Steve Freeman Summary of Debate

[Melissa G]

I had asked Steve Freeman for a summary of the Mitofsky debate. He gave me permission to reprint an email he sent me here..

Summary: All in all, the event went well. Thanks to the excellent suggestions and help by people who may not wish to be identified, I was able to prepare excellent text and slides. http://www.appliedresearch.us/sf/epdiscrep.htm> There is the record of the talk and it gave the issue some new visibility. The few statisticians I have heard from seemed appropriately alarmed by what I had to say, and were dismayed by Mitofsky’s failure to respond to the points I raised.

Details of the Presentations: The talk was not as well attended as we hoped. I believe that Mitofsky brought a table’s worth of people including Lenski. I probably had about 20 people there, and another half a dozen with whom I’d communicated by email but never met.

My research center has a publicist who tried to get TV and print coverage. Neither CSPAN nor other TV stations came, but one of the most prominent writers on polls and polling, Richard Morin of the Washington Post did come and I did speak with him briefly. I only know of one news story published. I’ll send it, and my response, in another message.


I do not have Mitofsky’s talk, but as Morin put it, it was largely his “exit poll stump speech.” He did not address any of the 17 improbabilities and neglected correlations that comprised the central body of my presentation, but rather calmly and confidently described how they make their Election Day estimates. Josh Mitteldorf pointed out that he seemed to be making our point for us that he conducts the polls very professionally, and there is good reason for confidence in the original estimates and the data.


Only towards the very end he addressed my talk, and then simply that he didn’t know where I came up with my numbers. He concluded with a slide that he claimed “kills the fraud argument,” although I don’t think anyone in the room understood why it would do so. He didn’t even identify the axes.


Rebuttal: I made a few clarifications after his presentation:


(1) All the exit poll data I used in my study came from Edison/Mitofsky, most of it from the 77 page Inauguration eve report. For most of my analyses, I used Precinct Level Disparity (PLD), which is the difference between whom people said they voted for as they walked out of the voting booth, and the way those votes were officially recorded.


(2) Be careful with pollster-speak. Not only WPE/PLD; but “correcting” the exit poll data means adjusting it to conform to the count; and “Democratic Overstatement” may well be, and certainly to at least some degree would more accurately be called ““Democratic Undercount.”



(3) Warren suggests that “selection interval” helps explain the discrepancy, but “selection interval” may well be a spurious correlation, in that it likely co-varies with actual predictors, notably “paper” and probably “rural”



(4) Likewise for “co-located precincts” which allows for the Ballot rotation mischief that has been documented in Cleveland. And also “interviewer distance from polling place.” Ken Blackwell likely had good reason why he didn’t want observers close to the polling place.



(5) It’s hard to know what to make of any chart of residuals. For example, are outliers excluded (as they are for most of the Edison/Mitofsky analyses)? Even the Albert Einsteins of the world have to show methods and data.



(6) The data that has been made available is not the data that is useful for important analyses. Everyone knew what data was important:



(a) Precinct level summary data so we could identify precincts with high PLD, some of which have impossibly high PLD: Why were they so high? Some we would want to physically investigate.



(b) County identifiers to statistically examine effects of voting technology and partisan control.



But instead, Edison/Mitofsky released individual level data and now claim they cannot identify precincts because it conceivably could compromise respondent confidentiality by linking it to this individual level data.



Of course, even the most private, personal census data – which this isn’t – is studied, as is the most sensitive national security data (which this, in fact, may be) using procedures such as clearances, background checks, and on-site inspection. Since shortly after the election I offered to work on-site at Mitofsky International along with whatever team of statisticians and social scientists they chose, and to raise funds to reimburse Mitofsky International for whatever costs were incurred from our conduct of this research. He said he could not permit even this.



(7) When they speak about “improvements” for next year, note that they will not permit any release of “uncorrected” exit poll data. Be careful about such an “improvement.” Is that something we consider desirable? To never again receive exit poll information that can shed perspective on official election results?



Some Lessons:



(1) Try to be crystal clear on the data, in particular, the importance of PLD: Where it comes from. Why it is such an important measure. What it means.



(2) Do not let them dismiss anything. At the point in which I was explaining my use of PLD, Lenski said from the back of the room that you can’t do that, “that’s not how we used it.” I said I know that’s not how you used it, but that I was using the data to clarify the discrepancy, to which he threw up his arms and snorted dismissively. I don’t think I let him get away with it, but next time, I’d be even tougher, and put him on the spot.



(3) Several people came away confused about “early results.” Add to the list of misleading pollster-speak that when Lenski or Mitofsky speaks of “early exit poll data,” or the “half-time score” that although the implication is of early-in-the-day results, they actually mean end-of-day data that is not yet adjusted to conform to the count.



(4) Continue to hammer on the proprietary data. In fact, we really should press more on this confidentially argument. Not all pollsters support Mitofsky’s position, and we should clarify those arguments.

[rdmccur]

1. Good to see Freeman

decry the absence of supporting data from Mitofsky
(or supporting mathematics). I think Mitofsky might get laughed out of science conventions with these kind of presentations.

[Melissa G]

4. Kinda funny how rigorous Mitofsky is about some things and

how faith based he expects us to be about some others...like his assertions of no fraud for instance...

[OnTheOtherHand]

14. source, please?

If we're going to talk about being rigorous, have you any evidence that Mitofsky made "assertions of no fraud"?

I missed the first part of his presentation, but he has sent me his slides, and slide #3 reads: "This is not a discussion about whether there was fraud in the 2004 election. There may or may not have been fraud. This is a discussion about the exit polls and whether they shed any light on the claims of fraud."

(I was there by the time Mitofsky presented his last graph, and he certainly did identify the axes -- but we were all pretty tired by then, even those of us who hadn't gotten stuck driving in the rain for over four hours /whimper/, so it's not shocking that Freeman missed it.)

[Time for change]

30. Perhaps Melissa is referring to Mitofsky's January report

The report says that their analysis of the exit poll data does not support the allegations of fraud due to rigging of voting equipment. The only evidence put forward to support that statement is that they found no statistical difference in WPE in precincts that used optical scan versus electronic voting.

IMO that is awfully thin evidence on which to make such an assertion.

[OnTheOtherHand]

46. I don't know what to say

To say that a particular data set "does not support" a certain hypothesis isn't, to my mind, a very strong claim requiring a high standard of evidence. I can't help but feel that some distinction between "does not support" and, say, "utterly and unequivocally trashes" tends to get lost on this board. ;) But yes, I know (more or less) what you think E/M should have said about this issue.

[Time for change]

49. Here's what I think

Scientifically speaking, I absolutely agree with you. Well, I agree with you at least partially anyhow.

But there are two issues here. When you say that your analysis "does not support" something, there is at least the implication there that an attempt was made to seriously assess the situation. It's like if my wife asks me to find her keys because she's got something real important to do. I completely forget about her request, and she comes back an hour later to ask me the status of my search for the keys, and I tell her that I didn't find them. Absolutely true statement, but misleading nevertheless.

In this case, there is also a very important national interest at stake. When Mitofsky says that his analysis does not support fraud, millions of Americans take that to imply that he seriously looked for it, and furthermore that the fact that his analysis didn't support it means that it's highly unlikely that it's there.

So, what I'm saying is that what he said is very misleading. Whether he meant it to be misleading I cannot say. I do believe that our MSM wants to mislead us on this issue. Since he works for them, perhaps he was just doing what his bosses wanted him to do. But whether he intended his statement to be misleading or not, I believe that he should have known better.

[OnTheOtherHand]

57. OK, I take your point, and here's what I think

I don't think that E/M conducting that test and reporting the result is much like your forgetting to look for your wife's keys. E/M has said all along that their exit polls aren't intended to be used to audit elections, which is true. If their statements about exit polls are interpreted as statements about the election, I frankly think that their critics are more to blame than they are.

I don't mind disagreeing about this, but I'm sharing my view because I don't think many social scientists can live up to the standard of never saying anything about the 2004 election that couldn't be construed as trivializing the need for reform. I don't think you would find many survey analysts who would agree with your assessment that the E/M report is "very misleading" on the issue of voting technology -- because the analysts, like E/M, take for granted that the report is about the exit polls, not about the election. That's as succinctly as I can frame the difference in perspective.

Maybe this is a quibble, but I don't think you could find millions of Americans who have any clue who Mitofsky is. But indeed some reporters may have wondered, "Does this story has legs?", and asked survey experts about it, and got those experts' candid and considered opinions: Not. The emphasis on exit polls is absolutely self-defeating. The reporters should be asking different experts different questions. (Of course the media coverage would probably still be pretty poor -- but I still think that the e-voting coverage is distinctly more sympathetic, in part because the experts are.)

[Time for change]

62. I suppose it's a matter of perspective

You say that survey analysts take for granted that the report is about the exit polls and not about the election. You may very well be right about that.

But the great majority of people who read the report were not survey analysts. It was released to the American public, is that not right? If so, then I think that it should have been written with that in mind. As it was, I believe that the great majority of people who read the report took it to mean that election fraud had been considered and dismissed (not-withstanding the fact that the report didn't actually say that specifically).

Also, with regard to the comment about E/M saying that exit polls aren't intended to be used to audit elections: But the report didn't say that. If the report would have said something to the effect "Since our exit polls were not meant to audit elections, this report will not comment on the integrity of the election", that would have clarified the matter and not been misleading IMO.

[OnTheOtherHand]

66. actually, let's consider the title page of the document

Evaluation of Edison/Mitofsky Election System 2004
prepared by Edison Media Research and Mitofsky International
for the National Election Pool (NEP)

It was released to the American public, yes, but it wasn't written for the American public. And the title perfectly frames my point that the report is about the exit polls and not about the election.

Arguably that was a missed opportunity. It's nice to imagine how things might be better if Mitofsky were a bit more like, say, Febble. (Well, I guess not everyone around here would agree about that!)

(I rather imagine that the great majority of the people who read the report were already convinced about fraud one way or the other, so they weren't unduly -- or perhaps even duly -- influenced either way. But that may be cynical of me.)

So, we're not really disagreeing, although from my standpoint it is rather as if you are complaining that the Frisbee wasn't a toaster.

[Time for change]

67. Ok, that's a good point about the title -- at least partially

Remember that my original post in this argument simply noted that when Melissa made the comment about Mitofsky expecting us to take it on faith that biased exit polls rather than fraud was mainly responsible for the red shift, perhaps what she had in mind was this January report. I think that that's a good guess.

Your point that the report was released to the American public but was not written for them is a subtle one, and honestly something that I had never considered. Maybe that's because in my line of work anything that we've ever released to the public is for the public. AND, if the public doesn't understand what we write, it's our fault, not theirs, because it is our job to write things in a way that they will understand. So, I guess that that's the way I was looking at it.

I do know that when I read the report, I read it with intense interest, mainly with the assumption that it would tell us something about the likelihood of fraud. So to the extent that I saw the title, that didn't change my view in the least that what I was reading was an assessment of election fraud vs. exit poll bias. I'm sure that the great majority of DUers (and possibly the majority of other Americans as well) who read the report did so with a similar intent and viewpoint as mine.

Like I said, the idea that this was not written for the American public never crossed my mind, and I'm still having trouble getting a grasp on that. I guess that the idea that something would be released to the public but not written for them is something that I just can't comprehend right now, perhaps because of the type of work that I'm in. So, I guess that's where our main difference lies. If Mitofsky in fact wrote this for a select body of exit poll researchers or whatever, and if it was understood by the people for whom it was written that this was not meant to be an assessment of election fraud vs. exit poll bias, then you are right that it wasn't misleading. But I just can't grasp that concept.

[OnTheOtherHand]

74. well, I don't think that the report was written in "code"

(I should say at the outset that this post isn't arguing against anything you stated; it is mostly trying to lay out some general points about why two parties in the debate are largely talking past each other and/or just annoying the heck out of each other.)

The report wasn't exactly geared to hard-core exit poll researchers, although by ordinary standards it isn't a page-turner, either. (You read the report thoughtfully and carefully, but it cannot have escaped your notice that most DUers haven't -- and I'm not even saying that they should have.) But I think this is basically the right way to think of the document, as counterintuitive as it may seem. E/M and the NEP wanted to know what went wrong with the exit polls; in my professional opinion, they did not assign a very high prior probability to fraud as an explanation, they did conduct some tests that did not alter their priors, and indeed subsequent investigation including Freeman's is unlikely to alter their priors. And so, very little if any of the report is directed toward educating the public about the fraud debate, or anything else; most of it is directed toward figuring out what, in the (possibly mistaken) judgment of the authors, actually went wrong.

One consequence is that the report doesn't really explain what all the subscribers know very well: that the NEP operation isn't designed to "call" competitive states based on exit poll interviews alone. The possibility of bias in the interviews is built into the system. The notion that the interviews should be assumed bias-free unless proven otherwise -- well, survey researchers just don't think that way. We do not kneel at the altar of MoE.

I think that if either Mitofsky or the NEP subscribers took the Exit Poll Error Indicates Vote Count Error argument very seriously, the report would have spent more time assessing it. I also think that if, say, the WPE for touch-screen precincts had come back six points higher than for precincts using the other technologies -- or if the swing state relationship had come back, say, twice as strong as the "not able to approach every voter" relationship, rather than slightly weaker -- E/M might have gotten more interested.

But by and large, survey researchers including Mitofsky are inclined to assume that the error is in the polls, not (primarily) in the vote count, because a lot of their professional training and experience is dedicated to studying all the ways that polls can be in error. When challenged to prove their nutty idea that the polls were wrong (by an average of maybe 3 percent of vote share), they tend to react with perplexity or impatience. And if they report the result of a particular hypothesis test that might have supported fraud, but didn't -- if they report it the same way they would report any other result that comes back negative -- they often get clobbered for claiming certainty that the hypothesis and wider theory are wrong. This attribution is likely to strike them as either scientifically illiterate or willfully unfair.

I've spent a lot of time on this issue in part because it seemed possible to me that precisely because of that training and experience, they (we, i.e., folks trained in survey research) might miss something that outsiders might pick up on. But what is just as annoying as all get-out is when some outsiders (not you) who have no relevant training or experience pronounce that Obviously the exit polls Prove massive fraud and that the survey researchers must be Covering Up. It reflects a fundamental unwillingness to try to see the world as someone else sees it. (It is also damn weird to treat the exit polls as gospel, and the people who conduct the exit polls as demons.) And if we (people concerned about election fraud and working for election reform) aren't willing to see the world as Mitofsky and his peers see it, I think we have neither reason nor right to expect that they will pay attention to us, either.

So, that's what I think about all that. ;)

(Edit to try to clarify the "We"s in the last long graf. I seem to be one of relatively few people who identifies both sides in this debate as "we.")

[Melissa G]

32. Hi OTOH, okay it was late...I was not complete in my post

Warren Mitofsky says several versions of the below in various places others are pdfs and I am not good at copying them...

But mostly the short version as I read it is that the exit polls have nothing to tell us about election fraud move along folks nothing to see here....Do you read what he says differently?
-----Original Message-----

From: Warren Mitofsky

Sent: Tuesday, May 17, 2005 00:41

Subject: Re: For those of us who didn't go to Miami



snipThe exit polls do nothing to establish or disprove the claim of fraud in Ohio. The so-called unexplained questions I thought were answered in Miami.

snip warren mitofsky
http://www.appliedresearch.us/sf/aapor_response_to_mitofsky.htm

[OnTheOtherHand]

47. umm, I would skip the part that is sort of like innuendo--

"move along folks nothing to see here" -- well, he might say, "folks, get a life!"

If Mitofsky believed that his data pointed to election fraud, I think he would act a lot differently. As far as I can tell, he believes precisely that the exit polls have nothing to tell us about election fraud -- although conceivably, under the force of argument, he might alter that to "very little." (Of course, I'm not exactly a Mitofsky confidante -- I've met him once, saw him in Philly but didn't talk with him, and we now have swapped a few dozen e-mails.)

Whether he is right is a whole 'nother issue. I think Mitofsky and Freeman both believe what they say.

[Melissa G]

50. There was also this ESI press release that Mitofsky coordinated with

which I found pretty much saying exactly what you called innuendo..
Plus Bill Bored did a quote where I don't have the exact link but it's in a current running thread so if you want it I'll get it..

Citation: Mitofsky said that because there's no correlation between Red Shift and Bush's improved vote count in the polled precincts in 2004, that this "kills the fraud argument."


I have other stuff if you really want to talk about it pm me..

http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=203&topic_id=369824

[Febble]

53. I do think there is a potential for misunderstanding here

Freeman's argument is that the exit poll discrepancy is evidence of fraud, and that fraud was the cause of the exit poll discrepancy.

Mitofsky's argument is that the exit poll discrepancy is not evidence of fraud, and that fraud was not the cause of the exit poll discrepancy.

His comments about the ESI finding, and his own, similar finding, presented in Philadelphia, need to be read in the that context. I agree with Mitofsky: both analyses kill the argument that fraud was the cause of the exit poll discrepancy.

Whether they kill the argument that fraud occurred is a quite different issue. I believe Mitofsky is defending his interpretation of his poll, not the election. After all, it's his poll he is interested in. There may be forms of fraud that do not show up in the poll. I believe there are.

[OnTheOtherHand]

58. "kills the fraud argument"

I really need to write my FAQ.... I totally agree with Febble on this. In context, "the fraud argument" clearly meant "the argument that the exit polls demonstrate fraud." It was an exit poll debate. In my personal, political, and professional opinion, the election reform debate should not be an exit poll debate, but it was not my choice.

Folks, expecting survey folks to speak eloquently about the need for election reform is sort of like expecting George Steinbrenner to galvanize the cause of racial reconciliation. He isn't opposed to it, as far as I know -- it just isn't his thing. It is just plain nutty to prod survey folks to talk about surveys, and then complain that their statements don't take election reform seriously enough. Wrong people, wrong arguments, wrong outcome. Over, and over, and over again. :shrug:

[Land Shark]

43. OK, so Mitofsky admits there "May or may not have been fraud"

[OnTheOtherHand]

48. sure -- he is Mr. Exit Poll, not Mr. Election

From a professional standpoint, it's not a matter of "admitting" something about the election, it's a matter of recognizing one's own area of expertise. Mitofsky is qualified to have an opinion on what the exit poll data do and don't indicate, and his opinion is widely respected (and shared) in the profession. If Mitofsky expresses an opinion on fraud in general, well, it may or may not be better than Britney Spears's political analysis, but it's not the sort of thing that professional training predisposes us to do. At the risk of sounding like Rumsfeld, it's important to know what one doesn't know.

[Land Shark]

60. But then, why does Mitofsky either use himself or allow himself to

be used in support of a "there's no fraud" argument? If he knows what he doesn't know, why does he have such a cocksure attitude about what he doesn't know ?

[Melissa G]

61. example of cocksure Mitofsky post in advance of OTOH asking..

-----Original Message-----

From: Warren Mitofsky

Sent: Tuesday, May 17, 2005 00:41

Subject: Re: For those of us who didn't go to Miami



Marc Sapir once again is talking without knowing what was said in Miami. The exit polls do nothing to establish or disprove the claim of fraud in Ohio. The so-called unexplained questions I thought were answered in Miami.



The other fraud arguments were never discussed. Only five of the original 20+ something who made up the group that raised the questions about the 20+ exit polls signed their current release. Their own members don't believe their arguments any longer -- just Marc.



warren mitofsky
http://www.appliedresearch.us/sf/aapor_response_to_mitofsky.htm
I've seen other remarks that have the same tone....this was just handy...

[OnTheOtherHand]

65. but... but...

Just try reversing the order of two words, and see if the meaning registers then: The exit polls do nothing to disprove -- or establish -- the claim of fraud in Ohio.

That isn't cocksure, that's the truth as he saw it, and sees it. (Not that I can read his mind to know for sure, but that's the dominant position in the field as far as I can tell, so there is no reason for me to think otherwise.)

Mitofsky overreached in speculating about the views of people who didn't sign the May USCV paper. If we're gonna start parsing tone, I can think of some other folks nearer at hand who might profitably be addressed, but whatever.

[Febble]

63. Land Shark

I myself have been accused many times of "allowing myself" to be "used" in support of a no fraud argument.

Worse - I have been accused of being complicit in the death of US soldiers in Iraq, and likened to a Nazi collaborator.

However, the fact is that knowledge, analyses, scientific findings can all be used to support arguments that the author did not make.

Actually that is as it should be.

But I myself am not aware that Mitofsky has ever claimed that any of his investigations demonstrate that there was no fraud. What he claims is that fraud was not the reason for the discrepancy in the exit polls. Those two statements are very different.

I am now pretty convinced, following the recent findings, that fraud is highly unlikely to have had any more than a minimal impact on the exit polls. That does NOT mean there was no fraud, although it may set fairly strict constraints on the kind of fraud there could have been. But if, for example, every voting machine in every strongly Democratic precinct was set to record extra overvotes for a randomly selected proportion of voters, this would have any effect on the exit polls at all. They could be as wrong or as right as they liked.

But it could still result in a massive loss of votes for Kerry, at the cost of a relatively small loss of votes for Bush.

I'm not saying this happened. Merely that the linkage between exit polls and fraud is not necessary to postulate fraud. On the other hand the evidence for massive fraud is thin without the exit poll argument. But the fact the machines were so insecure certainly makes it worth investigating.

[Melissa G]

2. Link to article mentioned in Summary

Was 2004 presidential election stolen?

Bucks County Courier Times

Sen. John Kerry won Bucks County by two points and took Pennsylvania by three, but lost the national race to George W. Bush.

Like me, you probably thought this was old news and that Election '04 was settled.

Not Stephen Freeman, a visiting professor at the University of Pennsylvania, whose soon-to-be-published book will make a case that the Bush/Kerry election was riddled with "corrupted counts" that deserve high scrutiny, perhaps a recount.

Freeman stops short of saying Bush stole the race, but he comes close. His case is this. On the afternoon of Election Day, exit polling - which Freeman said is quite accurate - showed Kerry was winning with 51 percent of the vote, to Bush's 49 percent.

http://www.phillyburbs.com/pb-dyn/news/219-10202005-557661.html

[Melissa G]

3. Freeman lettter to editor in response to article

reprinted with permission of author so here in full..

I would like to thank the Bucks County Courier Times for covering my presentation to the American Statistical Association (Re: Was 2004 presidential election stolen? by J.D. Mullane http://www.phillyburbs.com/pb-dyn/news/219-10202005-557661.html ).



But what kind of journalism is it for a reporter who acknowledges that he “is lousy at math” to render a pronouncement on a presentation to statisticians without even taking the trouble to try to understand what “correlation” means? Wouldn’t your readers be better served by learning what the statisticians and others in the audience thought?



Just to set the facts in order, the thrust of my presentation was:



We had an election in which (1) extensive malfeasance and count corruption has been documented, and far more has been alleged, but has gone uninvestigated; (2) in Ohio, a state in which Kerry had a decisive edge in the exit poll numbers, a manual count which could have verified the official count was obstructed by the Secretary of State, Ken Blackwell, who also served as chair of the Ohio Bush/Cheney campaign; and (3) 30% of Americans cast their ballots on electronic voting machines providing absolutely no assurance or verifiability that votes were counted as cast.



Given such a context, an exit poll discrepancy of 6 percentage points between the official count and exit poll projections warrants investigation. Especially given a nationwide 6.5 percentage point Precinct-Level Deviation. This is the difference between how 114,559 voters in selected precincts across the nation reported casting their ballots in confidential questionnaires as they exited the polls and the official counts in those very same precincts where the exit polls were conducted. Moreover, these are the final numbers, not the “half-time score” as Mitofsky and Mullane imply.



My talk to the American Statistical Association showed that the official explanation that Bush voters disproportionately – and by a large margin – refused to fill out the questionnaires offered by pollsters is unsubstantiated by the facts. I documented 17 statistical improbabilities and mathematical impossibilities that result from this hypothesis. Readers can read the text and view the slides in the presentation at a web site I established for this research: http://www.appliedresearch.us/sf/epdiscrep.htm



By the way, Mullane should brush up on metaphor as well as math. “Sour grapes” refers to a rationalization made by a hungry fox as he walked away from succulent grapes that eluded his reach.” I am saying exactly the opposite, that nothing could be more important than a clean election. As John Roberts acknowledged in his confirmation hearings, the right to vote and have that voted counted is “preservative of all other rights.” Because of that, we can’t just move on and walk away from damning evidence of a stolen election.



Steve Freeman

Center for Organizational Dynamics

University of Pennsylvania

[Peace Patriot]

12. GREAT LETTER! Go, Stephen! Where can we get the book?

I looked for it at Amazon.com. Not there yet. Is there a pre-order site? Can't wait! Profound thanks to Stephen, and to you, Melissa, for all your amazing work. You have been like a beacon of light to me, since that awful postelection week. The clear, brilliant explanations that Stephen has given, in the papers I've read, and your passionate commitment to the re-enfranchisement of the American people, have been tremendously inspiring. You are among the great patriots of American Revolution II. I hope you both know that. And, Stephen, thanks for standing up for us to those DOCTORERS of the exit polls, those FUDGERS of statistics, those LIARS, those DESTROYERS of honest elections and DAMNABLE JERKS. I emailed C-SPAN to broadcast it. I guess they aren't interested in our right to vote. Was it taped? Any chance it could be posted?

[Melissa G]

20. Hi PP!

The book is scheduled to be out first of the year. I'll check about the preorder option on the book. The debate was taped. When I find out if it can be released for hosting i will post a link.

Glad you like the letter. I will repost that as a separate thread as you suggested.
Thanks for all you do!!!
Best,
Melissa

[althecat]

13. This Would Be A Disaster..... there is an election in 13 Months...

(7) When they speak about “improvements” for next year, note that they will not permit any release of “uncorrected” exit poll data. Be careful about such an “improvement.” Is that something we consider desirable? To never again receive exit poll information that can shed perspective on official election results?

[bleever]

25. That's a great letter, easy to understand, and very persuasive.

Thank you, Dr. Freeman!

[blm]

31. Excellent. I hope more journos pick up on his work.

Will be sending to my media list.

[Carolab]

5. Good.

"The few statisticians I have heard from seemed appropriately alarmed by what I had to say, and were dismayed by Mitofsky’s failure to respond to the points I raised."

Does this mean they'll help to look into it further?

[Melissa G]

7. Hi Carolab! Hope so! Hope a whole bunch of folks start looking...

It took Fitzgerald two years but look at all he is uncovering now!
Let's hope the same is true about the sElection only sooner!!!:hi:

[Carolab]

8. Hi, Melissa G!

And Ojai Person reminded me: Nominated!

[Helga Scow Stern]

6. Nominated. Thanks for posting this!

I am glad to see that Steve Freeman is still devoted to uncovering the truth.

[Melissa G]

9. Hi Ojai Person! Writing a book is a big commitment!

:hi:

[texpatriot2004]

10. Thank you Mr. Freeman. nm

[Nothing Without Hope]

11. Recommended with thanks. Here's a thread announcing the debate:

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=104x5030755
thread title: ASK C-SPAN TO AIR DEBATE - Evidence of 2004 ELECTION FRAUD !!

Anybody got a link to transcripts and/or video files?

[freedomfries]

15. KICK & REC

thank you so much Melissa and Steve for all you are doing for the sake of democracy!

[CalmMan]

18. Transcript

http://www.appliedresearch.us/sf/epdiscrep.htm

[Melissa G]

22. Here is a link to the Text and slides and DU discussion from Freeman

http://www.appliedresearch.us/sf/epdiscrep.htm

DU discussion
http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=397502&mesg_id=397502



Have no Mitofsky info yet. Heard rumors it might be being hosted on Mystery Pollster soon....

[Chi]

16. Thanx Melissa!

'Since shortly after the election I offered to work on-site at Mitofsky International along with whatever team of statisticians and social scientists they chose, and to raise funds to reimburse Mitofsky International for whatever costs were incurred from our conduct of this research. He said he could not permit even this. '

Funny how he has no problem with sharing his data with people that already agree with him.

[OnTheOtherHand]

17. I sorta wish that last part were true

My life would be a lot easier if Mitofsky shared "his" data with me. Not that I keep tabs on what we agree or disagree about, but we certainly agree that the exit polls come nowhere near proving fraud. Of course, that is a mainstream position among survey analysts, and as far as I know they generally don't have access to the data with precinct identifiers either.

[Land Shark]

44. Of course then the exit polls don't buttress a Fair Election conclusion

either. Mitofsky undermines the election, in reality. He jujst won't make a tunnel all the way to the No Man's Land of Stolen Elections. Besides all the arguments about data, etc., some people just lack courage.

[OnTheOtherHand]

45. huh?

Well, I suppose Mitofsky lacks the "courage" to come here and try to explain, over and over and over again, why the exit polls don't buttress much of anything.

I was talking with an old friend who has a Ph.D. in social psychology, and she literally laughed in disbelief when I told her that some people thought non-response bias was a contrived explanation for exit poll error. Once she believed me, she said, "But if you don't know about non-response bias, how can you say anything about exit polls?"

Yeah, well.

[Joe Chi Minh]

24. "Funny how he has no problem with sharing

his data with people that already agree with him".

They are so transparent, i.e. the fact of their secrecy and its obvious intent, it's almost scary. No wonder Christ called them "children of darkness".

Neocon "officials", overt or covert, seem to share an almost infantile, rather than juvenile, level of moral and intellectual development. It's as if they were infant-school or primary school children at play. "If you won't let me win, I'll take away the bat and ball I brought, so you won't be able to play!" That kind of level.

I dare say it's common in unambiguously totalitarian states, such as the former Soviet Union, but to find it in a historical democracy, the latter, in some significant ways highly developed, is kind of mind-blowing. Not really so far removed from Alice's Wonderland. "Words have whatever meaning I want them to have!" That kind of thing. I think I'll call it "hostile infantilism"! Has a certain ring to it, doesn't it? "Politics are the executive expression of human immaturity", Vera Brittain once observed - only they took it to a whole new depth.

A few little epigrams from a compendium, which seem particularly relevant to our respective kleptocracies:

"Politics is the art of preventing people from taking part in affairs which properly concern them".

"Treason doth never prosper: what's the reason? For if it prosper none dare call it treason". Sir John Harington.

"The poor and ignorant will continue to lie and steal, as long as the rich and educated show them how". Elbert Hubbard.

"There is no worse heresy than that the office sanctifies the holder of it". J E E Dalberg.

That which is called firmness in a king is called stubbornness in a donkey". Thomas Erskine.

"Nothing is impossible for the man who doesn't have to do it himself".

There are tons more, of course.

[Febble]

26. And your meaning is?

[Chi]

39. Rephrase

Funny how the only people who get to see the data, are those who agree with him.

But I doubt that will sit with you any better.

[Febble]

41. Well, who?

Who has seen the data?

ESI saw "blurred" data.

And I've seen the data, obviously, if you look at my sig.

But you can't infer causality from a correlation. It could be that only people who agree with Mitofsky get to see the data. It could be that people who get to see the data find themselves agreeing with Mitofsky.

And I can only speak for me. First of all, I didn't exactly "get to see the data" in the way you mean. I was hired to run some analyses on it, using my proposed measure of precinct level discrepancy. At the time, I didn't know whether I agreed with Mitofsky or not - I didn't think there was any basis for belief or disbelief. The whole point of my geeky paper was that I thought the conclusions in the E-M report were unsound because they had used what I considered (and still consider) a very poor measure of precinct level discrepancy, the "WPE". In my paper, I called for a reanalysis, using multiple regression, with my own, or similar measure as a dependent variable. Mitofsky saw my paper, agreed that my measure had greater merit, and hired me to run some analyses.

One result was presented in Philadelphia.

The authors of the ESI paper concluded that their analysis did not support the hypothesis of major fraud. I don't know what they thought beforehand, although I did talk with one author who at least thought the study was worth doing - in other words that Ohio was worth investigating.

I don't know who else has seen the data, what they thought beforehand, or what they thought afterwards.

[Chi]

51. There ya go

Well, he chose you after you wrote a paper not only agreeing with his conclusions, but also rebuffing his biggest critics.

Then he chose ESI, which wouldn't be anywhere on my list of best picks for that honor.
If you would like to lobby on ESI's behalf, as to why they deserved that honor over others...such as Freeman, have at it.
I have seen enough not to trust them.

You say one of the authors said that Ohio was worth investigating.
Out of curiosity, did that point of view make it into their report?

[Febble]

52. Correction:

You say:

"Well, he chose you after you wrote a paper not only agreeing with his conclusions, but also rebuffing his biggest critics."

This is not true.

You can start by reading my paper.

http://www.geocities.com/lizzielid/WPEpaper.pdf

Here's the abstract:

During the 2004 US presidential election, exit polls initially appeared to overestimate the proportion of votes cast for John Kerry, the Democratic challenger, relative to the proportion cast for the Republican president, George W. Bush. Because the margin was so close, the discrepancy between the early exit poll indicators and the result made the difference between indications that was Kerry ahead, and a result that gave Bush a victory, and helped fuel widespread suspicion that the error had arisen not from errors in the exit poll, but from errors in the counting of the votes.

In January, 2005 the polling companies, Edison Media Research and Mitofsky International, issued a report in which they concluded that the errors arose from differential response rates in the two groups of voters, exacerbated by problems with voter sampling methodology. In March 2005, a group of academics working for the US Counts Votes (USCV) National Election Data Archive Project, issued a paper on the internet detailing analyses of data provided in the Edison- Mitofsky report which they claim indicate that the patterns of error reported are more consistent with a hypothesis of vote-corruption than one of differential non-response.

One of their claims is rooted in the apparent differentials between “Within-Precinct Errors” (WPEs) found in highly partisan precincts and those found in more evenly divided precincts. However, the WPE as a measure is itself confounded by precinct partisanship. This paper addresses problems presented by the use of the WPE as a dependent measure, particularly when precinct partisanship is a predictor variable. It describes a computation whereby an unconfounded index of sampling bias can be retrieved from the data, models the nature of the confound introduced into analyses by the WPE, discusses implications for the inferences drawn by USCV from the Edison-Mitofsky data, and makes a recommendation for further investigations into the factors that led to discrepancies between exit poll predictions and election results.

And from later in the paper:

In the case of the 2004 exit polls, it would therefore be of interest to know what proportion of total variance in genuine within-precinct “bias” could be accounted for by the factors postulated in the E-M report, and whether, after thus accounting for known methodological factors, any precincts/states proved to be statistical outliers that might indicate the possible contribution of vote-count corruption to the exit-poll error.

In other words, I did NOT agree with the conclusions of the E-M report. I argued that there was no way of knowing whether the conclusions were correct or not, nor whether USCV's conclusions were correct or not, unless the data was re-analysed using a better measure of bias.

As it turned out USCV interpreted my findings as an attack on the overall case for fraud, which I always maintained it was not. Many of my papers investigating fraud are still on the USCV website. My only disagreement with USCV was with the logic of the USCV position - they were making a strong case that their analysis indicated that fraud (vote count corruption in Bush strongholds) was the most likely explanation for a particulat pattern. I disagreed, on mathematical grounds. It's a horribly geeky paper. But by exactly the same argument I maintained that we could not accept the E-M report's conclusion, because their conclusions were based on a faulty measure. In other words, I wasn't saying that E-M were right, and that USCV were wrong, but that we couldn't know which until E-M reanalysed their data with a better measure.

Regarding the ESI report. So far we only have a short article and some slides and figures (badly labelled). The study is currently undergoing peer-review, according the the ESI website. So I will be interested to read it when it is published. Peer-review is a healthy process.

[Chi]

55. As I understood it...

Mitofsky explained that the WPE was a result of a negative response bias from Bush voters.

USCV responded by saying that his own published data (in his report) demonstrated that there was no negative response bias from Bush voters.

You responded by saying that the Mitofsky's data did demonstrate that negative bias.

You agreed with Mitofsky, and refuted USCV.
Did I mis-interpret this?

(Nice barb insinuating I didn't read your paper....that's the second time you did that, but who's counting.)
Gotta run out, but I'll be back later for more flogging.

[Febble]

56. Well, maybe you did

read it, in which case, apologies. It does appear you misinterpreted it. Many people did.


Mitofsky explained that the WPE was a result of a negative response bias from Bush voters.

Yes.

USCV responded by saying that his own published data (in his report) demonstrated that there was no negative response bias from Bush voters.

My reading of the USCV report was that they presented two arguments against Mitofsky's explanation.

One was that if Bush voters were generally less inclined to respond, completion rates should be lower in high Bush precincts. And in fact, the means were very slightly higher (the E-M say, not significant). This is a good argument, although there is a decent counter-argument. But that was not the subject of my paper.

USCV also argued that the evidence showed that Bush Strongholds had More Vote-count Corruption (BSmVCC) - on the basis of the mean WPE in each of five categories of precincts "partisanship" - defined by the proportions of votes counted for Bush. Mean WPE was near zero at the "high Kerry" end, and highest at the "high Bush" end. From this, USCV inferred that something suspicious was going on.

You responded by saying that the Mitofsky's data did demonstrate that negative bias.

My paper showed that the apparent slope of the mean WPE from the low at the Kerry end to the high at the Bush end could (not was, could) be an artefact of the WPE which is a crap measure of the discrepancy. i said it needed to be recalculated using my proposed measure. Mitofsky did this (having read my paper online) showed that there was no linear correlation between my measure of the discrepancy and Bush's share of the vote. In other words, the slope that USCV were basing their BSmVCC hypothesis on was not supported by the evidence once an appropriate measure was used.

Ron Baiman and Kathy Dopp (and I think now David Dodge) still maintain that there IS some kind of correlation. I don't know what all the other authors of the original papers now think, apart from Bruce O'Dell, who seems to agree with me.

I never understood why it was such a big deal. I didn't know why it was so important for the fraud story that the discrepancy should be concentrated in high Bush precincts (although clearly fraud would result in more precincts ending up in that category). But it seemed to me that more random fraud was just as likely, actually, more likely. It seemed to me to be a geeky storm in a teacup.

But the debate got pretty nasty anyway. That flat correlation line was never a slam dunk for rBr any more than a slope would have been a slam dunk for fraud. Although the more I thought about it, the harder it seemed to be to explain the flat correlation with a pervasive fraud theory.

So, in short: what I thought my paper had done was cast doubt on a very specific fraud hypothesis. It certainly did not rule out all fraud hypotheses, though it did infirm a couple. The main target of my paper was E-M, I called on them to re-do all their analyses, as I didn't think their dependent measure was valid.

And they did.

[Chi]

59. Here ya go...

'your work, supported his work'.
(I removed all those distractions such as 'agreed with' and 'his conclusions')

If it didn't, I would bet dollars to donuts you would have never gotten an invite.

[Febble]

64. Sorry, I don't understand

are you quoting me?

If so, can you give context? Or is context a distraction?

[Chi]

68. That's OK...I can live with that.

I'm sure most got my meaning from my first post.

(Oh, and another nice barb, BTW 'Or is context a distraction?')

[Febble]

69. Chi

the barbs, as I see it, are coming from you.

I don't know, honestly, what you are saying I am saying, so I can't really clarify.

You seem to be saying that I am saying that I agreed with Mitofsky when I wrote my paper. That is not the case. The paper indicated that we could have no basis for agreement or disagreement until the data was reanalysed.

If you think I wasn't saying that, then I'd like you to cite, with context, the place you think I wasn't saying it. If I inadvertantly said what you think I said, I'd like to correct it.

Thanks.

[Chi]

70. Febble

Mitofsky found your conclusions useful, so you got to see the data.
If he didn't find your conclusions useful, he would have never done so (like he refused Freeman).
That's been my opinion from my first post on.

My opinion wasn't about you, it was about Mitofsky, I have changed the wording several times (for you), but you seem to rather make this about you.
I'm not biting.

"the barbs, as I see it, are coming from you."
Please cite these barbs.

(be back after work)

[Febble]

71. OK, peace, then

I certainly don't want it to be about me.

But I'd still point out that my only "conclusion" was that there could be no "conclusions" until the data was reanalysed using my measure (or something comparable). That in fact was what I was contracted to do (presumably because Mitofsky figured I was the one who knew how it worked).

Neither Mitofsky, nor I, knew what we'd find until I'd looked.

And forget the barbs. Sorry. I felt attacked. I sometimes am on this forum, and occasionally see offence where none was intended. I did not mean to attack you. I simply want to make it clear that what my paper was saying - and what it was not saying.

[Melissa G]

72. Febble, for the record I have never seen Chi attack a person

only what they write.. He is as much a gentleman as I have seen on this board. He is funny and incisive and he reads all the reports he comments on and what's more he seems to be able to follow the math..
He is one of my favorite DUers for these reasons..
Best,
Melissa

[Febble]

73. Sure

I shouldn't have said attack. But he did imply that only people who agreed with Mitofsky were allowed to see the data. That is what I was arguing with him about. But I shouldn't have taken it personally.

Thanks. Apologies to both of you.

[OnTheOtherHand]

54. I have to say

First of all, Febble's paper does speak pretty clearly for itself, it is by no means a naysaying screed, and if researchers aren't allowed even to raise methodological questions without being scored as "agreeing with Mitofsky," then -- well, as St. Teresa of Avila reputedly said to God, "If this is how you treat your friends, no wonder you have so few." Obviously, Chi, I don't mean "you" personally, but there really is a prevailing pattern here. The political scientists, survey researchers and such who _ignore_ this debate overwhelmingly remain happily unaware of its existence (many of them are happily cranking out and publishing papers about why Bush won); the shit storm is reserved for those of us who actually try to engage it.

I'm not gonna carry water for ESI -- I don't know most of the principals -- but Fritz Scheuren, who worked on the Ohio exit poll report, is president of the American Statistical Association, and he wrote the book (well, OK, the article) on data blurring. He would obviously be on the short list of people to do that work.

Mary Batcher was also in on that report, and while I'm not qualified to comment on her CV in the field of financial auditing, I can say that both in Minneapolis (Joint Statistical Meetings) and in Philly, she did a better job articulating an agenda for auditable, transparent elections than anyone else in the room. I had no idea who Mary Batcher was until I heard her at Minneapolis in August, but I can tell you, we need more of her. Financial Auditors for Transparent Elections -- has a ring to it, doesn't it?

[helderheid]

19. kick and recommended!

[PATRICK]

21. This raises an obvious question

in our competitive free enterprise system is there no other exit poll to rival Mitovsky and be unshakably out in the open and bi-partisan? I take it there were polls done by the parties, notably before the election, but in plain self-defense one would think, given the possible bias of pollsters that a competent party would be fore-armed with their own measures to keep the system honest.

All the while- all the while- there were adjusting the polls to the counts on the TV screen election, all the while- all the while they were spewing "values" memes and assorted preposterous covers for the adjustments- the Dems were busy in mourning and self-berating searches for their own mea culpas. Of course the bigger picture was even worse, the inability to roll back fight or protest the obvious theft maneuvers, even the battles fought before in Florida where the GOP alone made "improvements" in strategy. As some campaign workers confessed, they had no strategy for digital fraud, and continue to this day in an atmosphere os quiet surrender or unbelievable ignorance. Competence seemed to center on the naive notion that working to get something out of the HAVA scam was actually productive.

[IndyOp]

23. I think we have to raise funds to do our own exit polls -

We have to have raw exit poll data to evaluate the vote tabulation.

Any chance that Freeman's talk will bring a few more statisticians to speak out in public?

:nuke:

:grr:

:kick:

[Febble]

27. Quite a few statisticians

have spoken out in public.

Trouble is, you don't believe them.

[IndyOp]

28. Trouble is, I don't believe them - Huh?

I am aware of three statisticians who have published on this issue: Stephen Freeman, Dennis Loo, and Michael Hout. Is that 'quite a few'? Who am I missing?

By 'speaking out in public' I mean have published under their own names - publishing with an online username does not count in the 'real world'.

On the issue of belief. I don't believe people because they have a degree or a title, I believe them when they provide credible evidence and are able to withstand scrutiny of their peers - which includes me. I believe them only after I have had adequate opportunity to read and try to poke holes in their arguments and make sure that my concerns have been addressed. I will admit to not having taken sufficient time to do this.

The biggest question I still have is about Mitofsky's methodology -- and I am disappointed to learn that Mitosfky did not address Freeman's points at the 'debate'. I was hoping the debate would resolve some of those issues.

:(

[Febble]

34. Fritz Scheuren

(president of the American Statistical Association) and Walter Mebane, who published studies demonstrating that the butterfly ballot was single-handedly responsible for Gore's loss of Florida, as were undervotes in largely African American precincts. Jasjeet Sekhon who worked on the same issues, and who has done ground-breaking work in this area of analysis.

All three are statisticians of huge standing. Mebane and Sekhon are specifically experts in voting, and in the statistical analysis of voting patterns, and are not exactly Bush shills - they argue that he should not have been president. One of Mebane's papers was actually entitled "The Wrong Man is President". He came to similar conclusions to my own about disenfranchisement of African American voters in Franklin County in 2004.

Stephen Freeman is not a statistician, as he says in his talk. Dunno about Dennis Loo. Michael Hout certainly has standing, but his conclusions about Florida were widely criticized. I do think that study made a fairly fundamental error. We are all prone to error. That is the virtue of peer review.

Oh, and as for my own name:

It is Elizabeth Liddle. You can google me, and if you also google Febble, you will find I am the same person. I do not claim anonymity.

[IndyOp]

35. Febble - EL - I am delighted to meeet you!

I read your initial study on Florida and was concerned about the Dems that vote Republican (my father is one). I was very, very angry that the corporate media would not report on your revisions of that analysis that took the Republocrats into account.

Your paper comes closest to providing a 'how to' use statistical analysis to reveal electronic vote fraud - that is what I think we need most (because that is where I think the greatest number of votes can be manipulated).

I read at least one of Jasjeet Sekhon's papers on the butterfly ballot. I think one title was "The Butterfly Did It". I was devastated at his dismissive comments after the 2004 Election -- he dissed everyone who was trying to determine if there was fraud. His paper on bad ballot design is not the most important issue at hand - even though the 'caterpillar' ballot appeared in the 2004 election and accounts for some lost stolen votes.

I've not read Scheuren or Mebane -- I should look at Mebane's paper on disenfranchisement of AfAm voters in Franklin County - that sounds very interesting and important.

Are Mebane and Sekhon associated with US Count Votes?

Thanks for all of your work. Thanks for remaining standing after having been attacked by the press last November.

:yourock:

[IndyOp]

29. Febble -

I've just read your sig line. You run analyses for USCV and work for Mitosfky as a consultant.

It seems to me that your voice would be particularly powerful -- when/where are you speaking out?

[Febble]

33. Well I have and do

What are you hoping I will say?

I've researched machine fraud in Florida, undervotes in New Mexico, machine rationing in Ohio, and exit poll data. I think the electoral system in America is broken, has been broken for years, and needs fixing. In some very real sense I believe the election was stolen, as it elections are regularly stolen, by systematic disenfranchisement of the poor, particularly the poor who are also ethnic minorities. I can be passionate about that, especially as the leader you elect styles himself "leader of the free world", and has undue influence over the lives of those of us who live in the rest of it.

But I'm a scientist, not a politician. I'm not going to draw the conclusions I want from the data - I'm going to draw the conclusions supported by the data.

Of the work I've done I think that the evidence is inconclusive in Florida, but I wouldn't rule out some fraud. I think Kerry may have lost New Mexico due to machine related undervotes in largely Hispanic and African American precincts. I think Kerry lost large numbers of votes in Franklin county because of unjust machine allocation. I sent my work on the last issue to many people including the media, lawyers for Kerry Edwards, Cliff Arnebeck, and John Conyers.

But my current, considered opinion is that the exit poll discrepancy was not primarily due to vote-switching fraud. Not even largely, although I wouldn't rule out vote-spoilage as a substantial contributor. It was a perfectly plausible hypothesis, and worth serious investigation. But the more I have discovered, the more convinced I have become that it was not due to fraud. Mitofsky's presentation in Philadelphia contained an analysis that in my view virtually rules out fraud as any more than, at most, a minor contributor to the exit poll discrepancy. Which means I no longer believe there is much of a probability that Kerry won the popular vote.

From my short exchanges with him, I think Steve Freeman is a good man, and I respect what he is saying. But I think that his analysis is flawed in parts, actually wrong in part, and though apparently plausible in other parts, inconsistent with the findings from the analysis that Mitofsky presented on Friday.

But your democracy still needs fixing.

[IndyOp]

36. The exit poll discrepancy was not primarily due to vote-switching fraud...

Mitofsky's presentation in Philadelphia contained an analysis that in my view virtually rules out fraud as any more than, at most, a minor contributor to the exit poll discrepancy.

I am very, very interested in hearing what you think about this... Could you elaborate?

Has Mitofsky published a reply to Freeman?

The very existence of the National Election Pool -- which 'mixes' data from the exit polls and data from the incoming vote tabulation -- makes me exceptionally suspicious. Not convinced, just very suspicious.

Freeman's book is coming out soon -- is he going to be roasted on a spit? I feel protective towards him and everyone else who is working to discover fraud.

As for 'my democracy needs fixing' -- I am not sure how to reply. Briefest reply: Of course, in very, very, very many ways - campaign funding, media, campaign conduct, vote casting and counting, abuse of power, incestuous relationships within government and between government and corporations.

I took the immigration test for Canada last November - I got a 70 (I think) more than enough points needed to get Canadian citizenship. It would be a cop out, so I will stay here.

I am active politically in many ways - including giving presentations on election fraud to a variety of local groups.

Best wishes!

[Febble]

37. Well, thanks, IndyOp

I will elaborate, but not right now. It is a bit incomprehensible without the Mitofsky's plot, which I hope will be linkable to soon. But in essence - the increase in Bush's vote share relative to 2000 was completely uncorrelated with the Precinct Level Discprepancy. And as the PLD was near zero in 2000, it is pretty hard to see how fraud could account for the discrepancy.

I think Mitofsky and Freeman are largely talking past each other. I don't think either have really responded to the other. If Freeman doesn't change quite a lot of stuff, some people will roast him. But lots of people will love it. And some people will roast Mitofsky whatever happens.

I agree, I feel a bit protective of my fellow fraud investigators as well - I think the circular firing squad has gone on long enough. I think the exit poll stuff is a dog, but there is plenty of other evidence - enough for prosecutions I would have thought, and definitely a case for radical reform.

Funnily enough, I never had any problem with the "mixing" of data, because it's exactly what happens in the UK and is completely explicit. But it's all clearer for two reasons - we are all on the same time zone, and our elections are completely transparent - paper ballots hand counted under public scrutiny. For two hours after the polls close we only have the exit poll projections, then as the results come in the projections are gradually weighted in light of the new data.

But then we don't expect to use exit polls to audit our count. If one of our constituency results is contested we do an immediate full recount.

More soon, I hope!

PS I live in the UK now, but I recommend Canada.

[rdmccur]

38. Febble

The more I read your posts the more I respect your knowledge about this.
What would be great if maybe you and Freeman could have a discussion on du (or another web location). I lean very much toward Freeman's arguments based on the arguments I've seen thus far and a strong intuitive element as well (scientists can be intuitive). And I have to continue to maintain that, without supporting data and, usually, a supportive theoretical framework (couched in mathematics normally), Mitofsky's arguments can't be seriously considered as valid.

[Febble]

40. Thanks

I'm quite overcome today with the nice things people have said today, even if they don't believe my conclusions!

I have wondered whether it might be a good idea to set up an internet debate at a particular time and place, where some of these things could be properly discussed. Certainly it could be useful to establish some kind of consensus about what is compatible with what. For example, eomer has suggested some interesting fraud hypotheses that could theoretically be compatible with the new analysis, and perhaps could be tested - but while a number of hypotheses might be compatible with that plot, they might not be compatible with each other.

So it would be good to set some parameters as to what is still plausible, what is compatible with what, and what there is abundant evidence for. The "Brain Storm" thread was supposed to be a start, but of course it didn't have Steve on board.

Herding cats is not easy!

[IndyOp]

42. Online debate - Interesting!

Whoever sets this up would need to publish a list of papers to be read in advance so that most of the people who log on would have adequate background knowledge.

Neat idea!