Is the "Exit Poll Debate" a "Waste of Time"?

[anaxarchos]

In order to find out, I did an exit poll of the exit poll debate.

Before I begin, I want to freely admit that my exit poll of the exit polls was entirely unscientific, largely unintentional (I was actually looking for where TIA went to), and had a margin of error (MOE) of at least 75%.

My method was simply to search 3 or 4 of the leading Web search engines for various search combinations ("Election fraud", "Exit Poll", etc.) and to add up the hits. What I found was that there were between 700,000 and 1.6 million hits on Democratic Underground. Between one third and one half of those were on election fraud of some type (this very forum). Two thirds to three quarters of those involved the exit poll debate. And, the lion's share of those (> 80%) centered on TIA's analysis.

By way of comparison, the Wall Street Journal and NBC News achieved approximately one fifth as many hits as TIA, FOX less than half as many and the Washington Post, about the same on precisely the same topics.

In addition to quantity, I found quality. There are literally tens (and maybe, hundreds) of TIA and DU exit poll compilations in cyberspace. The quality of some of these is striking. For example, there is this one by PI/DU poster DrDebug:

http://p216.ezboard.com/frigorousintuitionfrm9.showMessage?topicID=47.topic

There is this by PI/DU poster byronius:

http://www.truthisall.net/

Byronius modestly "fesses up" here:

http://www.progressiveindependent.com/dc/dcboard.php?az=show_mesg&forum=120&topic_id=81&mesg_id=83

And, there are many, MANY others...

Now, I am painfully aware that there are 10 or 12 people on this board who don't like TIA or the "exit poll debate" and I am also aware that my "poll" is pure whimsy (although it is significantly more accurate than those run by CNN). So before I get bombarded with objections, let me cheerfully renounce my crap poll. Let's say I am off by a factor of 10 or even 100, that only one fifth instead of 5 times as many election fraud links on the web point to TIA versus the Wall Street Journal.... how bad is that?

My conclusion?

The exit poll of the "exit poll debate" offers incontrovertible proof that DU, this forum, the exit poll debate, and TIA have been a caustic irritant which continues to get under the skin of Republicans, freepers, and fundies everywhere. With a little luck, some day it may be a full body rash.

To the last, I say:

Remember, if it itches... don't scratch!

[applegrove]

1. By "exit poll debate" do you mean Diebold? Cause usually it is

called the Diebold debate.

There are more than one issue with the "exit polls". One is they are used to prove Diebold stole votes.

Two is they could have been easily jerked around, and show that the polls themselves were used to encourage voters to stay home (they were leaked across the country on the internet).

Diebold itself creates a wedge. Cause people only believe in that depending on how much they mistrust George Bush & the GOP. So it smashes our Democratic tent. And we don't really know what the truth is and will not find out - so that is a wedge that will keep paying off to the Repukes.

If the exit polls themselves were jerked around - and indeed the fact that some studies have shown them to be off in places where no diebold machines existed - then we need to train Democrats to "resist the temptation to stay home - despite having to call a babysitter - on election nights - because someone leaked a poll that said DEMS ARE WINNING".

So there are many issues involved in the whole debate.

I find your use of the term "exit poll debate" very vague. If people mean Diebold - they usu sally say Diebold.

And the debate does not benefit anyone but the repukes. It creates apathy (my vote will not count), anger (how dare you not agree my vote & yours will be stolen), which together will suppress the vote of Democrats.

It does not help us fight Repukes. It helps us fight amongst ourselves. Transparency in elections is one the the President & his party will purposely not be working on to solve.

[merwin]

3. I think he means what TIA was dealing with.

Proving that the exit polls were, in fact, accurate.

[applegrove]

5. And some proved they were not accurate. It is a big mess the

whole election thing. And not something we have specific answers to. We should keep our minds open. Cause something is up.


We do know however that there was voter suppression in other ways. So we can agree that there needs to be voter transparency laws and the like and laws specifically directed against any form of voter suppression with quick & easy ways to prosecute that.

[petgoat]

20. "some proved they were not accurate"

Who proved that? You mean the Mitofsky "Chatty Dem" hypothesis?

I thought that was completely discredited by USCountVotes. The fact
was, voters in Bush strongholds were more likely to participate in the
exit poll survey than voters in Kerry strongholds were.

[autorank]

26. Here you go...this is a great summary thread with direct links

http://www.scoop.co.nz/features/usacoup.html

OR look at this, something that just came to my attention

http://www.indybay.org/uploads/preelectoralpolls.pdf

[applegrove]

46. In the second paragraph they claim that 104% of Bush voters voted

In the second paragraph they claim that 104% of Bush voters voted compared to 100% in 2000. Well that right there is a lie. If you remember that Bush was the poster boy for DUI the week before the 2000 vote you will remember that a percentage of the Christian Right stayed home and were unhappy. A small percentage. But there you go.

This guy assumes that there is something wrong when the percentage of voters is over 100% in the next election. That makes absolutely no sense since the turnout in most AMerican elections is about 65%. The lowest in the free world.

The assumptions so far rival Donald Rumsfield's on troop requirements for the occupation.

I have read two paragraph's and found two glaring errors of logic. I will not read any further.

Thanks for the heads up.

Like I said - I like my information to be of a high quality.

[anaxarchos]

6. The Iraq War debate is worse...

... according to your criteria. That "creates a wedge", "smashes our Democratic tent", results in "apathy" and "anger", and "helps us fight amongst ourselves".

Come to think of it, so does every "debate" (torture, Roberts, Katrina, etc.).

You are right. We should never debate anything. There might be Democrats who disagree and that would be "divisive".

[autorank]

9. You know what he means...exit polls.

Diebold is discussed in this forum and in General Discussion. If you're so supportive of that area of research I encourage you to go support this thread. The thread provides a demonstration concerning the security problems of Diebold tabulators and also mentions the fact that a large portion of the votes in this country are tabulated by a foreign owned concern, Sequoia, which has significant market presence.

You say of the exit polls, "they could have been jerked around" and then a few lines down you say "If the exit polls themselves were jerked around". You introduce a question, treat it as though it were a fact, and then follow up with another question treated as though it were a fact. Your argument is unsupported.

Exit polls were very accurate in the last two presidential elections. The firm of Edison Mitofsky was highly reputed. Take a look at this site and come back with specific criticisms. As of right now, you're just spreading information as fact when you offer nothing more than rhetorical questions.

[autorank]

13. Site to review statistical evidence below

:dem: http://truthisall.net/ :dem:

[applegrove]

15. i'm old fashioned. I need consultant and all.

[autorank]

27. Well, I'll consult for you...or you could review this excellent summary

resource of links...



http://www.scoop.co.nz/features/usacoup.html

and/or

check this out, something that just came to my attention

http://www.indybay.org/uploads/preelectoralpolls.pdf

[applegrove]

32. No. I studied as an economist and forgot a thing or two about stats.

I do know that unless the study is done properly and tested and tested for issues - they simply don't mean anything.

So - I'm going to have to go with the people who are doing regressions on this topic and do so for a living. It simply isn't enough to know the basics of how stats are to come to any conclusions.

I only have so much room in my head. There is all sorts of stuff out there on the net. Some of it amazing. Some of it not exactly what it advertises. I'm just going to have to stay judicious.

Sorry. I'll try reading. But I'll need to find out who they are first.

[applegrove]

14. I'm not into Diebold. And the firms in California said they were off.

So don't say it is no substance to my question.

[autorank]

22. There is substance to questions about Diebold and other voting machine cos

We agree. I disagree with the dismissal of exit polls as a key elements in "the hunt for presumed fraud"(this is an after the fact phrase based on all the *co corruption of the last five years). In CA, the Dem Secretary of State had to resign for some personal reason, I think, and was replaced by an Arnie appointee. Even the Arnie appointee de certified Diebold after tests where their vote count was approaching 20% inaccurate. We agree, they have highly political management and a lousy product. Paper marked ballots stored forever is the only way. We can do that immediately or any time up to about 30 days (printing requirements) prior to any election. The Boards of Elections have not had much scrutiny. They gripe about hand counts. They can all quit and we'll replace them.

[kster]

2. So these people are paid to get it wrong?

At 5:00 p.m. Eastern time on Election Day, I checked the sportsbook odds in Las Vegas and via the offshore bookmakers to see the odds as of that moment on the Presidential election. John Kerry was a two-to-one favorite. You can look it up.

People who have lived in the sports world as I have, bettors in particular, have a feel for what I am about to say about this: these people are extremely scientific in their assessments. These people understand which information to trust and which indicators to consult in determining where to place a dividing line to influence bets, and they are not in the business of being completely wrong. Oddsmakers consulted exit polling and knew what it meant and acknowledged in their oddsmaking at that moment that John Kerry was winning the election.

And he most certainly was, at least if the votes had been fairly and legally counted. What happened instead was the biggest crime in the history of the nation, and the collective media silence which has followed is the greatest fourth-estate failure ever on our soil.

Many of the participants in this blog have graduate school educations. It is damned near impossible to go to graduate school in any but the most artistic disciplines without having to learn about the basics of social research and its uncanny accuracy and validity. We know that professionally conceived samples simply do not yield results which vary six, eight, ten points from eventual data returns, thaty's why there are identifiable margins for error. We know that margins for error are valid, and that results have fallen within the error range for every Presidential election for the past fifty years prior to last fall. NEVER have exit polls varied by beyond-error margins in a single state, not since 1948 when this kind of polling began. In this past election it happened in ten states, all of them swing states, all of them in Bush's favor. Coincidence? Of course not

http://www.huffingtonpost.com/theblog/archive/2005/05/b...

[anaxarchos]

7. I saw that too, kster (the bookies)....

... and came to precisely the same conclusion.

[autorank]

12. Excellent!!!

This is the real world. The sports book cannot afford to be wrong. They only get taken down in the movies. Was this their big mistake. I don't think so. These folks are also wired and privy to all sorts of good information we have to fight for. Great post.

[OnTheOtherHand]

17. huh?

Anyone who went to grad school and learned about the "uncanny accuracy and validity" of social research should demand his or her money back -- although maybe these folks could learn a thing or two from the offshore bookmakers.

Your link is broken, so I couldn't check what it had to say about the 1948 exit polls.

I don't happen to have a table of error margins by state in previous exit polls. But as TIA pointed out, in 1988, one major national exit poll (actually, TIA presented it as _the_ national exit poll) had Dukakis beating Bush in the popular vote. (Bush won the official returns by 7.7 points.) So I fail to understand the basis of your confidence.

TIA, I think, was arguing that the exit polls had been getting steadily more accurate until 2004 -- but AFAIK he never claimed that the exit polls had always been accurate.

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x341940

[petgoat]

21. What's the significance of 1948? And yes I know "Dewey WIns!"

Were the exit polls inaccurate in 1948?

The pre-election polls were inaccurate because they
stopped taking data six weeks (relying on my memory)
before the election.

Using the failure of the 1948 polls to impugn modern
scientifically-conducted polls is not reasonable.

[OnTheOtherHand]

23. fair question -- if you look back

at the post to which I was replying, it says, "NEVER have exit polls varied by beyond-error margins in a single state, not since 1948 when this kind of polling began."

AFAIK there were no exit polls in 1948. I'm pretty sure the rest of the assertion is just wrong, although as I said, I don't have a table that would let me tally how many state exit poll results have been outside the margins of error.

Yes, one reason the pre-election polls were off in 1948 is that they stopped too soon. (According to the Roper Center database, there was a Gallup-AIPO poll, apparently fielded October 7-12, that showed Dewey up just 5 points -- in retrospect one wonders why they wouldn't rush to field another poll.) The 1948 results also spurred some methodological improvements, as I imagine will be true of the 2004 exit poll results.

I agree that it would be silly to use the outcome of a 1948 poll to impugn modern polls. I hope it is clear in context that that wasn't my point. However, it does seem worth noting that even very old polls can illustrate some common fallacies. The 1936 Literary Digest poll -- a pre-election poll, to be sure -- had over two million respondents, for a nominal margin of error of, umm, I think that's 7 hundredths of a percentage point. The editors sagaciously wrote, "We make no claim to infallibility. We did not coin the phrase 'uncanny accuracy' which has been so freely-applied to our Polls. We know only too well the limitations of every straw vote, however enormous the sample gathered, however scientific the method. It would be a miracle if every State of the forty-eight behaved on Election day exactly as forecast by the poll." (http://historymatters.gmu.edu/d/5168 -- I don't know if the quirky capitalization is accurate.) The poll gave Landon 57% of the two-party vote; he got 37.5%. Someone could have some fun with that P value.... Granted, we wouldn't consider the Literary Digest survey as a "random sample" at all -- but there never was an opinion survey in which random sampling error was the only source of error.

OK, where were we? I like modern polls very much. I just don't think that the 2004 exit poll provides much reason to think that the election was stolen. Whether one thinks the election was stolen or not, I don't think the exit poll sheds much light. The 1988 exit poll that put Mike Dukakis ahead was also a "modern scientifically-conducted poll," as were IIRC at least two others that gave Bush the edge. Those 1988 polls weren't junk, but obviously at least one of them was quite wrong. As Febble suggests in #16, the 2004 exit poll may provide some useful clues about where and how fraud may have occurred. But as far as I can tell, most people who work with surveys -- people who also like modern polls very much -- just don't think that the observed exit poll discrepancy is beyond the realm of plausible error. They could all be wrong, but I'm certainly not intending to disparage their work when I suggest that they may be right.

[Awsi Dooger]

50. Sorry, but that's just not true

How did you check sportsbook odds in Las Vegas? Gambling on non-sporting events is prohibited in Nevada. In fact, you can't even wager on so-called sporting elections like the Heisman Trophy or MVP awards. It's a high hilarity myth that "you can bet on anything in Las Vegas, honey." I've seen couples confidently walk up to sportsbooks and try to wager on the Oscars or Emmys and been miffed when the ticket writers have to restrain a chuckle. The restrictions are severe and the new breed of corporate bean counters who run the sportsbooks are heinously afraid of taking a bet and have eliminated all but the meat and potatoes. I've lived in Las Vegas since the '80s and worked in sportsbooks and for companies that help establish the odds.

Moreover, I've wagered bigtime on political elections since 1996. I have offshore accounts at all the biggies. My friends have accounts at dozens more when you total them. We keep in touch so very little escapes our attention. You need those edges because wagering is a brutal business, trying to reverse the house edge. My neighbor Paul stays home all day checking one offshore venue after another and playing small advantages.

At 5 PM Eastern Time on November 2, EVERY offshore wagering outfit was closed and had been closed for hours and hours. Many of them closed the presidential wagering the night before. That is standard. I checked Ladbrokes, Olympic, Pinnacle and Cris and none of them were open Tuesday morning. These outfits realize they are vulnerable on special events like political elections and don't want to take risks. Bookies thrive by not taking risks.

Bush was a 5/9 favorite on election eve at the major offshore books. That's bet 9 to win 5 or -1.80. Kerry was NEVER FAVORED ALL YEAR. It ranged from -1.40 (7 to win 5) on Bush early in the year to as high as -2.60 (13 to win 5) during September after Bush's convention speech and his surge in the polls.

There is one kernel of truth to what you are reporting. But it was hardly extremely scientific. HA! Further from that would not be possible. It was identical to what happened here, a whirlwind overreaction to early exit poll numbers. There were a couple of man-to-man sites, Tradesports for one, that stayed open for betting long after the polls opened. Again, that is completely standard. They are mere portals and the bettors essentially take on each other, not the house. Likewise, Bush was favored throughout the year at those sites by similar margin to offshore books. There is also the Iowa Futures Market. I'm not sure when it closed. It's also man-to-man.

I was working GOTV in Nevada on election day. During an early afternoon break I took a peak at DU and saw the glee regarding early exit polls, but didn't have time to take a close examination of them. I contacted my friend Paul and he wagered on Kerry for us at Tradesports. Bigtime. The odds soared in Kerry's favor within minutes of our bets. He became virtually a 1/2 (bet 2 to win 1) favorite, as you stated. Paul told me later he checked all his offshore accounts trying to get down more on Kerry, but each of them had closed the presidential wagering long before the exit poll numbers were leaked. Bush closed as substantial favorite at every single one.

Then in late afternoon I came home and realized the disastrous truth within minutes. I looked at the early exit numbers and saw looney tune examples like up high double digits in New Hampshire, up 10 points in Pennsylvania, down only a handful of points in North Carolina. Those were blatantly ridiculous given the historical tendencies I've charted and studied for nearly a decade, going back to 1988 with the analysis. Then I slowly peaked at the vital states and collapsed onto my bed in a seizure of resignation. Kerry was "up' only a point or two in Florida and Ohio. Immediately I knew the Bush global nightmare would continue. Even as DUers were celebrating our certain win and reporters were touting Kerry's lead in the exit polls, I turned off my computer and TV. Ohio and Florida were certain to deviate in the same direction as the states where Kerry's numbers were obscene. There was no way the supposed one and two point bulges in Ohio and Florida would hold up since the exit polls were obviously flawed. Everything was tilted in Kerry's favor.

Here is one paragraph I found a minute ago on Mystery Pollster, verifying the scenario I outlined above regarding the Tradesports man-to-man betting site: http://www.mysterypollster.com/main/2004/12/so_why_were_the.html

"I made $150 betting on tradesports election option contracts (I bet on specific states and that bush would get at least 260 electoral votes). There are people on tradesports who bet amounts in the many thousands of dollars on the election. When the exit poll results were leaked on drudge and slate, as everyone knew they would be, the contracts on bush winning fell drastically, and the kerry options rose."

[OnTheOtherHand]

54. I might just mention

that that last paragraph you quote wasn't written by Mystery Pollster. You made absolutely fair use of it IMHO -- I just don't want anyone to get the impression that Mystery Pollster made $150 betting on Bush. He gets enough bad raps around here.

[Chi]

55. Are you saying there were no odds given...

by the oddsmakers in Vegas, on the election?

[wtmusic]

4. yes

1) it's over and makes Democrats look like sore losers
2) exit polls are unscientific and more prone to fraud than the election itself
3) I have been over the statistical analyses you mention many times and the methodology is seriously flawed and biased.

Next year, let's make paper ballots the issue, and ignore verification schemes which carry not a damn bit of weight anyway

[autorank]

10. Three strikes

1) What if we're not losers? We certainly have huge questions on that score when you look at the cumulative evidence: systematic voter suppression in key areas of the country; highly vulnerable voting machines and central tabulators made by a few corporations owned by extreme right wing companies; all the strangeness in Ohio and select other states; the exit polls; and the manifest willingness of those accused of violating domestic and international norms of behavior, regulations, and laws. If that's not enough to raise some serious questions for you, I don't know what would...video taped confessions by Rove et al.

2) "exit polls are unscientific and prone to fraud" Two sweeping statements. If you mean "unscientific" in the sense that these measures, like all other behavioral and social science measures, are done in the real world rather than the laboratory and hence lack the design integrity of lest say a physics or chemistry experiment, then the answer is you're partially right. The term for non-laboratory science, in the biz, is quasi-scientific research. This is an open admission that much of what's done lacks the purity of laboratory research characteristic of the hard sciences. Nevertheless, the methodology is rigorous, scientific in structure, and the results of the research are verifiable and replicated when you get it right. I don't think that's what you meant. On the "fraud" issue, a little evidence. You mean the polling companies stack the deck in order to look bad. Why would they do that? In the case of 2004, the three polls through 1:00am were real polls. In this instance, there was an alteration. While the 13600 approx. respondents in the three initial polls had Kerry winning, the final adjusted poll had Bush winning. The initial polls were re weighted with the final vote count in mind, a subject of great debate in some detail. Was that fraud? I wouldn't use that term, maybe politically correct in the current environment is better. Nevertheless, you must have some evidence for your assertions. Why don't you share it.

3) I assume this is the material you reviewed www.truthisall.net . Please share the results of your extensive study of the statistical analysis you mention.

Paper ballots are the issue but nobody will go to paper or verifiable paper/electronic systems unless there is a reason. We happen to have a great reason. The damn election sure looks like it was stolen.

[wtmusic]

18. We're not losers

1) I believe, as you apparently do, that it's possible the election was stolen. Nothing will ever, ever happen about it, except perhaps there will be a footnote in history books fifty years from now. The reason has a lot more to do with a timid congress than anything in law or a dearth of evidence.

2) What makes you think that polling data is less immune to stacking the deck than the actual election results? At least state election boards are ostensibly impartial. Gallup, Harris, Zogby and Pew are corporations with shareholders who demand profits and operate with zero oversight. What if Kerry had won the election, but Harris came back with exit polls which showed Bush winning? Would you be willing to concede defeat? I sure the hell wouldn't (based on that).

3) I didn't do any extensive statistical analysis at truthisall.net. It would be an extensive waste of time. Maybe you don't remember, but TIA presented numerous posts before the election crowing about a 99.9999% chance of winning. This started months before the election based on *current* data, taking no account of the idea that future events could alter the outcome. That's not only statistically irresponsible but asinine, and showed that TIA was far more interested in cheerleading than actually creating some meaningful data. If the posts achieved anything, of course, it was making DU activists rest on their laurels, confident in their sweeping victory.

"Lies, damn lies, and statistics".

[OnTheOtherHand]

19. zowie!

Man, that was fun to read. But I have some minor differences with it.

(3) You speak my mind about the basic pitfalls of TIA's projections, although I don't know if they actually harmed anything. He has a new one that is pretty funny where he fits a fourth-degree polynomial to Bush approval data in order to project how it might decline by next February -- umm, OK, my idea of "funny" may be sorta esoteric. It definitely seems like cheerleading to me. But anyway, TIA isn't here, so whatever.

(1) It's really unlikely that Bush will get kicked out of office for stealing the election, even if someone presents objectively compelling proof that the election was stolen. But if that happened, it would probably create a lot more support for radical election reform. (So I don't wanna say "Nothing will ever, ever happen about it....")

Meanwhile, presenting the best arguments for even thinking that the election might have been stolen, or could have been stolen, makes a good case for election reform. Exit polls can be a piece of that effort. But I agree with you that a lot of the exit poll arguments haven't been good ones. Oh, do I agree with you about that.

(2) The corporations you mentioned don't conduct exit polls, as far as I know (although I'm not sure how the LA Times poll operates). So there's one less thing we have to worry about, sort of.

Actually, I think the basic situation there is that the media corporations that pay for the exit polls want to be able to say fairly early who won and why. So they really want the exit polls to be "accurate," i.e., close to the official returns. Of course they also want them to be fast and cheap. (Fast, cheap, good: choose two?)

Anyway, I believe it's true that one well-placed hacker could have messed up the entire exit poll and it would be almost impossible to prove. And the exit polls are subject to many other kinds of errors, too. I won't get hung up on "unscientific," but they're certainly error-prone.

When we get stuck debating "which is right, the election returns or the exit poll results?"... well, we're toast. The pollsters don't think their polls are inherently more accurate than the election returns, and if we can't even convince them, how far is the argument gonna go?

[autorank]

24. Losers lay down...IMHO

1) "Nothing will ever happen about it..." Yes, that's true if that continues to be the prevailing opinion. No oncer there is a lack of confidence in Democrats when we lay down, I'm talking about the leadership, and take it two presidential elections in a row.
2) Polling dada is not immune from 'stacking.' I mentioned how the three national exit polls were "re done" at the end and changed fundamentally. Because one bank gets robbed, you cannot infer that all banks are robbed. Harris and Zogby, I believe, are privately held. Corporate polls would be those commissioned by the networks, ets. Non corporate polls would be those from privately held firms. There can be differences.
3) OK, you didn't look at TIA's work, as you implied, but then you accuse him of doing bad work. Go look at the pre and post election work or stop the generalizations.

The election in 2004 was a mess. The 2000 election was a mess. Look at the GA Senate election, mess. How many messes do we have to lay down and roll around in before we decide enough is enough.

[autorank]

28. You want a variety of statistics, narrative, etc. check this out

It's a summary in chronological order and it's got many voices, not just DUers...



http://www.scoop.co.nz/features/usacoup.html


and/or

check this out, something that just came to my attention

http://www.indybay.org/uploads/preelectoralpolls.pdf

[wtmusic]

33. thanks...will do nt

[Stevepol]

30. Bravo Autorank. Statistical sampling (exit polling) is about as scientific

as you can get. The universe is based on it. The structure and disintegration of atoms takes place according to statistical principles. That is, you know that a certain number of atomes will break down under certain conditions but you don't know which ones will do it. The percentage is always about the same. It's about as scientific as you can get and it's part of the fundamental law that undergirds all of life.

As for the polls, a lot of the criticism is aimed at the pre-election polls, which are admittedly less accurate than the exit polls. I don't know how you could have a better way of testing the accuracy of the election results. As Steven Freeman notes in his first ms about the discrepancy between the exit polls and the alleged vote in 04 in some of the key states, in Germany exit polls are used to determine the winners of the election until the actual winner can be verified in a couple weeks. They use paper ballots and take a few weeks to complete the count evidently. I doubt that the exit polls in Germany have been wrong 1 in a hundred times (about the margin of error in other words)

Exit polls are the best and right now about the only way of verifying the accuracy of votes that are counted 80-85% by electronic voting machines that either cannot or are not audited to check for accuracy. And exit polls are exceedingly accurate. If they're not, why aren't they? They're scientifically and mathematically true in principle and if there's some social or political reason for why they might in one case or another have some weakness, let's hear why that's so, not just a bunch of allegations that they're not trustworthy. It's just nonsense to argue they're not accurate, and I suspect it's freepereze used to squelch the truth. Just as well deny the validity of mathematics itself.

So don't believe any of that crap that denies the validity of exit polls. If they're wrong, let's investigate the election to see which one was wrong, the exit polls or the vote count, because one of them was damned sure wrong.

[OnTheOtherHand]

31. exit polls are obviously not that accurate in the United States

unless you consider TIA a freeper?!

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x341940

Like it or not -- whether it's true or not -- the view that the exit polls, not the election results, were wrong in 2004 is absolutely mainstream (in fact dominant) among survey practitioners and analysts. E.g. Ruy Teixeira of DonkeyRising, since he doesn't seem to have gotten flamed much here yet:

http://www.alternet.org/story/20530

Teixeira was writing very soon after the election, so you're certainly entitled to think that he wasn't taking account of all the evidence. There's no hint that he has changed his mind since -- and this is a fellow who spent most of the run-up to the election going through surveys looking for every sign that Kerry was going to win. Mark Blumenthal was more even-handed about it, but I don't know anyone who was calling him a Bush apologist on November 1st. The whole freepereze thing is really, really unfortunate, IMHO, unless one is actually setting out to alienate most people with survey expertise.

Maybe both the exit polls and the vote count were wrong. Working for transparency in both past and future elections is important.

Sincere question: would it be worth my time to write a shortish article explaining why survey folks themselves generally don't consider the exit polls convincing evidence of fraud in 2004? This is for me one of the most bewildering disconnects in the debate -- almost as if someone announced that doctors must have discovered a cure for cancer because, after all, doctors are scientific, so if anyone says they haven't, probably they are trying to cover it up. I am always happy to try to explain stuff. If there's an article that you think has made the single best case that the exit polls prove fraud, I am happy to respond to it, but I won't bother unless you actually care.

I mean, don't you ever wonder why most of the people who insist that the exit polls prove fraud are mathematicians, or political economists, or geologists, or humanities professors, or one is not quite sure what -- and a lot of them are very, very smart and good people -- but the survey folks are few and far between? Again, I'm not saying to assume that the survey folks are right, but it seems to me that if most of them don't agree with your interpretation of survey evidence, that might be a riddle for you.

[anaxarchos]

25. You'll have to bear with me....

...because I am a little frustrated by your post.

The fault is entirely mine because I created the title to this thread in the form of a question. That seems to have made it a lightning rod for everyone who DOES believe that "the exit poll debate" was a "waste of time" (a poll of a poll of a poll, so to speak).

What I should have titled this thread was: "The exit poll debate was NOT a waste of time BECAUSE it touched a breadth and number of people which no one in their right minds who posts on Internet boards has any right to expect AND became one of the most often linked arguments as to why the 2004 presidential election was stolen".

But, I still don't get your post...

You could have said "The Internets don't matter" - Many have.

You could have said, "I don't care how many people were touched. Debates are divisive" - Someone up thread just did say that.

You could have said, "Shucks, I wanted to argue with TIA but, despite the fact that he posted here 5 times a week for a year, somehow I missed him" - It's never too late. Here he is: http://www.progressiveindependent.com/dc/dcboard.php?az=show_topic&forum=120&topic_id=168

You could have said, "Your 'poll' is crap. You might have overestimated the reach of the debate by a factor of 100" - OK... Guilty! So what?

You could have said "He should have exclusively argued for 'paper ballots', instead" - Why don't you? I googled on you (and paper ballots) and got 3 hits...

Instead, you say "I think the exit poll debate was a waste of time".

What do I do with that? Tally it?

OK..... it's 158,289 to 1.

[wtmusic]

34. is vs. was

Quite a difference. At one point in time, there was virtually nothing more important. Our reps dropped the ball.

[mestup]

8. How can we ever get an accurate vote, doing it in just ONE day.

That probably worked fine when the US population in 1800 was under 6 million. But we're pushing 300 million now.

Jeez, you get 'til April 15th to file your taxes!

Why are we pretending we can tally an accurate vote from just one day? Exit poll or not, the whole process is a sham.

[autorank]

11. Right, why not have it extend over a weekend. Would be great...

Give a Friday off, and Saturday/Sunday. Awesome. The idea has been suggested. The goal of some, mostly Republicans, is to create barriers to voting: keep elections on a work day; no mail in ballots; no morot-voter (easy registration with your license); separate voter ID cards (hard for the indigent and poor to get); etc. They used to have "literacy" and other tests to register voters but those are gone thankfully. They don't give up so good ideas, like yours, are hard to sell.

[Febble]

16. It depends what you want to achieve

if you want to include it as circumstantial evidence to back up harder evidence of malpractice, then further debate is a waste of time.

So if your purpose is to trigger an investigation - my vote would be leave well alone. Leave it as a suggestive footnote.

However, if you want to use the poll evidence to help focus an actual investigation, then I think there may still be some juice to be squeezed out of the debate. My own reading of the exit poll evidence is that it is evidence against some kinds of fraud in some places e.g. vote-switching fraud in Ohio, which means that the thing to look for might be ballot destruction in Dem precincts (for which there is good evidence) plus possibly ballot multiplication in Rep precincts (for which as far as I know there is no evidence yet, except that the obstructed recount etc is suggestive of malpractice).

But that runs the risk of getting the whole thing up people's noses, and in any case I think the the exit poll data is too coarse-grained to help much.

BTW I think your poll is entirely scientific. If you want to know who is linking to what on the internet, a Google search is the right methodology. Your conclusions seem sustainable to me.

[autorank]

29. Great Resource for Polling Debate....PDF and it's worth reviewing

:dem:

Using Pre-Election Polls to Check the Validity of Final Election Polls
Click here for PDF download of the article

or
http://www.indybay.org/uploads/preelectoralpolls.pdf

Tim Lohrentz,
Mathematician, Researcher, National Economic Development and Law Center
http://www.nedlc.org/index.htm
:dem:

[byronius]

35. Intense. And, in a sub-mathist response to Febble and Anaxarchos --

I found this forum the day after the election. I read it whenever I can, and generally wish I could contribute more to the effort and debate. I've read enough to know the main personalities, and the main body of evidence reasonably well. Autorank's sharp and astute replies, Febble's cautiousness, OTOH the Devil's Advocate, TIA the Absolute, the Irascible and the Probably Correct.
This was a perfect crime, except for the exit polls. A group of people, with access to the circuitry, and a dedication to some lesser ideal than the America I understand and love, has done this thing. They thought it out carefully. They coordinated. They ran scenarios. Not once did they step back and consider their actions as anything less than necessary to the survival of the Ur-America, the true basis of which we can all now so clearly see in the behavior of FEMA, and their blockade of New Orleans.
Anyone that can even barely get the tip of their teeth around any clue they left behind gets my admiration. They were thorough, dammit. Thorough and cunning and full of thinking that makes me sick at heart. AND THEY ARE STILL THERE, AND STILL ACTING.
I first saw that bar chart showing the differences between exit polls in electronic-versus-paper states that someone posted the Day After, and knew it, from a lifetime of reading history and political science, to be fact. Even a mainstream knowledge of modern American history must eventually refer to the people willing to commit these acts, the ideology-over-liberty crew. They exist, they now count the votes, the exit polls defied them until the 'adjustment', and they ruthlessly continue to attack and suppress almost any effort to investigate the facts.

I saw MY Senator, with tears in her eyes, as if she held the weight of the world on her shoulders, and could not bear it --

Whatever works. Whatever we can do. And the Exit Polls are, at this point, the only real clue to what WE ALL KNOW AS FACT. Except, for, of course , the Terminally Sophist.

Just my take. I love you guys.

[Febble]

36. cautious, hmmm, I suppose so....

Although I disagree with you about the perfect crime. I think there was a very messy crime, and digging around Ohio has already revealed a lot of the mess. More digging may reveal more.

What I am cautious about is reading too much into survey evidence, because I know only that survey evidence is also messy.

Here's my take:

I am confident the election was corrupt. I am confident that Kerry lost many many votes through disenfranchisement of all sorts, some of which may well have been deliberate fraud, and some of which may have been structural neglect of the rights of some largely Democratic voting groups, particularly African American and Hispanic voters.

I am confident that evidence for the above is robust. "To the extent that" (weasel words) that the above was criminal, I think the crime was not a "perfect" crime at all, except in the sense that when crime/neglect becomes endemic, it's easy to get away with murder because nobody in power gives a damn. Look at NO.

But the evidence for a "perfect crime, except for the exit polls" remains, IMO, weak. I think the crime was messy, and so were the polls.

"Voter suppression" may well, IMO, have cost Kerry the election. I think it's far more "proveable" than fraud, and may yet cost more Democrats more elections unless it is fixed, regardless of whether DRE's are outlawed or random audits made mandatory.

"Felon" purges; rationing of voting machines based on turnout; discriminatory issuing of provisional ballots; differential ballot "spoilage", and even carelessly designed ballots, are all ways in which a president can be elected who is not the citizens' valid choice.

The siren call of perfect crimes with perfect polls is sweet - but misleading, in my cautious view.



(edited to add this link to a new post on DKos by John Conyers:
http://www.dailykos.com/story/2005/9/19/83255/0297 and his action page here: http://www.johnconyers.com/)

[mgr]

37. Exit poll vs. vote suppression problem

I think you are touching on the basis of my criticism, that I think the exit poll debate helped to prevent timely evaluation of how great the vote suppression was in Ohio. If one argues, as the exit poll argument seems, that all precincts and BOEs were subject to fraudulent vote switching, then no baseline can be obtained by which the magnitude of suppression can be evaluated against. It is essentially nihilistic.

As an example, if one compares the voter turnout of Cleveland (53%) against Cincinatti (59%), almost 50,000 votes are lost; include the calculation you made for Franklin, and the total jumps to 100,000.

Factor in the ~15,000 undervotes from Cleveland, (93,000 total for Ohio?)and one can visual the Bush margin melting away. If the pattern targeted minorities, then it is prima facie illegal.

Mike

[Febble]

39. What do you mean by "targetted"?

Suppression clearly affected minorities more. Does it have to be deliberate to be illegal?

Rationing machines on the basis of turnout in 2000 meant that precincts with low turnout in 2000 had a greater pool of active voters to flood the polls in a high turnout year - and turnout was lower in predominantly African American precincts in Franklin County.

Is there any chance of prosecuting the BoE in Franklin County?

[mgr]

40. It follows the environmental justice standard,

it does not have to be deliberate to violate their civil rights, it only has to do so. I'm not an attorney, but apparently case law on this developed from New Mexico. Things must be applied equibly.

I'm surprised no one has considered it since last year. My guess is that there needs to be stonger legal proof, or it has a zero sum effect that the democratic party is aware of. It is easy to demonstrate with nominal categories, but it's hell with incremental, and it may be because the event is unique, and pattern requires history.

The thing that struck me is that other African American communities, other than Cleveland and Columbus, report long lines (Toledo and Akron in particular). However, my recent discussion with OTOH regarding the EIRS database has me looking at the records askance. The answer will probably rest within decisions between the BOE and SoS.

Mike

[mgr]

41. Elaboration

The case law from New Mexico pertains to having hispanic-americans running the local government but putting into place policies that still violate the civil rights of hispanic-americans.

mike

[OnTheOtherHand]

42. since you mentioned my "name," let me just say

I don't want anyone to think that I am trying to discount long lines in Toledo and Akron -- honestly I don't know much about them. The only thing I pointed out was that the EIRS database has many, many reports attributed to "Delaware County" that are quite clearly Franklin County reports.

Lucas County (Toledo) has 10 long-line reports which appear to be in five unique precincts, but of course there are sure to have been more that weren't reported to EIRS. Summit County (Akron) has 49 long-line reports -- I didn't try to count the uniques, but there are many, although certainly there are some repeats. (In both cases there may actually be more long-line reports; I am relying on the EIRS codings via pull-down menu.) It's nothing like the weird problem with Delaware County.

My mom worked with unions all over the state of Ohio, and I don't want anyone to get the impression that I'm pooh-poohing vote suppression anywhere in the state (or anywhere else, for that matter, but especially there).

Certainly when the Supreme Court in Bush v. Gore shut down the Florida recounts, it wasn't alleging intentional discrimination. But I'm not a lawyer either.

(Edit for missing words and clarification on long-line reports)

[mgr]

61. Two things

1.) I'm certainly not implying that you are pooh-poohing long lines as reported in the EIRS database. What I am looking at requires that the EIRS database be as accurate as possible, and frankly, I now don't trust the filing or classification schema. Delaware County is the obvious issue, but if you consider that a report of possible vote switching from Colorado is not included because the name Kerry is not mentioned, we start to lose our focus.

2.) I'm not sure of your reference to BvG. So bear with me. The Supreme's decision was based upon the equal protection clause, the fourteenth amendment, and this applied to the fact that equivalent state wide standards for how a manual recount was to be conducted, and what constituted voter intent, were not in place, and the court doubted that they could be in place by the deadline to certify the vote. The thing that makes your statement confusing to me is that the class of voter suppression, specifically targeting minorities, that predominantly vote democrat occurred in Florida as well, separate from the 'suppression' of not counting all undervotes.

What I think may be up with Ohio is that if one estimates the number of african american voters who could not vote due to long lines is possibly equivalent to the margin between Kerry and Bush. The fact that greater numbers of undervotes appear to correlate with this pattern is another element that needs consideration--and, I would suspect falls under BvG. I assume this is where your statement is pointing.

Mike

[OnTheOtherHand]

69. OK

1) Nope, I didn't even mean to hint that you were implying any such thing. And yeah, the EIRS database is kinda dirty, although still useful.

2) My point was only to suggest that voters are entitled to equal protection regardless of discriminatory intent. Bush v. Gore is such a mess that I shouldn't even have mentioned it. But certainly, if equal protection could be applied in the way the Supremes applied it there, it "should" apply to the huge differences in wait times in Ohio.

Of course, even if the Supremes accepted that analysis, that in itself wouldn't fix the facts on the ground.

[autorank]

38. Carter-Baker: What REAL Voting Rights Activists Have to SAY!!!

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x394051#394102

[foo_bar]

43. only if you "expect results"

Psychic hotlines usually bear a disclaimer, "For Entertainment Purposes Only". If you present numerology in that context, then it's a perfectly valid hobby, like Astrology or Civil War reenactments.

how bad is that?

Or performance art.

"In the future everyone will be world famous for 15 minutes."
- Andy Warhol

[anaxarchos]

45. Bugs you.... don't it? n/t

[foo_bar]

48. just enough to write a reply at 2am

Watch that reverse psychology stuff, or I'll recommend this thread.

[anaxarchos]

67. f_b, why are we on opposite sides of this thing?

You're basically a contrarian and you get what I'm sayin'...
So why, huh?

If you are going to pitch the "abstract truth" thing, save your breath...

But if you can REALLY tell me, I'm curious...

What's our beef?

[foo_bar]

71. it's because Febble reminds me of my mum

Or maybe because I fell for the TIA hype in 2004. I could "visualize" President Kerry, "believed" it was in the bag. A couple naysayers asked TIA to pipe down before 11/2, something about Pollyanna optimism becoming a self-negating prophesy on election day. But they were just jealous of Shadow President Kerry, boo-yeah!

So it's one part projection ("don't be as gullible as I was"), displacement ("Bush sucks, let's go kick the family dog"), and of course transference ("not in front of Mommy!")

Group hug?

[anaxarchos]

75. Nah... you got the wrong guy for hugs.

But thanks for the explanation.

[Chi]

56. Great analogies.

Exit polls equate to Psychic hotlines and Astrology....

So that would make Warren Mitofsky... Dionne Warwick, or Jeane Dixon?

[foo_bar]

57. "the exit poll debate", not "exit polls"

Q.E.D.

[Chi]

58. The exit poll debate, is based on the accuracy of exit polls....

So which is it...
Warwick or Dixon

D.U.H.

[foo_bar]

59. excluded middle: fallacy or way of life?

I wouldn't compare you to either person.

[Chi]

60. Nice retort.

It's about as appropriate as your analogies.

8)

[foo_bar]

62. I differ with your analogy of my analogies to my retorts

</infinitely recursive flamewar>

[Chi]

63. No flamewar...

I just pointed out how ridiculous your analogies were...
Defend them, or not, it's up to you.

"Psychic hotlines usually bear a disclaimer..."

Remember?
Frankly, I can't see how a bright person like you has not distanced yourself from the analogy by now.
I would have done it on my first response.

[foo_bar]

64. I must have missed the post where you pointed that out

I read some kind of Mitofsky:Miss Cleo strawman, but that was debunked two posts ago.

[Chi]

65. I'm sorry...are you defending the analogy, or not?

(in my best Rob Schneider voice>
'C'mon little guy, you cang do it'

[foo_bar]

66. absolutely

I'm waiting for that rebuttal from the Chi-man. The Chi-meister, makin' copies!

[Chi]

68. Wow

That must mean you think of anyone who believes the exit polls, as a valid indicator of possible fraud, is akin....to someone who believes in the validity of the psychic hotline.
That's a pretty derogatory assessment.

I'm surprised you feel that way about so many of us here.

(Kudos on the Schneider lines, I got a good chuckle out of that)

[Bill Bored]

44. The real question is this:

Since Nov. 3, 2004, what ELSE could you, TIA, and the others at the top of the hit parade have been doing to advance the cause of election reform instead of analyzing the exit polls, or the exit-poll poll? And what effect could THAT work have had in the grand scheme of things?

For example, let's say TIA spent as much time auditing precinct tallies to rule out central tabulator fraud, a la the Bev Harris/Howard Dean/as seen on TV/Diebold GEMS hack? How many counties or states do you think could have been fully canvassed by now?

And...suppose all those arguing with TIA, reading his posts, questioning his numbers and assumptions, posting those little "Atta boy TIA!" posts, or "kick nt"s, or those "I don't understand a word of this crap, but it sure looks like the election was stolen!" posts, or thread nominations, etc. were ALL doing precinct canvasses to find the dreaded GEMS hack that could have swung the election for sure if it were used without anyone bothering to check for it by adding up those precinct totals! IMHO, that's a LOT OF ELECTION REFORM!

Do you think we'd have all 3000 counties audited by now? Certainly those in the swing states. After all, there were only 3 or 4 states that made the difference, right?

And of course, this is just one hack that could have been used to steal the election, so I'm not saying we'd know the whole truth by now. But at least by devoting all those exit-poll person-hours to REAL election reform, we might be a lot closer by now.

After all, Dieb Throat said this was "the greatest threat to our democracy!" Not that we should be worrying so much about the exit polls.

Surely this is a better way to judge whether the exit polls stuff is a waste of time than just looking for Google hits. What do ya say Anax?

[anaxarchos]

49. I say I respect your posts, Bill....

... and your sincerity but you presume a level of agreement which we do not have. You might as well tell me that working for Kerry or Animal Rights is the way to achieve "REAL" reform.

You assume that I give a rat's ass about Bev Harris, the Diebold GEMs hack, Curtis, Madsen, or paper ballots. I don't. You further assume that I depend on some guy who works for Diebold to tell me what "the greatest threat to our democracy" is? I don't do that either.

I believe machine fraud occurred in 2004 (although I still can't gauge the extent or exact nature of it all) but this is the one case in which I agree with the NRA: machines don't steal elections, people do.

I think "the greatest threat to our democracy" is Republicans. I think that they have created a truly "big tent" starting with Nixon's southern strategy in '68 and Reagan's Western strategy, which relies on combining every rotten constituency in America with every rotten idea that has ever been available here. I think it was inevitable that they would eventually start stealing elections and by golly they have. In 2000, before DREs, they used registration suppression and ballot "spoilage". Before that they used felony disenfranchisement, registration policy, vote suppression,.... you name it.... and they still do.

It follows from this that I don't think the problem is just a criminal problem to be addressed by cops, or a technical problem to be addressed by programmers, or a juridical problem to be exclusively solved by lawyers. I think it is a democratic problem to be solved by "the people". That's why I think pointing continuously to the stolen elections and explaining why you think they were stolen matters. That's also why I think "hits" matter, although this thread was largely done tongue-in-cheek.

If you were to ask me what I thought the most important things to do now were, I would say (in this order):

1) Fight to preserve the Voting Rights & Civil Rights Acts of 1964 and 1965.
2) Support Congressman Jackson's Voter's Rights Amendment
3) Start pointing at the anti-Democratic nature of the current state of voter registration, voter suppression, local voter jurisdiction, and felony disenfranchisement.
4) Keep pounding on the stolen election.

But, hey, that's what I think. I have no idea what TIA or autorank or anyone else thinks because I haven't asked them.

....but, then again, I'm not pushing one agenda as the only way to achieve "real reform", am I?

[Febble]

51. FWIW

These are my views almost exactly.

My only difference with you is in respect to (4) and then only in the sense if you are going to pound on the stolen election, I would pound on the hard evidence of shenanigans in Ohio (including voter-suppression) rather than the very soft exit poll evidence because IMHO pounding on the exit poll evidence is simply presenting anyone who wants to oppose your excellent reforms with a ready-made straw man.

Because whether TIA is right or whether he is wrong (and of course you know that I consider that on many important counts he is wrong) the fact is that smarter and better informed people also do not think that the exit polls point to a "stolen election" in TIA's sense, and these include people who have done and are doing sterling work to achieve your other goals, especially your (3).

Walter Mebane for example.

[Febble]

52. Oh, and I'd add one more thing

the other threat to democracy is the lack of free and independent news media.

Voters need good information to make an informed choice.

So I'd add a (5) - or better a (1) - keep blogging.

[Bill Bored]

77. This is where you argument doesn't hold water IMO.

Here's why:

You listed the pounding thing last, and that's what you say is the value of the exit poll debate. OK, even if you meant to put it higher up on the list, I happen to think that your items 1-3 are much more important than your item 4.

The ironic thing is that items 1-3 have NOTHING to do with Exit Polls. In order to be counted in an exit poll, you first have to be allowed to enter the polls and cast your vote. And this is exactly the problem that you are addressing in items 1-3. So the exit polls will NOT help you achieve or even argue for those goals. Voting rights and vote-counting rights are like apples and oranges!

Exit polls have only to do with vote counting, and so does the Bev Harris hack. Those who are EPTBs, including TIA, have said here many times that because the exit polls did not correlate well with machine type, the hacking MUST have been done at the central tabulators! So if this conclusion is truly believed by the true believers, why are they STILL having the exit poll debate instead of spending their time canvassing precincts?

You see, in order to believe that the exit polls are correct, you have to believe that the central tabulators were hacked. But this is something that can be proven by precinct canvassing. So if you are an EPTB, you should be canvassing precinct totals instead of debating the exit polls.

Like you, I also think there are more effective ways of stealing the election than the particular hack in question. But according to those with true belief in the exit polls, hacking of the central tabulators is all it would take. But that can be checked by canvassing precinct totals if those who are so convinced of the idea would allocate more of their time to investigating it.

Again, it's all a matter of person-hours and how they are spent and the excessive debate about these polls takes time away from any form of election reform, including that which the exit polls themselves indicate is necessary!

So, while I agree that your priorities 1-3 for election reform are EXTREMELY important, and again are NOT quantifiable at all by looking at the exit polls, I fail to understand why anyone would apparently spend so much time on analyzing the polls instead of working on items 1-3, canvassing precincts to prove that the tabulators were hacked (as may be indicated by the exit poll discrepancies) and discovering other ways that the vote count could have been manipulated.

That's why I think the exit poll debate can be a waste of time.

All exit polls all the time makes Jack a dull election reformer. And what we need are sharp ones!

[anaxarchos]

80. Bill, you continue to presume way too much...

1) I DO spend MUCH more time on 1 to 3. I just don't do that here. Look at my post count. I have less then 500 posts (in a year). I had started precisely 1 thread previous to TIA's banning (now I've done 4 or 5). Having said that, if I spent ALL of my time on number 4 that would be OK too. It's my decision...

2) I don't get any of the rest of your point. "You see, in order to believe that the exit polls are correct, you have to believe that the central tabulators were hacked", etc. I don't get that at all nor do I even come close to agreeing.

3) "the excessive debate about these polls takes time away from any form of election reform"... This is so wrong, I don't even know where to begin. For now, let me say that if I don't agree with your assessment, I am not likely to let you police my time.

The handful of people who argue for the exit polls here, generally believe the election was stolen. You will have to decide whether they are natural allies or "enemies of the people".

One caution: there are a few people in these forums who wrap the flag of unity around the lamest of agendas. Thus, anyone who disagrees with Kerry is a "freeper". Jane Fonda is a "tool" of the right for coming out against the war when she obviously makes "leftists look bad". Karl Rove planted the court case on the Pledge of Allegiance in order to create a wedge issue and make Democrats look bad. TIA's prominence on the web is a plot because his arguments are "weak" and make "real experts look bad" (almost all of these have something to do with making someone "look bad").

Etcetera, etcetera...

Every time I read this kind of shit, I laugh and laugh and laugh. Talk about sheep in wolves' clothing....

[Bill Bored]

81. You are taking my comments too personally.

Let me try this again:

This is not about you per se, Anax, but you asked the question of whether the EP debate is a waste of time and I'm trying to answer it.

Exit polls attempt to measure vote counting -- not how many people make it to the polls or who is allowed to vote and who isn't, or how long the lines are, or whether there are enough machines. (There have been questions in some polls about the voting EXPERIENCE but those have been mostly dismissed here on DU because they didn't "prove" that the election was stolen.) So when you say you're concerned about the RIGHT to vote, ex-felons, suppression of the vote, etc., THIS HAS NOTHING TO DO WITH EXIT POLLS OR EXIT POLL DISCREPANCIES! In fact, the debate about the latter detracts from the efforts to address the former. Do you really not understand how this could happen? Those who are not allowed to vote do NOT participate in exit polls.

Those who have devoted hours upon hours to studying, debating, and otherwise focusing on these exit polls could NOT have been engaged in any other election reform activities during those periods of time, could they? And from the number of exit poll threads here, and your Google hits, there were many people devoting many hours to this.

Nevertheless, what they found, way back in Nov. or Dec., thanks to TIA's work in fact, was that voting machine types did not correlate well with the exit poll discrepancy at least at the state level. This is why NY's exit poll was way off, even though we don't have e-voting and yet Georgia, who has had e-voting exclusively, had pretty accurate exit polls. And there are other such examples.

Therefore, in order to continue the exit poll debate, the conventional EPTB wisdom goes like this: The vote count HAD to be corrupted at the "central tabulators" which are used with every type of voting machine. Now, do you not understand that if this were so, the way to detect it is to do a PRECINCT CANVASS in which precinct tallies are added up and compared to "central tabulator" totals to see if they match?

So, why aren't the exit poll true believers, and there are more of them than you give them credit for, doing this work instead of trying to convince everyone else that the exit polls alone prove that the election was stolen????

Answer: because they have been wasting so much of their time on the exit poll debate itself, that's why. I'm sorry to say they are spinning their wheels.

If one has a theory, the next step is to try to prove the theory -- not just repeat the theory over and over again, right? Yet this is exactly what we've been reading here since last Nov. about the exit polls, the central tabulators, etc.

I confess that I too have been guilty of this. But I've tried to rehabilitate myself. I've joined Exit Polls Anonymous! I'm on a frickin' 12-step program! When I start looking at the exit poll data, I e-mail my sponsor!

And now, after many months, I see just how much of a waste of time it was to keep going over and over the same old arguments ad infinitum with and without TIA's help, especially where the popular vote is concerned, which has no legal standing in a Presidential election in the first place!

So to answer your question, yes, I think the exit poll debate has been for the most part, a waste of time, except for those who do public opinion research for a living, and those who aspire to. Febble, for example has come up with a better way to measure within precinct error, taking into account the partisanship of the precinct. Good for her! But she's done other work on election reform too. Did you know she wrote a paper that showed there were excessive undervotes in NM in minority precincts? Maybe not because most of the attention paid to her here has been due to the fact that she just doesn't believe the exit polls. Meanwhile, she and Josh Mittledorf may have proved that a whole state was decided incorrectly due to excessive undervotes, which should have gone to Kerry. And, the exit poll in that state wasn't that far off BTW!

I guess if you don't get this by now, you probably will never be swayed. I'm glad that you don't spend too much time on this though, although that little game you started a while back was an exception to that rule, wasn't it? It was fun while it lasted though.

[kiwi_expat]

91. Atta boy BB!

kick ;-)

[mgr]

93. Excellent point Bill, one small point

"If one has a theory, the next step is to try to prove the theory -- not just repeat the theory over and over again, right? Yet this is exactly what we've been reading here since last Nov. about the exit polls, the central tabulators, etc."

The idea is to disprove the hypothesis (it really is not a theory since no relevant data in its support existed, as you point out).

The fact that the red shift did not correlate well to vote counting technology, would disprove the hypothesis, but may not have been sufficient to do so. In my opinion, what killed the exit poll argument for fraud by central tabulator was the Washington State recount of the governor's race. Or is it now that its only the central tabulators in battleground states?

We know that the game in past elections was vote suppression.

What I still don't get, is why it is so hard for a group of essentially white progressives to understand that the republicans targeted the votes of folks of color and other minorities, and left them alone. That is the question I would like to see some of us try to answer, because that would have been the basis for taking it to the streets back in November or December.

Mike

[Febble]

108. Absolutely.

The damning correlations that I can see are with ethnicity not with technology.

Technology matters, but all the tabulator security in the world won't help if some people are systematically prevented from reaching the precinct. Or have their votes invalidated when they get there.

[Bill Bored]

148. But you can suppress the vote electronically now.

Because Ohio was mostly paper (punch cards), they had to use other means. It's much easier to switch votes electronically than with paper and someday, there won't be a need for voter suppression at all! The suppressors can just go out and play golf on Election Day.

Of course, if there's enough auditing and paper trails and so on, this won't work, but that's the reason why verifiable voting is being fought against so hard by those who don't want it.

[mgr]

171. Precinct level suppression is hard to pull off state wide

By your use of suppression, help me out here, I'm not sure what your definition is, but I think this refers to those who think that they may have voted, but their vote did not count, but were not included in the undercount. One can falsify signatures in precinct poll books, but you need all the workers complicit in it, or do it a the BoE (as is likely occurred in Warren, but remember Andy Stephenson's finding). However, the limited recount would have triggered greater negative commentary--in other words we know where to probably look.

However this class of suppression is already possible with punch cards, one just keeps two books; and simulates signatures to the book; and removes from the deck those punchcards that are straight party line. It would seem to require about the same amount of work

Mike

[Bill Bored]

179. Well, I didn't mean suppress turnout, just the vote itself.

You can generate undervotes but they may not be seen unless someone does the math as part of an auditing process. Diebold, for example, usually calls undervotes "blank voted races" or "blank votes" and they are not correctly reported.

[anaxarchos]

102. Nah...

This isn’t personal... It's political (although you really have to stop calling me a "true believer").

And No... I didn't ask if the exit poll debate was a waste of time. You volunteered it. My question was obviously rhetorical.

And Yes, I am aware of your "zeal of the newly converted" which is precisely why we are having this friendly chat instead of triggering an instant polemic.

But, our chat is almost finished… It’s time for the naysay stalker chorus to show up and I don’t have the same patience for it as TIA did.

…So, let me say this plainly:

Your logic is all twisted up into a pretzel.

It does not follow that because I think Republicans steal elections through a dozen methods, legal, quasi-legal, and illegal which mainly have their impact before people can get to the polls, that I can not also care about fraud after they get there. That’s just some shit that you made up.

It also does not follow that because I think the exit polls indicate fraud I MUST think that the vote was corrupted at the "central tabulators" let alone that “the way to detect it is to do a PRECINCT CANVASS in which precinct tallies are added up and compared to "central tabulator" totals to see if they match”. That is just some more shit you made up.

Now, we still don’t have a problem. If you think this is a plausible thesis, go for it. If you prove it, you’re a hero. Where this whole thing gets funky is at the next step…

It does not follow that because YOU think all of the above, that this means that everyone else is “spinning in place” or doing the wrong thing or wasting their time or (gimme a break) “hindering”, “holding back”, “standing in the way”, “working against”, etc. “REAL REFORM”… now you are making up dangerous shit.

Politics is essentially simple: you are for some things and against other things. The people who are for the same things as you are your allies. The others… well, let me tell you a story.

Someone told me that there are a small group of people following Cindy Sheehan around who claim to support her “general objectives” but disagree with her call to immediately withdraw from Iraq. So they’ve convinced themselves that their task is to follow Sheehan around with signs critical of her message. Someone else told me the story was archetypical but, true or not, I’ve seen similar situations a hundred times before.

If you ask those people what they think they are doing, you will be amazed at how coherent, seemingly logical, and “left” their thinking is. But to everybody else in the world, they are just counter-demonstrators… “Hmm, they must be for the war.”

Just a final note.

Since you liked my “game”, let me explain it to you further.

I made up the game because I got tired of watching TIA “defend”. Oh, he was defending the goal just fine. But after the 2000th puck bounced off his face shield, the fundamental unfairness of it all came to me.

You see, logically and rhetorically, it is far easier to attack then to defend. And by the time he had done 14,000 posts, it was easy for anybody to point to a mistake he had made in post #1846, or to raise some completely tangential issue, or to get under his well known skin.

So I decided that it was time for someone else to defend… thus the “game”. But, I kept the real secret of the game to myself. And I am going to tell you that secret now, if you promise not to tell anyone else…

The game is rigged. It is as rigged as any game of 3-card Monty played on the street. And when we “played”, OTOH was the “patsy” (don’t feel bad for him… he volunteered).

Here is how the scam works. The exit polls were adjusted to the actual vote through a 43/37 split of 2000 Bush/Gore voters. That split is impossible. But it is also essential. There is no way to reconcile a different split. Now, theoretically, you can make up anything you want but there is no way to do it in a way that is remotely believable. The problem is not in the Exit Poll but in the 122 million people who voted. There is no Bush victory there just as there is no Jack (target card) in the 3 cards you have to choose from in Monty.

If you don’t believe that, I am here for you. The “game” lives! Step right up (although we’ll have to do it somewhere where TIA can play too).

If, on the other hand, you do believe that, then it is entirely possible to view your responsibility as a citizen or as a progressive, in making that knowledge as widely known as possible – in explaining and repeating that in a million different ways and thus spreading “the truth”.

TIA was unusually good at that.

[foo_bar]

104. parable of a rigged game

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x272294

1) Check assumptions.
2) Double-check #1.

The 43/37 hypothesis assumes people tell the truth, on average. In yesterday's headlines:

The groups also reported in a separate survey a suggestion that neither men nor women are all that truthful about their bathroom hygiene. When asked in a telephone survey whether they washed their hands after using a public restroom, 91% said yes, a contrast with the observational survey's finding of 82%.

http://www.medpagetoday.com/PublicHealthPolicy/PublicHealth/tb/1778

Is it false memory or is it Memorex?

I thought your game was rigged too, until the thunderbolt deus ex machina finale.

[anaxarchos]

105. The issue is irrelevant...

Make them all "liars". You make the split any way you want. Then convince yourself it is plausible and "call" me.

But I ain't takin' no responsibility for giving you the blues...

[Febble]

107. The game wasn't rigged

except in the sense that it looked like a math game, but was actually about assumptions.

There were no winners, deus ex machina notwithstanding. The result of the game was this algorithm:

If you find it plausible that:

1) people misreport their previous vote, and have a greater tendency to so in the direction of their current vote than otherwise

and/or

2) the weighting used to calibrate the exit poll projection in line with the count was computed by an algorithm that simply chose the most efficient (and not necessarily the most plausible) weights to apply to which demographic variables (and it would not have chosen "previous vote" as the relevant demographic variable as this data was only available for a minority of respondents)

then the 43/37 split is no bar to finding it plausible that the exit poll discrepancy may have been due to some form of sampling bias.

However (else),

if you find it implausible that

1) people misreport their past vote

and/or

2) no reweighting of the projections to match the vote returns could have been done that would not have resulted in "impossible" Gore/Bush proportions

Then the 43/37 split is impossible unless fraud occurred.

No rigging of the game was therefore required. TIA's math was, as usual, correct, according to my own calcs, anyway, but I would trust his. OTOH's math was also correct (again, I checked).

The answer depends on the assumptions, not the math.

If Anax rigged the game, he only rigged it by not stating that the answer depended on using his assumptions. Or, if you like, his belief that the election was rigged.

But even making the first set of assumptions does not allow anyone to conclude that fraud did not occur. It merely allows us (me) to believe that fraud was not necessarily the cause of the exit poll discrepancy.

[foo_bar]

111. depends who the judges are

The "subjective boundaries" of anax's first rule are ineffable qualia. Since the winning perceptions are decided by "he who counts the emotes", the game has the rigging potential of an Ohio recount.

[Febble]

112. Well that's why the answer

was an algorithm.

Plug in your qualia, get the answer.

There were certainly more emotes for assumption set 2.

But there should have been a corollary to Rule 10 that said: The first person to say "the real vote count has to be wrong because no-one in their right mind would have voted for Bush", loses.

Although to my mind, it's still one of the strongest argument for a stolen election.

[mgr]

113. You can't get there from here

One of the more significant articles when I was studying vegetation ecosystems and patch dynamics made reference to models, mechanisms, and pathways. What it refers to is a hierarchical nesting of theories and hypotheses based upon the scale and scope of inference. But what needs to follow is consistency between the hierarchical levels. An excellent example of this is the link between the theory of island biogeography, its application to park and conservation design, and how extinction dynamics in individual parks perform.

If the exit poll discrepancy is an indicator of fraud (model), then there must be a coherent hypothesis on how that fraud was carried out (mechanism), with specific examples (pathways). What Bill indicates is that the class of vote theft most likely to follow from TIA's analysis has been experimentally shown to not stand up. It was the responsibility of adherents of this position to provide alternate hypotheses as these ones failed. Arguing for a large patchwork of various fraud scenarios dependent upon local circumstance is not an articulate defense given numerous constraints upon the lack of a coherent mechanism.

The fact that no testable hypotheses are forthcoming that stand up to inspection, but one thinks the exit poll discrepancy is an accurate indicator of fraud is a statement of faith.

The Bush/Gore vote to the Bush/Kerry vote debate is a distraction from the fact that the emperor is naked, by looking at the rings on his fingers. It should become very clear to the exit poll discrepancy ayesayers to recognize that until they can explain the 'how' of vote stealing, they are true believers, and ascientific.

As long as there is life to this argument, the longer it is until other potentially more plausible models can be addressed and evaluated. Because what the exit poll debate in essence implies is that there is no basis to take any number on trust.

Mike

[Chi]

115. question

"If the exit poll discrepancy is an indicator of fraud (model), then there must be a coherent hypothesis on how that fraud was carried out (mechanism), with specific examples (pathways)."

If I had money disappear out of my vault,
but could come up with no coherent hypothesis of how it was done,
could I claim it was stolen?

[mgr]

132. Problem with the parallelism and with example

This is what Chi suggests:

Model mechanism pathway

Red Shift=fraud n/a n/a
Money in vault n/a money misplaced


In the first place your argument is a syllogism, not a nested hierarchy:

P1: There was money in my vault
P2: The money is missing
C1: It was stolen

Secondly, there are insufficient scientific constraints to your syllogism to afford your conclusion. There has to be a shared observation or verification of your propositions. Is the vault yours alone, do you suffer memory lapses, does your significant other have access to the vault, etc. Is your example sufficient to go to the police; or would you have to show that 1.) the money was actually in your possession; 2.)You did not hide it for insurance purposes? What if the money disappeared in a period of ten minutes while you were in the room? That makes for a nice Agatha Christie or Wendell Wilkie plot, but I don't think either is required reading for the police these days.

You may also want to consider that the red shift argument follows a deductive argument, rather than inductive. This the reason many of us take issue with it, is that it's a (ab?)use of descriptive statistical measures, not inferential. The votes have to be in your possession to begin with, and that has not been demonstrated to the satisfaction of many, nor can you provide a clear mechanism of how the observation could be accurate.

The problem with the red shift argument is that you need a universal pattern of irregularities at precincts throughout the country. If you argue the it is only with the battleground states, then what is up with the red shift elsewhere? No one has explained this, or was it that NEP was right when it came to non battleground states, but not with the battleground?

So where do you go?

To continue to cling to the position is nihilistic for other more probable and feasible explanations of how Bush won the election by less than legitimate means.

Specifically, how can I develop a baseline for modeling voter turnout to test the voter suppression question, if the red shift implies that all precincts are suspect for their precision and accuracy?

What I am saying is nothing new, it's much the same points I and others made back in December and January to TIA, fully expecting that with time, she/he might come to realize the overweening implications, and step aside.

Mike

[Chi]

135. I guess thats what I get for asking a yes, or no question......

:eyes:

Chi didn't suggest anything, Chi just asked if he could claim his money is stolen, even though he didn't know how it was accomplished.
Chi also doesn't think he got an answer. 8)

That's OK, we both know it was rhetorical.

[mgr]

140. Simple answer to what was not a simple question:

If in regards to yourself you can assume anything, but recall Descartes' evil deceiver argument. In regards to obtaining consensus from anyone else, given that your conclusion does not necessarily follow from your assumptions, no. The question you ask is how can one know with certainty is at the center of the mind body problem, and what drives the meta logic of why science is done in the manner that it is done.

Mike

[Chi]

173. I really don't think it could have been any simpler.....

No, is not correct.
Of coarse I could claim the money was stolen, and both the police and insurance companies would come and investigate.
I would have presented sufficient evidence to call for investigation (which should be the target point of the EP debate).

I never said it would prove it was stolen, just as you never stipulated the indicator has to prove fraud.
(In ecology, a plant or animal whose existence in an area is strongly indicative of specific environmental conditions, but the presence of said plant or animal does not prove those conditions exist.)

"If the exit poll discrepancy is an indicator of fraud (model), then there must be a coherent hypothesis on how that fraud was carried out (mechanism), with specific examples (pathways)."
I still see this as inherently false, in it's currant form.

(edit - spelling)

[mgr]

200. Are we persuading you, or persuading others?

You are splitting hairs.

But you did not present sufficient evidence in your example, and you are now making that hidden assumption understood (that you had evidence, or maybe it's just the veracity of your word). One has the right to believe anything. I approached your statement much as I approach work done by myself. I think I very clearly suggest that if a coherent heirarchy can be constructed, it would show fraud.

However, other progressives, the public at large, and the government require more than a claim that something was stolen. The exit poll discrepancy has numerous predictive and explanatory problems, in and of itself it present no sufficient evidence. The failure to associate a pattern with a particular type of vote counting machinery is quite damning as to it being a distinct indication of fraud, separate from what occurred at a particular precinct, BoE, or state. Pattern is there, but it does not link well to the prediction of the exit poll red shift that the popular vote was stolen nation wide. You cannot infer a universal condition from an existential, and you make the same error with your ecology example.

One should never associate individual plant or animal species to environmental conditions, but an assemblage (see Robert Whittaker's Ordination and Classification of Plant Communities, Springer Verlag, I think 1974). An assemblage is a combination of steno- and eury- tolerant species, that when one employs a Venn diagram (I'm being simplistic here)the particular environmental parameters intersect. There are several reasons one does this.

First is that a particular indicator species may not be present--much like the American Bison not occurring in the Willamette Valley (Its re-expansion of its range did not reach the valley following the megafaunal extinctions during the Ice Age), or that parallel evolution has had insufficient time, as in the case of ultramafic soil colonization by species of Penstemon, Phacelia, and Mimulis.

Second, there is always the probabability for individual species to become acclimatized to conditions at, or just beyond the published experimental limits.

Third, environmental conditions change over time. So one may have a bristlecone pine living at 8,000 foot, it may do quite well, but it would not be at an evolutionary advantage since none of its seedlings will survive.

The specific problem that I do not think you are getting is that the exit poll red shift and the argument that it indicates fraud is a hypothesis, and a weak one since the analysis is descriptive (explanation in the form of patterns, and hypotheses must be forthcoming) rather than inferential (can be directly tested). As a hypothesis it must have predictive power, and I am suggesting the framework in which that may work. May be I am missing the predictions, do you have any?

By the way, what were the calculated odds for life developing and evolving on the planet? Would you take these odds as impossible, and thus you do not exist?

Mike

[Chi]

203. Who the heck is "we"?

While you're at it, who the heck is "others"?


As far as the ecology lesson goes, thanx, but no thanks.
If you want to argue the definition of 'indicator'....
You can do it with American Heritage dictionary.

"3. Ecology A plant or animal whose existence in an area is strongly indicative of specific environmental conditions."
http://education.yahoo.com/reference/dictionary/entry/indicator

Let me know how that turns out for ya. 8)

[mgr]

208. There are rigged games, and there are traps

I favor chess over tennis.

I apologize for assuming you knew a little ecology. I try to speak at an individual's level of expertise to work to consensus and understanding--the devil is in the details. Since you were working from a dictionary definition (non-specialist) rather than a learning or knowledge situation, I give you the US EPA (specialist)to defend mine:

http://www.epa.gov/bioindicators/html/indicator.html

I assume we are still discussing persuading others that the election was likely stolen. That would imply that personal standards of acceptance or rejection of a proposition may not apply when discussing with others.

We = those of us that think something went awry in the election. This is the only we, you and I fall in, as I currently understand it, since you seem to take opprobrium to being in consensus with me on anything.

Others = those that do not understand or agree. The police or insurance agents in your and my examples.

The specific point I was getting at is that presence or absence of a single barometer of a condition is never sufficient to make a determination. All factors need to be considered. The point you made was that an indicator species is absent, therefore a condition can still exist. The point I was making is that you don't use single indicator species to make an evaluation of widespread conditions, but with all the species present evaluated. Examples where indicator species have led investigators to false conclusions are legion.

I could see you thinking analogously that the exit poll functions as an indicator species in the last election of fraud, which is why I went into the gory detail of why the term indicator species was wrong, and why. Regardless of model, mechanism, pathway--the use, as understood by the specialists, contradicts your analogy.

The problem was that one cannot build up wards to a grand theory that the red shift indicates widespread fraud without any data at all other than dubious measures of supposedly raw exit poll responses, which are applied circularly. You seem comfortable with that conundrum. Rejecting that a mechanism for vote stealing after a person enters the voting booth is not required, or even evidence at the precinct level that something like this happened is very thin gruel to persuade others.

The underlying basis of this ongoing discussion on exit polls is how much trust one places in specialist knowledge. Ignoring, and failure to assimilate the viewpoints of those with specific work in the area of polling and polling dynamics is on par with the Bush administration ignoring the Army Corps of Engineers. Or do you wish to argue that I paint with too large a brush

Mike

[OnTheOtherHand]

210. might I interject --

I don't think I've seen either of you deliberately misunderstand someone on the board. And I think you both are having trouble understanding each other. So, please be patient. Me, I'm neutral, because I'm having trouble understanding both of you, although I feel closer than I did at this time yesterday.

If Chi wants to say that the exit polls are suspicious, I'm fine with that; I can't tell how much more than that he is saying. I'd say that the back-and-forth-and-sideways starting with #175 on this thread is one of the more serious and respectful conversations DU has mustered on the subject of false recall, so thanks to Chi for being part of it. I certainly agree with mgr that the exit polls by themselves are pretty thin gruel. I haven't yet been able to identify an assertion about the exit polls that I feel sure the two of you disagree about.

Chi, I have often found it hard to understand what you are trying to say. For instance, back in #58, you wrote, "The exit poll debate, is based on the accuracy of exit polls...", but I have little idea what you meant. When people like me talk about the accuracy of exit polls, it is a different discourse than when wikipedia cites some mysterious source that seems to have claimed that 'exit polls are usually accurate to within a fraction of a percentage point,' or than when someone argues that anyone who questions exit poll results is being nihilistic (at least, I think that was the argument I encountered once). And so on. So if you think foo misread you, he probably did, but not on purpose. Just like me, you are at risk of being misunderstood in light of other people's months-old quarrels. Or so it seems to me.

[Chi]

219. Absolutely

(readers digest version)
Actually, I understand mgr, I just don't agree with him.
I've quoted the sentence I disagreed with, and gave an example as to why.
He knows how I make my points, so he refuses to answer candidly.
When I gave him an example of 'indicative' used in a sentence, he chose to debate the accuracy of the example, instead of the point.
He has now even suggested I represent others not present, as well as sending barbs about expertise or education.

I find it fascinating to see what extent he's willing to go to not concede anything,
but I'm not willing to put everyone else through the nonsense any further.

But thanks for stopping by OTOH 8)
and thanks for the compliments.

[OnTheOtherHand]

224. tell me about it...

mgr suggested that you represent others not present? umm, maybe. Well, did he actually state that you were probably being paid for it? That happened here in the last week. Not that I want to turn this into some sick competition!

You understand mgr's arguments, but I think you misunderstand his motives.

If someone can look at something Febble says about exit polls and take it as evidence that she is paid to defend electronic voting -- and I'm going to assume that that reading was absolutely sincere, partly because to me it seems too silly to be sneaky -- then why can't mgr be confused about what point you are actually trying to make when you talk about exit polls as an indicator?

Applegrove is right, the "exit poll debate" has been far too stupidly polarized for far too long. It's hard for anyone to say anything without having fifty other views imputed, or at least suspected. But mgr did something excellent, actually: he gave you an opening to say without rancor, 'Yeah, I think you are painting with too large a brush. Let me explain....' Trust me, that level of courtesy is not to be taken for granted around here.

OK, if it's time for you to walk away from this thread for a while, do it. But please cut mgr some slack -- we've all been through the wars.

[mgr]

227. In my defense:

Chi:

Quite frankly, you and I go around and around. I don't understand your method of argumentation (although you think that I do), as you seem to leave several jumps in your logic unexplained, and expect me to be smart enough to fill in the gaps in a pattern consistent with what you are thinking. Honestly, it is unfair, and because so, I have to infer motives. You seem willing to acknowledge some class of my expertise, but do not agree with my position without staking out why. I would suggest that you may be less than candid, and you leave it for me to construct straw men, once you provide more details after the fact to show that I am wrong.

Your argument follows the logic of that old prosecutorial question: When did you stop beating your wife. Essentially I pointed out why that is an unfair question to ask. To expect me to do otherwise, and then criticize it as being less than candid, would confirm my inference as to your motives--you were trying to set a trap.

This is what you posted:

"'If the exit poll discrepancy is an indicator of fraud (model), then there must be a coherent hypothesis on how that fraud was carried out (mechanism), with specific examples (pathways).'

If I had money disappear out of my vault,
but could come up with no coherent hypothesis of how it was done,
could I claim it was stolen?"

What I responded with was that I disagree with the example and the implied parallelism. My answer was clearly 'no', although you think the answer is 'yes'. To agree with your assertion is to lend support to a viewpoint I do not agree with--that the exit poll discrepancy is stand alone evidence of fraud.

That would imply that I agree that the discrepancy is inferential rather than descriptive. Which I do not. The fact that I think it is descriptive implies by necessity that inferential statements can be derived from the observation--whether these are models or mechanisms is besides the point--for it to be a scientifically valid explication of a phenomena.

Your use of the term indicator species was unfortunate--its mention suggests knowledge of the field of autecology (had you stated that you took it from a dictionary, my response would not have been the same), and I responded at a time when I was rushed to be in the field, and was incomplete. The concept is a red flag--all species are environmental indicators, what the hell is a niche anyway? Pretty much the concept is an abstraction, and embraces a thumbnail sketch of what occurs within the environment, that cannot be confirmed or denied unless compared with other like indicators. I was running through examples of false positives, and false negatives. The indicator species concept is never sufficient on its own to support or refute a observation, but it is a hypothesis that leads to further investigation.

To say that the exit polling discrepancy is indicative of fraud is logically equivalent to TIA's claim that it is circumstantial evidence. Common sense, the root of both science and jurisprudence, would not allow circumstantial evidence alone to try or convict anyone if inconsistent with other salient details.

So again, I am not sure what you are arguing, if this was not a trap.

Mike

[OnTheOtherHand]

143. hmm, I'm not quite as lofty as mgr

I would have said -- yeah, of course, you can claim anything, but folks may stand by alternative hypotheses, e.g. that the money never existed, or that you took it yourself.

But y'know, the way this plays out -- there are some fraud theories, and some fraud evidence, that don't fit together especially well, and mgr's framework offers one way of talking about that mismatch. But I don't think either (or any) of us would argue that unless someone presents a seamless soup-to-nuts fraud narrative, all the arguments are worthless.

(EDIT: Hmm, not sure what happened to my last point. Probably just as well. Anyway, no game allusion intended.)

[Chi]

174. There ya go 8)

As phrased, the answer to my question was yes....good job.

I don't think you can dismiss an indicator, simply because you can't figure out the method the affect was achieved through.

But maybe my parents had to much fun in the 60's.
:hippie:

[petgoat]

137. "there is no basis to take any number on trust."

You got it.

As to your "patchwork of scenarios," consider that 80% of the votes
were counted on machines made by two companies, ES&S and Diebold, both
of which have records of dishonesty.

The mere fact that registration fraud occurred here or there does not
in any way discredit the notion that electronic fraud occurred
elsewhere.

As to your belief that the presence of non-battleground anomalies
discredits the assertion of fraud, consider that the popular vote
"mandate" conferred legitimacy on the 2004 results. Disputing
electoral votes in this state or that was unseemly and legalistic.

The easiest votes to steal were those of the Republicans-for-
anyone-but-Bush, because nobody was going to fight for them. Inflated
Bush percentages in red states could easily account for the putative
popular margin.

[mgr]

144. Reply

Paragraph 1: Great factoid, that piece of speculative argument. Consider the constraints as to how those machines might steal votes, change votes in light of recounts (Washington), firewalls (the lame Bev Harris hack from inside a Florida BOE), and citizen oversight (Liam Laddies' as yet unreported inspection of Cinncinnati ballots)

Paragraph 2: I don't know what you mean by registration fraud. Voter suppression refers specifically to the long lines experienced by voters in predominantly democratic party leaning precincts; it may also include the undercount of ballots. Where's the beef for the electronic fraud--it's all tissue thin speculation.

Paragraph 3: The prediction of the exit poll red shift implies that vote stealing occurred within both battleground and non-battleground states, thus all of the states. It does not refer to electoral votes, but popular votes The problem is that some are now arguing that it only applies to battleground states (Simon's paper in particular), you cannot argue for an existential <delete of> with special antecedants from phenomena that are universal wiht the same consequent.

My position very clearly is that Kerry likely won based on voter intent in Ohio, possibly Florida. That intent was foiled by vote suppression. Some one who does not appear at the precinct to vote will not be measured by an exit poll. However, if I cannot reliably measure turnout, I cannot measure how many in Cuyahoga or Franklin County walked away without voting. You seem to think this is acceptable.

However, there is no basis to accept the presumptions of the exit polling debate, since no empirical data is forthcoming, and no acceptable mechanisms for how it was accomplished are provided.

Recognize that your argument is more pertinent to reform of existing vote counting technologies, but has limited to no applicability to what we are discussing, is how the current election may have been stolen. There is no evidence, and there is no pattern consistent with voting technology.

The proof otherwise is on you.

Mike

deleted 'of' to make sense.

[kiwi_expat]

118. I am happy to try to play the "game", now.

I stopped proposing alternative hypotheses when autorank convinced me that all eyes were on the ERD forum and that any counter-arguments would be picked up by the opposition. But I gather from autorank's post #103, below, that he thinks that the coast is clear now. Nobody is paying any attention to us.

I have a couple of scenarios I would like to try, but they both involve stipulating a *narrow* Bush popular-vote win. I'm afraid that I can not think of any scenario where Bush actually won the popular-vote by 3 million votes.

Also one of my scenarios involves assuming a specialized form of respondent-bias. Is that allowed?

Finally, I don't know how to play tennis. Is that a problem. :-)

[OnTheOtherHand]

126. (smile)

I think it's worth noting that the opposition already has arguments. AFAICS all they can glean by watching us is greater or lesser respect. (Personally, I never find conservatives more appealing than when they are arguing, reasonably, with each other -- it makes them seem, you know, serious.)

I think a better argument against serious debate here is that this is an activist community and if we don't agree, we should keep it to ourselves and not be divisive. But I hate that argument, so I probably didn't represent it very fairly. We need to cultivate the ability to distinguish between disagreement and divisiveness, or else IMHO we aren't democratic any more.

OK, I'm clambering down from the soap box now. I think any idea you believe could be taken seriously is "allowed." Specialized form of respondent-bias: yes, unless it entails that voters are receiving microwave signals via implanted chips, in which case perhaps not (grin).

My wife can testify that I really don't know how to play tennis either. (Of course some folks think I was lousy at the "game," too.)

FWIW I still think that Bush most likely won the popular vote by millions of votes, although I'm not adamant about the "3" (nor about the electoral vote). I was hoping anax had something better than (IMHO) he came up with. But of course the point of the game shouldn't be to win, but to learn something useful.

[anaxarchos]

138. I got something for you, Mr. Hand....

I got it right here.

But you don't get it without doing some work first...

Don't you watch television? This is America. There are no free handouts here, ever since Reagan and Bush and Bush. You are going to have to earn it. Which means, sooner or later you are going to have to make a case.

Your game was lame, but I have hope for you... There is always a second chance. I'll play witcha again. Pickup up where we left off or start anew... But this time:

1) I ain't playin' here. It's disconcerting when your tennis partner get's arrested in mid-swing. You tell me to set it up and I will. The web is a big place.

Of course, you can play here without me or TIA. Maybe you could play with mgr... and then, grade his paper afterwards...

2) Next time bring your case, Mr. Prosecuter. You say "I think Y2000 Gore voters voted for Bush"... so? Prove that it is plausible. Personally, I think 20 million martians teleported down to vote for Bush. Where is the evidence, supporting either one of us? This ain't about the polls... it's about life. 122 million people don't stomp into the polling booths without leaving some evidence of where they have been and what they have done...

You've been talking to sophmores too long. You state what you "think", peer into the audience, and then repeat what you "think" again. That will only win you one kind of game and it isn't this one. Who cares what you "think"... Prove it.

3) You still don't get that this isn't about you... You were whining the last time about no one showing you somethin' that would change your mind, etc.... You don't get it. It is YOU who are showing - your job to convince.

I think you should bring along 75 friends to help you the next time - those same people that you continuously claim you want to convince. There is no better way to reach them then to make them prove what they claim to believe. You could "trash talk" them like I am you: "Hey, there's this guy on the Internet that says you're a punk...", or you could remind them of their civic duty as an expert, yadda yadda...

One last thing... If we play again, I want foo_bar on my side even if he doesn't want to come. I need to prove to him that intervention by gods on a stick (or a wire) doesn't mean the same thing if you think your gods are less than you are (as the Greeks most certainly did). In that case, it ain't so bad.

[OnTheOtherHand]

139. huh?

AFAIK I did exactly what you asked me to do. I accepted your rules and I played by them. If you think DU is unfriendly turf, imagine what I thought about it. But in the end, to my astonishment, you were the one who fled.

"You don't get it. It is YOU who are showing - your job to convince."

Umm, sorry, which rule was that? Wow, I guess the game was three-card monte after all. But you left before you could collect your "winnings." So now you are trying for double or nothing?

I don't really care whether I convince anyone here that Bush won the 2004 popular vote -- in fact, I wish that someone here could convince me otherwise. Call that "whining" if you like. I never signed up for that game or that task, and I still haven't.

But the case for false reporting of past votes is pretty well supported by the NES, the GSS, and even past exit polls (as you probably know by now if you are actually intellectually curious). I have no idea why you believe or contend that 43 Bush/37 Gore is a clinching argument, or what proportion of Gore2K->Bush04 defection you regard as unthinkable and why. At some point that can't really be my problem.

Congratulations on finding a safe place where you can revel in your certainties, since this one didn't work out.

[anaxarchos]

142. Yup, that's what I thought.... eom.

[kiwi_expat]

152. I'd like to try to play right here. Your partner has already responded.

>8) This game is only indirectly about Exit Polls and Polling
>in general. We are actually trying to reconstruct the
>election of GW Bush, in a "plausible" way.

Here is/was my serve:
11.5% Gore voters and 94% Bush voters voting for the "War President".


Here was your partner's response:

http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=203&topic_id=352989#353071


I'm not sure if that response was a drop-shot or a complaint to the referee. (I had some help with my tennis terminology.)

[kiwi_expat]

230. Your partner agreed: "Kerry lost" that scenario.

http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=203&topic_id=352989#353010


(I just want to move things along here.)

[Chi]

128. Suggestions

Start a new thread....forget the term 'the game'.

Present your scenario, and explanations.

Ask like minded participants to not be condescending (if they start being that way).
If you get someone in there comparing people that disagree, to Ms. Cleo fans, the thread goes F.U.B.A.R. fast.

I'm pretty sure you'll get participants on both sides.

[OnTheOtherHand]

129. seem like good suggestions to me

(FWIW, the whole "game" concept/semantics weren't mine to begin with, although I felt that I understood the rationale for them)

(oh, I should also note that there can be more than two "sides")

[Chi]

130. Agreed, on both counts. n/t

[Febble]

131. What actually might be useful

as OTOH suggests, is to construct the debate in such a way that we might actually learn something.

Maybe not a debate but a brainstorm:

What do we know, and what does it mean?

We might argue about what a given piece of evidence means, but (a suggestion) if we rank evidence like this:

1. evidence suggests votes for Kerry were lost
2. evidence suggests votes for Kerry were deliberately stolen
3. evidence suggests that loss of these votes cost Kerry the electoral college vote
4. evidence suggests that loss of these votes cost Kerry the popular vote.

we can perhaps come to some broad consensus as to what part of the case is a slam dunk, and what is perhaps a suggestive footnote.

The reason this might be useful is that it may present us with a ranking of answerable further questions that we can go out and find the answers to.

And a plea: don't forget suppression, including the structural suppression of ethnic minority votes through spoilage! That alone cost Gore Florida.

And a final plea: I'm not a troll. I can vouch for OTOH. No second guessing what people's motives are. Critique all contributions merely on their factual accuracy and logical coherence.

[Chi]

134. I would love to see that happen.....

With all the intellectual capital, and 'hands on election experience' on this forum, the possibilities for discovery are impressive.

Hold your thoughts on how to organize that 'brainstorm'....cause I surely wouldn't be the one to negotiate/brainstorm with.
There are many better suited for that than I.

Plea 1, I agree, it's also the most acceptable issue of fraud to outsiders (and the most obvious).

Plea 2, I personally don't think either of you are trolls.
But do understand one thing....there have been, and will continue to be trolls coming in here with one objective in mind, to disrupt....
and by nature, they will always be on the same side of the issue, the side that says there is no evidence of election fraud.

[mgr]

141. Minor suggestion:

I think we might seek to separate the endeavor of a brainstorming thread from a critiquing the brain storming thread. There should be no criticism at all during the brain storming, but freeplay of one idea to another. All for it, and let the trolls come.

Mike

[kiwi_expat]

153. I love the idea of a true brainstorming thread.

Would you like to start it, Mike?

[Febble]

160. Yes, please do!

I'd rather not, for a number of reason, number one being that I'm not the right person to start a non-divisive thread.

Not that I want to be divisive, but I seem to have that effect on people here...

[mgr]

172. And I don't?

Such an observation, I suspect Andy Stewart (Silly Wizard) may know a kirk for thee, lassie.

Give me some time to address others that fall within my "don't suffer fools gladly" mentality (there's is one individual in particular I suspect I need to address) to take it lightly; and I would love to do so.

Mike

[mgr]

228. Done n/t

[kiwi_expat]

231. Good on yer, mgr ! n/t

[mgr]

145. Glad to see you are up to it. n/t

[Bill Bored]

151. Well I must confess:

I thought OTOH's memory hole thing was a bit off the mark, but the other day, it took me quite a while to recall who Gore's running mate was in 2000! I eventually did, but if I were filling out one of those 17-page questionaires with 99 other questions on it, I might not have been able to recall it was Lieberman. I also don't think of him as a Democrat nowadays, but that's beside the point, or maybe it's not.

Anyhow, there will soon be a public dabate between Freeman and Mitofsky in case you haven't heard. This is not exactly public knowledge yet, but it's not exactly a secret either.

So I'll just sit back and try to get a transcript of that somewhere, or maybe even a web video like Conyers' Ohio hearing! Perhaps you'd like to ask TIA if he'd like to attend or better yet, bring a camcorder, or at least the popcorn!

I'd be interested in your choice of a moderator too. Are you available?

It'll be the Thrilly in Philly!

[Chi]

175. Would you have any trouble with this.....

Add 20 more years to it, would you have a problem picking?

Did you vote in the presidential election in
1980?
1 No, I did not vote
2 Yes, for Jimmy Carter
3 Yes, for Ronald Reagan
4 Yes, for John B. Anderson
5 Yes, for another candidate

[OnTheOtherHand]

176. we can actually try that one from the 1984 ABC exit poll

which had something over 20,000 respondents, of whom apparently all were asked whom they voted for in 2000 (although they didn't all answer).

In the unweighted poll, 57.0% of respondents said they voted for Reagan and 42.6% for Mondale, a 14.4% difference. (Reagan actually won by 18.2% in the official returns. But I refuse to calculate a P value for that error.)

Now, in 1980, Reagan got 50.75% of the popular vote, Carter 41.01%, and Anderson 6.61% (per Leip's Atlas), with other candidates splitting the other 1-point-something. So, Reagan's actual margin was about 9.7 points, or about 9.8 if we look at just the three leading candidates.

Would you care to venture a guess, among the people in 1984 who recalled having voted for one of those three candidates in 1980, how much Reagan "retrospectively" beat Carter by? Bear in mind that the exit poll actually overstated Mondale's support.

Warning: Bill Bored may not be a representative American voter.

[OnTheOtherHand]

177. (oops, of course I meant "in 1980" in the first line!) n/t

[Chi]

197. Couple of items...

First, my point was about people misremembering, your example includes the possibility of both misremembering and intentionally answering something that isn't honest.

Now, as far as examples go, I'm gonna pick an election that should give very interesting results, I have NOT seen the results of the exit poll that I am choosing.

I would like to see how people responded to the question...
" IN 1972, FOR WHOM DID YOU VOTE? (54)
1. < > NIXON
2. < > MCGOVERN
3. < > SOMEONE ELSE
4. < > DID NOT VOTE"

I would believe that if people would intentionally lie about who they voted for (to be on the winning side), they would also lie about voting for Nixon, a national disgrace.

So, did the 1976 exit poll indicate more people voted for McGovern than actually did?

[OnTheOtherHand]

198. excellent point

I do know the result in 1984, but I do not know the correct explanation(s) of the result in 1984.

The result is that among people in 1984 who recalled having voted for one of the three major candidates in 1980, 57.7% recalled having voted for Reagan and 36.3% for Carter, a difference of 21.4% (compared to the actual difference of 9.7%). Some of that difference could be actual -- differential turnout. Some of it could be misremembering, and some of it could be deliberate dishonesty. (I tend to downplay deliberate dishonesty, but I have no grounds for estimating it in 1984.)

The one thing I will emphasize, for the benefit of anyone else who might be reading, is that it demonstrates that the unreliability of the recalled-vote question does extend to exit polls. I can demonstrate this in other exit polls including 2000. It doesn't tell us who won in 2004, but I think it pretty much kills the argument that the recalled-2000 43% Bush/37% Gore result is impossible and therefore proof of fraud.

I think ICPSR has the 1976 exit poll study, but I don't have it right now. I should be able to get it tomorrow. I can tell you that in the General Social Survey, McGovern basically held his own. The actual 1972 result was Nixon winning by 23. That was down to about 18 points in the 1973 and 1974 GSSs (for a while there was a GSS every year), then 25.6 points in 1975, 21.6 points in 1976. Of course there is sampling error in here (the effective sample sizes were only about 1000 per year).

So, folks didn't exactly retrospectively repudiate Nixon en masse, but he certainly didn't get the retrospective bounce that most incumbents apparently do. I would venture that Nixon is a fairly extreme case. But at any rate, there isn't any iron law here.

[Febble]

199. FWIW, in the UK

some time after John Major's unexpected but meagre victory in 1992, over Neil Kinnock's Labour party, polls were giving Labour a retrospective victory.

And of course Tony Blair took Labour to a landslide victory in 1997. So it would appear that people don't necessarily simply forget having voted for the loser - they also forget having voted for a winning party they no longer support.

http://www.icmresearch.co.uk/reviews/2002/failure-of-the-polls-1997.htm

Although things may be different in Britain because we vote for parties not Prime Ministers, so at least people don't have to struggle to recall the name of the loser, and they also don't have to struggle to recall the name of the anticipated next winner.

[Chi]

201. Another thought...

In and election where there is a strong third party candidate, AND an incumbent with a high approval rating,
I would expect the 'who voted for' discrepancy to be at it's highest.
Yet when I chose an exit survey that I thought those conditions favorable (1996 VNS), I find no such bandwagon jumping to the incumbent.

Clinton 43
Bush 35
Perot 12
Other 1
Didn't Vote 9

(Actual - 43, 37.4, 18.9)

Are there any exit polls (that you know of) that illustrate 'who voted for' discrepancy, favoring a Democratic incumbent?

[OnTheOtherHand]

202. uh oh, my results don't match yours...

(EDIT: As I note below, I don't think anything is wrong with your results -- the header just reflects my first rush of panic at the apparent replication failure. Our results actually seem to be interchangeable.)

I just extracted only the '96 and '92 presidential vote fields from the 1996 survey. I got, as unweighted results:

1996
1 Clinton 8550 = 52.2%
2 Dole 6133 = 37.5%
3 Perot 1407 = 8.6%
9 Other 285 = 1.7%
(margin 14.7 points, note the error favoring Clinton; Clinton won by 8.5 points in the official returns)

1992
1 Clinton 7023 = 45.6%
2 Bush 4949 = 32.1%
3 Perot 1901 = 12.3%
4 Other 131 = 0.9%
5 DNV 1396 = 9.1%
or refactoring that to remove the DNVs, Clinton's retrospective margin (based on 14,004 respondents in categories 1 through 4) is 14.8 points. Again note the error favoring Clinton, who won in 1992 by 5.6 points in the official returns.

So, the error favoring Clinton in 1996 is 6.2 points, and the retrospective error favoring him in 1992 is 9.2 points. Subject to minor rounding error (and perhaps big stupid mistakes, dunno). With 14,000 respondents, that is well beyond margin of error.

I think the difference between our results is probably that (1) you are using weighted returns, and (2) you aren't factoring out the DNVs. Your 43/34/12/1, with the DNVs removed, gives Clinton a retrospective 8.8% margin over Bush, which is again about three points larger than his actual margin of about 5.6 points -- and again strongly statistically significant.

But the gap is less than in 2004 recalling 2000, which might be significant, I guess.

I can't think of any other popular Democratic incumbents running for reelection in the exit poll era. However, in the 2000 exit poll recalling 1996, in unweighted results (which overstate Gore's margin by 1.7 points), Clinton retrospectively defeats Dole by 18.0 points, limiting the calc to people who did recall a 1996 vote. Since Clinton actually beat Dole by 8.5 points, that is a huge "who voted for" discrepancy favoring the Dem incumbent, although the incumbent wasn't running.

I'm not sure that result has ever been reported before. Folks can compare the weighted results in the archived 2000 exit poll on cnn.com (http://www.cnn.com/ELECTION/2000/epolls/US/P000.html -- Clinton 46, Dole 31).

(EDIT: Clinton's approval in 1996 was mid-50s, so not _that_ popular. We would have to go back to LBJ in 1964, who hardly counts. But I bet if we pull the retrospective on JFK in the 1964 NES, assuming there was one -- there was no exit poll -- it will be quite high, and won't prove much.)

[Chi]

204. I should be more wary of where I travel....

My disagreement of this 'who voted for' topic has been, who could of possibly forgotten who they voted for in 2000.
The most dragged out, hotly contested election (M.O.) in US history, and only 4 years ago.
I haven't seen anything to date that would convince me that many people would actually 'forget' who they voted for in 2000.

Now, as far as the argument that it's impossible for 41% of 04 voters, voted for Bush in 2000 cause it doesn't add up is not an argument I came prepared to represent.
But let me just point out one thing...
Even though Clinton got a higher percentage (2.5 %) of 'who voted for' respondents, the number it actually equates to is possible.
(43% of 96,275,401 voters = 41,398,422, compared to the 44,909,806 who actually got their votes counted for him in 92)
Which can not be said of the Bush numbers from 2004.
(41% of 122m = 50,140,000 compared to the original 50,460,110 votes)
Which also contradicts the RBR/dem oversampling assertion.

I'm still curious on the Nixon numbers....but DON'T put alot of time into it.
I'm sure you're busy enough with other things.

[OnTheOtherHand]

205. OK, I want to make sure we don't get stuck in point/counterpoint

You are apparently hypothesizing that the extent of false recall should (or at least might) have been lower in 2004 recalling 2000 than in previous elections because the 2000 election had been so contentious.

I think that would be a pretty good argument about why false recall of the 1992 vote might be lower in 1996, because the 1992 _campaign_ was contentious -- or more specifically, as E/M point out, according to the CBS News polls, it had the highest level of interest among the last five, with 2004 a close second place. That might be why the 96-recalling-92 gap is smaller; if so, we can look forward to a smaller gap in 2008, maybe.

I don't think it's a great argument about the 2000 election, because most people weren't paying especially great attention _before_ the election. In fact, I'm not sure most people cared all that much who won, weird as that seems. But I don't have a strong opinion about that, much less strong evidence.

I do think -- not to be contentious, just for the benefit of lurkers -- that at some point the accumulated evidence warrants the presumption that a substantial "false recall" was at least possible, and indeed likely, even in 2004. I am open to serious contrary arguments, even speculative ones.

I think the argument about which numbers are possible is (taken literally) probably a red herring. Once we accept the premise that False Recall Happens, then AFAICS all the numbers are possible. 92-to-96 numbers are easier to accommodate because the turnout dropped by something like 8 million votes. But your point does raise interesting questions. For instance, maybe the 92-to-96 turnout retention was significantly larger in one party than the other (i.e., maybe relatively more Clinton '92, or Bush '92, voters stayed home), which would obviously affect our estimates of false recall in 1996. In 2004, I think we all agreed to assume that turnout retention was equal.

It should be quick and easy to get the Nixon numbers, if I can just download the file. We'll see.

(edit to fix a typo)

[OnTheOtherHand]

232. the Nixon numbers, for Chi

Recall that in 1976, Jimmy Carter narrowly defeated Gerald Ford, 50.1% - 48.0% per Dave Leip's atlas -- and in 1972, Richard Nixon had crushed George McGovern, 60.7% - 37.5%. In the interim, of course, Nixon was impeached and resigned in disgrace. So we would expect 1976 to be an exception to what seems to be a rule, that exit poll participants consistently overstate their support for the incumbent -- or, in this case, the would-have-been-incumbent.

The 1976 CBS/NYT exit poll interviewed 15,300 people. Of these, 6416 or 41.9% reported voting for Gerald Ford, and 8485 or 55.5% reported voting for Jimmy Carter. (The percentages should be slightly higher to compensate for a small number who reported not voting for president at all.) Yes, that's right, a double-digit error on the margin -- favoring the Democrat, once again. However, as with other unweighted results, these percentages don't account for possible bias in precinct selection or the effects of differential response rates.

Among respondents who reported having voted in 1972, 5470 reported having voted for Nixon, 3793 for McGovern, and 739 someone else. Those numbers conveniently total 10,002, so the percentages are 54.7, 37.9, and 7.4 respectively. The "someone else" percentage is sort of hilariously high, and perhaps reflects people being caught between Nixon, whom many knew and didn't like, and McGovern, whom many didn't know or like.

Still, Nixon held his own retrospectively, although he didn't actually increase his retrospective vote margin as other past winners have done. If we upweight Ford respondents and downweight Carter (and other) respondents to match the official returns, of course Nixon does better in the retrospective vote. He gets 58.1% to McGovern's 34.9% -- a 23.12% margin, just shy of his 23.15% margin in the 1972 official returns.

I'm going to continue to work my way through exit polls to see if I can find any in which (1) the Democrat didn't do better than in the official returns, and/or (2) an actual incumbent's retrospective vote margin wasn't greater than his official vote margin four years earlier. I figure there ought to be at least one.

[OnTheOtherHand]

233. oh, did I report on the 1980 poll yet?

In the 1980 exit poll (unweighted), Carter beat Reagan by 2.5 percentage points (actually, he lost by almost 10 points). And, retrospectively, he beat Ford by 18.4 points before factoring out non-voters etc., about 23.3 points among those who recalled a vote. So, even accounting for the Carter bias in 1980, looks to me like there will still be a hefty Carter retrospective bias in 1976 -- and this is for an incumbent who actually (or officially) lost.

Not surprising, since I'm pretty sure Bush did the same thing against Dukakis in 1992 (widened his retrospective victory four years earlier, even while losing). Of course, Dukakis hadn't actually been president, unlike Ford.

[Chi]

234. Thanx

Before you get to immersed in searching other exit polls....
The question should be why didn't the 'memory hole' appear in this exit poll?

Just to stay focused though, let's remember why it's an issue.
The 2004 exit poll was supposedly in error because of Dem oversampling
Yet the numbers reflect (in the 'who did you vote for in 2000') that it was actually an oversampling of Republicans.
The 'memory hole' was the reason given for why those numbers were not accurate, and therefore do not demonstrate sampling error.
So other polls that show an imbalance of returning voters, won't give us any answers.

Which goes back to the question....
If a 'memory hole', that causes voters to misreport their previous votes, is the result of people wanting to feel better about their previous vote (whether conscience or un-conscience), is a factor at work here....
Then why don't we see an influx of McGovern voters on this poll?

I chose this exit poll specifically for the reason, that if this memory hole exists, then this would be the poll that illustrates it the most.....yet it shows nothing, curious.

[OnTheOtherHand]

236. it doesn't seem so curious to me in context -- let's compare notes

While "wanting to feel better about their previous vote" might be a source of misreporting, I've never thought it was a decisive one. The pattern from the NES and GSS (which seems to be borne out in the exit polls) is that people, collectively, generally overstate their support for the incumbent, regardless of whom they support in the current election. So, of the people in the 2000-2004 NES panel who said in 2000 that they had voted for Gore, and said in 2004 that they had voted for Bush in 2000, about a quarter of them also said that they had voted for Kerry in 2004.

I think a lot of people just plain forget who they voted for (or, they conveniently forget that they didn't vote at all -- there "feeling better" is probably more important). The incumbent is in their faces, the loser isn't, so they end up remembering that they voted for the incumbent.

But I would expect the size of this effect to be influenced by the popularity of the incumbent (and the loser). What's anomalous about the 1976 case is that it's the only exit poll I've found where the previous winner didn't actually widen his margin -- and, given that the previous winner in question was Richard Nixon, that doesn't seem surprising. (Arguably we could throw it out since Nixon wasn't even an incumbent, but hey, I hate to throw out data.)

Otherwise, the exaggeration of Bush's 2000 support in the 2004 survey seems pretty much in line with other surveys, although it might be possible to argue that it is somewhat larger (or smaller?) than it "should" have been.

I think your statement about an "oversampling of Republicans" is problematic. Party identification is confusing enough without mixing in recalled vote. (The weighted results indicate an equal percentage of Dems and Reps; unweighted, the Dems have a 2-point edge. Since party ID is subjective, I don't know how we could figure out what the true percentages should be.)

[Chi]

237. Let me ask a simple question...

Do you think it likely that a person who did not vote in 2000, but showed up to vote against Bush in 2004, would misreport voting for Bush in 2000?

To me....that makes no sense.
If you believe it's likely, make a case for that scenario.

("oversampling of Republicans"....since 91% of Bush2k voters were Republicans, I don't see a big problem there. But if you wish, change "oversampling of Republicans" to "oversampling of Bush 2000 voters")

[OnTheOtherHand]

238. my starting point is fairly direct evidence that it happened!

In the 2000-04 NES panel, there were 757 people who stated
-- their 2000 vote (or non-vote) in 2000
-- their 2000 vote in 2004
-- their 2004 vote in 2004

So that's the limited sample on which I am drawing. I think it's safe to assume that some of the people who reported in 2000 that they had voted, didn't -- and probably vice versa. But it's the best we can do AFAIK.

So, among this group, 111 people said in 2000 that they hadn't voted. Among those 111, 26 of them reported in 2004 that they had voted for Bush in 2000. And of those 26, 5 of them said they had voted for Kerry in 2004; 1 reported voting for someone else.

Whatever else we do, we shouldn't rationalize those people out of existence. The confidence interval on 5 or 6 out of 26 is quite wide, but it doesn't include 0. It doesn't really matter whether you or I consider those people "likely" -- there they are.

That said, their existence doesn't surprise me. We've been discussing a propensity of a small but significant minority of survey respondents to report having voted for the incumbent when they didn't. If, as I believe, this is largely because people just plain remember the incumbent better than the last challenger, then I don't see why it is any more surprising for people who _misremember_ voting for the incumbent 4 years ago to vote against him, than for people who _correctly_ remember voting for the incumbent 4 years ago to vote against him. In fact, since they didn't actually vote for the incumbent in the first place, they should be more likely to defect.

[Chi]

239. What happened...is the issue.

Right off the bat a question would be....how many respondents said they voted for Gore in 2k, then misremembered and said they voted for Bush?

Here's another, how many that claimed they voted for Bush2k, voted for Kerry in 2004?

If your data indicates the majority of these people voted for Bush 04, it doesn't contradict the 2004 oversampling of Bush voters/Republicans...it supports it.
(Not that I am real hip on randomly selected, paid respondents, filtered 4-5 times)

[OnTheOtherHand]

240. answers

"how many respondents said they voted for Gore in 2k, then misremembered and said they voted for Bush?"

23 out of 312.

"how many that claimed they voted for Bush2k, voted for Kerry in 2004?"

The question is ambiguous. Of the 23 from the first question, 6 said they voted for Kerry, and 15 for Bush.

"If your data indicates the majority of these people voted for Bush 04, it doesn't contradict the 2004 oversampling of Bush voters/Republicans...it supports it."

I don't follow your argument here. (Again, I caution you not to treat those terms as synonymous. The kind of Bush voters who can't even remember whether they were Bush voters or not are not the kind of voters _I_ usually think of when thinking of "Republicans.")

The panel study indicates that Bush did retrospectively better, among 2004 respondents recalling 2000, than he did in 2000. This finding dovetails with other surveys (exit polls and otherwise) from other years. Whom the 'faux retrospective Bushies' then vote for is a separate question, isn't it?

I'm not enamored of the panel data either, but if we are stuck on the question whether false recall/reporting favoring Bush is intrinsically plausible, then I think the panel provides additional evidence.

[Chi]

241. OK

I agree with you on the point that out of the people who misremember who they voted for, that most probably didn't vote that year.

At some point in time I concluded that in order to disqualify the 'WVF 2k' results as an indicator of sampling,
you would have to make the case that those who misremembered had to be Kerry voters.

At this point I can find no flaw in your assessment, that the two do not relate in that way. 8/
That was pretty much my basis for disagreement on this point.

I think my Id has a front row seat on my Ego having a case of cognitive dissonance on this one....either that or that bad sleep I had is effecting my brain.
How's that for honesty 8)

[OnTheOtherHand]

243. it's excellent (but one thing...)

Me, I think this stuff is confusing -- as long as we are straight with each other about that, I hope I will never bust your chops about Id and Ego.

I just don't know whether your first statement is true, and I didn't mean to say that, or at least I don't think I meant to say that. Just using raw numbers from the panel, among people who change their answers from 2000 to 2004 (not the same as those who misremember, because some of them must have gotten it wrong the first time), Bush picks up a few more votes from "Gore" and "other" than he does from "did not vote." I haven't tried to collate the rest of the panel numbers, and I sure won't venture an opinion about real life. I'm pretty confident that the _proportion_ of actual non-voters who report having voted for Bush is higher than the _proportion_ of actual Gore voters who report having voted for Bush -- yeah, I know, crazy stuff ;). Whether most of the false Bush voters are non-voters, not so clear; I think probably not (at least among the people who actually voted in 2004).

[Bill Bored]

183. Probably not, but I don't remember actually doing it.

I recall watching the debates and I would have voted for Jimmy, but I can't recall pulling the lever that year or seeing his name on the actual ballot on the machine, or anything like that. So it's possible that I didn't vote that year, but not likely. Perhaps some would say they voted in 2000 when in fact they did not, rather than forgetting who they would have voted for.

But I don't like that whole "who voted for in 2000" argument anyway because in my mind, if there is no memory hole, it actually proves that the exit polls suck because they came up with the wrong answer about an election the outcome of which was known as far as vote counting goes.

As I say later down the thread, my problem with this debate is that it diverts time and attention away from developing other better evidence of 2004 election fraud -- not that it actually delays election reform per se. So I have softened my position a little.

On the other hand, if there is stronger evidence that isn't being uncovered, due to this diversion of investigative resources, that otherwise would have been uncovered in this or even a future election, it could delay at least some election reform too!

Let's let Freeman and Mitofsky have at it and see what they come up with. Who knows, maybe they'll actually agree on something!

[Time for change]

53. I think exit poll issues AND the issues that you discuss are important

Nobody can argue that the issues that you are working on are very important.

I feel strongly that making the case for a stolen election is also important, because we need the political support of a lot more people in this country in order to give meaningful election reform a better chance of becoming a reality.

Exit polls are part of that argument. Not the whole case, but certainly part of it. Do they prove that the election was stolen, all by themselves? Of course not. But do they substantially raise the suspicion level for a stolen election from what many of us consider to be an already high level? I think that that is also hard to argue with.

And now I'll repeat something I heard on DU a few weeks ago, because I can't think of a better analogy myself: Saying that people ought to stop trying to show that the election was stolen so that they can devote more time to meaningful election reform is like telling someone that they should be a doctor rather than a lawyer (or vice versa) because the one is more important than the other. They're both important, and different people have different capabilities and talents.

But I'm awfully glad that people like you are working on the things that you are, because it's sorely needed, and I don't know enough about it to make a meaningful contribution.

[kster]

47. Its a waste of time, because we won that debate.

hands down. Time to get rid of the vote stealing machines,Paper Ballots Hand Counted By The People. O6

[Nederland]

70. Won the debate?

By what measure?

If you did a poll of Americans and asked them if exit polls proved election fraud, what would the results be? I humbly suggest that the answer would be "exit wah?"

Or perhaps you'd like to do a poll of people with Phd's in statistics? Do you think the fact the USCountVotes latest paper had only two signatories instead of the original twelve reflects at all on how many people still believe that there is conclusive evidence that exit polls proved fraud?

Please, the only place that the debate was "won" was amongst a handful of people in an isolated dead forum on a website called DU.

[m.standridge]

72. Is it possible the focus was too broad?

in other words, the Ohio exit polls may have been useful, as they occur in the midst of widespread irregularities that may have been massive enough to have actually affected the outcome in a previously "red" state.
There were some irregularities in GOP favor in some other states--Nevada, New Mexico, apparently Florida and Colorado, but it wasn't clear they were massive enough to turn them around this time, although New Mexicso seemed possible.

But in the Electoral College, Ohio was the big one. This time around, the Red state seemed to turn white, as Professor Phillips noted.

To try to determine how accurate every exit poll in every state was, seems a daunting task, doesn't it?
Surely we can say there was fraud going on. But how massive--and was it massive enough to change the outcome of the election? Clearly, in Ohio, we're talking about a situation involving probably the turn of 50,000 votes or less. And enough Electoral votes to give Kerry at least an E Electoral victory.
So, in a situation like that, inaccurate state-level exit polls, on top of other evidence of numerous types of fraud and state-level government malfeasance, become more significant, having a kind of "straw that broke the camel's back" quality, do they not?

[m.standridge]

74. "a situation involving probably the turn of 50,000 votes or less"--er, 59,

anyway, half of the 118,000 current "official" Bush lead in Ohio, i.e., Kerry needs 59,001 of Bush's votes, right? And that's assuming the August trial doesn't call into question the recount methods and change the official total in some way, to a smaller Bush margin.

[kster]

76. TV Networks Officially Refuse to Release Exit Poll Raw Data

Heres what gave me my first clue that we won.

http://www.truthout.org/docs_04/123004B.shtml#1

[Febble]

83. FYI

First: the "raw exit poll data" is always released, in enormous detail. And it was released as usual, in January.

It is available for download here:

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

What has not been released, and never is, is precinct identifiers, or the precinct vote count totals because releasing these would allow the precincts to be identified.

And the reason this has not been released, and never is released, is that the data are collected under the ethical guidelines of the professional body that governs survey research, AAPOR. And like ethical guidelines that govern all social and medical research, these includes stringent confidentiality rules. If I collect data from participants for my psychology research I can publish it. I can even publish my raw data. What I cannot do is publish any information that would allow my participants to be identified.

The data that is downloadable at the above link contains the responses of every single participant in the exit polls, plus demographic details like age, gender and ethnicity. This means that in many cases actual respondents could be identified and their vote determined. People who participated in the poll did so on the understanding that their response would be kept confidential. Only if the data was subpoena'd should this be violated.

However, it is possible for the vote-count data to be prepared in such a way - "blurred" a little - so that analyses can be done by third parties. It is true that the networks own the data. However this was not a bar to the preparation and release of a "blurred" dataset from the exit polls in Ohio to ESI, and their findings have been released, although not yet the form of a peer-reviewed paper.

http://www.votewatch.us/Members/stevenhertzberg/report.2005-07-19.7420722886/view?searchterm=None

It would appear therefore, that there is no bar in principle to further datasets being prepared and released.

Finally, the "raw" aggregated data, in other words the unweighted state estimates, were released back in January by E-M in their published evaluation:

http://www.exit-poll.net/election-night/EvaluationJan192005.pdf

None of these releases may satisfy your (or my) desire to know more. They are all flawed in their way (although I am hopeful that ESI will eventually publish details of their methodology).

But the idea that the election must have been stolen otherwise why would the TV networks be refusing to release the data is simply unsustainable.

a) it WAS released, as usual, in January
b) what was not released is bound by stringent rules regarding confidentiality that have nothing to do with ownership of the data.
c) notwithstanding (b) blurred precinct totals for Ohio were released to ESI and an important analysis has been done.

(edited to revise irritable subject line)

[Chi]

85. Quick question

So, to your knowledge, precinct identifiers have never been seen by anyone other than Warren Mitofsky (close associates included)?

[Febble]

86. To my knowledge, no.

My understanding is that for ESI's "blurred" dataset, vote count totals for precincts in Ohio had a small random number added or subtracted to them so that they could not be identified. However, they must have had county IDs as some of their analysis is done by county.

We could really do with a methods section for their report.

[Chi]

87. I failed to phrase my question properly...

Historically speaking, past elections included, has anyone other than the people who run the exit poll, ever been allowed access to the precinct identifiers?

Thanx in Advance

[Febble]

88. Ah.

Not that I know of. Try OTOH.

But I do know that AAPOR ethical guidelines have stringent confidentiality guidelines, and also that demographic details of respondents have been published in previous years.



(on edit: just spotted an error in my post above: I should have said that the E-M evaluation has the un-re-weighted state estimates i.e. the estimates not re-weighted to the vote returns.)

[Chi]

90. Thanks n/t

[Bill Bored]

92. Releasing precinct IDs would violate whose confidentiality exactly?

I'm surprised they even ask respondents for their names in the first place. Do they? Or are we talking about Mitofsky's "trade secret" precinct selection algorithm?

Give me a BREAK! If you don't believe the election was stolen, it means the polls sucked anyway, so who would want ot steal his useless methodology? That would be like stealing Diebold's "trade secret" software because you wanted to count votes!

[OnTheOtherHand]

94. no, neither

Heck, the interview forms are on line, so you can verify that they don't ask for names. (And there is no ethical standard about protecting Mitofsky's trade secrets, no matter how lousy they are!)

The big problem is that the interviews contain a lot of demographic information, probably enough to serve as a unique identifier in many precincts.

People think this problem is contrived (and the way social security #s get flung around these days, maybe the confidentiality concerns are anachronistic), but the Census Bureau and Federal Reserve and other government agencies work hard on making sure that no one can identify individual respondents from their survey data.

On the other side of the coin, lots of people have told me that we could get to the bottom of election fraud if all the exit poll data were released, but they generally trail off when I ask them how. (As I think I've written before, the CAPE folks in San Diego have all the information from their parallel election -- which does include names, by the way -- but as far as I can see, it doesn't get them any closer to proof of fraud.)

I don't know whether it is a waste of time to address this particular exit poll question again, but since you asked....

[Bill Bored]

98. Not a waste of time because it goes to the issue of why the debate

has still apparently not been settled.

But there is a group of DUers who have identified many of the Ohio exit poll precincts, if not all, and this is more useful than some of the other stuff that keeps being debated. Perhaps Kiwi Expat can produce a summary report of their findings to date, since they actually had a man on the ground in Ohio (liam_laddie) looking at stuff.

[OnTheOtherHand]

109. yeah, I'd like to know more about what was learned in Ohio

My feeling has been that at most, the exit polls might give you some ideas where to look -- and we already have ideas where to look. So then the follow-through becomes crucial. Of course a lot of that happens behind the scenes, but if (for instance) liam_laddie didn't get cooperation from local election officials, are we missing a chance to follow up with legal support?

[kiwi_expat]

116. Ohio Exit Poll Raw Data thread

Hi Bill

I just kicked minvis' Ohio Exit Poll Raw Data thread to the ERD front page, again, so it would not die.

By early July, we had managed to identify 16 NEP precincts, before getting bogged down with Summit County. (The model just didn't seem to fit exactly.)

Of the precincts we did identify, the precinct with the largest WPE (-28%) was Cincinnati 4-M.
http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=329749&mesg_id=381571

Liam-laddie made several trips to the Hamilton County BoE to try to audit Cincy4M and a couple of other precincts. The staff were very courteous and moderately helpful to him, but they pleaded that they were too busy to actually let him count the ballots.

Liam was also considering trying to look at ballots in neighboring Clermont County. However, his experience at Hamilton County, where the staff were at least friendly, put him off even trying to approach the Clermont BoE staff.

I'm afraid that, as OTOH has suggested, some form of legal intervention is necessary in order to get a manual recount under the Ohio Sunshine laws.

The county most in need of a proper manual recount is Cuyahoga county. I suggested to Time_for_Change that someone on DU might have an "in" with Kucinich's office and might be able to persuade them to donate a staff member or a keen volunteer to at least try to inspect the Cuyahoga poll books' ballot counts. I had earlier considered going to Cleveland, myself, but realized it would be impossible to arrange cheap trans-Pacific fares with enough flexibility to accommodate the whims of BoE-staff scheduling.

Perhaps some retired/unemployed DUer might have the flexibility to hop on a (cheap?) domestic flight to Cleveland, at short notice.

As Liam has demonstrated, it takes a lot of determination and persistence to have any success at all. Are there any other Liams out there?

[OnTheOtherHand]

136. whew

OK, there's a lot in that thread, I don't claim to have mastered it.

WRT Cuyahoga, per Minvis in #149 in that thread, there were six exit poll precincts of which one was off by about 10% on the Kerry share (that would be a -20 WPE). That's Bedford Heights 4B, exit poll 47K-4B-1 DNV-1 blank, official result 380K-93B-2Peroutka.

That appears to match precinct #8 in the ESI report, so it looks like a good match.

By visual inspection (with a digital loupe) of the ESI scatterplot of Bush vote proportion in 2000 vs. 2004, it looks like that precinct was about 18% Bush in 2000, vs 20% in 2004. From the 2000 canvass report, Bedford Hts 4B was 17% Bush. Actually a bit more error than I would expect, but that could be the blurring. BedHts 4B in 2000 went 362 Gore, 78 Bush, 19 other.

I would hazard that none of the exit poll precincts in Cuya looks like a top candidate for fraud (although BedHts 4B is possible if it was hacked similarly in 2000 and 2004 -- and of course the exit poll samples aren't big enough actually to rule out fraud anywhere). In Cuya, you might want to identify suspicious outliers by some other means (I know this isn't utterly uncharted ground!).

Looking at the ESI scatters, if one had to choose one precinct, it would be the one with what appears to be a 17% error in Kerry proportion (most likely ESI's precinct identifier #6, vote proportion 36%, exit poll proportion 53%) and a 7% change in Bush vote proportion from 2000 to 2004 (which, if this is the right precinct, would mean 43% Kerry in 2000). There are actually rather few precincts where the Bush 2004 vote proportion is appreciably higher than both the 2000 vote proportion and the 2004 exit poll result; this is the most striking of them. I have no idea where it is (although it may already be identified in that thread).

(EDIT: BTW I spent a long time staring at that ESI scatterplot trying to reverse-engineer crucial data points in view of the ambiguity of the axis labels, but I could still have it wrong. I'm assuming that these precincts where 2004 Bush vote proportion is high versus both 2000 vote and 2004 poll are at UPPER LEFT in Figure 3.)

[Bill Bored]

149. Glad to see you guys talking about this

instead of the usual exit poll debate stuff.

I would suggest as another criterion for outliers, to ignore the 2000/2004 correlations in the precincts where there were collocated ballot order rotations between Bush and Kerry, esp. if this also existed between Bush and Gore if that's known. That way, you can assume that a exit poll discrepancy in 2004 might have existed in 2000 for the same reason. Instead of ruling the precinct out, it becomes a potential outlier.

I believe in order for Kerry to have lost Ohio due to counting fraud, there had to have been some vote switching to Bush instead of just undervotes and switches to third party candidates.

[OnTheOtherHand]

150. that's sort of an un-criterion, isn't it?

But if we can get info about the ballot order rotations in 2000, that would be excellent -- and otherwise, we can at least see whether rotations are correlated with odd "shifts" that didn't come out in the DNC analysis (that's 2004 vs. 2002, of course -- would there also be contrasts with the 2004 Senate race, where there were apparently only two candidates? of course that wasn't very competitive...).

I've got no money on that hypothesis, but it is crying out for a decent test.

[kiwi_expat]

156. I'm afraid I have to agree with your statement in #109.

"My feeling has been that at most, the exit polls might give you some ideas where to look -- and we already have ideas where to look. So then the follow-through becomes crucial." -OTOH


My reasons for saying that Cuyahoga is the county most in need of a proper manual recount have much more to do with the various documented anomalies, than with the exit polls.

Cheers.

[mgr]

95. The right to vote in secrecy is the basis of this. n/t

[Bill Bored]

99. Well, that right is given up whenever someone submits to Early Voting

because they need to link individual voters with their votes so they don't vote again, or even often!

However, I do agree that this is a legitimate right despite the fact that folks are told it's OK to forfeit it in a number of states, or aren't told they are actually doing so when they vote early.

In any case, exit pollsters don't take names, so unless a certain 6 foot 4 inch Arabic gentleman who has benefited a great deal from our current administration is interviewed. and non reluctantly admits to voting for Bush, I don't think it's so easy to figure out who's who.

[mgr]

114. A precinct is small place, as is a census plot.

What I think you may fail to accept is that government is entrusted to protect our privacy, but not so other third parties. One chooses how one's privacy can be invaded. If I disclose that I am Jewish, male, and make $60,000 a year to an exit pollster, contingent on their maintaining my autonomy; and that response is made available to others along with knowing where the precinct is, all one would need to do is cross reference it with census data and a phone book to come to my door. That is a serious broach in my privacy even if the outcome is a mass mailing tailored to my ethnicity and income.

Mike

[Bill Bored]

147. Well, you could always just ask for a raise! nt

[mgr]

170. LOL!

[kiwi_expat]

89. Precinct identifiers can be painstakingly determined ....

and have been done for 15 NEP precincts in Ohio (Ohio Exit Poll Raw Data thread).

The painstaking part is the hounding of BoEs for lists of the exit poll precinct names. The rest is easy - thanks to Skids.

[Chi]

101. Kudos again for that task...

(Chi points upward)
I think post 98 is partially for you.

[kiwi_expat]

117. Thanks from the OEPRD gang (minvis, skids, blue22, liam_laddy....) n/t

[mgr]

146. Post the thread, its one to be proud of n/t

[kster]

121. HERES WHERE WE DISAGREE,

"But the idea that the election must have been stolen otherwise why would the TV networks be refusing to release the data is simply unsustainable".

If I was the media and the American people was accusing me of the biggest crime in America's history, My damn *ss would be laying everything out to prove to the American people that they are wrong. But thats just me.

Your explanation of what is released or not.

a) it WAS released, as usual, in January
b) what was not released is bound by stringent rules regarding confidentiality that have nothing to do with ownership of the data.
c) notwithstanding (b) blurred precinct totals for Ohio were released to ESI and an important analysis has been done

Imagine this, when you get stopped by a cop and he wants to search your car, try telling him this.

A)Yes, officer you can search my vehicle

B)But you cannot search the glove compartment.

C)you can also search the trunk officer.

Is there any RED flags that the police officer should wonder about at this point or should he just be happy because he was able to search most of the car.

I figure I can explain away (B) I'll tell the officer that its my sisters car and I want to protect her confidentiality.

Think the officer will buy it. According to you he should.

[Febble]

123. I don't expect the officer to buy it

but nor do I expect the driver to allow the officer to search the glove compartment if it contains people's confidential medical records.

Especially if he isn't actually a police officer.

By all means subpoena the information, and have it examined by whoever you nominate under some form of non-disclosure agreement. In effect, that happened with the Ohio data.

I'm not saying that there was no fraud. I am not saying the data should not be examined.

I am saying that the fact that the driver did not allow the officer to search the glove compartment is not in itself suggestive of fraud. It simply follows from the fact that the data contained in the glove compartment (as opposed to the data contained in the trunk) is bound by fairly stringent confidentiality constraints. As it should be.

Ironically, had the stuff in the trunk not already been released - in line with annual practice, the stuff in the glove compartment could be. It's just you can't release BOTH demographic details of the respondents AND precinct identifiers because otherwise respondents could be identified.

Yes, it would be good to get more data released
No, the fact that more has not been released is not in itself suspicious.

[kster]

124. PLEASE !

"Yes, it would be good to get more data released
No, the fact that more has not been released is not in itself suspicious".

Maybe in your own little world.

Come on you can not believe what you are saying.

[Febble]

125. Explain please

Why is it suspicious that a polling organisation has not released confidential data in violation of the ethical guidelines of its own professional body (AAPOR)?

Is AAPOR a "little world"?

[kster]

155. If you were in their business

You have 3 choices,

A) Break confidentiality and release.

B) Arbitrate a solution so that you can release without breaking confidentiality.

C)Explain it away in the hopes that nobody forces you to pick A or B.

If you were in their business what would you do?




Oh, I apologize for the "little world" comment.

[Febble]

163. B

And, essentially, that's what they've done.

But it's time-consuming and expensive so if people want to raise money for further release it would be worth brain-storming exactly what kind of questions would be worth asking of the data.

Which is largely what I've been doing for the last few months! It's not obvious, for example, what kind of hypothesis would be worth testing. The thing about statistical analyses is that you have to have a hypothesis to test. You can't just look at the data and see the answer. Well, sometimes you can, but on the basis of what's been released so far there is not an obvious answer. Two fraud hypotheses have been advanced, to my knowledge, and both have been tested and come up negative. That's not to say there aren't more.

But one of my purposes in trying to do some brainstorming on DU is to try and thrash out exactly this kind of issue: what would the data look like if the red-shift was due to fraud?

No worries re "little world"! We are all in little worlds, and the miraculous thing about blogs like this is that it's a way of widening the world we live in. Finding out what other people think and why they think it. Even if we find we still disagree!

Cheers!

[OnTheOtherHand]

127. calling all civil libertarians!

For heaven's sake, let's not hurtle down the slope of insistence that privacy concerns are a scam to protect the guilty.

kster, if someone flags down my car and says that "the American people" insist on searching my glove compartment -- well, I don't think there is anything interesting in there, but don't count on me sticking around to try to assuage The People's suspicions. And if "the American people" insist on strip-searching my daughter, well, they would be well advised to step away from the car.

(EDIT: To emphasize: When I say "someone," I mean "someone." I and my daughter have constitutional rights even against police officers. But AFAIK no one who is demanding more information from E/M has an iota of legal authority for the demand.)

I agree with Febble: if someone really thinks that E/M has evidence that would prove fraud, then by all means they should subpoena it -- or make some other arrangement. But frankly, there aren't a lot of people thinking and arguing seriously about what that evidence could possibly be. Mostly there's the argument that E/M's unwillingness to disclose everything preemptively, just in case, is a cover-up.

[kster]

157. I kinda agree

"And if "the American people" insist on strip-searching my daughter, well, they would be well advised to step away from the car". BUT

when she gets drafted because we didn't do everything in our power to be sure that the people who decided the draft were not selected by a vote stealing machine.

Do you tell the United States Military to step away from he car? What do you tell your daughter then ?

[OnTheOtherHand]

159. bless your soul!

As long as we understand each other -- I hate to get stuck in a dialectic of disagreeing about everything because, well, we disagree about everything. (Which we certainly don't.)

Again, if someone can explain how releasing more exit poll data will reveal whether the election was stolen, then we should find a way to get it done.

I don't see it. I'm not invested in arguing against more data release, but I won't argue for it if I don't see a specific argument. Just a general argument that this is a national emergency... well, we all know the problem with that.

(EDIT: see also kiwi_expat in #156 -- although obviously kiwi_expat isn't arguing against further release, cf. #154.)

[kiwi_expat]

154. I think people who say they want the NEP "raw data" usually mean...

that they want the pure demographically(etc.)-weighted data released.

The Univ.of Mich. data (ftp://ftp.icpsr.umich.edu/pub/FastTrack/General_Election_Exit_Polls2004/) contains the respondents' raw data and the FINAL weighting for each respondent record - a weighting which includes vote count adjustments.

What most of us would like to see are weightings that do not include vote count adjustments. I don't even know if Mitofsky has that data. He might have overlaid the demographic-weightings with the the combined demographic-votecount weightings.

For example, I know that the interviewers noted the age, ethnicity and gender of missed-respondents and refusals. Respondent records with the same age, ethnicity and gender would have been weighted accordingly. And I assume that Mitofsky would have weighted respondents from some precincts more heavily than respondents from others. (To offset the fact that he had polled more Democratic than Republican precincts, for example.) The list goes on.

Does Mitofsky have a detailed model that he used for those weightings? Would it be possible for us to figure out the "pure" weightings (no vote counts) if we could see that model - plus records of the missed/refusal respondents etc.?


Unlike precinct identifiers, this is NOT a privacy issue. (Except E-M commercial privacy from rival polling-companies.)

[OnTheOtherHand]

158. ooh, that's a good question

(I'm not at all convinced that that's what people "usually" mean, but it surely depends on the people!)

I don't know whether it is possible to reconstruct the "original" weights (or whether they still exist, no reconstruction necessary).

Forgive me for probably an obtuse question (I'm just drinking my coffee over here), but why would this be important? I sort of gave up on thinking that any permutation of weights would shed much more light, but that could well be laziness on my part. I agree that AFAICS releasing the weights, and/or the model, wouldn't violate privacy.

[Febble]

161. I think you just want it raw

frankly. This is because E-M found that the precinct selection was pretty representative, so precinct weights would be minimal.

The "bias" was between respondent and count.

What you need are the counts for each precinct, which are what ESI got in blurred form for Ohio.

Then you can see where the gaps between poll and count are.

Now, how do you measure the gap?

Weeeeeelllll......

I thought I knew.

Ask OTOH.

[kiwi_expat]

162. Wouldn't we at least want to correct for bias from refusals?

"When a voter refuses to participate, the interviewer records their gender, race and approximate age. This data allow the exit pollsters to do statistical corrections for any bias in gender, race and age that might result from refusals." -Mystery Pollster

Isn't it important to get the refusal-related weightings, in order to more accurately compare precinct exit polls with vote counts?

* * *

"What you need are the counts for each precinct, which are what ESI got in blurred form for Ohio." -Febble

Thanks very much for the suggestion re. ESI data!


Cheers.

[Febble]

164. I think that might have been something

that MP got wrong. About the weighting anyway.

But yes, that data would be useful, as would response rates, completion rates etc.

What I really meant was that the weights they used are not that important, seeing as we know pretty well that it wasn't the weights that were wrong - it was that the polled responses simply didn't match the count.

But if MP is right (and I can't find where I seem to remember him correcting that statement) then yes, those data would be useful as well.

[OnTheOtherHand]

168. I always assumed MP (and kiwi) were right about the weights

p. 9 of the E/M report states baldly that the survey weights incorporate an age-race-sex adjustment for non-interviews. But I'm not sure what is the basis of that adjustment.

It does seem to be true that the geographic weights wouldn't matter so much.

I'm not sure how important the age-race-sex adjustments would be. One place to start that inquiry is on p. 54 of the E/M report -- estimated completion rates for various demographics -- the most startling one IMO is an average 10 points lower among 60+. Upweighting those folks would give Bush an appreciable boost. And the apparent gender gap in response rates is not quite trivial.

Oh, by the way, I should clean up some misinformation I posted elsewhere -- too distracted to figure out where. The E/M report says (p. 52) that they got back about 95% of the questionnaires. I was confused on that point.

[Febble]

169. So it is.

Maybe I dreamt MP had said that. O boy, dreaming about MP....

Okay. Yup, that stuff might be useful.

[kiwi_expat]

178. skids has said that all kinds of stuff can be found on the NEP data file.

http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=203&topic_id=346440

I wonder if missed/refusal respondent records are already on the file - along with the age, gender and race (and NEP precinct) of each. It is entirely possible.


I have a question for you and Febble: am I correct in inferring that the NEP questionnaires were self-administered (i.e., the respondent had to read the questions and fill out the form)? If that is true, I would expect the poorly educated and less intelligent to be under-represented. Those refusal-respondent characteristics would not necessarily be reflected in adjustments based on age/race/gender.

[OnTheOtherHand]

181. a few answers

The files have lots of info -- but AFAICS only for actual respondents (and not all of those, but the ones whose questionnaires were phoned in from the field -- also the phone survey folks, I suppose). I don't think there is a file from which one could recreate the completion rate info by demographic that E/M give in their report. (I can't rule out being just plain wrong about this, but if there were one record per refusal, that I would have noticed for sure!)

Yes, the questionnaire is self-administered, and I agree that that raises questions about representativeness. That might be a wash, since the candidates seem to have run practically even among the relatively small number of respondents who reported not having a high school degree -- not that I take that as a very reliable test.

I don't know if you know that if you go to the NEP web site, you can actually download all the questionnaires, which may be of interest. And of course all the data are still available, in ASCII and SPSS (.por and .sav) formats, via ftp://ftp.icpsr.umich.edu/pub/FastTrack/General_Election_Exit_Polls2004/ . I could, say, exhale the national file to Excel if that helps you get started with something. (You can do everything with the ASCII files, but that is a hassle.)

[kiwi_expat]

182. Thanks.

I have no idea what percentage of voters did not graduate from high school. I do know that (by definition) half of the population have I.Q.s of less than 100. I wonder how many people with an I.Q. of 80-90 could fill out a questionnaire.

I'll ask skids if he has encountered anything on the file that might look like refusal information.

[Bill Bored]

184. Useful might be an understatement. nt

[OnTheOtherHand]

186. why -- what are you looking for? n/t

[Bill Bored]

187. Well, if you had the demogs. of the non-responders, you could

take an educated guess as to what their responses would have been, based on responders with the same demogs., instead of just saying they would have been Bushies because the vote count said Bush won.

If this has already been quantified in this way, then I'm just not up to speed on the latest nuances of the exit poll debate. Sorry.

[OnTheOtherHand]

188. what, you didn't get the memo?

Exit polls are what the naysayers want us to talk about instead of paper ballots, hand counted.

;)

Umm, if we had complete national results, with the demographic weights but without any election results, those would be E/M's best effort (at the time) to answer that question.

We could do this, back of the envelope. I'm not sure it would be hugely helpful, because it's not really a brilliant assumption that, say, because respondents over 60 broke 53-47 (or whatever) for Bush, non-respondents over 60 did the same thing. But it's not a terrible assumption, either. Mostly I'm just burned out on exit poll hypotheticals. Maybe kiwi_expat would like to run with this for a while?

[kiwi_expat]

191. E-M did weight the responders with the same demogs as the nonresponders

....but the results obviously did not match the vote counts.

So Mitofsky then further re-weighted all the responders as the vote counts came in. (His job was to predict the vote count winner.)


If we could reconstruct his original weightings it would give us a somewhat better exit poll picture - better than just the "raw" data.

[OnTheOtherHand]

192. I can imagine two lines of "attack"

One is to ask Mitofsky: hey, wouldya mind sharing the weights? (Shrug.)

If he minds, then -- well, it depends on how close we want to come. It shouldn't be hard to come up with weights for the national survey that match the reported completion rates on p. 54, to get some idea of how much it affects the overall results. (Personally, I don't care so much, but that could well be a failure of imagination.) --The weights presumably would and should vary by geography, of course, so anything we did based on the national rates would at best be a first cut.

[kiwi_expat]

215. I once tried using the Ohio gender/age completion rates.

Weighting for the higher refusal rates of males did not have a dramatic effect. I gave up trying to weight for the oldies. It was too complicated (two sexes were much easier to adjust). However it did NOT look like the oldies' weighting would make a big difference - because not that many oldies voted.

Nevertheless, I think the WPE scatterplot really should include demographically weighted data - and the comparison should be with polling place results instead of precinct results. (It appears that, normally, all of the polling place voters were sampled.)

I assume that would apply, to some extent, to Febble's alpha scatterplot also.

[Febble]

221. the weighting question is probably more interesting

than I gave it credit for.

The reason I think it is interesting to look at "my" scatterplot, though, is that it really is "raw" - weighting, if it was any good, will tend to reduce the error. The unweighted data really does tell us how bad the error was - how big the bridge is that you have to cross. And it is a big bridge.

However, I can see that if you could explain away the some of the gap by appropriate weighting, to compensate for non-representative sampling, the residual error ought to tell you what was left to be explained - how much of the bridge is left to cross. The problem is that it will only tell you whether broad demographics were under-over-representated. It won't tell you whether, within each demographic (say middle-aged white women, the ones who are present are more likely to be Kerry voters than the ones who are missing.

But I agree that completion rates by demographic would be a useful way to go. However there are other potential factors that may also bridge the gap, but which will be invisible to the weighting by age, race and gender. So I still think it's useful to start with the unbridged chasm. It's certainly larger (in effect size)than the gap between the early projections and the count.

[Bill Bored]

165. This might go back to what I suggested earlier

about Mitofsky's proprietary system being the "confidentiality" he can't violate.

If he believes his exit polls were so bad, why would he care if his competition copied his technique? After all, he thinks the elections was fair and square and his polls just sucked right?

[Febble]

166. Well that may be true

but I don't think his algorithm will tell us much. The ingredients would be more useful than the recipe. It wasn't the recipe that was wrong.

Although, far more important, as OTOH, is good questions to ask of the data.

Once you have a good question, you are nearly there. Instead of demanding "release of the data!" you can ask your question.

Or the wherewithal to answer it yourself. But the questions have been slow in being formulated. ESI's was a good one, I think, as was Josh Mitteldorf's. But the answers to those appears to be negative. So what next?

[Peace Patriot]

73. I have to laugh at these guys saying the exit polls are not accurate.

At least they are independent from Diebold and ES&S's secret formula "results."

Why don't y'all spend some time discussing how to check the accuracy of vote tabulation that was done with SECRET, PROPRIETARY programming code, by Bushite corporations, who insisted (along with their good buddy Tom Delay) that no paper trail was needed for their election theft machines?

We had a non-transparent, unverifiable election, which hardly deserves the dignity of the word "election," and the exit polls are one of the few external gages we have, to figure out what really happened, since WE CAN'T DO IT FROM THE VOTES THEMSELVES.

[Bill Bored]

78. PLEASE See post 44 and 77!

And the odds of both of them having double repeating integers must be a trillion to 1 at least! :)

[kster]

79. (Post 44) We'd have all 3000 counties audited by now

thats alot of work, how about this, we start putting people in jail for breaking the common sense laws that we have on the books already.

What do they mean "TV Networks Officially Refuse to Release Exit Poll Raw Data" start putting these people in jail, until they do release it.

After they release it and the numbers don't add up,on to the vote
stealing machine makers, after we tell them to stick their SECRET PROPRIETARY programming code up their *ss. We start opening up machines.

After we prove our case we get a hold of local area printing companies and create new jobs in our own communities by giving them huge orders to print PAPER BALLOTS for 06!

[kster]

82. TV Networks Officially Refuse to Release Exit Poll Raw Data

Could someone please tell me why, it could end this 11 month exit poll debate.

http://www.truthout.org/docs_04/123004B.shtml#1

But for some reason they won't release the Raw data.

Why?

[Febble]

84. See reply 83

[petgoat]

96. Is the demographic info necessary for USCountVotes to recap

Mitofsky's WPE analysis?

If not, why are confidentiality concerns preventing the release of
raw data for WPE recapitulation?

[OnTheOtherHand]

97. because the demog data have already been released

as in past years. (PMJI, it's after midnight in Febbleland, so I thought I might be able to answer sooner.) Most if not all the exit polls from 1988 on (and some older ones) are available via ICPSR, raw data galore, complete with demographics -- but not precinct identifiers that would let you identify the original precinct.

There is no way to unrelease the demographic data now. However, it is possible to "blur" the precinct-level data, as was done for Election Science Institute's Ohio analysis. This is expensive, and it would be useful for more people actually to look at the ESI work and see whether they think there is any point to doing it for other states (and, if so, which ones). Febble has posted several good explanations of this, but I don't have links handy right now.

(But if you don't feel like looking at the ESI stuff, which is fine, Bill Bored may have some other suggestions!)

[petgoat]

100. "There is no way to unrelease the demographic data now."

Okay, I get you. The demographic data is released, so the precinct
data would violate confidentiality.

Then obviously we need an expert commission under non-disclosure
agreements or whose report justifies the expense of the data smearing
so they can study the error.

Of course there's no point because in the last 11 months the data could
have been subverted.

[OnTheOtherHand]

110. yes, the data aren't exactly bulletproof

I've read the claim that the exit polls have a paper trail, but I'm not at all sure that it is true -- I don't know whether NEP even tries to collect all the surveys. (VNS in 2002 did try to collect all the surveys, and apparently got about 80% of them.) And if they do, I doubt the security protocols are up to spec.

Yes, inspection under non-disclosure agreements, or further data blurring, both are viable. AFAICT, most survey folks don't think that close inspection of the exit polls would reveal much more about the possibility of fraud, so there isn't a collective sense of urgency about this. And there isn't much concern about data subversion, because we generally don't think the data could be important enough to be worth subverting. But....

There's a matter of degree here that is hard to convey in DU discourse. Some folks here really do seem to think that the "raw data" have some conclusive proof of fraud, and so Mitofsky and/or the networks have conspired to cover them up. Very few survey folks believe any part of that, AFAICT. But the exit polls still could contain further suggestive evidence. So it's totally legitimate to try to figure out, for instance, is it worth it to try to get X dollars to blur some more state exit poll data, or would we be better off spending that money on FOIA work (or whatever)? (And of course if some people want to spend their money only on one of those, then they can do it -- no need to reach consensus on priorities.)

[kster]

122. See reply 121

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=393939&mesg_id=394538

[autorank]

103. My position...

This is an anonymous forum. What people do with their time is their business. The level of self-revelation is optional but not necessary, or even desirable on a frequent basis. Nor is it desirable to engage in criticism of anonymous users for the use of their time, particularly when they're anonymous (and the critic has not idea how they use their time).

It's alternatively amusing and tedious to be attacked as some sort of slacker, dilettante, or fifth columnist for expressing a strong interest in the power of exit polls to point to fraud and generate public awareness. Don't even respond to my post if this is what you have in mind because I'm through repeating myself. You won't convince me of anything and nobody's watching or influenced by this. I won't convince you. Now that "The Big Show" is gone, the lurking is infinitesimal, the attention on the exit poll issue is elsewhere, and none of us is doing much other than speaking to each other when debating this issue. Consider this an announced "virtual ignore" by me on "engagement" on the issue; and consider what it means to post to someone who has already said they won't respond on the issue, tacky. I'm too busy, as it were, for the nonsense.

What are my priorities?

1) Free and fair elections, transparent and open to all.

a) Motor voter registration
b) Other easy access registrations
c) Any ID will do that has a picture or is deemed appropriate.
d) No "new" ID's, no voter ID's
e) Early voting starting the Saturday and Sunday prior to the constitutionally mandated date all over the country.

3) To meet that goal, there is a simple solution:

a) Paper ballots marked with a stamp in a closed voting booth.
b) Hand counts of paper ballots by volunteers without a strongly identifiable partisan bias; with partisan representatives looking on.
c) Public announcements of precinct total published in a variety of media.
d) Public transmittal of the announced results to tabulation center(s).
e) Visible presentation of the results received at the central tabulator facility
f) Totals that everyone can live with, verify, know took place.
g) Preservation of the ballots indefinitely in appropriately monitored facilities.
h) The use of symbols or pictures for those who can't read and the ready availability of those ballots and the use of voting machines to accommodate the handicapped (we certainly have enough now.

No levers, no punch cards, no scan tron ballots, no DRE's, no mediation at all...just paper, an ink stamp, and a private place.

4) General political awareness through public awareness of the need for change NOW including the use of any and all available means to raise that consciousness.

5) Specific political awareness of the fact that election fraud in 2000 was a race crime ("felons" list) and resulted in a stolen election and that it's been a race crime as long as we've had "spoilage" and continues to be one right through Ohio. It's not exclusively a race crime by any means but it is primarily. The use of any and all available means to make this point.

6) Support of all types of election monitoring including: private foundations or other forums to evaluate elections, equipment, etc.; political parties or non partisan groups to implement audits; exit polls designed to catch fraud (although the paper ballot process, with full participation would make this an after thought.

7) A recognition that voting rights is a potent issue that crosses the political spectrum.

Now, that's where I arrived after spending time on this forum and talking to a lot of people off the forum. There is a limited, tiny constituency for voting machines and an even smaller one, fractional, supporting election fraud.

The larger constituency is the one I'm interested in. The American people are restive and disgusted as reflected by the plummeting approval ratings for *. They want answers to many questions including, "How did this guy get in this job twice?" As we get to a 35% consensus approval rating, the public will decide on their own answers, one of which may well be probable election fraud. The public reasoning will have little or anything to do with debates on minutia. Once they've decided * stole it, the action will the same as Nixon's a crook and a cook...adios * and co.

That's the point where the solution is vital and necessary...it must be clear and simple, able to be grasped quickly, and easy to implement. It's an end run, Flutie's Hail Mary, MacArthur’s Inchon Invasion...it by passes the bull shit, entirely, and provides a solution that all can trust because with participation (and there is incentive for that) it will work.

[Febble]

106. Just wanted to say....

That this is a fantastic list.

Have you disseminated it elsewhere?

Also, that I think what you do with the daily news post is fantastica essential (I mostly try to recommend if I get there).

Seeing as we've had our differences over the EPs, I thought I'd get this on record.

For whoever's still reading.

[Ellipsis]

119. ...you rule!

If I may... we all plug along doing what we do... reading the same lines walking the same path and and just when things seem mundane... a glimmer of light... a bit of electricity gets injected into the forum.

In my book... you sir get ANOTHER gold star next to your name.

I have said it before... many thanks Autorank.

And, a side bar note: Clinton breaks tradition, Carter calls Gore the winner in 2000, Fitzgerald Indictments coming..."Me thinks" the show is just beginning.

PEACE.

[autorank]

120. btmindfrmr...Thanks for the star!

It's like it all happened in 24-48 hours. You've got the feel for this, I can tell, so you'll understand this. Carter supported a really odious provision in his report, a National Voter ID card (which you probably know about). I figured, oh boy, that's Jim Crow; he's gone on over. Then a couple of days later, Carter says this!!! They're using Carter to ask the ontological question: does Bush exist as a legitimate President. The answer is implied in the question and provided by Carter, so directly. Of course not, he was never elected in the first place; hence, arguments over his re election are both mute (he should not have run as President) and germain (he probably stole it again). Thanks for the thought provoking comment, in addition to the star.

:hi:

[Ellipsis]

133. Remember rob Georgia?

What a "gem" that was.


Jimmy's gotta reason to be pissed.

[autorank]

180. Jimmy should have gotten the word about Gwinnett Co's voter ID Card

Damn, Bulton Gwinnett signed the Declaration of Independence for Georgia. Now the county that bears his name pulls a stunt with the five day old Voter ID card and the GA civil rights community is up in arms. Two - three weeks later, Jimmy comes out with Carter - Baker for a National Voter ID card. I'm getting dizzy.

But, three days after the National Voter ID recommendation, Jimmy recovered quickly and said "Gore lost, Hell Yeah! Cuz Jimmy Carter says so!" (well Stone Cold said that but Jimmy said something like that).

Things are moving quicker than we know when Carter turns on a dime like this.

I want a Carter so pissed off he'll start kicking in doors. Like this Carter:

"Former U.S. President Jimmy Carter, whose Carter Center works with 65 nations on various democratization projects, said that he was "taken aback and embarrassed" by what happened in the Florida election. In a January 9, 2001 interview aired on National Public Radio's "Morning Edition," former President Carter said that, if the Carter Center had been asked to monitor the Florida election, it would have refused to participate. The Center will not take part in election monitoring when a country fails to follow certain procedures to guarantee the fairness and integrity of elections. "
http://www.disabilityworld.org/01-02_01/gov/carter.htm

[Bill Bored]

185. I just heard him say Gore WON in 2000 Auto.

When did he say otherwise?

[autorank]

189. He didn't say otherwise, he did just affirm that clearly. His timing

is both helpful and interesting. I want him to keep saying it. The truth is always helpful.

[Bill Bored]

194. But but but, you said:

"Gore lost, Hell Yeah! Cuz Jimmy Carter says so!"

[kiwi_expat]

190. The big news is that he is now saying Gore won FLORIDA in 2000.

And thus would have won the Electoral College, if the Supremes had not intervened.

Even Michael Moore has never ever really said that.

[Ellipsis]

193. Let alone an ex-president.

...and one of the more soft spoken ones.

[Bill Bored]

195. In F 911 he did say that Gore won, didn't he? nt

[kiwi_expat]

196. I think MM has only said that Gore won the popular vote.

I was so miffed by his overlooking the results of the newspaper consortium's overvote count that I sent him an e-mail. He didn't respond - or change his ways, as far as I know.

I wish the Dems hadn't made such a big deal of Gore's popular vote win. It set us up for Bush claiming a "mandate" based on the popular vote in 2004.

[Bill Bored]

206. I'm almost sure Moore mentioned the consortium's findings in F 911.

It was near the beginning after the scene in Congress when they couldn't find a Senator to challenge FL. Does anyone else remember this?

[OnTheOtherHand]

207. thanks, that was driving me nuts!

I had convinced myself I was confusing F911 with Unprecedented.

This is not the source I would have chosen, but I googled a blow-by-blow critique of the movie --
http://www.davekopel.com/Terror/Fiftysix-Deceits-in-Fahrenheit-911.htm
-- that describes some of this under "deceit 3."

[kiwi_expat]

209. Thanks for tracking that down, OTOH

"A little while later, Fahrenheit shows Jeffrey Toobin (a sometime talking head lawyer for CNN) claiming that if the Supreme Court had allowed a third recount to proceed past the legal deadline, 'under every scenario Gore won the election.'

"Fahrenheit shows only a snippet of Toobin's remarks on CNN. What Fahrenheit does not show is that Toobin admitted on CNN that the only scenarios for a Gore victory involved a type of recount which Gore had never requested in his lawsuits, and which would have been in violation of Florida law...." -Dave Kopel


Sorry, Bill. I hadn't remembered the Toobin comment. I'm pretty sure that when I saw the movie, I did not update my opinion of Moore's position because it was Toobin talking, not Moore.

I am not yet convinced the Moore now claims that Gore would have won Florida if the Supremes had not intervened.


* * *

While we're on the subject, the contention that Gore had not requested the type of recount that would have given him a victory is a red herring. It is true, Gore had initially requested only undervotes be counted and only in certain counties. However, the Florida Supremes had already expanded the count to include the undervores in all counties. The OVERvotes would have been counted too, according to Judge Terry Lewis, who was assigned by the Florida Supremes to oversee the count.

[OnTheOtherHand]

211. on Moore, I don't know if this is germane, but

http://www.michaelmoore.com/warroom/f911notes/

The flat statement there is that "Gore got the most votes in 2000," which in context has to be applied to Florida. But I don't know whether Moore directly addresses the counterfactual of whether Gore would have won if the Supremes had not intervened.

This citation from Moore's site might be germane: “The review found that the result would have been different if every canvassing board in every county had examined every undervote, a situation that no election or court authority had ordered. Gore had called for such a statewide manual recount if Bush would agree, but Bush rejected the idea and there was no mechanism in place to conduct one.” Martin Merzer, “Review of Ballots Finds Bush's Win Would Have Endured Manual Recount,” Miami Herald, April 4, 2001.

Umm, I'm going to stop thinking about this now. ;)

[Bill Bored]

212. No, there's one where Moore narrates it too.

And I believe the news clipping about the U. Chi. study is shown, which actually came out AFTER 9/11 itself on Nov. 11, 2001, contrary to some popular belief.

I know Moore said "under every scenario" and then went on to say that because Bush's brother was the Gov. and his dad had appointed Supreme Ct. justices, he was awarded the Presidency.

I may have this a bit wrong, but I'm pretty sure this is how it was explained in the film.

And BTW, only the U. Chi study showed Gore winning in every possible scenario.

The one thing I'm still not sure about is whether the overvotes were ever counted. If they were, and it's hard to be 100% sure of the voters' intent with overvotes, I think Gore would have won by a landslide, relatively speaking. So the butterfly ballot cost him the election along with the bogus FL felons list.

If anyone wants to correct me, please do, but frankly, a lot of folks still get a lot of this wrong, unfortunately.

[kiwi_expat]

214. If MM did say "under every scenario", he was definitely wrong.

>I know Moore said "under every scenario" and then went
>on to say that because Bush's brother was the Gov. and
>his dad had appointed Supreme Ct. justices, he was
>awarded the Presidency.


If MM did say "under every scenario", he was definitely wrong. The consortium stacked the scenarios. But the number of scenarios you win does not matter, in politics or in exit poll spreadsheets. :-)


>The one thing I'm still not sure about is whether the
>overvotes were ever counted. If they were, and it's
>hard to be 100% sure of the voters' intent with overvotes,
>I think Gore would have won by a landslide, relatively
>speaking.

The consortium counted the overvotes. The intent was very obvious: people ticked Gore AND wrote in his name ! Gore won the overvotes by tens of thousands.


>So the butterfly ballot cost him the election along
>with the bogus FL felons list.

If the overvotes had been included, OR if the butterfly ballot had not existed OR if the felons list had not existed, Gore would have won the count.


>...frankly, a lot of folks still get a lot of this wrong, >unfortunately.

Gore won the VOTES CAST. He would have been President, if the Supremes had not stepped in.

But the key to his election would have been Terry Lewis' handling of the uncounted votes. Not many people realize that.

[Bill Bored]

216. OK, I'll bite; Who's Terry Lewis?

Yes "write in candidate's name" and "write-in candidate's name" are two different things. I wonder if they had that printed correctly on the ballot, with the hyphen.

But I was thinking more of the overvotes in which Buchanan and Gore were voted for on the butterflies. Didn't this also happen?

I think Gore won under every scenario in the U. Chi. study. This is not the same as the consortium one that was completed months before. It was completed in Nov., 2001.

[kiwi_expat]

217. butterfly ballots; Judge Terry Lewis

"But I was thinking more of the overvotes in which Buchanan and Gore were voted for on the butterflies. Didn't this also happen?" -BB

I think, in the case of the butterfly ballots, there was only one vote for President per ballot. The Gore voters accidentally voted for Buchanan INSTEAD of Gore. But I guess there could have been some ballots where the voters voted for both. If so, the voter intention would not have been certain, and those ballots would not have been included in the tens of thousands of overvotes determined to be for Gore.

Please see below regarding the role of Judge Terry Lewis:
__________________________
consortiumnews.com

So Bush Did Steal the White House

By Robert Parry
November 22, 2001

George W. Bush now appears to have claimed the most powerful office in the
world by blocking a court-ordered recount of votes in Florida that likely
would have elected Al Gore to be president of the United States.

A document, revealed by Newsweek magazine, indicates that the Florida
recount that was stopped last year by five Republicans on the U.S. Supreme
Court would have taken into account so-called "overvotes" that heavily
favored Gore.

If those "overvotes" were counted, as now it appears they would have been,
Gore would have carried Florida regardless of what standard of chad -
dimpled, hanging, punched-through - was used in counting the so-called
"undervotes," according to an examination of those ballots by a group of
leading news organizations.

In other words, Bush lost not only the national popular vote by more than a
half million ballots, but he would have lost the key state of Florida and
thus the presidency, if Florida's authorities had been allowed to count the
votes that met the state's legal requirement of demonstrating the clear
intent of the voter.

The Newsweek disclosure - a memo that the presiding judge in the state
recount sent to a county canvassing board - shows that the judge was
instructing the county boards to collect "overvotes" that had been rejected
for indicating two choices for president when, in reality, the voters had
made clear their one choice.

"If you would segregate 'overvotes' as you describe and indicate in your
final report how many where you determined the clear intent of the voter,"
wrote Judge Terry Lewis, who had been named by the Florida Supreme Court to
oversee the statewide recount, "I will rule on the issue for all counties."

Lewis's memo to the chairman of the Charlotte County canvassing board was
written on Dec. 9, 2000, just hours before Bush succeeded in getting five
conservative justices on the U.S. Supreme Court to stop the Florida recount.

Lewis has said in more recent interviews that he might well have expanded
the recount to include those "overvotes." Indeed, it would be hard to
imagine that he wouldn't count those legitimate votes once they were
recovered by the counties and were submitted to Lewis.

The "overvotes" in which voters marked the name of their choice and also
wrote in his name would be even more clearly legal votes than the so-called
"undervotes" which were kicked out for failing to register a choice that
could be read by voting machines.

<snip>

[Bill Bored]

218. Link please?

I have always thought the overvotes were the key, and in the cases where the write-ins matched the Gore choice, voter intent was clear. But this could have also happened with Bush. Do you have a link to the Newsweek story? Looks like a good one.

[Febble]

222. Not to Newsweek, but try these:

The Butterfly Did It: The Aberrant Vote for Buchanan in Palm Beach County, Florida by Wand, Jonathan N., Kenneth W. Shotts, Jasjeet S. Sekhon, Walter R. Mebane, Jr., Michael C. Herron, Henry E. Brady. 2001.

and

The Wrong Man is President! Overvotes in the 2000 Presidential Election in Florida by Mebane, Walter R., Jr. 2003.

See Walter Mebane's web page and Jasjeet Sekhon's web page for more.

[petgoat]

242. Here's the Butterfly Ballot

It looks like you're supposed to vote for both President and Vice
President. Except if you're voting for Bush, or for the Natural Law
candidate, and then it's obvious that you select just one.

[kiwi_expat]

220. My mistake: Gore overvote win is small - unless third party votes examined

I should never have relied on my memory! -kiwi
________________________________________________

www.gainsvillesun.com

Saturday, November 24, 2001

Al Gore was the genuine choice of a majority of Floridians
By Mark Mayfield
Mark V. Mayfield is a Gainesville attorney.

'Bush wins media recount" was the headline, in various forms, in nearly
every newspaper after the National Opinion Research Center's (NORC) recount
of votes in Florida. The headlines imply that Bush actually received more
votes in Florida than Gore. Nothing could be farther from the truth.

Under any scenario where all of the votes are counted, Gore won. The only
scenarios where Bush won were those where significant numbers of votes were
simply not counted.

The Associated Press made the point clearly: "Under any standard that
tabulated all disputed votes statewide, however, Gore erased Bush's
advantage and emerged with a tiny lead that ranged from 42 to 171 votes."

Gore's margin of victory swells to over 46,000 if the overvotes are
carefully examined. Such an examination of the 113,000 overvotes shows that
more than 75,000 chose Gore and a minor candidate and just 29,000 chose
Bush. Common sense demands that we admit that most of these voters were not
supporters of either Patrick Buchanan or the Socialist Workers' Party.

Furthermore, many overvotes were entirely legal. They simply weren't counted
because a voter may have punched in Gore's name and written it down to be
certain the counter got the message. If all of Florida's counties had
error-checking machines in the precincts to prevent such overvotes, Gore's
margin of victory would have been beyond any doubt! The media consortium
paid practically no attention to these ballots. Why? To conceal the
strongest evidence by far that Florida's voters preferred Gore.

Beneath the deliberately misleading headlines is the inescapable fact that
Al Gore was the genuine choice of a majority of Floridians as well as the
victor by more than 540,000 votes nationally. As the Associated Press report
states, "In the review of all the state's disputed ballots, Gore edged ahead
under all six scenarios for counting all undervotes and overvotes
statewide." In other words, Gore got more votes than George Bush.

Gore won under a strict-counting scenario and he won under a loose-counting
scenario. No matter how you count it, if everyone who legally voted in
Florida had a chance to see his or her vote matter, Al Gore would be sitting
in the Oval Office today.

After Sept. 11, many seemed to feel it was their patriotic duty not to do
anything to call into question the authority of the commander-in-chief.
However, if we're supposed to support a war effort against an enemy that
hates us because of our free, democratic process, shouldn't we first ensure
that we really are living by that democratic process?

We've already witnessed the shredding of the Constitution in the name of the
war on terrorism (secret searches, indefinite detentions, military tribunals
with no judicial oversight, attorney-client privilege reduced), must we also
endure the dismantling of the very foundations of our democracy? It should
never be considered unpatriotic or treasonous to demand that our leaders
live up to the rule of law. The right to question the fairness and accuracy
of an election, even more so a Presidential election, should be sacrosanct.

Instead, we are witnessing the media at best giving short shrift to the
results of the recount and, at worst, deliberately misleading the public
about the results.

There were, at a minimum, serious irregularities in the election of 2000. At
worst, there may have been illegalities committed by a number of state and
national government officials. Yet all we hear are, "Get over it" or "It's
time to move on."

While we must surely try to return our lives to the best semblance of normal
in the wake of Sept. 11, we must not let those events go without recourse.
We must fully investigate and understand the events that lead to the fiasco
that became the Florida recount of 2000. If that points to the fact that the
wrong man is occupying the White House, we must face that and deal with it
appropriately. We do our principles and ourselves no good by hiding or
"spinning" the facts, ignoring or misstating the truth.

George W. Bush was not the choice of a majority of Floridians (or of all
Americans). Therefore, to "restore honor and integrity to the office" he
should either resign the presidency, allowing Al Gore to take over
immediately, or he should be removed. This would be unprecedented, no doubt.
It would likely cause great turmoil too. The country will survive, as it has
after other crises, because it is the right thing to do.

[Febble]

223. But check out the links

in my reply to Bill. Mebane says that undervotes did it.

[kiwi_expat]

226. I agree: the overvotes won Gore Florida.

I'm just saying that I had remembered the overvote win as being much bigger than it was.

[Bill Bored]

213. Who is the Kopel guy? He's got the Aug 6 PDB COMPLETELY WRONG!

Evidently neither he NOR Bush read Page 2 of that document which said,
"Nevertheless, FBI information since that time indicates patterns of suspicious activity consistent with preparations for hijackings or other types of attacks, including recent surveillance of federal buildings in New York."

Close enough to Moore's interpretation Kopel? See, it's that "Nevertheless" word that trips people up who don't know how to think or read critically.

And the part before that, said they couldn't confirm plans for hijackings airplanes to gain the release the blind shaykh -- not that they weren't planning to fly planes into buildings in NY. Get it?

Anyone who knows NYC knows that the WTC was a stone's throw from the federal buildings and the 1st WTC attack was mentioned on Page 1 of the PDB where it's stated that Al Caida wanted to "follow the example of World Trade Center bomber Ramzi Yousef and bring the fighting to America."

This Kople guy is a Bush apologist extraordinaire! I wouldn't feel comfortable citing any of his work, OTOH.

Here's the PDB for your reference. Make up your own mind about Moore's interpretation of it:

http://www.cnn.com/2004/images/04/10/whitehouse.pdf

It really pisses me off when people make excuses for Shrub not reacting to this warning which was about as clear as it could have been without handing it to him on a silver platter.

[OnTheOtherHand]

225. I don't trust Kopel as an analyst

but I thought it was likely that he hadn't actually invented entire scenes in the movie, especially when his account jogged my memory.

I couldn't rely on the michaelmoore.com response, because it didn't really make clear what the movie had said.

I think there are extensive synopses -- probably even complete scripts -- of the movie out in cyberspace, but I didn't manage to find one quickly.

The surveillance of federal buildings in NYC seems like a red herring, since obviously Atta et al. didn't have to know anything about the federal buildings in order to fly planes into the Twin Towers. That aside, I think Bush's pre-9/11 record is pretty clear and damning. (For that matter, I think his post-9/11 record is pretty clear and damning.)

[Bill Bored]

229. And don't forget his record ON 9/11! nt

[Chi]

235. This guy errors on his first point...

Why even bother reading on.

"In fact, Fox did not retract its claim that Gore had won Florida until 2 a.m.--four hours after other networks had withdrawn the call."

http://archives.cnn.com/2001/ALLPOLITICS/stories/02/02/cnn.report/cnn.pdf
"9:54 p.m.: The Decision Team recommends that CNN and CBS retract
their Florida calls for Gore. The two do so and are the first networks to retract
(AP retracted approximately four minutes earlier). VNS retracts at 10:16 p.m.,
and all networks have retracted by 10:18.

[petgoat]

244. I got as far as davekopel's Deceit 3.

He claims that Jeffrey Toobin admitted "that the only scenarios for a
Gore victory involved a type of recount which Gore had never
requested...."

If you go to the CNN transcript kopel cites you find Toobin admitted nothing.

[OnTheOtherHand]

245. Does anyone know a neutral synopsis of F9/11?

(As I said, I wasn't endorsing kopel's critique -- just looking for a prod to my memory about what was in which movie.)

A fair summary would be good to have as reference.

[Bill Bored]

167. OK Auto, let me change my answer then:

It's not necessarily that election reform per se is being impaired by the exit poll debate, but I do think the development of additional evidence that the election was stolen sometimes takes a back seat to that debate, which has not yielded anything new in a long time. That's what I mean by wheel-spinning.

Elections can be stolen in so many ways that would not be apparent in the exit polls, so why give this so much focus for so long? Why perpetuate a myth that with paperless DREs, exit polls are the only way to tell if the vote count is incorrect? They are not and in fact, they may not. Why not do the precinct canvassing, the FOIA requests, etc. instead of just repeating the same old argument?

But true believers in this exit poll hypothesis seem not to require any additional evidence. That's a shame because anyone outside their ranks, remains unconvinced by the exit polls after all this time, no matter how many times TIA has tried. I guess you'd call them naysayers.

So thanks for moving election reform forward! But let's not forget there is still evidence out there to be found, besides the exit polls, if we look for it. And we can't look for it if we're spending so much time defending the exit polls.