to voter suppression that helps Bush.

Exit polls are not scientific samples. We need to be aware.

[Febble]

39. Intelligent Design versus Natural Selection

is, as you point out, false equivalency.

But there is another false equivalency here too: that the "debate" between ID and Natural Selection is equivalent to the debate as to whether Bush won the election by fraud.

These are not equivalent debates. You may be convinced that Kerry won. You may know it in your gut. I felt it in mine. Proponents of ID probably "know" it in their guts too. They sure as heck don't have the evidence. And I'd love to think that more Americans voted for Kerry than voted for Bush. But we need more than a gut feeling to convince anyone else. Heck, I need more than a gut feeling to convince myself.

So let's talk about evidence. I'd like to know about your evidence.

I agree the exit poll talk is a red herring. I have been saying this for months. So can you source your numbers? What was the count when, with what proportion of which demographics reporting?

And can't we have this discussion without some people here assuming that those of us who still think it possible that Bush may have won, wanted him to win? Believe it or not, it is possible for good, intelligent and informed people to come to the conclusion that while fraud and/or voter suppression may have cost Kerry votes, they may not have cost him the election - and still wish he had won.

I would be delighted to be confronted with convincing evidence that the scale of the fraud was enough to swing it for Bush. I'd like to see him impeached on any grounds. But nothing I have read on DU has yet convinced me that the case for large scale election theft is irrefutable.

{edit for typos}
Which is why I'd like to see the source of your numbers.

And I'd also love it if this debate could be about evidence, not about the assumed motivations of the posters.

[Griffy]

41. This site is full of links to evidence...

I have read many articles linked from the election forum, I also read Conyers report. I view this more like a court case... means, motive, opportunity are all in place for a rigged election. The voting machines are owned by republican companies and the software was non-disclosed. Clint Curtis(who I meet and spoke with at a rally in DC) is very clear that controling the vote was underway back in 2000.

I can not list all the evidence that is out there, and I dont expect you to take my word for it, but I promise you it is real, the election was rigged as sure as the facts were being fixed around the Iraq war policy.

Its about seeing the forest for the trees!

I dont think it is possible for an intelegent INFORMED person to think bush won. The evidence is overwhelming if you believe in statistics, not voodoo. and if you know there was fraud, but not sure if it was enough to change the outcome, then at BEST you cant say who won... but no one on DU, with all the data available, should think for a second that bush won.

I try not to assume too much about other posters, I am only intested in helping them break the lies propagated in our nation. I suggest to you the reason you even think bush might have won is because they told you he won. They posted completely unverifable number that when compared to registered voters didnt add up... in some cases there were more votes than voters, and not just Ohio!

The question I have is what is enough evidence for you... you need rove to publicly admit it? Some crimes leave only traces, but carefull analysis can still provide solid proof.


you read nothing that tells you large scale fraud is irrefutable? Read Conyers report and get back to me!

[Febble]

42. I've read the Conyers report

I even sent analyses to Conyers.

I have analysed data from Ohio, New Mexico and Florida, and I think that the evidence is overwhelming that voter suppression, at the very least, deprived Kerry of votes in Ohio (see my paper here). I think it is possible that other fraud went on in Ohio, though I am not as convinced as I was, precisely because I have looked at some of the evidence in consierable detail. I think that if fraud swung Ohio, it must have been fraud of a particular type - and was largely unreflected in the exit polls. I think that some polling factors probably primarily account for the exit poll discrepancy in Ohio.

I think New Mexico probably would have gone to Kerry if voting had been fair (see my paper here). I think that the machines recorded disproportionate undervotes in New Mexico in largely Hispanic precincts, and the margin was so tight, I think it probably swung it for Bush. I do not know whether this was deliberate fraud, although I am open to the possibility that it may have been.

I am no longer convinced by my own conclusions here. re machine fraud in Florida, although I am certainly open to the possibility that fraud occurred in Florida.

I certainly do not believe things I am told by the US media. My default, frankly, is to disbelieve it. That's why I have spent so much time since last November crunching numbers.

But I am no longer convinced that the exit poll discrepancy was primarily a result of fraud, and I therefore no longer believe that fraud necessarily occurred on a scale that is implied by the magnitude of that discrepancy. I do not think the "statistics" provide overwhelming evidence of large scale fraud. I certainly do not believe in voodoo. I do think that some of the "statistics" cited in support of the fraud story verge on voodoo. Some don't. I check everything.

[Griffy]

51. ok.. read this and tell me how its not stolen?!?! link included!!

http://www.harpers.org/ExcerptNoneDare.html

its a nice article summing up the fraud using many examples of the irregularities.

Kerry won.. bush cheated... please try to understand this, we need you out there spreading the truth. We cant afford to be wishy-washy about this and say "perhaps bush did win"... sure theres a chance he did... theres a chance Saddam buried WMDs in the sand, or shipped them to Syria before the invasion, but I dont believe these far fetched ideas till I see more proof... like the election, Kerry won and I wont give the "bush won" theory much credibilty unless more evidence is presented.

[Febble]

52. I've read that too.

I agree. There is strong evidence in Ohio of a plot to steal the election. Quite apart from anything else, Blackwell had a clear conflict of interest. Frankly, I don't think it matters whether the plot succeeded or not, the fact is that if Blackwell acted illegally then he needs to be prosecuted. I hope he will be.

Nixon tried to steal an election too. The burglary failed and he won anyway. It didn't stop impeachment proceedings and it didn't stop his resignation.

If you think the evidence is overwhelming that a plot to steal the election succeeded it, fair enough. I don't find the evidence overwhelming. I am convinced that Kerry lost large numbers of votes in Ohio through voter suppression of various kinds (check my links). I don't know whether it cost him the election, although I believe it may have done. What I am sure of is that it was completely unacceptable, and, among other things, racist. African Americans were disproportionately disenfranchised in Ohio whether through deliberate means or sheer neglect of their interests. I suspect the former. Blackwell's behaviour over the recount also makes me suspect fraud of some sort.

But my considered view is that I think it is unlikely that Kerry won the popular vote, though I consider it possible, if not probable, that he would have won Ohio had voter suppression/fraud not occurred. I'm not saying Bush won the presidency. But I'm not yet prepared to say that Kerry did.

What I am prepared to say is that the 2004 election in Ohio was a travesty of democracy, and that the fact that electronic voting does not allow your elections to be transparently audited is not only a scandal but dangerous.

[FreedomAngel82]

102. The same people who did our exit polling

did Ukraine's as well. They found out fraud from their exit polling.

[Febble]

13. How do you know what Applegrove wants?

Too many people on this forum make inferences about what people want based on what they think.

There is a legitimate argument that the exit polls were right and the count was wrong. There is also a legitimate argument that the exit poll discrepancy was due to a phenomenon that has occurred many times in many exit polls in many places - over-polling of progressive voters relative to conservative voters.

I know most people here don't believe it. But you don't have to want Republicans to be honest to believe it. I know that Bush has verifiably lied on countless occasions. It doesn't stop me finding a polling explanation for the exit poll discrepancy plausible. And it doesn't stop me thinking that the election may have been stolen by a combination of voter suppression (massive evidence, probably massive scale) and/or fraud.

[Old and In the Way]

16. How about some common sense?

Bush lost by 500,000 votes in 2000.

In the 4 years since he was selected, he:

* Took a $400BB surplus and turned it into a $400BB deficit.
* Ignored warnings on 9/11 and was criminally negligent letting 3000 people die.
* Lied about Iraq's involvement with WMD in Iraq and started an immoral war of occupation for no good reason.
* Clearly lost the debates.
* Was under 50% in polls just prior to election....undecideds always break against an incumbant.

And for this, he gets 4MM or so more votes than Kerry? Kerry got huge pre-election crowds...Bush couldn't fill an auditorium. Sorry no sale.

So what can explain? Hmmmmmm.....maybe this?


<>

I'll take TruthIsAll's statistical analysis over a half-baked theory that "people thought Kerry was going to win, so they fell asleep on the counch". Here's what you have to believe if you think Bush won:

http://www.democraticunderground.com/discuss/duboard.php?az=view_all%20&address=203x22581

[applegrove]

18. I am not saying there was not a discrepancy. I am just saying that

it could be the exit polls were fooled with. It does not necessarily have to be the machines. It doesn't. That is your preference. Many agree.

And yes of course Gore won in 2000. But - not realizing the scope of the aggressive Repuke attack - he ceeded the election in a characteristically gentlemanly way. He would not do that again.

[Old and In the Way]

20. OK, let's deconstruct this.

"it could be the exit polls were fooled with"

Would the Democrats puff up Kerry's numbers? From your early theory, that would have kept Democratic voter's complacent and away from the polls. Rather counter-intuittive, no?

Would the Republicans do it? Perhaps they wanted to make the stealing of the election even more spectacular and improbable?

Or maybe it was all those mysterious "Reluctant Bush Voters"...the ones who are too shy to tell the truth or too embarassed to take a poll. Because they were ashamed to admit that they voted for Bush. You know what shrinking violets Bush supporters are. And they certainly don't want to share their political opinions. Right.

I took my own nonscientific poll up here in rural Maine. A place where independent Republican-types are the norm. I asked, over the course of a couple of weeks, a dozen Republicans that I know, in very tactful and casual manner, "who'd you vote for"?

My non-scientific, non-random results:

2 - Overtly stated Bush
2 - "none of my business" Assumed Bush
1 - Did not vote
1 - Nader
6 - Kerry

Know whose vote was the easiest to steal? Registered Republicans.

[applegrove]

22. If you cannot follow the logic that there is more than one explanation

that is fine. But it would make great sense for the exit polls to be puffed up early in the day, leaked to the press, with the resulting message sent to harried Dems (oh - we won - forget the babysitter).

Could happen. No reason why not.

[Febble]

24. Risky, though

I know a number of Republicans who think that the early exit polls were leaked to discourage Western Republicans - "Kerry won, no point voting".

I'm not saying either happened. Just that it would not be clear what result puffing the polls would have. Could easily backfire.

[applegrove]

34. Not when Repukes had their ministers telling them they had to vote.

[applegrove]

19. Go back and make charts of the exit poll discrepancies in Texas

for Bush Governorship - before diebold came into play.

[Old and In the Way]

25. So your point is that where Bush goes, crooked elections follow?

I have no interest in analyzing his "win" in Texas. But I have no doubts that they cooked that election as well. Rove = Dirty Elections.

[applegrove]

38. No my point is that where Bush goes, crooket elections lead the way.

And a whole host of other funny stuff.

[Febble]

21. I don't know the provenance of these plots

but don't seem to bear much relationship to machines actually used in each state.

Maine, Illinois and Wisconsin are supposed to have paper ballots.

Maine also had a large proportion of optiscan ballots.
Illinois had punchcard and optiscan ballots
Wisconsin had a few paper ballots, a few lever machines, and mostly optiscan ballots.

Ohio, N Caroline, New Hampshire, Florida, New Mexico, and Pennsylvania are supposed to be "electronic voting".

Ohio was mostly punchcards with some DREs and optiscans
N Carolina had DREs and optiscans
New Hampshire had a few paper ballots and mostly optiscans
Florida was all DREs or optiscans
New Mexico was DREs
Pennsylvania was mostly levers, with some DREs, a few punchcards and a few optiscans.

If you want to compare paper ballots with other forms of voting, you have to do just that. That is not what has been done here.

And there is a problem with comparing paper ballots with other kinds of machine, as paper ballots were almost exclusively used in rural areas, and a very small proportion of precincts used them.

I agree with you that it is hard to believe that people actually voted for Bush. But those plots simply do not make the argument that fraud occurred in precincts with electronic voting and not in precincts with paper ballots. Another analysis might, but not this one.

And it does not address Applegrove's point (and mine) that exit polls can be biased. There is plenty of precedent.

[Old and In the Way]

27. I live in rural central Maine

You are correct, we had paper ballots and optiscan.

But rural central Maine has always been rock solid Republican. Maine has always been more Democratic, primarily due to the Southern Maine population that is more urbanized. The 2 have canceled each other out to a great degree. Take another look at the clear advantage Kerry had. It was in the exit polls and in the results. Look at New Hampshire's deltas....hmmmmmm.

[Febble]

28. Glad to know my information was sound!

But I've seen these plots before, and I just don't think they demonstrate that paper ballots precincts had less fraud compared to electronic voting. That could well be true, but I don't see that comparing state to state is a way of establishing the case, because of a) the range of voting methods in each state and b) the fact that paper ballots were used largely in rural areas - so there is a confound with demographics.

And what do you mean by "New Hampshire's deltas?" Sorry, don't know what you are referring to here.

I think, as I've said, that arguments can be made that election theft was attempted, and that it may have succeeded. But I think every argument needs to be examined on its merits. A lot of poor arguments don't add up to a good argument - and one good argument stands on its own. Actually I think that a lot of poor arguments can dilute the impact of a good argument. And I think there are some excellent arguments. But, IMO, these do not include the exit poll evidence (for reasons I've laid out elsewhere) and in particular, they do not include the exit poll evidence vis a vis voting method.

But I still think Ohio stinks.

[Old and In the Way]

32. Refer to the chart I posted in 13

That's way too hard to explain. I remember wondering why it took NH so long to get called election night. There's only 400K voters, I think. It was real late that night. I think they were holding it back. It was damn close in 2000. Had Gore won NH, Florida wouldn't have mattered. Lots of dirty tricks in NH, they've nailed one GOP operative with jamming Democratic phone banks.

[Febble]

40. I assume you mean 16

13 appears to be my post!

Yes, New Hampshire was odd. It was so odd it was recounted. And the recount satisfied at least one Kerry voter.

Now there may be ways in which the NH recount results are compatible with a fraud hypothesis, but it's starting to be a stretch. There are is also an factor that might account for the exit polls in New Hampshire being more than usually "off": a substnatial number of "wards" (precincts) in NH are huge

http://www.invisibleida.com/New_Hampshire.htm

- way larger than precincts elsewhere in the country. As interviewing rate for the exit polls was based on precinct size, it would have tended to be low in these large NH precincts. And we know that the exit polls discrepancy tended to be greater where interviewing rate was low (only a small proportion of voters polled) possibly because it is more difficult to get a truly random sample under these circumstances(Edison-Mitofsky Evaluation January 2005, page 35.)

So although I would not completely rule out fraud as an explanation for the exit poll discrepancy in NH, we have both an apparently clean recount and a legitimate mechanism by which polling bias could have been greater in NH to argue for polling error as an explanation for your plot for NH.

[Time for change]

43. NH recount

Please correct me if I'm wrong, but my understanding of the NH recount was that it was only a 3% sample and not random. I don't see why we should ever accept the results of a 3% non-random sample.

[Febble]

44. I'm not saying you should

all I'm saying is that at least some observers were satisfied, including those who asked for the recount, unlike in Ohio. "Invisible Ida" actually recommends that: New Hampshire should definitely be considered the model for all future recount scenarios, with their systems as standard practice.

My understanding was that they actually selected precinct where they had cause for concern - I THINK following analysis of voting patterns in 2000.

If so, the recount does seem to offer some support for the hypothesis that polls can be well outside their margin of error even when fraud has been largely excluded as a factor.

I said largely.

[Time for change]

45. I think it all depends on how the precincts were selected

You believe that "they" selected precincts where they had cause for concern. But who did the selecting? Was it the same officials who presided over the original vote count? Or, was it those who made the request for the recount? I think that makes a big difference.

But I agree that there may be something different about NH that could have led to more exit poll bias than elsewhere.

[Febble]

47. Check out the links

It's not something that I have gone into in great detail. But "Invisible Ida" claimed on 8th November 2004:

I am also advocating for a MANUAL recount of ALL 2004 presidential ballots due to an ongoing study showing problems with the software systems. On behalf of this goal, I am currently the coordinator of EFDAT: the "Election Fraud Data Analysis Team," a team of grassroots voting activists.

And she links to her study dated 5th November 2004 that appears to have found some precincts whose vote count totals are out of whack with totals from 2000.

That study recommends:

RECOMMENDATION:

At a minimum, a MANUAL (hand) re-count where paper trails are available in a sampling of each category where Non-Trending occurs is HIGHLY RECOMMENDED. No vote total should be accepted by any candidate without this basic verification of RAW DATA versus REPORTED DATA, as “assigned votes” appear to be inconsistent with projected voting patterns.

So the implication is that she was asking for something that would investigate what the 5th November study had raised as an issue.

Invisible Ida appears to be Ida Briggs, and there is more info here:

http://www.invisibleida.com/Press_Release_1.htm

And this piece:

http://www.thenation.com/doc/20041206/baker

implies that the precincts chosen were the ones that were out of whack according to Briggs's analyses.

Which seems fairly convincing, if so. Actually better than random, if you home in on the precincts you think are out of line - better, IF custody of the ballots is secure. One advantage of random is that it can be done immediately before the recount, so no preparation of ballots can be done. I don't know about custody of ballots in NH. It might be worth contacting Invisible Ida to find out. It was the weak link in the San Diego story as well. Otherwise, both studies suggest that polling bias is a real phenomenon.

[Time for change]

53. Thanks for the links

I haven't been following NH very closely at all.

It may be true that the article implies that the precincts were chosen wisely, but it's hard to know for sure. Anyhow, I'm more interested in Ohio than NH -- sorry to change the subject.

It will take me a little longer to answer your other post, but I'll do it :)

[dchill]

33. "I'll take TruthIsAll's statistical analysis..."

So will I.

Some think that somehow, what the media says is true - Bush won. The media also said Iraq was a good idea. The media has repeatedly quacked out ALL of Bush's lies, pumped out talking points as "facts" - practically inventing our Heroic War President, creating out of thin air and Bush's real record of cowardice and avoidance of responsibility an image of "Reaganesque" vision combined with a blunt-spoken cowboy macho - even though the real Reagan was far from visionary, and the cowboy is from Connecticut.

So if the media says Bush won and there must be good reason to have paperless (therefore un-recountable) voting machines and that Gore wouldn't have won anyway (even though he did), then it must be true that Kerry lost, too, because the media wouldn't lie. Unless they're Dan Rather or Bill Moyers.

No, my friends, too many real stories have been ignored or flatly labeled as "conspiracy theories" to expect a sane, rational American citizen to even toy with the idea that "Bush won." There is no way that exit polls could have deviated to Kerry in seventeen battleground states. It is far easier to infer that there was massive, multi-faceted vote fraud across the entire country than to find a way to "believe" that Bush won.

From their own real record of "achievements" it is unbelievable that the Republican party wouldn't cheat. In order to win any national-level election, they'd have to.

[Febble]

72. Just tried to find the provenance

of these plots, and found them in a Wikipedia piece. All there was was links to my own work, work by Steve Freeman on something else that doesn't mention them, a dead link, and an old BradBlog piece about Florida.

So I just deleted the paragraph with the plots. Do you have another source?

The labelling of states by voting method seems to be simply wrong.

[texpatriot2004]

91. I always like your good old common sense. Thanks for this post

I like the nice graphs and I have always loved TIA's list of bullshit you have to believe to think * won.

[applegrove]

17. And Exit polls are not scientific. Pollster have to stay 100 feet from

the site of the vote. Easy enough to fluff up the numbers as there is no way to verify that the people being polled actually voted in that voting station or that they are telling the truth when they answer pollsters questions. It is simply not a scientific poll.

So there are many explanations for the discrepancy with the exit polls and the vote.

Also - didn't the exit polls not match with votes when Bush was running for governor? I don't think Diebold was in place then.

The fact of the matter is that "leaked skewed exit polls" would do wonders to the mind of a hurried mom or poor dad when it came time to vote after work. As would huge line-ups and threats of one kind or another.

And as for the scientific polls over the phone on election day - well if only 65% of Americans vote and 100% are put into the kitty of random phone calls - you have a way the phone polls will not match with the vote.

This is not all exactly figured out. One thing for sure... we need to train ourselves and each other to vote, vote, vote no matter what on election day. Ignore "leaked exit polls". Ignore long line ups and line up. Take the day off work. Think of the people of Louisiana on the day we vote and be willing to walk through 8 feet of water (or the equivalent) to get to the polls.

We all agree there is voter suppression of one sort or another. No doubt about that. So we will just have to show up in numbers such that if the telephone polls at the end of the day say Bush only had 45% - then he only gets 45%.

[Old and In the Way]

29. Election exit polls have ALWAYS been used to assure honest results.

Suddenly, in this election, they don't hold. But we can't see the raw data, either. We have to take the Republican Corporate Media's word for it that all those results were correct. We also have to take the word of the Republican owners of the electronic voting systems that their proprietary/secret code was counting the votes correctly.

Right.

[Febble]

30. What makes you think this?

That exit polls have been used to assure honest results?

It isn't that "suddenly in this election they don't hold". The exit polls have overstated the Democratic candidate's official count for years. Maybe that means there has been fraud for years. But in that case you can't say that this time they suddenly "don't hold".

And in the UK, where we have pretty transparent elections, we can judge the quality of the poll against the official result. And our polls are NOT weighted to the result. The polls are frequently wrong. It doesn't worry us (it's good for a laugh sometimes) as we have reason to trust the count. But in 1992 it was heartbreaking, just as in 2004 in the US. The polls predicted a Labour win, and the Tories got a narrow but workable majority.

So it's not even true in other countries. And in general, it seems to be the case that it is the progressive, rather than conservative, vote that is over-stated. Maybe conservatives cheat everywhere, even in the UK. Or maybe conservatives don't like being polled.

[applegrove]

35. They didn't hold in Texas elections when Bush was governor.

[texpatriot2004]

92. That's WAY too much trust for this corrupt administration. n/t

[liam_laddie]

48. ...not scientific? So what?

Exit-polling techniques and sampling methods have been refined through fifty years, or more, of experience.
"Big business" certainly wouldn't throw away big money on polling, or market research (the two are similar) if they didn't see
accurate results. So...no, exit-polling isn't "scientific" in the narrow definition that matching results from re-tests are possible. But this has nothing to do with accurate sampling. Exit-polls are accurate to a fine degree, depending on the sample size, and have been used by this administration (I should say cabal) to question results in Ukraine and other elections. But not here? Puh-leeze - you know why Edison-Mitofsky, hired by the MSM, clouded the exit-poll waters in 2004...
I'll go with common sense, actions of participants, collected and observed evidence! Doesn't evidence become "proof" when a court, judge, jury, whatever, accepts and applies it to a verdict?
Disclaimer: I am not an attorney, statistician or analyst. Just
a keen reader and observer of the Cheney-Bush-Rove team's behavior throughout its public "career." 'Nuff said.

edit verbs and tense

[Febble]

50. You are right

that exit polls should be used to question results, especially where there is reason to doubt the probity of the election, as in Ukraine, and as in the US.

I take no issue with that at all.

But to maintain they are accurate - and that accuracy is simply dependent on sample size - is not borne out by evidence. There is strong evidence that exit polls can be "biased" ie. well outside the "Margin of Error" computed from "sampling error" which is, as you say, dependent on sample size. It happened in the UK in 1992 (we have hand-counted paper ballots, counted under public scrutiniy); it probably happened in New Hampshire in 2004; it may have happened in the San Diego parallel election.

Sampling bias, aka non-response bias, is a well researched problem in all surveys. And yes, the networks throw big money at the exit polls and want to get them right - especially after the Florida 2000 debacle. And E-M got the "right" answer this year. The polls correctly predicted the count. But they got it right not because the poll responses were an accurate reflection of the vote count proportions, but because, as in previous years, the responses were weighted to the actual vote returns.

Because all the networks want to do is predict the count - not audit the election. If you want an exit poll to audit the election, you have to design a better poll - one that does not incorporate the vote count into its predictions.

E-M did not "cloud the waters". They did what they've done before, and what they said they would do. It was all there on their website. All they claimed to do was predict the count.

The discrepancy remains suggestive of fraud. But it is not proof of fraud, and there are other well-documented precedents for that kind of discrepancy arising from polling error.

Having said that, I do believe that a) the election should be investigated (especially in Ohio) and that b) your elections should be made fair, secure and auditable. No democrat or Democrat should ask for less.

[FreedomAngel82]

101. Don't forget

EVERY SINGLE error FAVORED Bush. None favored Kerry.

[aquart]

6. I have no idea where you are getting your numbers from. None.

[texpatriot2004]

7. Must be that FUZZY MATH those Republicans are famous for.

[texpatriot2004]

8. Or maybe they just made it up as they went along. n/t

[applegrove]

12. How do you explain "exit polls" being "off" when Bush ran in Texas.

Before diebold?

[texpatriot2004]

23. You know I haven't delved into that. Shame on me cause I'm

from Texas and I admired Ann Richards but you know, she lost, somehow and I had a life. I didn't really have time or more likely, make time to follow up on WHAT WENT WRONG. How selfish is that? I just wanted to go on with my own little life and have my Democracy too. Now I see that I must be much more vigilant and skeptical regarding my Government. Tough lesson. My apologies Ann - You were doing a wonderful job as Gov. UNLIKE Bush who left us bankrupt and riddled with the wreckage he brings to the table.

[Old and In the Way]

26. Don't feel bad Tex....

Lots of us coasted through the 90s. Times were good, I was raising a family and working on a career. I was pretty damn complacent...lots of people were. Why not? We had a real President that actually believed in competency and public service. I just assumed everyone would see through the bullshit of the Republican Party and their phony political Inquisition. I wasn't expecting a Supreme Court to hi-jack the election, nor was I expecting Republican control of electronic voting systems. I got no one to blame but myself.

2000 was a wake-up call and I've become more politically aware than at any time in my life (and I graduated with a PoliSci degree in the 70s).

[texpatriot2004]

55. Wow, a PoliSci degree in the 70s...Okay, thanks for your words

here. I am a Clinton Era political person. I was born in 1969. I worked very hard for Clinton and have been involved ever since. 2000 was a wake-up call. I think something that bothers me now is that I was so naive. You know, even in 2004, even being aware of the threats of electronic voting, when people expressed to me their fears and skepticism about the possibility of defeating * I would say, "I think we can do it." How naive of me. I think many Americans are naive. You know, we don't want to believe that far-out-there-tin-foil-hat conspiracy theory stuff. Then again, it's not a conspiracy if it's true. With everything that is going on, now is the time for America to wake-up from the dangerous slumber it has been in.

PS I have a minor in PoliSci :patriot:

[applegrove]

37. Yes. Don't feel bad. I am so ready for a "fight" in these them threads

over what we know..that I ignored you.

We all assume people around us are human beings like us. So we take them at face value - even long after we should be aware. And then we fight to keep the world in our hearts the same place it used to be.

Realization takes much longer than it should.

But it is the same journey for all adults who find themselves fooled. You are not alone. Nor is your journey a solitary one. We have the ones we love. We have so much more than them..for just being human.

[applegrove]

11. Exit polls were wrong in Bush Texas elections. At times. They didn't

have diebold.

[autorank]

73. This claim is disinformation...unless you can document widespread public..

knowledge...and you can't. Look at the polls, "not for on air use...embargoed etc." The release of the polls was a nice piece of corporate whistle blowing and it compromised the entire stolen election by allowing a few key people to do screen captures; one in particular (Simon who shared them widely on the net AFTER the election). This allowed the analysis of the biggest rip off in American history since the Louisiana purchase where we acquired 1/3 of the country from the French for a song.

This is not helpful because it's disinformation. Sorry, that's the truth. So there you go...

[Febble]

79. Whoa!

Although I agree that it doesn't look as though the polls suppressed the Kerry vote (look at those lines in Ohio) the networks were certainly reporting that Kerry was ahead in the polls - you didn't need to catch that particular screen.

Dammit I nearly drank my champagne and went to bed. Then I just thought I'd stay up a little longer to be sure...and a little longer... and a little longer ... then I drank the bottle anyway.

So I don't think much whistle-blowing was involved, except maybe that screen with the numbers on it.

And eventually E-M published the table anyway (or their "call three" results, which are even more informative).

[organik]

103. No need for bickering, it's time for HAND COUNTED PAPER BALLOTS

It's good enough for several REAL democracies around the world.

Doesn't everyone agree that there were massive irregularities in the 2004 election, that can't be accounted for? Wouldn't it be better to have actual paper ballots cast for every vote, that are randomly (not ohio random, real random) audited to detect fraud.

I personally think Kerry won. Think of the Kerry vs. Bush campaigning. Bush with closed campaign stops where people had to sign a loyalty oath just to attend! He knew he wasn't going to win legitimately. And what newscaster was it on election day that said, "i don't know what happened...the numbers just...changed".

Think of wrong paper weight in ohio registrations, not enough machines, challenging voters at the polls, white stickers over Kerry ballots in ohio, documented touch screen vote switching, the list just goes on and on and on and on...

So, whether or not you believe the 2004 election was stolen, you must agree that someone (or several people) TRIED to steal it with every trick in the book. All these things favoring Bush didn't happen accidentally.

So, let's stop arguing about if or how much fraud took place and just agree that it needs to be fixed!

HAND COUNTED PAPER BALLOTS!!!!!!!!

[Cocoa]

14. your source also exposed the "Clinton body count"

http://www.whatreallyhappened.com/RANCHO/POLITICS/BODIES.html

The following is a partial list of a large number of persons who have recently met their demise in suspicious circumstances who appear to have some connection to the Clintons. I stress partial because new additions are coming in faster than closets can be found to hide the bodies in!

more...

[Time for change]

46. Here are reasons that I consider to be good evidence of fraud

This is something that I posted recently on GDP, with links to this Forum:

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=132&topic_id=2052179&mesg_id=2052179

Many of these are also discussed in the main link from this thread, by texpatriot2004.

If those who believe that Bush really won the 2004 election, such as Febble and applegrove, could tell me which if any of the 10 points that I discuss they disagree with, I believe it might help to clarify the basis of our disagreemtn.

But Febble isn't allowed to object to #7, because she did the most important research in that area. In fact, her conclusion (Correct me if I'm wrong Febble) was that the witholding of voting machines from Democratic precincts in a single county in Ohio cost Kerry 18,500 votes.

[Febble]

49. A response:

1. Secret vote counting

Don't disagree with this one at all. The fact that theft is possible is in itself an outrage. Doesn't prove it happened, but shows that it could have - in 2004 and in the future. Needs fixing as an urgent priority.

2. Evidence of vote switching machines

Again, I agree Curtis's testimony was interesting, to say the least, and the Lemme story worrying. I agree both need to be investigated.

3. The Edison-Mitofsky (E-M) exit polls

I agree (who doesn't?) that the exit poll discrepancy cannot have been due to chance. And, as you say, that "exit poll bias cannot be completely ruled out."

To address your points against this possibility:

a) The E-M report on the exit poll discrepancies, although claiming that the discrepancies are due to exit poll bias, presents little or no evidence to substantiate that claim

No, the report does present evidence. In page after page, the report shows that bias was greater in precincts where some factor which might be expected compromise random sampling, and thus allow for the introduction of bias. These, importantly, include where interviewing rate was low, and where the interviewer was stationed far from the poll. Unfortunately they do not quantify these effects, nor do they give statistical details. The evidence presented thus remains suggestive rather than conclusive.

b) The exit poll discrepancies in 2004 were greater than in any other year since E-M began conducting Presidential exit polls in 1988;

Yes, but not that much greater, and actually, according to my calcs, less significantly greater than in 1992. But I agree, they were greater. E-M offer the suggestion that years in which the discrepancy was greatest have been the years in which the electorate showed most awareness (according to surveys) of the election. 1992 had a high profile because of Perot's candidacy - and as I said, the WPE at any rate was more significantly non-zero than in 2004. But it merely over-stated Clinton's win, so possible is less well known. In the same year, in the UK, UK exit polls predicted a Labour win and the Tories won with a working majority. Fraud is an unlikely factor in the UK. Exit polls do have a history of being wrong, and seem to tend to over-state the progressive, rather than the conservative vote.

c) The exit poll discrepancy was outside of the margin of error in 5 of the 11 crucial swing states (OH, FL, PA, NH, MN) and only 12 of the other 39 states (not to mention the fact that the discrepancy was very close to being outside of the margin of error in three other swing states – CO, NM, and NV).

Not if you use the design effect. And I'm not convinced by NH, for reasons given elsewhere!

d) The only type of bias that was capable of being measured (between precinct error) was found to be in Bush’s favor, not Kerry’s

Don't understand this point. The E-M report measure two sources of error: precinct selection and within-precinct error (i.e. voter selection/vote counting). The net error in precinct selection was not significantly different from zero (is my understanding) but marginally in Kerry's favour. The net error in WPE was in Bush's favour, but the variance was enormous. Many precincts had a "blue shift" that was greater than the net "red shift".


4. Pre-election trends

Well, I was following these desperately, and hoping for a Kerry win. but I was never very confident! Not a slam dunk.


5. Election day turnout

With the polls so close, it seemed that the “ground game”, i.e. the get out the vote effort by the respective parties, would determine the winner. A high overall turnout is almost always a very good sign for the Democratic candidate in such a close race.

Yes, that gave me hope too. But I'd like to know if it's really true - and even if true, these kinds of heuristics have a way of not being true next time round.... Not convinced per se.

But I agree re Ohio. Those lines of umbrellas really made me think that Kerry had won.


6. Fraudulent Ohio recount

Agree. And I think I can take credit for the Georgia10 link, no?


7. Voter suppression


A study that looked at voting machine allocation per voter by precinct partisanship showed that machine allocation was far less adequate in precincts that voted for Kerry. In fact, it appears from looking at the scatterplot that there were about 30 Kerry precincts where there was less than one machine per 440 registered voters, while there were no Bush precincts in this category. This same study showed that “voter turnout” decreased substantially in Franklin County as machine allocation decreased. And an extensive analysis by Elizabeth Liddle came to a similar conclusion. Furthermore, as Bob Fitrakis reveals, all this happened while 68 voting machines were available in Franklin County but held back.

Yes, a similar conclusion, but I also concluded that quantification must remain difficult because of evidence that machines were allocated on the basis of past turnout, thus confounding the analysis. Even after controlling for this, I concluded that turnout had been suppressed - and in any case, this basis for machine allocation should be illegal (perhaps it is) as it effectively caps turnout at previous levels, and amounts to simple disenfranchisement of those who do not vote regularly.

8. Electronic deletion of votes from Democratic precincts – Cuyahoga County, Ohio

There are numerous strange findings surrounding Cuyahoga County in the 2004 election.

I agree, Cuyahoga stinks.


9. Electronic addition of votes in Republican precincts

Agree this is possible, and wouldn't show up in exit polls.

10. Apparently fraudulent voter registration figures

Interesting.

See, I'm not a freeper!

[Time for change]

54. Another response

I certainly would not think that you're a Freeper!

And yes, you can take credit for the Georgia10 link. An awful lot of good information there -- and she put it together in such a short time.

So it looks like you either agree with or at least express no disagreement with all my points except # 3. Well, maybe a little disagreement with points 4 and 5 -- but those were probably the weakest two anyhow.

But it seems that you basically agree with points 1, 6, 7, and 8 -- and it seems to me that those alone add up to a very good probability of a Kerry victory. But I guess you don't see it that way. I suppose that you're more cautious than I am about these kinds of things.

Anyhow, I do have some disagreement with you about the exit polls, but I don't have time now, so I'll just post this and come back to that later.

[Time for change]

56. The exit polls

I made four points against believing that exit poll bias explains most or all of the red shift. It sounds to me that you agree with me factually about points b. and d. 2004 exhibited the largest red shift (discrepancy between the polls and the official vote) since Mitofsky began polling in 1988, though there was almost as much discrepancy in 1992, as you point out. And the precinct selection favored Bush, which means that there would have to be even more exit poll bias within precincts to account for the difference between the official vote and the polls -- though as you point out the precinct selection factor was small.

I don't understand your response to my point c. I don't know what you mean by "design effect", or what that has to do with the fact that 5 of 11 swing states were outside the margin of error with regard to red shift.

Now, let's consider my first point about the exit polls, which is that Mitofsky in his report presents little or no evidence that exit poll bias accounted for the red shift. You disagree with that statement, and to make that point you say "bias was greater in precincts where some factor which might be expected compromise random sampling, and thus allow for the introduction of bias". Then you give two examples, including distance from the polls and low interviewing rate.

I have three problems with this explanation.

First, as you correctly point out, there was no information given that would allow the calculation of statistical significance.

Secondly, in the two examples that you give (and I believe that this applies to any other example that could be given), there nevertheless remains a red shift in every single category, including the ones that supposedly would be most favorable for accurate interviewing. For example, the median WPE with the smallest interviewing rate (1) is still -4.5, and the red shift for the smallest distance from the polls (inside the polling place) is -4.2. In either case, the red shift even under the most favorable of circumstance is still great enough to overturn the election if we postulate that the red shift is due to fraud rather than to exit poll bias.

And thirdly, not all the red shifts are even in the direction that you might expect to be consistent with your hypothesis that factors that might compromise random sampling might lead to bias. For example, highly educated interviewers were actually associated with a greater red shift, and I would expect that the more educated the interviewer is the less likely it would be that the interviewer wouldn't be able to conduct the interview properly.

[Febble]

57. OK

Point a)

1. I agree. The report does not allow us to quantify the effect. However, I frequently read the assertion that the E-M report offers no evidence for its hypothesis. It may not quantify the evidence, but in fact it gives details of the evidence the authors believe support the hypothesis. We certainly do not have enough information in the report to weigh that evidence. But evidence it is. If you think your sample was biased, you would predict that your results would be most biased where there was most opportunity for biased sampling, and this appears to be the case. Not all evidence in favour of fraud has been quantified either and in many ways the best evidence - the sheer insecurity of the tabulating software - has not been quantified at all. It is still evidence.

2. Yes, this is a point made by USCV, but I think is fallacious. It assumes that only one factor applied to any one precinct. A multifactorial statistical model is required to determine whether collectively, the factors cited account for all the shift. Clearly no one factor does. More homework required from E-M!

3. Good point, about the education factor. This puzzled the report authors too. It may be a genuine stumbling block. However I don't find it intrinsically implausible that more highly educated interviewers might introduce greater bias. Precisely what is required for good random sampling is an unswerving adherence to blind selection protocol. In my experience, the more highly educated a person is the less likely they are to adhere unswervingly to anything! But that is a post hoc rationalisation. I agree it is odd. It would be good to know the size of the effect, and whether education was collinear with some other factor. If the highly educated interviewers were students, and the less highly educated interviewers were sensible 50 somethings, then that might account for the effect.

Point c)

refers to calculation of MoE. The "Design Effect" has been much talked about, and is a factor that accounts for the greater degree of sampling error introduced by the fact that the voters were not randomly selected from an entire state, but were clustered within selected precincts. It can be calculated, though not by me. Steve Freeman used a "Design Effect" of 30% in his paper, and there was an extensive discussion on Mystery Pollster about the appropriateness of this figure. Its value depends on details of the survey design, and it appears that values of between 50% and 80% were computed for the 2004 election, depending on the number of precincts in each state. When the appropriate Design Effect term is used, no state is outside its MoE (I verified this myself with the screenshot data at the time).

This of course does not mean that the polls were accurate - what you also have to consider is that so many states were out in the same direction. Because, indeed, the overall red shift was highly signficant. However, with the issuing of the E-M report we can move on from this - we know that the important shift was at precinct level, and in fact the effect size was far greater when WPE is considered. However, the ranking of states now changes somewhat, when "noise" from precinct selection error is effectively removed. IIRC, Ohio moves up the list. However, I think New Hampshire, also near the top of the list, needs to be borne in mind, as I've said elsewhere in this thread. You need to go through quite a lot of hoops to account for fraud as an explanation of the large precinct-level discrepancy in NH - so the NH example, if anything, weighs on the side of polling error as a plausible explanatory factor, at least in principle. And in both Ida Briggs' NH study, and ESI's Ohio study, vote in 2000 was used as a control. In NH deviation of precincts from vote in 2000 does not appear to have been associated with fraud (recount) and in OH, deviation of precincts from vote in 2000 was not associated with degree of red shift in the exit polls (ESI).

However, E-M note that previous years in which a large red-shift was observed, public awareness of the election was high. IF red-shift is affected by public awareness of the election (presumably by affecting voters' attitude to pollsters) then you might expect greater red-shift in the states where the campaigns were focussed, i.e. swing states.

I am offering this merely as an alternative explanation. Your interpretation may well be right. But I am really just trying to explain why the exit poll evidence is not a slam dunk. And in some instances - NH, OH, and now San Diego, we do seem to have SOME evidence that "red-shift" may be a real polling phenomenon.

BTW - I do appreciate the civilized tone of this discussion!


(edited for typo)

[Time for change]

58. Swing states

Let's suppose that there is something about the far northeastern states that makes them highly susceptable to poll bias. Of the 17 states that M-E found to be outside the MOE with regard to WPE (not counting design effect), a whole bunch of them were from the far northeast (VT, NH, RI, CT, NY, DE). So, we have 6 of the 9 most far northeastern states outside the MOE. That leaves 31 non-swing states and 10 swing states. Out of the 11 additional states that are outside the MOE, 4 are from the 10 swings states (OH, MN, FL, PA). Design effect or no, these states stick out from the rest regard to red shift, no matter how they are looked at.

Sure, one can postulate that the strikingly greater red shift in swing states is due to more attention being given to those states, but that really seems like a stretch to me. Possible, sure. But the hypothesis that fraud would be more likely to be committed in those states seems a to me to be much more plausible -- especially given everything that we know about Florida, and even more especially Ohio.

And certainly you are correct that multi-variate analysis would need to be done in order to adequately assess the factors that may have produced bias in the exit polls. But my only point was that the E-M report did not show that analysis, and the univariate analysis that they did show demonstrated such weak (in the sense that even the most favorable categories showed enough red shift to swing the election) and unconvincing (in the sense that some variables even went in the opposite direction as expected) effects that I think it's fair to say that the E-M report provided little or no evidence of exit poll bias.

[Febble]

59. We may have to agree to differ!

I read the E-M report as soon as it came out. My mindset up till that point had been that there was a lot of evidence of fraud, certainly in Ohio, possibly in Florida and New Mexico, and that the exit poll discrepancy might really have been due to massive fraud. My first reaction, therefore was actually disappointment - page after page of "evidence" that factors likely to bias the polls were, in fact associated with greater bias. I am not sure what you mean by variables that went in the opposite direction - I agree that interviewer's education was an odd one but the only other thing was completion rate and partisanship which I certainly won't rehash here!

So I certainly wouldn't call that no evidence. But it made me doubly frustrated by lack of quantification - it seemed to be real evidence but without any indication of standard errors, you couldn't tell how strong it was.

Regarding the North-Eastern states - not sure what you mean by the 17 states that E-M (?) "found to be outside the MoE". I just did a search of the E-M report for any reference to the MoE and couldn't find it. Do you have a page reference? I am fairly sure they don't give an MoE for WPE, as it would be different for each precinct.

Regarding the swing state story - I plotted mean WPE by state by state margin and got this:



I think either BB or TIA did something similar. It doesn't to me look as though swing states have worse WPE, certainly regarding the means (for the full explanation see here. "Bias" is plotted with positive = red shift.

It looks as though WPE was consistently where states were bluer. It's interesting. But it doesn't make the swing state point.

If there IS a swing state hidden in there, it's a small effect. If so, it could mean fraud - but it could also be that other thing. Maybe.

Cheers!

[Time for change]

60. E-M report

Look at the "composite estimator" on pages 21 and 22.

There are 14 states where the t score is -2 or less. These states include the 6 far northeastern states I mentioned above, plus the 4 swing states I mentioned, plus NC, SC, UT, and MS. So, it's even more pronounced than what I said. Of the remaining 41 states, 4 of 10 swing states exceed the margin of error, and only 4 of 31 non-swing states. If you look at the "best geo estimator" on the same page the states are slightly different, but you get almost the same thing.

I realize that this doesn't take into account the design effect that you mentioned (for which there seems to be controversy on exactly how to measure it.) But in any event, whatever mechanism you use to assess outliers, the swing states are over-represented -- especially when you discount the far northeastern states. And the same thing applies if you use raw exit poll numbers.

[OnTheOtherHand]

61. maybe I get to poo-poo design effect just this once

Design effect is really important when someone is attempting to make naive calculations of MoEs based on sample size alone. (As you know, "design effect" refers to the loss of efficiency in a cluster sample -- e.g., drawing all interviews from about 1500 precincts instead of being able to draw a true random sample of voters.)

We don't know what assumptions are going into E/M's standard error calculations ("Estimated Error (SEDF)") in the tables on pp. 22 ff. But I would be shocked if they don't attempt to incorporate design effect.

However... when you are considering t values, you also have to consider context. The WPE in Wyoming is larger than in Ohio, so how come Ohio is on your list and Wyoming isn't? Well, the sample was a lot larger in Wyoming. Why? Because Wyoming wasn't a battleground state.

If you don't like that example, try comparing New Hampshire and New York. Similar (and huge) WPEs, and both do have t scores less than -2, but you'll notice that the standard error is much smaller for New Hampshire. That's partly (not entirely) because E/M interviewed about 30% more people in NH than in NY, even though NY had more than 10 times as many voters! Again, NY wasn't a battleground state.

In short, we should expect standard errors to be smaller in the swing states -- and, therefore, t values to be more extreme -- because of the way the survey was designed, where or not there was fraud in those states.

As a resident of New York, I'm not quite sure how one can discount it as being in "the far northeast." We border Pennsylvania and almost border Ohio. (And Delaware, of course, is actually mostly south of Pennsylvania.)

(EDIT: removed typo)

[Time for change]

62. Good point about the sample sizes --

Let's just consider the overall red shift: 4.6 in the swing states, vs. 3.1 in the other states. I can't figure out how to compute a precise p value on that, but it's got to be statistically significant, with a sample size for the swing states of almost 25,000, right?

[OnTheOtherHand]

63. Well, there's a problem with those swing states

...that I really wasn't expecting. I didn't see your list of swing states, but I figured I could find one of those almost anywhere.

So I googled, and the first link I clicked on was a CBS News article from September 22, 2004.

http://www.cbsnews.com/stories/2004/09/22/politics/main...

It listed 14 states, but it also said, "Some of the battlegrounds have left the competitive mix – for now, at least. Washington State's 11 electoral votes and Maine's 4 look solidly in Kerry's camp; while Arizona's 10 and New Hampshire's 4 look to be in President Bush's electoral coffer."

New Hampshire in Bush's coffer on September 22. How's that for beginner's luck? (I think there was a poll in early September that put Bush up 9 or so, which must have thrown them off.)

After some more random clicking, I decided to settle for the Wikipedia article on swing states. It has several lists, so I decided to try the main numbered list (19 states, including Oregon for which there isn't an NEP exit poll result) and a list attributed to the Washington Post in April 2004 (12 states). Then I went back and used the CBS list, because at least it was more contemporaneous. And finally, I realized that I had been staring at a list of 10 swing states the entire time, because the electoral-vote.com spreadsheet I was entering data into had color-coded the states. So I had four lists of swing states.

For the dependent variable, I figured I would choose between the best geo estimator and the composite estimator in the E/M report, and I used the composite estimator because its mean error better matched the numbers you cited -- -3.6 (actually, -3.62) -- and because this generally should have been the initial election-night estimate. (As the report says, "The Composite Estimator is most often the estimate used in the survey weighting process to create the exit poll analysis data during election day before the actual vote is reported." You can see this in action at exitpollz.org: check out Delaware, where the Best Geo had Kerry winning by 23.6 and the Composite had him winning by 16.5. The gender crosstab on exitpollz.org puts Kerry up by roughly 16.54 depending on rounding. http://www.exitpollz.org/cnn2004epolls/Pres_epolls/DE_P... -- although the page then perhaps confuses the issue by using a different gender weighting and getting a slightly closer result.)

So, here are the swing-state Composite averages for my four lists:

Wiki long: -3.99
WaPo via Wiki: -3.69
CBS: -3.24
electoral-vote.com: -3.63

all states: -3.62

I figured the composite estimator might obscure the effect (whatever the effect is) because it incorporates pre-election surveys, and who trusts surveys, anyway, unless they're exit polls? (Sorry, not directed at you, I've just been staring at these numbers a bit too long!) So I went back and put in the Best Geo numbers, which yield a much higher average for all states. This time I got:

Wiki long: -5.26
WaPo: -5.74
CBS: -4.22
e-v.com: -6.92

all states: -4.99

OK, finally got that to work, or to not-work. If we use the "right" lists and numbers, anyway. That e-v number is sort of impressive. I notice, however, that if I remove New Jersey (which was ultimately less close than three states not on the e-v list) and New Hampshire (which doubles as a state from the "far northeast" -- no, seriously, New Hampshire is where exit polls go to die: worst result in 1996 -- only WPE in double digits -- and third worst in 1992, not to mention a few famous pri... ahem), the e-v mean is reduced to -5.51.

So, it seems fair to say that the result is sensitive to plausible variations in definition. I certainly didn't pick out the CBS News list from late September with an eye to reducing the mean WPE (although of course any list that eliminates NH will tend to do that -- but with NH added to the CBS list, the mean is -4.89, still below the overall average).

This is maddening, because I was all set to concede the point after some perfunctory tire-kicking, and explain why I didn't think it was conclusive. (If E/M is right in conjecturing that exit poll bias is higher in high-interest campaigns, then surely it should be higher in the states where the campaign is actually taking place?) But actually, after two hours of work, I'm not sure whether red shift was really higher in the swing states.

I do know that my wife is pretty mad at me, however. So that's something.

[Time for change]

64. So, let me see if I understand what you're saying

If you use the composite estimator, the "swing states" come out with just a very slightly greater red shift than the other states.

If you use the best geo estimator the "swing states" come out with a somewhat more substantive greater red shift than the other states.

In either case, CBS' definition of swing states make them appear to have no greater red shift at all than the other states.

And how did you come up with the totals? Just add up the numbers and divide by the number of states?

[TheStates]

65. Here is what especially doesn't make sense....

In the exit poll completion results, they go down instead of up.

In the swing states, the exit poll completion rates actually go down for Kerry while going up for Bush...

In numerous states where the poll bias is supposedly strong, the exit poll completion results are getting smaller, or going down...

It would take an extremely unusual exit poll defect to produce those results.....While we know the machine counters can produce those results with just a string of code...

[Time for change]

66. What do you mean about completion rates going down for Kerry and up for *?

Completion rates aren't defined in terms of one candidate or the other. At least, that's not Mitofsky's definition. What are you defining as completion rates?

[TheStates]

67. Percentages of completion rates.....

....Went higher in the red areas and went lower in the blue areas. There's even a scientific report graph which explains it. I'm surprised nobody is factoring that into their little equation......




If the "chatty blue socities" is scientifically correct, the completion rates should be going the other way.......with more swaying towards the side of the democrat, than the republican....This unforseen discrepancy weighs more into the debate and axes the usual arguments, than anything I know of.

[Febble]

71. Loads of people have literally

factored it into their equation. It's what the USCV equations are all about here:

http://uscountvotes.org/index.php?option=com_content&task=view&id=86&Itemid=43

and here:


http://uscountvotes.org/index.php?option=com_content&task=view&id=91&Itemid=43

I don't myself happen to believe it's very relevant, although I once did, the reason being that the variance in refusal rates, which are the biggest and most relevant component of completion rates, is colossal.

See the plots here:


http://uscountvotes.org/index.php?option=com_content&task=view&id=86&Itemid=43

They may be the ones you are referring to. Refusal rate is the third plot. It shows that although you are right, and refusal rate is slightly lower at the Bush end of the plot, the slope is a long way from significant.

A counter point to this is that if "rBr" is correct it should be slightly higher, not lower or flat. A counter point to that is that if the bias in the poll was due to non-random voter selection rather than non-random voter refusal, it wouldn't show up in the refusal rates.

However, it is difficult to conclude anything from that mess. Recorded refusal rates range from near zero to near 80% throughout the range of precinct "colour".

Butit certainly isn't true to say that nobody is factoring it in. Check out Ron Baiman's equations. I think they are wrong, but we won't start that flame war again....

[TheStates]

74. Yes, those are factored in. What it shows me is a confusing mess...

You can't have refusal rates at the same time having completion rates. It doesn't work that way, and its why I'm quite disturbed with the exit poll analysis.

I think its time for the whole truth to come out, the fraud was obviously there. Just how much deliberate fraud, that needs a total reassessment and audit I believe.

[Febble]

75. I don't know what you are saying

What do you mean by "you can't have refusal rates at the same time having completion rates"?

Completion rates = 1 minus (refusal rate + miss rate).

Miss rates were lower than refusal rates, and in theory at least should be random, so refusal rates are another way of considering completion rates.

High refusal rates mean low completion rates and vice versa.

The plot shows that refusal rates were very variable, and not significantly different across the range of precinct "colour".

[TheStates]

83. Here's what it means.

You can not have high refusal rates at the same time you have high completion rates, in an entire state or city (or precinct, however you prefer).

Such a massive discrepancy does not happen because, it would defy the rules of science.

If the sample was really so ridiculously skewed, there shouldn't be higher completion rates in most of the states that have high refusal rates (specifically the swing states) and vice versa, there should not be low refusal rates, or over-sampling on purpose in the larger cities which have the lowest completion rates.

It doesn't take a scinetist to know that, any person can look at the graph and understand its not possible. What I don't see is how this can happen across at least 6 states, and buck the trend of normal deviation unless the vote counting was filled with errors. I'm now inclined to see it was indeed full of errors.

[Febble]

85. No, of course you cant!

Who do you think is saying this?

Ah - I see what you are looking at - TIA's plot.

I think what he has done (although it would be nice to have a link to an explanation) is that he has demonstrated that completion rates (or response rates, if you like) are not related to greater accuracy.

I think he is right, as I have argued elsewhere

And it is perfectly possible (probable, given the other plot I linked to, the OTOH reference) that polling bias, if there was any, was not actually due to differential response rates, but to biased sampling. This is partially supported by E-M's finding that bias was greater where interviewing rate was low - and therefore where there was more opportunity for the interviewer to depart from strict random selection. If Kerry voters were overall slightly more likely to be selected for interviewer (perhaps they tended to look more friendly, more likely to respond) then bias could still occur, but it wouldn't show up in the response rates.

So yes, you are right (of course) that refusal rates and completion rates cannot both be high or both be low - that would be like saying that states could be both strongly blue and strongly red. And you are also right that TIA's plot seems to indicate that higher completion rate was not associated with greater accuracy (although a scatterplot would have been a rather more informative way of representing the relationship), at least at state level. Precinct level analysis may or may not show something different.

[TheStates]

86. Right, and that's why I can't understand professors saying....

All is well because the exit polls are simply "inaccurate".

That is just totally untrue, and a statement of failure in our electoral process.

This is not a procedural or moral issue regarding the response samples. Even if the polling is not perfect. This to me is a legal and ethics issue, as we currently have no way of auditing the machine errors aside from precinct level reports.

Therefore, there is a most definite legal need to audit the machine results by precinct sample, or by state sample like they're trying to do in Ohio. We need to sieze whatever errors and fraud complicated the last election, if we are to permanently fix the process or save the electoral votes from an even worse calamity coming down the line.

[Febble]

87. I completely agree

The exit polls do not allow us to say that "all is well" and you definitely need to be able to audit your elections.

But, nonetheless, exit polls are "inaccurate" - though some are more accurate than others. So they can never be as substitute for a proper transparent auditable process. All they can do is set off alarm bells. And sometimes they may behave like our smoke alarms, which often go off when someone just has a steamy shower. But in lieu of something better, they are still worth having!

However, I don't actually know of any professors who have said "all is well" on the basis of the exit polls. All the professors I know have said that all is not well - even if the exit polls can't tell us much about it.

You only had to look at the queues in Ohio to know that all was not well.

Actually, you only have to look at an unauditable voting machine.

Cheers!

[OnTheOtherHand]

77. graphics check

In #71, when Febble refers to plots of which the third one shows refusal rates, I think she meant to link to this page:

http://inside.bard.edu/~lindeman/

-- see the third plot, "Exit poll recorded refusal rate by 'True' Bush vote proportion," which indeed shows no significant relationship between Bush vote proportion and refusal rates.

[Febble]

78. Thanks OTOH!

Didn't notice I'd pasted the wrong link. Doh.


(Edit to add comment)

[TheStates]

82. Rebuttal.

What it still shows, and what I can attest to is a deviation that is beyond normal in higher red areas versus urban blue areas and overall states.

This is where I see less of a correlation between what Mitofsky was explaining, and more in the area of the vote totals simply being off- there would be no other simple explanation.

Since vote counting machines do not have an open-auditable record for ballots cast, there does not seem to be any other easy way to audit, than using parallel elections or exit-poll canvasses.

And while these methods may not be lax free, they show a higher degree of concentrated uncertainty and obvious fraud or counting errors, than do the other arguments for defective polls.

That's what needs to be unearthed and documented instead of speculated.

[OnTheOtherHand]

89. I'm not sure that I am parsing this correctly

"What it still shows, and what I can attest to is a deviation that is beyond normal in higher red areas versus urban blue areas and overall states."

What is "it"? Are you saying that the deviation between the exit polls and the official counts was larger in red areas and states? That seems to depend on where and how one looks. (Upthread Febble posted a graphic that seems to indicate that there was more red shift in blue states.) I've spent a lot of time trying to figure out what it all means.

I'm also not sure what you mean by "concentrated uncertainty and obvious fraud or counting errors."

But I agree that with the DREs, we seem to be stuck between pretty bad auditing (exit polls and parallel elections) and no audits at all, which is not much of a choice.

[OnTheOtherHand]

76. I think that's more or less it

assuming that the electoral-vote.com list of swing states is more or less the one you had in mind (CO, HI, IA, NH, NJ, NM, OH, PA, VA, WI). (With respect to the composite estimator, I would say that even "just a very slightly greater red shift" verges on hyperbole for the e-v list -- about a hundredth of a point.)

Yes, these numbers are raw state averages. I spot-checked weighting on sample size for the e-v list and the best geo estimator; both the e-v average and the all-states average decrease, but the gap is about the same, and the effect of removing NH and NK is about the same.

Truth be told, if we are trying to discriminate between fraud and bias in the exit polls, I would think that if anything we would want to weight the pre-election surveys more heavily than the exit polls. So the fact that any apparent "swing state effect" is smaller with the composite estimator tends to infirm a fraud hypothesis.

But these are aggregations, and while one might gain statistical power by aggregating things, one might also miss the actual signal. Ohio's exit poll error was not far out of whack with other states' (it's within one standard error of the national average, using either estimator), but Ohio was certainly close enough that vote suppression and possible vote rigging could have made a difference.

[Time for change]

80. That list of swing states doesn't look right to me at all

It's missing FL, MN, NV, and MI. I suppose it's possible to legitimately leave out NV and MI, but no legitimate list of swing states could leave out FL and MN. And if NV and MI is left out, what is HI, NJ, and VA doing in there?

[OnTheOtherHand]

81. I was juggling so many lists...

I didn't even notice the absence of Florida. But it is missing for a specific reason, now that I think about it. As you may know if you followed the electoral-vote.com website during the campaign, it categorizes states on a seven-point scale from strong Kerry to strong Bush. It relies on a simple, objective, but volatile rule: look at the margin from the last available state poll. States with less than a five-point margin were categorized as "barely Kerry/Bush" (or "exactly tied"), and those are the states that show up as swing states in the spreadsheet. The last Florida poll that e-v had processed on November 2nd (Opinion Dynamics, 10/31) showed Kerry up 49-44, so it wasn't considered a swing state. And yet there was another poll at the same time, by e-v's rules, that showed the state tied, and one that came in a bit later that showed Kerry up 2. So certainly it should be on their list. Overall, according to their database, there were eight Florida state polls released on 10/31 or 11/1, and they ranged from Kerry + 5 to Bush +8. The average is Bush +1, obviously not robust.

Florida had a best geo error of -3.9 (about a point below the national average), a composite error of -4.3 (about 0.7 points above the national average).

Minnesota is a bit more complicated -- Kerry had the upper hand in most of the late polls (except for one from a Republican firm, and a slightly earlier Mason-Dixon poll) -- but certainly looks like a swing state. It had a best geo error of -10.8 (well above the average), and a composite error of -5.3 (1.7 points above the average).

(Weirdly enough, while several NJ polls gave Kerry double-digit leads, several others had it much closer -- and the very last survey to come in had it tied. Kerry won by about 7 points.)

If the presence of a swing-state effect depends on using the best geo estimator and including Minnesota (which Kerry won by over 3 points), well, that's a mess. But Florida deserves some attention even though Bush actually led in the exit poll (both estimators).

[TheStates]

84. What's off about this system here...

Is that the swing states especially, seem to exhibit much higher completion rates in areas where the refusal rates are supposed to be highest. And the lower completion rates towards the opposite trend, seem to go into areas and urban cities where oversampling is supposed to be magnified.

Nothing about these swing-states make sense, unless you take into account the result that the vote counting was full of errors. If not fraud than certainly a massive pattern of lost vote tallies, and of course secluded fraud in the larger districts.

[OnTheOtherHand]

90. I'm quite lost

Tfc and I have been discussing whether the exit poll disparities are greater in swing states than in other states -- basically, I think it seems to depend on the state.

But I'm not sure how one could tell whether the swing states "exhibit much higher completion rates in areas where the refusal rates are supposed to be highest." What information do we have about completion rates in various areas of swing states? and where are refusal rates supposed to be highest?

[texpatriot2004]

93. Thanks for your top 10 list about why we were robbed. You did

a great job here.

[Peggy Day]

68. I'm losing my family and some friends over this

because I can't believe people won't do anything about this. I refuse to act like all is well. These elections were stolen right under our own eyes. The fucking media just closed down shop the day after the elections. I'm pissed at Kerry, because he seems to be in on it (as well as some other so-called Democrats-I did appreciate Edwards effort) It is now so frustrating to be able to prove (with all the excellent work being done), yet still no one does anything. It goes on and on. Right now I'm watching c-span Ms Greenhouse is testifying about Halliburton. This kind of stuff has gone on and on. I sometimes want to move to another country, but everything (money) is tied up in this way of doing things. I don't know anyone in another country either! When I try to get involved (besides the numerous phone calls and emails I make), all the organizations want to spend their time on things that won't count if the votes are rigged. Don't they get it? I'm not that smart, surely they see this too.
Sorry, I don't mean to be so negative, but no one wants to hear about it. Everyone is correct; if this kind of stuff happened to Republicans, the shit would hit the fan.

[TheStates]

69. Kerry isn't in on it....

He's fighting very hard too, check recent articles. Kerry is to appear before the court August 26 2006 about the false recount and other charges underneath the ongoing criminal investigation.

[Peggy Day]

70. I either called or emailed Kerry about this (for him to please continue)

and hopefully his skills will come through, but I have been very discouraged as of late.
My best case scenario is:
Somehow (with Repub power don't know how) everyone understands that this was all rigged since 2000, kick all the bums out, put them in jail, and put in their place the real winners (ex:Max Cleland). I can't reconcile Gore vs Kerry, but Gore didn't get a chance -he has a lot of class in my opinion. Maybe Gore with Kerry vice (I don't like Lieberman one bit). Clinton could have some important function (he is working on something good now). All the laws that were passed illegally will be null and void.
And we can once again be the great country we are-something like everyone lives happily ever after. The end.

[TheStates]

88. Have you seen this?

http://www.nvri.org/troubled_elections.shtml

Blackwell to be deposed in court this year. Whatever people did, it must have been something right. "Are we having fun yet?" is the words people need to remember, because the statement is clear....


Leaving the backdoor to voting machines open is like leaving your own home's door completely unlocked or wide open so the neighbor can walk in and steal anything they like.

In a democracy, its completely illegal and should be banned.

[texpatriot2004]

94. Hey thanks for this info from the Nat. Vot. Rghts Ins. I had not seen

it. Thanks for the post.

[Peggy Day]

96. Thanks for the article.

Totally agree.
I've been a bit impatient, and discouraged lately, and it helps to know I'm not alone. Hopefully the window of opportunity won't close.
Is anything going to change in 2006? When would people ever take to the streets like they did in the Ukraine?

It gets down to the media and the vote counting. I'm asking for fair elections and a responsible media (we own the airwaves-at least we did). And a free internet-very important.

Also, remember when someone said something on TV and the other side had the chance to rebut it? Wasn't that knocked out with Reagan? That should be brought back too.

I have a lot of family in Toledo, Columbus, and southern Ohio. There's a big political difference between Toledo and say Gallipolis. It just seems like everyone would want a fair election process. Maybe a Democrat could override the cheating. We could end up with a Japanese teenager for president (just kidding).

Anyway, I hope we can start having fun now!!

[kster]

95. I think there are some people

in this thread that are paid to try and keep the million dollar vote stealing machines,in operation. LET ME REREAD THE THREAD AND SEE IF I CAN GUESS.

[texpatriot2004]

97. I think you are right. n/t

[OnTheOtherHand]

98. I think if you are gonna slime people anonymously...

you might at least state whom you are sliming.

I don't see anything in here that indicates that anyone is defending the machines, much less that they are getting paid to do it.

[petgoat]

99. This continual hectoring about "proof" and "smoking guns" gets

frustrating for two reasons: the system is designed to leave no proof
and anybody who watches "NYPD Blue" knows that people are commonly
convicted of capital crimes on the basis of motive, means, opportunity,
a bit of circumstantial evidence, and lack of an alibi.

The important new point from the "Dieb-Throat" report is that ONE person
can subvert the Diebold tabulators.

Exit poll veracity can never be determined until an expert commission is
permitted to examine the raw data under non-disclosure agreements.
Of course a year after the fact there's been plenty of opportunity to
tamper with the data.

At some point a coverup becomes evidence of guilt.

[OnTheOtherHand]

100. beg pardon?

If you are defending the posters who accuse other DUers of working for Diebold, then with due respect, we don't have much to discuss. Your standard of "circumstantial evidence" is in that case far too low.

I think you will find very few people in survey research who think there is an "exit poll coverup," whatever else they may believe about the 2004 election. If you have any idea how to prove fraud from the exit poll data, by all means share it.

(EDIT: I mean, really, what are you trying to say? Febble used the word "proof" once in this thread, in #50 above, and it is a pretty special reader who can interpret that post as "hectoring." No one used "smoking gun." It would be instructive to know whether there is anything in particular that you disagree with, or whether we are just engaging in some abstract shadowboxing.)