Pollster Pleads Guilty To Faking Results For Bush, Lieberman Campaigns...

[galloglas]

This seems to be big!!

As one poster states, it is easier to steal elections if the polls are altered.

It was the difference between the pre-election polls and the exit polls which first showed election thefts. So, what happens if the polls are rigged. IMHO, there are some fairly flaky numbers floating on the Senate races.



http://www.huffingtonpost.com/2006/09/07/pollster-pleads-guilty-to_n_28910.html

[wryter2000]

1. K&R

Wow. If they're altering the polls, too, it's harder to spot the vote tabulation fraud.

[Totally Committed]

For Bush and LIEBERMAN, you say?

Hear that, folks... Bush and LIEBERMAN together again, at last. And this time, they are using the same pollster to help give them the APPEARANCE of having won, so fewer questions would be asked once the elections were stolen.

LIEBERMAN. Remember that if you are tempted to vote for him in the GE.

TC

[MADem]

4. Well, we KNOW Karl sent MILLIONS of GOP dollars to Joementum

That is NO SECRET anymore.....

[Greeby]

2. The question is

Who else is doing it, and which candidates are they doing it for? Maybe trying to supress a Dem takeover by polling the races close? hmm

[galloglas]

6. That is possible, but so is

the idea that "tweaking" the polls could lead Democratic contributors to funnel their contributions to the wrong place at the wrong time.

In Missouri, Senator Talent (GOP) has outspent the challenger, Claire McCaskill by 3 to 1, so far. At the same time, Talent has built a huge warchest.

Why?
I figured it is because he really will need it. He is so freaking hapless, Alfred E. Neuman should beat him. Yet, he has Bush coming in for a fundraiser tomorrow. On the other hand, in that race, I have seen polls showing less than 1% difference, polls showing McCaskill down by as much as 9%, and up by 4%.

With Dems really needing at least six solid Senate seats, it would be pure Rovian to get the Democratic funds spent on races that the GOP has already written off.

If the Democrats could find a way to make sure that tons of money was spent on Katherine Harris or Rick Santorum, that would be lovely. Maybe polls showing those two races really tight could make that happen?

Knowing the GOP is doing this helps to determine their strategy. As Socrates said, "All knowledge is good", and IMHO, it can be damned helpful. Its what one doesn't know that usually bites one in the ass.

[SmokingJacket]

3. It's easier to steal elections if polls are considered invalid, too.

Just an observation.

Right now, polls show that the Dems have a good chance of taking the House. But if the pollsters are all crooks...

[galloglas]

9. But in this case

we are not led to believe all pollsters are crooks. We are in a position to say that these particular GOP pollsters are crooks.

My question would be whether the purpose is to:

1) buck up the GOP ranks, perhaps garnering more donations
2) befuddle the Democrats and get them to misallocate funds, or
3) discredit polling, in general, in order to make theft easier

[longship]

5. I knew it!!! The polls are "bollocks"

There's some considerable disconnect between the public opinion polls and the anecdotal evidence. I've been thinking for some number of weeks that there is something wrong here. Virtually all the anecdotal evidence puts the Democrats clearly in the lead for November. The shear number of Republican moderates bolting from the party, or taking stands clearly opposed to that of ChimpCo is telling. E.G., In KS, there are nine formerly Republican moderates running as Democrats.

So when I see these tightening races at a time when everything tells me that the pendulum is swinging the other way, I call "bollocks" on them.

This is very, very serious. Congress needs to launch an investigation, and pronto.

[galloglas]

10. That investigation will share space with Articles to Impeach

Neither will happen until there is a Democratic majority in one House or the other.

[diva77]

7. At least with parallel elections you have a signed affidavit from the

participant. Maybe we need more of these instead of straight polling :shrug:

[diva77]

8. and 1 more thing -- this is a good reason for candidates to run on what

they truly believe in rather than tailoring their messages to what the polls tell them :eyes:

[galloglas]

24. This came up yesterday with a state legislator

edited for clumsy spelling and grammar.:eyes:


Herr Bush visited our fine City of Kansas yesterday.

Many protesters, and when Herr Bush crawled, then flew, from our city, those many, many protesters withdrew to the comforts of comraderie and liquid comforts. It was, of course, prime campaigning territory for Democratic candidates on the prowl for votes.

Speaking to one whom has followed, with growing belief, our recitations of "You can't believe what they (the GOP) are stealing (your votes)!", the conversation took and interesting turn.

I was asked, "What should we do? We can't seem to win anything!"

So, I told 'em what I thought.

"Understand they haved you gamed. Join other 501s, NGOs, etc., and sue our state to rid it of machines. Stand four-square for HCPBs, and as Democrats, say that it is more moral to spend public funds to pay citizens to count ballots than to enrich corporations.

"After that, start heading left into Progressive territory, and stop trying to suck up to the corporations for donations. In the end, we peasants will fund you by the Grassroots, 'cause the corporations will give you just enough to run a close second.

"Lastly, like George Washington, never tell a lie. 'Cause if you do, we'll think you are the GOP, and we'll break your balls and start a third party."

Then I ordered up another round. :toast:

[applegrove]

11. Shows how important GOP thinks poll numbers are to manipulating the

vote and getting a "win". Poll numbers change people's behaviour.

[galloglas]

12. Another thread on this at General Discussion

Here is another, pile on, or combine?

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=364x2072576#2072959

[elehhhhna]

13. Can't pull it off this time unless poll fraud precedes vote fraud.

[galloglas]

14. Right! Not if we all put this election under a spyglass!

At every precinct and polling station, if we expect failure, fraud, deceit, and "unexpected results", they will not pull it off, IMHO.

"Glitches" can no longer be an acceptable answer to voting problems.
Dell Computers withdraws 1.5 million notebook batteries, everyone has Norton, McAfee, etc. Spamming, "phishing", and other hacking cost the economy billions per year. Yet the EDs and SoS's expect us to believe the machines are infallible. And when they aren't, it is always a "glitch"?!

Such computer incompetence by commercial vendors in the corporate field would result in beaucoup damages being paid under "failure to deliver" clauses in their contracts. Yet, in the election arena, the EDs and SoS's shrug their shoulders and say "glitch" and the vendors say "proprietary interests"? Who is to be held accountable if we need a head on a pike in November?

They can play with the polls, but our being aware of that, there is no history, no nothing, that should give us confidence in the ability of machines to deliver an honest and fair election. We should expect fraud and, when it happens, we should be in the streets!

HCPB, under the continuous scrutiny of the public and press, throughout the polling and counting process, is something we can trust. That way we can use our eyes and ears to validate reality, something the machines can never do.

The only thing which should keep us out of the streets in November would be a totally unexpected event: our finding that the "fix" had been taken out this time, that the Democrats sweep both Houses, and the country does an about face, with most incumbents swept from office.

The "polling of my own senses" tells me that there are no more clever Rove tricks, like "reluctant Bush respondents" or "unprecedented Christian Right voter turnout" to blame Democratic defeats upon. My senses tell me many GOP voters are reluctant to even tell friends they would vote for the GOP, thus validating Bush.

If there is no landslide, it is time to revolt. Peacefully, but revolt nonetheless.

The arrogant ineptitude of this Administration is so complete that a voter revolt in November is inevitable. If the Diebold Cheat-o-Matics say otherwise, it is time to change the manner of our voting to be this: "Controlled by the People, Counted by the People, and Enforced by the People"!

I think the first two are inherent in the Constitution. The "enforcement", if needed, is something we will just have to work out. We have a couple of months to plan for that, but we should be ready for it. For thieves will steal when they can. And, we know now that they are thieves.

[elehhhhna]

15. DEMS MUST ALSO NOT CONCEDE ANYTHING until we're all

satisfied with recounts, etc. If it takes 6 months, so be it.

[PATRICK]

16. Questions

Which, of course, the MSM may not in their wildest unfettered dreams imagine asking.

This is fraud by one person directly asking others to falsify and skew data. Arrogance encouraged by the fact no one has been holding polls to account directly for their methodology and possible biases. In other words, she took the next logical step or did it blatantly never dreaming of getting caught, so bad is the appearance of enforcement. Like "Duke" Cunningham and many others, impunity is a catching disease and the current state of affairs so embarrassing that crooks almost have to beg to be caught.

One question is: how reputable and unbiased ARE pollster owners and managers? How about workers and interviewers and their individual biases(which in many instances are at least subconsciously skewed toward eliciting desired responses).

Who has been accused, investigated in the past? Who has been held to account for persistent outlier results in crucial elections that have all gone, let's say, GOP?

What oversight and regulation of polls is there?

Any poll that goes beyond simply asking how a person actually voted from a large and fair representative sample is dabbling in opinion and prediction. Shouldn't there be standards across the board for this science of , sampling, query and interpretation? How often has the methodology shifted to reflect results that in fact are false? How often has the interpretation shifted to fit explanations for false results and pushed theories(by the winning side's propagandists)?

Politicians want REAL poll numbers for themselves and too often, fudged or kindly blurry polls for the electorate. Are there STILL secret polls to which the public is not privy- besides the ones they always
leak that show they are doing fantastically better than is the case?

Just how well policed ARE elections and all their facets to prevent fraudulent campaigning, corruption and vote suppression and stealing? Just maybe the PUBLIC should know how bad things are not the memes, "the greatest democracy", far better than others, not too bad, not "significant".

From campaign financing, influence by lobbyists, vote protection, the vast majority of the public without benefit of much encouragement from press or politicians wants real and pervasive reform. All we hear in the MSM are attempts to divert and argue against substantive reforms. Bad apples make for brief, stand alone news stories, dropped soon after without any natural follow up. The big story ends before the ink is dry on the bannerline.

Is she the only one? Are others more subtle, more with their finger on the e-tabulator and result editing, more subconsciously biased? Who is even bothering to find out?

Who is following through on the interpretative critiques of polling and its dual effect of interpretation and opinion creating which moves it immediately from observation to manipulation?

[Febble]

17. "What oversight and regulation of polls is there?"

From the AAPOR (American Association for Public Opinion Research) CODE OF PROFESSIONAL ETHICS AND PRACTICES:

1. Principles of Professional Practice in the Conduct of Our Work
   
   
   1. We shall exercise due care in developing research designs and survey instruments, and in collecting, processing, and analyzing data, taking all reasonable steps to assure the reliability and validity of results.
      
      
      1. We shall recommend and employ only those tools and methods of analysis that, in our professional judgment, are well suited to the research problem at hand.
         
      2. We shall not knowingly select research tools and methods of analysis that yield misleading conclusions.
         
      3. We shall not knowingly make interpretations of research results that are inconsistent with the data available, nor shall we tacitly permit such interpretations.
         
      4. We shall not knowingly imply that interpretations should be accorded greater confidence than the data actually warrant.
      
      
   2. We shall describe our methods and findings accurately and in appropriate detail in all research reports, adhering to the standards for minimal disclosure specified in Section III.
      
      
   3. If any of our work becomes the subject of a formal investigation of an alleged violation of this Code, undertaken with the approval of the AAPOR Executive Council, we shall provide additional information on the survey in such detail that a fellow survey practitioner would be able to conduct a professional evaluation of the survey.

Faking data is a serious professional offence in any science, and in this case, a criminal offence.

[PATRICK]

18. No offense to codes of ethics

Most professions have very admirable ones. Self-policing is another matter even in a competitive atmosphere. My question was wholly in the area outside the profession, namely government which in this case, unfortunately has impartiality problems even larger and they are important clients and beneficiaries of polls no matter how they are conducted. In fact, trusting in the professional ethics of statisticians and pollsters is far easier than in that of officials who might be regulating them, as evidenced by the attempts of the GOP after their "victory" in 2000 to influence and punish pollsters.

Then, beyond the ethics, there is the need for impartial examination from outside the actual industry as to the science and its usages. Again, who is to do the examining if, once more, the press are both important clients and beneficiaries of polling?

I know this is sort of chicken and egg and based on an old movie no less, but polling might have derived from market surveys originally and later applied to sophisticated political measurements? What code of ethics do advertising firms ghave when their actual interests and applications tempt them toward the pro-active and the subjective abuses?

[OnTheOtherHand]

19. fair questions

What we seem to know about the case under discussion is that a polling firm claimed to have conducted interviews that it hadn't conducted (or faked demographics on interviews it did conduct) in order to deliver "results" under time pressure. There is no reason to assume that the various campaigns were paying to receive fake information. "Internal" political polling indeed is a form of marketing research -- and the campaigns really want accurate, informative answers.* At the same time, they want good news, and that is why people are always suspicious of internal poll results. But the charge in this case is that the polling firm defrauded its clients by making up the data that it had contracted to gather for them.

*...which isn't to say that they will always want to ask accurate, informative questions!

Here's a link to the indictment: http://www.usdoj.gov/usao/ct/Documents/DataUSA%20Indictment.pdf

I'm not sure what to say next. I guess I can say that the people I've encountered who work on political polls for media give every indication of friendly competition with each other to get accurate answers. I think it would be hard to game the entire field. But I don't put too much faith in any one poll.

[Land Shark]

20. The interesting question, to me, is how did they PROVE this?

The AAPOR ethics principles, in addition to the seemingly self-disclosing item in C above, also state as follows:

"We understand that the use of our survey results in a legal proceeding does not relieve us of our ethical obligation to keep confidential all respondent identifiable information or lessen the importance of respondent anonymity."

I wonder where they "understand" this from? I know of no legal privilege applicable to survey data, so this appears to be a directive to keep respondent data secret unless and until ordered to do otherwise by a court. Sort of like insurance that the survey researcher doesn't cave the minute a subpoena is received, which is roughly like some other privileges in law.

The interesting thing to me about the confidentiality of survey data is that, just like the secret ballot in elections, it creates a radically unaccountable system in the sense that one can not go back to "source documents" to prove the accuracy of the survey or election (the respondents or the voters, respectively). For example, in elections it is illegal to be able to connect a particular ballot with a particular voter, because that capability would, of course, invade ballot secrecy.

Look around the law as well as ethics codes, and you find that the law and ethics are in many cases working against the auditability and verifiability of these systems.

Polls typically influence public or policymaker opinion.

Elections influence or control those who hold elective power.

From prior discussions, this "value" of respondent anonymity in polls is so important that it makes the accuracy of a presidential election look insignificant by comparison. I say this based on past discussions in which I attempted to say that a weighing process would be used in court to see which interest was the more important (anonymity of respondents or presidential accuracy) and I got the strident and emphatic objections from Febble asserting the unethical nature of such disclosures.

The ethics are presumed to exist by me, for these purposes. The question is which is more important, the presidential election or these ethics? In both the election and the ethics, the disconnect between the people and the election data or survey data creates huge problems of unverifiability. It makes the systems unaccountable/unauditable on the most basic level.

In the event Febble responds, I do hope it is not simply to assert the ethics or the purpose of the ethics because, once again, those are assumed. I'm wondering which is more important, the ethics or the presidential election's accuracy (or an exit poll's accuracy on said election)? There might be some interesting discussion (or it might be absent) when these ethical principles were adopted as to the tension and collision between subpart C above in Febble's post and the quote above concerning respondent confidentiality.

I'm also wondering, presuming the question above is answered, how the government proved their case of faking data. They probably had a whistleblower or else a stupid mistake such as in the form of self-incriminating documents.

[Febble]

21. Quite apart from the ethical issues

if you could not assure confidentiality, your potential for non-response bias would soar, and the accuracy of your data would plummet.

Research relies on the integrity of data. There are very few checks and balances - the check is that the penalty for being found out for falsifying data is professional death.

Peer-review is great - but reviewers rarely, if ever, ask to see raw data. Even if they did, there would be no guarantee that it was collected in the manner that was claimed. What reviewers do is check that the reported methodology was valid, and that the interpretation of the findings is warranted. They assume that both the raw data and the methodology are as reported. In this case, the problem does not appeared to have been with the validity of the methodology - but that the data and the methodology were not as claimed. The data was fake and the methodology was fraudulent.

I too would be interested to know how evidence for the fraud was collected. But it is, simply, fraud. The thing about fraud is that it can be difficult to detect. A pollster faking data is equivalent to a banker embezzling funds. Fortunately both are illegal. Unfortunately, both can be successfully hidden.

So there are two quite different issues here. One concerns the prevention and detection of fraudulent practice (which is exactly the same issue as professional misconduct in any profession, and has similar penalties) and the other is the issue of the confidentiality of data.

Yes, there are ethical principles involved in the latter. But the reason they are involved is that to get good data social science researchers give an undertaking of confidentiality. If they didn't, they wouldn't be bound by confidentiality constraints - but they'd get crap data. Given that they give an undertaking of confidentiality , they are ethically obliged to honour that undertaking.

[OnTheOtherHand]

22. I think the piece about "weighing" should also be addressed

Indeed, I thought we had gone over this issue quite a few times, but each time it is as if we are starting from scratch. I wonder why that is.

From prior discussions, this "value" of respondent anonymity in polls is so important that it makes the accuracy of a presidential election look insignificant by comparison. I say this based on past discussions in which I attempted to say that a weighing process would be used in court to see which interest was the more important (anonymity of respondents or presidential accuracy) and I got the strident and emphatic objections from Febble asserting the unethical nature of such disclosures.

This framing is, charitably, incorrect. There is no choice between "anonymity of respondents or presidential accuracy." The weighing is between the risk to respondent confidentiality and the value of the additional data for assessing the accuracy of the election.

Very few survey researchers or political scientists believe that the exit poll precinct identifiers are of any appreciable forensic value. If there is a strong argument for releasing the identifiers, to my knowledge no one has made it.

A few of us have spent considerable time attempting to explain the inherent limitations of the exit poll data.

[Febble]

23. Yeah, I managed to overlook that paragraph

When I saw it, I started to composed an edit on the lines of your post - then went off and had a beer instead.

Yes. Quite.

If you think it is important, Paul, please make your case. You will hear (and have heard) no "strident and emphatic objections" from me.

[Land Shark]

25. You're still primarily talking about the darned value of your secrets

That misses the point.

THere are lots of thinks that people think are totally confidential, that lose that confidentiality when the matter is litigated in court.

Some investigator, or court, may want to go back to the respondents and interview them as to their responses, and see if there are significant discrepancies with the reported confidentiality. This can happen in the context of a court order. The confidentiality is preserved, but the circle of those who know and can investigate has broadened somewhat to include the parties and the judge (the courtroom records are often sealed when this is the case).

So, with the above stipulation assuring that the information will never become widely known, what is your opinion on preserving confidentiality AGAINST EVEN LIMITED FURTHER CONFIDENTIAL "DISCLOSURE"?

[Land Shark]

26. Here's how DataUSA or Viewpoint USA apparently got busted

From the Connecticut Post. It appears that everyone surrounding this polling company was involved in criminal activity, but that's just an impression from this article. This guy Mackey was engaged in stealing $50,000 or more of computers off the desks of workers at a defense contractor. One wonders how they got so hooked up with Lieberman and Bush, eh?

Mackey suspected Mastri's ex-wife turned them in to the FBI.

Mastri countered by providing the government with information about his ex-wife, Tracy Costin, 46, now of Madison, and her polling company, DataUSA of West Haven and Guilford. That information led to Costin's indictment for wire fraud as a result of falsifying data for their clients. Her case is pending before Hall.

Mastri pleaded guilty to three wire fraud charges and awaits sentencing.

Smith pleaded guilty to conspiring to receive stolen computers and was sentenced to a month in prison followed by three years of supervision.

DOURCE: PC thief gets 2 years Connecticut Post (Bridgeport, CT) September 13, 2005 Tuesday.

The Connecticut post more recently said "A polling company owner admitted participating in a conspiracy to falsify data in order to meet deadlines for clients, which included the campaigns of President Bush, U.S. Sen. Joe Lieberman, U.S. Rep. Rosa DeLauro and Mayor John M. Fabrizi. Tracy Costin, 46, of Madison, admitted to U.S. District Judge Janet C. Hall that she participated in a conspiracy to commit wire fraud."

This CAN NOT be true, because pollsters don't engage in conspiracies to alter poll results! It's just absurd! It could NEVER happen! Not when involving a president or a senator! C'mon, we're not all tinfoil are we?

It appears from this that only when someone with inside knowledge has his life threatened (a nonquoted part of the boxed article above) does the scheme unravel. I wouldn't want to rely upon that "check and balance" on a consistent basis. One strange thing: their company name rarely appears in the news except for these indictments.

[Febble]

32. Of course it can happen

This CAN NOT be true, because pollsters don't engage in conspiracies to alter poll results! It's just absurd! It could NEVER happen! Not when involving a president or a senator! C'mon, we're not all tinfoil are we?

There are criminals in every profession. What are you alleging? That E-M faked their data?

[OnTheOtherHand]

27. you have a point?

Your hypothetical precisely dodges the issue we tried to pose. Under what conceivable actual circumstances would (e.g.) "go(ing) back to the respondents and interview(ing) them as to their responses" be especially useful to a forensic investigation of election fraud? I do not know, and evidently Febble doesn't either. You may know, but so far you aren't telling.

I won't speak for Febble, but my own opinion on preserving confidentiality depends upon the rationale for compromising it. Someone needs to state the rationale. And I mean an actual rationale. Hypothetically, if the information actually were crucial, I would certainly support getting it out to invetigators who could make good use of it.

[Land Shark]

28. one of several rationales or purposes for not preserving confidentiality

It would eliminate, for example, survey recordation error as a source of error. I thought this was obvious given this thread is all about faking the data/wrong data. There could be other reasons and benefits as well. In the context of investigating both an election and a exit poll, one would want to sample both the voters (probably not possible) and the respondents. (may be possible)

[OnTheOtherHand]

29. but... but...

First of all, we seem to have 'fallen' a long way, from election outcomes to "survey recordation error." I am not aware that any election inquiry hinges on the issue of recordation error. That would obviously be a concern for the pollsters, assuming (as one needn't) that the pollsters intend to get good results. But that isn't our subject here, is it?

Beyond that, it is likely possible to identify some respondents based on their demographic profiles, but it is not likely to be possible to identify all of them. I can't think why you would deem it simpler to sample the respondents than the voters.

Even if one could identify all respondents, it is hard to imagine how they could be compelled to report their votes accurately, especially in light of the evidence that people do not reliably report their votes in retrospect. (But I can certainly imagine circumstances in which wholesale canvasses could provide decent evidence of widespread fraud. I just don't see why the exit polls would be a crucial link.)

[galloglas]

30. Yes, we have fallen

So, I have a question OTOH.

Except for the perfectly obvious truth :sarcasm: stated by Paul in response #26,

"This CAN NOT be true, because pollsters don't engage in conspiracies to alter poll results! It's just absurd! It could NEVER happen! Not when involving a president or a senator!"

what is to prevent the polling industry from taking their rightful (I assume) commercial place, along with the Cheat-o-Matic manufacturers, in lining up to take their share of the gains, ill-gotten, from picking the bones of American Democracy through the wanton and willful undermining of the electoral pocess? A Code of Ethics? :eyes:

I do understand. But I am left troubled. And I now must make an admission to one and all, outing myself as it were.

For once, as a youth, I was a Boy Scout. And, as such, I swore the Scout's Oath, vowing to uphold the Scout's Law, among other things.

I promised to be:

1)Trustworthy
2)Loyal
3)Helpful
4)Friendly
5)Courteous
6)Kind
7)Obedient
8)Cheerful
9)Thrifty
10)Brave
11)Clean
12)Reverent

And, though I rue the day, I once broke # 8. Not only was I not cheerful, I was actually sad, if not overtly gloomy. I happened on November 3, 2004.

That said, will you please answer Paul's entry #26? The one you missed?

In particular, the part of # 26 quoted above, as I believe all at DU should do what we can to fulfill the Scout's Motto: "Be Prepared!"

Had we all done that in respect to 11-2-2004, perhaps I would not have broken Law # 8.

Alas!

[OnTheOtherHand]

34. what's to prevent pollsters from faking all their results?

Is that even the question? I have found all these threads pretty confusing.

First let me note that that isn't the topic I was discussing with Paul. I assume that anyone who thinks that the exit poll results were faked has no further interest in the 2004 exit poll controversy. (That is Lynn Landes's position as I understand it.)

There is nothing that can be counted on to prevent any and all poll results from being faked. That is one reason among many why both researchers and political observers treat individual poll results with caution, especially when they are "internal" results from campaigns.

I have seen no evidence that DataUSA faked poll results for the benefit of Bush, Lieberman, Rosa DeLauro, or anyone else. The indictment charges that DataUSA defrauded its clients by not conducting the work it had promised. That doesn't contradict my previous point.

If you think that pollsters generally are part of a conspiracy or conspiracies, then I would like to know why, but I don't see any way of bringing evidence to bear on the question. If you don't believe that, then I really don't know what we are talking about. Scientists do sometimes fake work in order to get published, but most people, most of the time, don't take an instance of scientific fraud as occasion to speculate that scientific research generally is faked.*

*EDIT TO ADD: I'm making an analogy here. Campaign-sponsored polling is, to borrow a theme from Patrick, more like market research than "science," although it does use some scientific principles. If you don't like that analogy, try another one.

[galloglas]

43. RE: Is that even the question?

"Is that even the question? I have found all these threads pretty confusing."


They have gone of on a tangent, I agree.

But the original post was about a polling firm pleading guilty to falsifying polling information. Since we have a guilty plea, no prosecutor has proven motive.

My concern, obviously, is that such a crime, using similar (if not such over-the-top) methods, could as easily be done for several different motives.

One might be, as this particular case might be, one of a polling company simply doing anything to complete their project in order to get paid.

But, (and I'm sure you can see the ramifications) a polling company might as easily use the same non-standard methods (like talking to dogs and cats) to skew, or even falsify, poll data in order to warp the public's vision of reality (as in, "Oh, look, this poll has Santorum up by 83%. I guess there is no reason to go vote now, Marge!".

At at some level of manipulated shift, it could alter an election. So, other than the code of ethics, what protects the public?

Knowing a pollster has been "shamed" by violating a Code of Ethics is damned cold comfort if they should be, IMHO, on a gallows for having manipulated the likes David Duke into the White House! So what protections are we afforded if this is the next card that Rove is dealing from his deck?

"First let me note that that isn't the topic I was discussing with Paul."

But Paul did state, "This CAN NOT be true, because pollsters don't engage in conspiracies to alter poll results! It's just absurd! It could NEVER happen! Not when involving a president or a senator! C'mon, we're not all tinfoil are we?"

And you never addressed the issue of poll manipulations used in order to subvert election results! That was my problem with your response. A Code of Ethics just doesn't cut it, IMHO.

After all, at some time or another, Josef Mengele must have taken the Hippocratic Oath!


"I assume that anyone who thinks that the exit poll results were faked has no further interest in the 2004 exit poll controversy."

You would assume so wrongly, I think. But it is not the issue at hand.

"There is nothing that can be counted on to prevent any and all poll results from being faked. That is one reason among many why both researchers and political observers treat individual poll results with caution, especially when they are "internal" results from campaigns."


I understand this. But, on the other hand, the "caution" would not keep a poll of lies from being part of a conspiracy, if the conspiracy required a falsified poll to be successful.

A Code of Ethics that implies, "Watch out, members may be liars", should not be a license to walk on any conspiracy charges.

"I have seen no evidence that DataUSA faked poll results for the benefit of Bush, Lieberman, Rosa DeLauro, or anyone else."

No one said they did. But, OTOH (pun optional), there is nothing to show that some other polling company might not do exactly that for Bush, or any other party.

"If you think that pollsters generally are part of a conspiracy or conspiracies, then I would like to know why,"

No, I don't think that.

However, I do think that any particular pollster could be helpful, if not necessary, to carry out a conspiracy. And I am more concerned that there be criminal penalties, with prison stretches attendant for them, than simply being drummed out of the Corps of Pollsters.

I am also concerned that ALL eyes be open for such a thing. If polls, true or false, were not significant in forming public opinion, and affecting human actions, they wouldn't be either sought out, or published.

Since they are significant in affecting public action, and each person's vote in November is "public action", falsification of polls, in order to affect those actions, is probably conspiracy.

"If you don't believe that, then I really don't know what we are talking about."

Well, since I stated that I don't believe that ("pollsters generally are part of a conspiracy or conspiracies"), but I still have the concerns stated above, I wouldn't understand why you fail to comprehend my reticence to rely of your "Professional Code of Ethics" to keep any pollsters from participation in Election Theft.

I mean, even Priests and Anchorites take vows! And,... well,... perhaps we should not go there.


"Scientists do sometimes fake work in order to get published, but most people, most of the time, don't take an instance of scientific fraud as occasion to speculate that scientific research generally is faked."

Granted. And I am not one of those who thus generalize.

But, as you stated "Scientists do sometimes fake work in order to get published". In so doing, they are breaking their professional codes for a damned site smaller prize than those that are generally awarded by K Street.

I am more worried about those polling "scientists" who are willing to tamper for Golden Parachutes, or cash in a back alley, rather than those who do so for publishing credit.

Polling professionals not seriously entertaining that as a possibility, and feeling the need for it to be addressed, is (to my mind) to be either fatuous, foolhardy, or complicit.

Bigger risks bring bigger gains. In polling, as anywhere else.

[OnTheOtherHand]

45. I don't think we're disagreeing about much

Is it possible that you are misreading the thread sequence? It's true that I didn't reply to Paul's gibe about pollsters. I didn't reply to that post at all. Febble did reply, to say, "Of course it can happen." That's that.

If you read my post #19, which precedes Paul's post, it should be clear that I never suggested that one could rely on a code of ethics to prevent pollsters from gaming their polls. I don't think that was Febble's point either.

I'm not sure what else to say. I'm not a pollster, and feel no need to vouch for the incorruptibility of pollsters. I will say that I have no reason to think that faking data is common, but that is not to say that I fail to comprehend your reticence etc. We may disagree about something, but I can't tell what it is.

[galloglas]

47. I think we AGREE about this!

Title edit: due to mind racing ahead of typing fingers!

That the thread got way too technical and confusing. Glad to know that we agree about the other aspects, also.

I suppose I'm most concerned about vigilance. If this Election Theft conspiracy was gamed out as long ago as I think it was (I exchanged with Febble recently on another thread about that subject), then we muct watch them carefully to keep from being Re-Roved.

One other Freudian/Rovian thing. The constanr retort of the Administration about "reality-based" critics is a worrying frame, I think, as the alternative is to believe the Rove/Goebbels line.

My concerning about polling is that it falls into Herr Goebbels territory, Orwell's Ministry of (Dis)Information.

[OnTheOtherHand]

49. vigilance is generally prudent!

I don't think the administration controls the polls -- for one thing, if it did, I think the approval ratings would be higher. Beyond that, the politics of polling is a complicated business, and I don't claim to have all the answers. Maybe we will come back to it.

[Febble]

50. Well, vigilance is good

I happen to think, personally, that it is extremely unlikely that the exit polls were gamed, and I don't think it is particularly reasonable to extrapolate from one small polling firm to the ethics of two companies, one of which was led by a man who has, in the last week, been the subject of glowing testimonies to his often cantankerous refusal to be bullied by his clients, who received a lifetime achievement award from the pollsters' professional body, and who was at one time its president.

But hey - he hired me.

As for political polls commissioned by political parties - well, it seems pretty clear that the political sympathies of pollsters are often reflected in their results. That doesn't necessarily mean they are dishonest - merely that their political sympathies may lead them to a different set of judgements as to what constitutes an appropriate weighting of their data. Although interestingly, in the UK, the ICM poll for the Guardian, the UK's big left-of-centre broadsheet, tends to downplay Labour polling - largely because of the disaster in 1992 when Neil Kinnock thought he was going to win the election, then didn't. We all still wonder whether he might have done if the polls hadn't put him in the lead beforehand.

FWIW, I did this yesterday from polling data that TIA came up with:



It's Bush's approval ratings since February 2001. The red line is ABC, the blue line Zogby, and the black line is the mean of:

Nwk Fox CNN Pew Harris CBS Time NBC

Here is TIA's plot of the same data:



It is clear that some pollsters get consistently higher/lower results than others.


On edit:

Link to TIA's DKos diary:

http://www.dailykos.com/storyonly/2006/9/10/10051/8074

Not that I agree with his take, exactly. ABC tend to be the highest, and Zogby tends to be the lowest. I'm not sure that that makes ABC a shill and Zogby right.

[Febble]

33. It is scarcely conceivable

that placing survey data with respondent details in the public domain would reduce recording error. I assume the mechanism you are suggesting is that it would allow independent investigators to contact the respondent and bug them about whether the details were accurate.

If so - just think about this:

The chances of getting a decent random sample of participants for any survey under those conditions would be vanishingly small. And even if your independent investigator managed to contact the respondents, how one would s/he judge whether the response given to the investigator post hoc were more or less accurate than those given at the time of the survey? In fact, substantial evidence would suggest that contemporaneous recording is more accurate, which is why contemporaneous notes carry more legal weight in court than impressions recorded after the event (or they do in the UK). And the price you'd pay for this dubious "check" on the accuracy of the contemporaneously recorded data, is that your data would be junk because you wouldn't have any reason to suppose that your sample bore any remote resemblance to a random sample. It would be a volunteer sample. It is hard enough persuading people to take part in research - and in my experience, many have serious concerns about the confidentiality issues. Of course, in my field, we are always able to reassure them, by detailing our rigorous confidentiality protocols. We are also, in the UK, bound by our Data Protection Act. And if you are systematically losing participants who have confidentiality concerns, your sample will be seriously skewed.

So your premise is wrong (it would not necessarily eliminate recording errors) and it would fatally jeopardise the validity of the survey.

Measurement error (which includes recording error) is indeed a source of error in surveys. It is one of the many sources of "non-sampling" error that must be considered when interpreting survey results. There are chapters and chapters of text-books on measurement error, just as there are chapters and chapters on selection bias and non-response bias. They are difficult to avoid, but good methodology should minimise them. But even a survey conducted with good methodology will be vulnerable to these sources of non-sampling error (and others). Part of the job of a good data analyst is to identify and control for these sources of error during data analysis.

Trying to eliminate recording error by allowing public access to respondent details would be not only absurdly ineffective, but would be appallingly poor methodology.

[Febble]

31. I have no problem

with the idea that a court could demand to see exit poll data that could breach confidentiality, if a case could be made that the data could shed light on the issue of election fraud, and if appropriate conditions were made that prevented public disclosure.

I am simply asking you to make that case.

I would be grateful if you would not ascribe to me positions I do not hold. I do NOT think that data that could breach confidentiality should simply be placed in the public domain on the off chance that some stray analyst might detect some fraud in it. I consider that would violate the confidentiality agreement made between the respondents and the pollsters.

I DO think that if a coherent case were made as to what the data could demonstrate (for example, the presentation of a testable hypothesis), then the data could and should be made available to analysts who could interrogate the data for that demonstration. As was done for ESI.

But I have, as I said, yet to see that case made.

(I should add that one important hypothesis was tested - by me - and the results placed in the public domain by Mitofsky. Steve Freeman's response was that he'd want to do the math himself. I don't think data should be made publicly available so that other analysts can repeat a reported analysis on the same data. The idea that findings need to be "replicated" by "independent analysts" is a canard. Replication studies are done on new data. In any case, if Freeman were to repeat my analysis on his own computer, he would have no guarantee that the data we were both working from was not fake. A replicated finding won't necessarily offer anyone any reassurance, if what is in question is the data itself. That should be an important take-home message from this case - and is why faking data is such a serious professional and criminal offence.)

[Land Shark]

36. For the most part all you needed to say was this:

"{I agree} with the idea that a court could demand to see exit poll data that could breach confidentiality, if a case could be made that the data could shed light on the issue of election fraud, and if appropriate conditions were made that prevented public disclosure."

One of the primary points I've been making is that courts can do precisely that which you've identified above.

You asked for the precise facts that might meet such a standard. I'm going to pass on that purely hypothetical exercise at this point.

Febble also said: "The idea that findings need to be "replicated" by "independent analysts" is a canard. Replication studies are done on new data."

Replication of results is a principle of SCIENCE, not a 'canard.' Though you attempt to distinguish it by saying that they are done on "new data." That's often true, but not when the population at issue is irreplaceable, like it is with exit poll data: One can't make up new exit poll data for the same election, or collect any more of it. You can't stand in the same river twice, as they say.

This condition where duplicate data or results can not be obtained any other way is one ground for nullifying the work product doctrine, which usually allows attorneys to keep their nonpublic experts a secret and away from the other side's use and benefit. But if it's 'the only game in town' you can be forced by the judge to share your little monopoly.

[Febble]

37. Of course

replication studies are a PRINCIPLE of science. I spend half my life trying to replicate what I did in the other half (or what someone else did).

But I am fed up with people implying that "replication" means doing the same analysis on the same data "independently". Sure, running the same analysis on the same data "independently" - i.e. on someone else's computer - would be a check on the trivial possibility that the first analyst actually did the math wrong. But that's not what scientists mean by independent replication. It wouldn't be "independent" in any real sense, anyway because, as with the ESI study, the data itself has to be assumed to be correct. I have heard the ESI study criticised because Mitofsky is an author. Of course he's an author - he collected the data.

What independent analysts could legitimately do on the data is run different analyses (as, indeed, ESI did). In other words, test new hypotheses. But that simply brings me to your first point. If you have a testable hypothesis - in other words, if you can say "if the election were stolen, such-and-such should be apparent in the data" - then you'd have a case for disclosure. Or at the very least, you could present it as a hypothesis to be tested. Warren Mitofsky actually offered to test any hypothesis Steve Freeman proposed. AFAIK, he didn't propose one, but I may be wrong.

I suggested a detailed series of hypotheses to Mitofsky. To my astonishment, he hired me to run them.

[Febble]

40. Clarification

by "the data has to be assumed correct" - I mean that there would be no way by which any independent analyst could check whether the data they were given to analyse was the data collected. All the analyst could do would be to check that the same calcs gave the same results. FWIW, when I first received the data, I did redo the analyses reported in the E-M report, and of course got exactly the same results. I also replicated my analysis on the publicly archived sub-samples - and found comparable results (not identical, of course, because the sample was smaller) from which I could conclude that the Roper subsample was a good random sample of the original questionnaires (because it had the same statistical properties).

But there is no way I, or any independent analyst could ascertain whether the data themselves are fake. As I said, there might be additional, key questions that could be asked of the data, on the assumption that the data are real. But you appear to regard those questions as "hypothetical".

Isn't that like calling a witness without knowing what they are supposed to have witnessed?

So I ask again, in good faith: what questions would you ask of the data? If you know, it might be worth putting them on the witness stand.

[OnTheOtherHand]

39. "purely hypothetical exercise"

Well, that's exactly it. I am beginning to think that the irrelevance of the precinct identifiers makes them all the more delightful as a conversation topic.

No one has ever contested that a court could demand to see exit poll data, and you know it. So what on earth are you trying to accomplish?

I ask this in the shadow of your eye-rubbing declaration to Febble of July 29, which you never did explain:

I don't think you can BE a serious social scientist and STAY HERE, engaged at the level you are. On the other hand, I can still be a lawyer and engage in this because people expect lawyers to fight once in a while.

I have a hard time not reading that as trashing Febble and all "serious" social scientists. So, when you raise hypothetical issues and then refuse to engage them yourself, I wonder. Is it actually your view that every time you argue with Febble, she is revealed as not a serious social scientist, and you are revealed as a fighter?

Hmm. I think it would explain a lot.

[Land Shark]

55. Let me spell it out for you

since you are, more and more often, having trouble understanding not only me but other people as well.

When "debate" here breaks down (and a remarkable number of people have you on ignore, OTOH) into some kind of rumble, the degraded nature of the analysis that takes place thereafter is in my estimation beneath the dignity of a serious social scientist. But, although lawyers can engage in some of the most ennobled discourse in our society, clearly society also expects that as trained "advocates" or fighters they are known to and perhaps suited to brawling at times.

Thus, the breakdown of the level of debate poses more of a problem for folks like professors and serious social scientists than it does for lawyers adopting the online persona of a shark. Make sense? In this sense, it probably means that social scientists and professors have higher social approval ratings than lawyers do (so in that sense my comments are self-deprecating) but they sustain more damage than lawyers do when they resort to intellectual brawling. Still, this kind of lowering is not of a permanent nature I don't think. It's more like the saying "it's beneath you."

That being said, this is still an overture toward a more civilized debate, but you'll probably call me out on something in this post, too. (see limitations on "civilized debate" in paragraph at bottom) If you just complain or call me out on something, then I'll probably add my name to the list of people who have put you on Ignore. You'll be my first. And, once you are in the future on Ignore, should that happen, if you say things about me or "Land Shark" that are not nice, just make sure you're on the right side of the law when you do so. OK? But that's contingent, as I said. It's along the lines of "let's not start fights we can't finish, OK?"

So, do you want to play nice, or not? Keep in mind that I don't have unlimited time to play with you, so it's often not fair to critique a response as if I should have spent a half hour or more researching and constructing it precisely and to your satisfaction. I trust you will provide me with that courtesy.

[autorank]

56. "In Praise of 'Ignore'" ...keeping priorities straight.

Use of ignore was suggested to me as a means of maintaining focus and perspective. After all, there are times when you simply want to put negative input on "extinction." The permanence of the ignore function is it's beauty. Think of all the time you will save and the aggravation you will jettison.

Of course, there are those you put on "ignore" who continue to show up even they know you can't see the response. On one level, that type of behavior is like walking up behind someone and punching them in the back knowing you'll get away with it every time. On a higher level, it's just another gesture that justifies the decision to ignore in the first place.

The rules allow you to tell someone they're on ignore as long as you follow through. Like all things in life, "follow through" is key.

Free up some serious time if don't like someones comments. It's your Admin given DU right.

[Name removed]

58. Deleted message

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

[Name removed]

59. Deleted message

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

[OnTheOtherHand]

60. no, I do not "rise above" baseless attacks on people's integrity

They were your attacks. They are your attacks. And I will not ignore them.

I'm almost flattered that you claim to consider me the most disruptive poster in the history of DU. People will have to score that for themselves.

[MissWaverly]

35. she told her employees to make stuff up

it wasn't manipulations of interviews, it was mythical responses, I believe she said something
like interview dogs and cats but get the numbers (pro-Bush) that we need. But, although,
I have believed for some time that Bush and Lieberman numbers are inflated, does not mean
that all pollsters are dirty tricksters, otherwise Bush would be polling at 90% instead of
the basement.

[Febble]

38. Do you have a reference for that?

I mean, the "pro-Bush" part. You may be right, but all I have read was that she made up interviews to fill her quota, not that she made the responses "pro-Bush", although that's perfectly possible.

[MissWaverly]

41. well, I concluded that she faked them for Bush

since the prosecutor stated that some of the surveys had all fabricated answers

FBI Special Agent Jeff Rovelli said 50 percent of information compiled by DataUSA and transmitted to Bush's campaign was falsified, the Connecticut Post reported Thursday.
Assistant U.S. Attorney Edward Chang said on several occasions when the company was running up against a deadline to complete a job, results were falsified. Sometimes, the respondent's gender or political affiliation were changed to meet a quota, other times all survey answers were fabricated.

http://www.syracuse.com/printer/printer.ssf?/base/politics-0/1157635817306640.xml&storylist=politics

[Febble]

42. Well she seems to have

defrauded her clients. But I don't see anything in the indictment that suggests she actually made the numbers go in their favour. The indictment says she made up responses to fill her quotas and meet deadlines.

Here it is, if you haven't read it:

http://www.usdoj.gov/usao/ct/Documents/DataUSA%20Indictment.pdf

There is nothing to suggest anything other than they paid for an honest survey and were given junk.

[OnTheOtherHand]

44. it might be worth pointing out

a basic reason for skepticism that the Bush campaign wanted faked results is that with so many different media outlets polling, a faked result is unlikely to have done it much good. Chances are that the campaign was doing some message-testing or looking for other info that it couldn't glean from other polls.

I don't know if it matters much, but if we are going to talk about the story, we might as well talk about the story. (Also, while Bush and Lieberman have been mentioned, another client apparently affected is Rosa DeLauro -- a pretty darn liberal Democrat who voted against the war.)

[MissWaverly]

46. well, if you were doing this for a client and faking all the answers

don't you think you would make them favorable to him?

[Febble]

48. Well, frankly, no

but I'm not arguing that they didn't, just that there's no suggestion that they did.

If I was faking results (since you ask) and I'd got a result in which 45% of my respondents said one thing and 55% said another, but my quota was too small, I'd fake some extra responses in the same proportion. That way, I'd bring down my MoE without changing the result. But it would be fraudulent, because I'd be implying that my results were more reliable than they were.

But I'd have less chance of being found out that way, because there'd be a greater chance that my initial results were correct, and in line with any corroborative polls, than if I not only made responses up, but skewed the proportions as well.

But whatever. I just think it's worth getting the facts right, and keeping speculation in the realm of speculation. What we know is that they made up data. What we don't know whether they also skewed their results. They might have done, but there's no reason to suppose they did.

What we know they did was bad enough.

[MissWaverly]

51. Yes, But we know that Bush and Liberman, have been higher

than they should be, after all Lieberman lost the primary, is it really being a moonbat to suppose
the extra false surveys were to prop them up. Otherwise, why fabricate them at all?

[Febble]

52. Well, the indictment says

to meet quotas and deadlines.

If you haven't got filled your quotas by your deadline, then you have an incentive to make them up.

But I'm not saying they didn't prop up the numbers. I'm just saying that wasn't what was in the indictment.

And in any case, the data would have been junk. It could have been junk in either direction. If Lieberman was paying for proper polls, and they told him he was doing well, then he lost, I expect he'd be furious. Unless he actually paid them to fake his results - in which case he'd be in the conspiracy.

And there just isn't any evidence for that. The indictment says they conspired to defraud their clients, not that their clients paid them to hype the numbers.

[MissWaverly]

53. well, maybe you have a point

I'd like to wait and see but I don't think that we will hear what happens.

[OnTheOtherHand]

54. to some extent we might be able to figure it out

If the numbers were in fact goosed to make Lieberman look better, then somewhere Out There there ought to be a media report of an internal Lieberman poll that highballs Lieberman support. It wouldn't be slam-dunk proof, but at least it would be data. (Of course, the absence of such a report wouldn't prove anything either.)

By the way, a pollster wouldn't necessarily need to fake data in order to highball a candidate's support. Other things can be done to torque the final result. Again, this is one reason why people tend not to take internal poll results very seriously.

[kster]

57. If the voters are not allowed to see the votes being counted

exit polls will always be off, unless of course you voted the same way as the guy/gal in charge of the secret vote counting machines.

They NEED us to change the subject from, THE VOTERS ARE NOT ALLOWED TO SEE THE VOTES BEING COUNTED! over to the exit poll, THE exit poll controversy, is their way to erase the faCt that the votes are being counted in SECRET.

THE VOTERS ARE NOT ALLOWED TO SEE THE VOTES BEING COUNTED!
THE VOTERS ARE NOT ALLOWED TO SEE THE VOTES BEING COUNTED!
THE VOTERS ARE NOT ALLOWED TO SEE THE VOTES BEING COUNTED!

Is this fair? Will the exit polls ever be accurate if,

THE VOTERS ARE NOT ALLOWED TO SEE THE VOTES BEING COUNTED!

DON'T LET THEM CHANGE THE SUBJECT, PLEASE,

[galloglas]

61. Yes. That was the point !

Right ON, my man!

[Febble]

62. It was?

The OP was about exit polling?


on edit:

Jeez, I got this OP and the other confused.

Are you seriously saying that it was the difference between the exit polls and the pre-election polls that made people think the election was stolen?

Funnily enough, the difference between the exit polls and the pre-election polls is one of the reasons one might conclude that the polls themselves can tell us little about whether the election was stolen. Mark Lindeman showed that the deviation of the state count from the pre-election polls is actually negatively (not positively) correlated with the deviation of the count from the exit polls. Which would suggest that what ever caused the deviations didn't have much to do with the vote count.

[galloglas]

63. No. This is what I thought.

"Are you seriously saying that it was the difference between the exit polls and the pre-election polls that made people think the election was stolen?"

While they may have raised some suspicions, or bolstered some people's already reached conclusions that the election was gamed, it is not my contention. I would only say that Mitofsky's handling of this issue, in the aftermath, contibuted to the look of something sinister being afoot.

I have my own reasons to believe it was stolen (maybe as Rove had his own "feelings", on Nov.2, that the "exit polls" were off).

I believe most people found the sudden reversal of the numbers around 12:23 AM (when they were probably sleeping), and their own gut feelings, to be their indicators of fraud. It is, after all, more egregious to go to bed with Kerry being declared the next President, and wake up to find it not so, than it was for the Chicago Times, in 1948, to print "Dewey Wins!"

You, Febble, are a "numbers" person. Others are "people readers" who see the signs, sniff the wind, or can tell by some one false statement, when something is amiss.

Caesar, a "people" person, should have followed up when noticing Cassius, and saying, "Yond Cassius has a lean and hungry look, He thinks too much; such men are dangerous."

I also think many had seen the POTUS race in 2000, the Saxby Chambliss fiasco, Chuck Hagel's amazing win, and others such things over the years.

They, like Cassius, they were thinking much. And, when 11-2-2004 happened, thought to themselves, "that is no coincidence!"

Statistical and numerical certainty, your bailiwick, is an entirely different form of human intelligence, or "intuition", if you will.

Such "people" people, referenced above, recheck their thinking, reconfirm their suspicions, and conclude "this is no coincidence!"

They then try to determine how the theft was done, and either prove it and prosecute it, or dedicate themselves to the prevention of the same "coincidence" happening in the furure.

For myself, it took a full two weeks for me to be convinced by my ownself. That said, I have decided HCPB are the most sensible way to prevent any more coincidences.

Above and beyond that, there are other reasons for HCPB.

1) It is a more moral dispensation of public funds to pay citizens, rather than "corporate persons", to count votes.

2) It is more environmentally sound.

3) It connects the people more directly to the Republic, making it more democratic, by nature.

4) It involves the people more directly in the functions of their government, much as jury duty or Selective Service do, and increases their connection to it.

There are more reasons, none which hinge on preventing election theft.



So, Febble, I ask you these two questions.

1) Do you think the exit polls were off because of "reluctant Bush responders".

2) Do you think that Bush actually won, fair and square, the 2004 race?

[Febble]

64. Well thanks for this clear statement

Sure, I'm a numbers person, but like yours, my intuition was that something was afoot in November 2004. It's why I got involved.

(And as a matter of fact, I'm a sort of people person too - I'm a psychologist, after all, and spent most of my first half century as a musician. I'd hate to be type cast).

So to answer your questions:

1. I think the exit polls were "off" because the sample was biased. I don't think that non-response bias was a very large factor, although I think it may have contributed. I think the main source of bias was at the level of respondent selection, and therefore in the interaction between a number of factors, including the interaction between the interviewer and potential interviewees; the opportunities afforded at the precinct for departures from strict Nth voter protocol; and an underlying tendency for Kerry voters to be more willing to participate in the polls than Bush voters. These conclusions are based on actual evidence within the data.
   
   
2. No, I don't think Bush won the 2004 fair and square. I don't know whether, had the playing field been level in 2004 whether he would be in the White House today. I do know, that had the playing field been level in 2000, he almost certainly wouldn't be anywhere near it. But I think a substantial amount of evidence (some of it actually from the exit poll data) suggests that it is unlikely that vote-switching occurred on a widespread and massive scale. I think that the most likely loss net loss of cast votes for Kerry was through overvotes, undervotes, provisional ballots, and possibly scams and confusion with punchcards - old-fashioned stuff. And I think he lost a great number of potential votes through various forms of voter suppression, some of it negligent, some of it malicious.

[galloglas]

65. And thanks for your very, open frank answer

EDIT: Clumsy grammar


It's nice when one gets the chance to see behind the names, and positions, to the quite real human beings behind them. Thanks!

[fooj]

66. Thank you for your assessment.

Much appreciated!:hi:

[troubleinwinter]

67. I throw in with that!

Feeble's posts are educational, clear and enlightening to me. A complicated subject that I had not understood well. But I'm learning much.

Appreciation here too!

[OnTheOtherHand]

68. excellent post

I have two stray thoughts that may not actually be distractions. One is that I don't think "statistical and numerical certainty" is really Febble's bailiwick. Statistics is about degrees of probability and confidence, not certainty. This isn't just a technical quibble about "p values"; quite apart from quantifiable probabilities, there is the task of assessing the assumptions that allow us to calculate these probabilities. A healthy respect for uncertainty could be world-saving, if we let it be. (Not that one needs statistics for that!)

Second, I think we all generally converge on preventing "this" (as you put it, "the same 'coincidence'") from happening again, even when we don't entirely agree on what "this" was. The fact that we can't agree on what "this" was is, in part, evidence of how messed up the election system is. Of course, our assessment of what happened in 2004 does influence our priorities for the future. In some respects it may be immaterial whether Bush stole ten million votes via some sort of vote tampering; in others, probably not.

[autorank]

69. gollogas, Here's some expert input from TruthIsAll on your question

This is an interesting thread. There was a lot of work done on pre-election and election day exit polls. The particular polling company sounds pretty rotten. I discussed with with TIA and got this response, which should be helpful. The key point is, one rotten poll, and these guys were rotten, does not invalidate the overall cumlative measurements of pre election polls. I don't know if these polls were included but I doubt it. Even if they were, the match with pre-election and exit polls is even more proof of the point given that these polls were skewing the results. When you look at the links to geocities sites, check and see if this company was included.

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

The OP premise is that the pre-election and final exit polls did not agree.
This is NOT true. As a group, they predicted a Kerry win.

In 2004, the final pre-election state and 18 national polls closely MATCHED the corresponding 12:22am exit polls. They did not agree with the (bogus) FINAL exit polls, which were forced to match the (bogus) recorded vote.

The distinction must be made between the near-pristime Nov.3 12:22am exit polls and the contaminated finals as displayed on CNN 12 hours later.

Assumption (pre-election polls): Kerry wins 60% of the late undecided vote





From the Interactive Election Model: http://www.geocities.com/electionmodel/InteractiveElectionSimulation.zip

Assumption (pre-election): Kerry wins 67% of the late undecided vote


Cheers

tia

[Febble]

70. People interested in this

might also be interested in another take on this data here:

http://inside.bard.edu/~lindeman/surprise01.pdf

The interesting finding is that the degree to which the pre-election polls deviated from the state count is not positively correlated with the degree to which the exit polls deviated from the state count, suggesting different causal factors for the two sets of deviations. The more one is inclined to attribute vote-theft to one, the more one is driven to ascribe biased sampling to the other. Although there is no logical reason not to ascribe both to biased sampling - one would merely have to postulate different sources of bias.

[OnTheOtherHand]

71. Febble, I just posted a short version

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

The PDF paper sloshes around trying to find some pattern in the distributions of "surprise" (vs. pre-election polls) and red shift. The short version just points out that the pre-election polls and exit polls don't really "match" at all. Treated as fraud measures, they tend to point at different states.

I wrote the text fast and sleepy, so I will probably have to go back and translate from the original StatSplat.

[Febble]

72. It might also be worth pointing out

that none of your exit poll estimates are based on the reweighted exit poll estimates (the ones TIA characterises as "bogus").

The magnitude of shift is calculated from the difference between the estimates made at "Call 3" - ie. when all the tallies were in, but no vote returns were included in the estimates - and the official count.

[OnTheOtherHand]

73. yeah, maybe I can go back and clarify that online

Of course, if the estimates were altogether reweighted, there would be no red shift.

I also did similar analyses using mean Within Precinct Error as the shift measure, and/or using other pre-election poll estimates. Qualitatively, they all yielded about the same results.

[autorank]

74. gollaglas, The difference is surprisingly small..check this one out (TIA!)

In the interests of free speech and open 
debate, I'm offering this gem...

Just assume 60% of unallocated votes 
going to kerry, calculate the Kerry 
2-party share, and voila 0.37%.  Once 
again, the similarity between the 
pre election polling and the exit 
polls is right on target.

When will people get on board and recognize 
that there is a wealth of information that 
points to serious problems in 2004. 
 
Kennedy noticed it.  
Freeman noticed it.  
Baiman noticed it.  
And from THE VERY START, TruthIsAll noticed it.  

The more work that's done 
the less chance there is, actually we're 
down to Zero chance, that there is an 
alternative explanation to fraud.  Sorry 
to disappoint those who'd like to think 
this can't happen but it did.  Look at 
the numbers, they tell the whole story.  
Of course, maybe the math is wrong :rofl:  
If it is let me know somehow but I seriously
doubt it.  If the math is right than the 
difference between the 
pre electoin and exit polls os 0.37%.

Sweet!
================================
TruthIs All:


The pre-election state polls matched the exit polls. Here's
proof.

Assumption:
60% of Undecided Vote allocated to Kerry				

Calculate:
Kerry 2-party vote shares		

Key result:
Net deviation between the Kerry average state pre-election
polls, 
(adjusted for the undecided vote) and the average state exit
polls is
0.37%.

Column headings:
KPoll: Kerry pre-election poll
Kproj: Kerry pre-election poll+undecided allocation
Kexit: Kerry Exit poll
Kact:  Kerry Actual vote

KpDev: Kproj-Kact 
KeDev: Kexit-Kact
NetDev:KeDev-KpDEv

Sensitivity of net average deviation to pre-election Kerry
undecided allocation							

Und %    50	 55	60	67	75		
NetDev % 0.26	-0.05	-0.37	-0.81	-1.31		

.	Kpoll	Kproj	Kexit	Kact	KpDev	KeDev	NetDev
Avg	48.7	49.4	49.0	47.3	2.07	1.70	-0.37
Median	49.7	50.1	49.5	47.6	1.91	1.78	-0.13

AL	40.6	41.4	41.1	37.1	4.30	3.98	-0.32
AK	34.5	37.8	40.1	36.8	1.03	3.37	2.34
AR	47.4	48.0	46.6	45.1	2.93	1.53	-1.40
AZ	48.9	49.6	46.9	44.7	4.88	2.21	-2.67
CA	53.8	54.4	55.7	55.0	-0.64	0.69	1.33
							
CO	49.5	50.0	49.1	47.6	2.37	1.44	-0.93
CT	55.3	55.6	58.5	55.3	0.33	3.20	2.87
DE	54.2	55.2	58.4	53.8	1.37	4.61	3.24
DC	87.6	84.6	91.6	90.5	-5.92	1.11	7.03
FL	50.0	50.6	49.9	47.5	3.12	2.45	-0.67
							
GA	44.7	45.6	43.1	41.6	3.95	1.46	-2.49
HI	50.0	51.0	53.3	54.4	-3.40	-1.08	2.32
ID	33.7	36.6	33.3	30.7	5.92	2.66	-3.27
IL	56.3	56.4	57.1	55.2	1.19	1.93	0.73
IN	40.2	40.8	41.0	39.6	1.22	1.39	0.17
							
IA	53.2	53.6	50.7	49.7	3.94	1.01	-2.93
KS	38.1	38.8	34.6	37.1	1.67	-2.53	-4.20
KY	41.1	42.0	40.8	40.0	2.01	0.76	-1.24
LA	45.5	47.2	44.5	42.7	4.53	1.83	-2.70
ME	56.2	56.6	54.8	54.6	2.02	0.25	-1.77
							
MD	55.7	55.8	57.0	56.6	-0.77	0.47	1.24
MA	70.3	69.4	66.5	62.7	6.66	3.72	-2.94
MI	53.6	53.8	52.6	51.7	2.07	0.83	-1.25
MN	54.2	54.4	54.6	51.8	2.64	2.85	0.21
MS	45.2	46.2	43.2	40.5	5.71	2.71	-3.00
							
MO	47.3	48.2	47.5	46.4	1.82	1.09	-0.73
MT	38.7	40.2	39.3	39.5	0.70	-0.22	-0.92
NE	34.4	36.2	36.5	33.2	3.05	3.39	0.34
NV	50.0	50.2	50.7	48.7	1.52	1.98	0.46
NH	50.0	50.6	55.5	50.7	-0.09	4.80	4.89
							
NJ	54.3	54.8	56.1	53.4	1.43	2.76	1.33
NM	50.0	50.2	51.3	49.6	0.60	1.74	1.14
NY	59.4	59.4	64.0	59.3	0.11	4.68	4.57
NC	48.5	48.8	47.3	43.8	5.04	3.55	-1.49
ND	38.9	41.0	33.6	36.1	4.91	-2.51	-7.42
							
OH	51.5	51.8	52.1	48.9	2.86	3.12	0.26
OK	31.5	34.6	34.7	34.4	0.17	0.30	0.13
OR	53.2	53.6	51.2	52.1	1.49	-0.89	-2.38
PA	52.6	53.0	54.4	51.3	1.74	3.15	1.41
RI	60.9	60.8	64.2	60.6	0.22	3.66	3.44
							
SC	43.3	43.8	45.8	41.4	2.44	4.42	1.98
SD	44.7	45.6	37.4	39.1	6.51	-1.67	-8.18
TN	48.5	48.8	41.2	42.8	5.99	-1.66	-7.65
TX	38.5	39.4	36.8	38.5	0.91	-1.65	-2.56
UT	25.8	28.2	29.9	26.7	1.55	3.28	1.73
							
VT	57.0	57.2	65.7	60.3	-3.10	5.39	8.49
VA	48.0	48.2	48.0	45.9	2.33	2.09	-0.24
WA	54.2	54.4	55.1	53.6	0.75	1.42	0.67
WV	47.9	48.6	45.2	43.5	5.08	1.67	-3.41
WI	53.7	54.0	50.2	50.2	3.81	0.02	-3.79
WY	30.9	32.6	32.1	29.7	2.91	2.38	-0.53

[Febble]

75. Well, for those who can see my posts:

The more work that's done the less chance there is, actually we're down to Zero chance, that there is an alternative explanation to fraud.

Fact check:

A great deal of work has been down on data from the 2004 election, including the exit polls, and there is a great deal of evidence to support alternative explanations to fraud.

That is not so say that fraud did not occur. But the exit polls discrepancies are strongly correlated with methodological factors likely to have been associated with biased sampling, and not at all with swing to Bush, either from 2000 vote results or from pre-election polls. It is therefore simply untrue to say that "we're down to Zero chance, that there is an alternative explanation to fraud". On the contrary, the statistics say that the chances are high, and that the chances that the fraud explanation accounts for much of the discrepancy is low.

I post this information not to be argumentative, but in the interests of honesty. It is (remotely, in my view) possible that fraud was a substantial contributor to the exit poll discrepancy. But there is certainly a very much greater than zero chance that it was due to "alternative" causes, because these are highly statistically significant correlates of discrepancy in the data.

[OnTheOtherHand]

76. maybe one can be even clearer

autorank cites "a wealth of information that points to serious problems in 2004," and of course that is true. It isn't helpful to set this up as a debate between people who think Kerry won by seven million votes and people who think the election was clean as a whistle (however clean that is).

It is possible to jigger the pre-election polls in such a way that on average they are close to the exit polls. However, the individual state pre-election polls do not closely coincide with the exit poll results. Again, I've tried to summarize the key points at http://inside.bard.edu/~lindeman/surprise.html .

Some people write about "free speech and open debate"; others actually practice it. Do we prefer a majoritarian movement, or is it crucial that we pretend that everyone who knows anything agrees with TruthIsAll about everything? Because my silence, or Febble's, would not make that true.

[Cameron27]

78. Thank you for the clarity of your explanations

Both Febble and OTOH, very impressive work. You both captured my interest in a subject that I find extraordinarily difficult to understand. As a matter of fact, I strolled over to Kos last night to read Febble's diary (wonderful) and peeked at OTOH's site (also wonderful)

Kudos and thanks from a mathematical dunce.

[OnTheOtherHand]

79. (blush)

Golly, if your interest survived a visit to my site, you may just have found a new hobby. ;) I keep thinking I should make it more user-friendly.

Thanks for the warm words. Sometimes it's easy to get cynical about attempts to communicate.

[Cameron27]

80. Don't get cynical OTOH

You're doing a fine job of communicating. I won't post much about it, but I'll continue reading and learning.

[Febble]

81. Thanks

Your comment is very much appreciated!

My DKos diaries are a bit old now. I've been thinking I should do an update.

[autorank]

77. Someone can't stand the Truth...