New USCV criticism Revised -- TIA what about the math part of this?

[davidgmills]

http://www.geocities.com/lizzielid/WPEpaper2.pdf

Somebody make sense of this if you can, please. It has some USCV personnel input.

[mgr]

1. Simply

David:

What it says is that NEP is correct in saying bias affected all sampled categories, was not greatest in the precincts with high percentages of republican registration. It contradicts the USCV report's contention regarding the reluctant Bush responder.

What the math shows is that the bias is not symmetrical in its effect, which is the source for the purported USCV error.

The implication for me, is that if one were to remodel the exit polling to attempt to remove the bias, a single correction term is not available, but that one which is dependent upon partisanship might. It's a step toward rehabbing the exit poll to where potentially it may detect voting irregularities.

Mike

[davidgmills]

3. Here is my email to Rick 0f Stones Cry Out

Well, with as little as I know about the subject, I was always a bit concerned that USCV's points about the Bush Non-Responders were pushing it a bit. Their proof didn't seem to be glaring evidence that there was no such thing as the Bush non-responder phenomenon.

However, I still don't find the Bush Non-Responder excuse credible in my mind, but I would never say it couldn't exist.

The reason I don't find it credible is that I have many conservative friends, far more so than liberal ones, (because around here I am in the real minority), and none of my conservative friends would be the slightest bit sheepish about being approached by a pollster, regardless of pollster age, race or dress.

On the other hand, I represent many of the poor and illiterate because of what I do for a living, and I can easily see these people being intimidated by someone of a higher status or educational level and see these people not wanting to respond. Many of these people also aren't able to go to the polls at their leisure and many would not have the extra time to fill out questionnaires. I sure don't see these people as being zealous responders.

Add to that the fact that the conservative districts have much better voting places, and more voting machines with shorter lines and shorter waits, I really have trouble buying it.

So the whole concept just goes against my personal experience and what I observe in the people I know and deal with all the time. It just strikes me as intuitively wrong.

I realize that intuition can sometimes fool you but it is my intuition that really makes me demand proof that this bias exists in reality and not that it is just a theoretical exercise.

If the exit polls are wrong, then what I am more inclined to believe is that the experts just overpolled Democratic districts and underpolled Republican ones. But no expert seems to be suggesting this is the case.

This is what my dad tends to believe by the way. That no polling expert, no matter how good he is, no matter how much data he has beforehand, can accurately pick representative precincts because Americans are too mobile and the Republicans are usually leaving Democratic areas and not the other way around. So what seemed like a representative precinct last election may not be representative at all of this one.

Take care Rick.

[davidgmills]

9. Rick's response to me was

That he understood that the BNR was not intuitive but that is what he data showed.

But since TIA's cited quote below shows that it can be either bias or fraud, will somebody please tell me why I should believe what is not intuitive (BNR) and disbelive what is intuitive (fraud)?

[TruthIsAll]

26. That is NOT what the data shows. The data SHOWS that Kerry WON.

Don't let them FRAME the issue. The data shows that the vote count was wrong.

There is no Burden of Proof required.
The Numbers cry out far louder than Stones ever will.

FRAME THE ISSUE:
Either you believe the RBR, which can not even be called conjecture, since it is pure off-the-wall nonsense (never used until now), or you believe the election was stolen.

Just like it was stolen in 2000.
Just like it was stolen in 2002.

Just like it will be stolen in 2006.

WHICH IS IT: RBR OR REAL BASIC RESEARCH?

[sunshinekathy]

13. Response by US Count Votes is coming in a few weeks

Hello Everone,

Lots of us at US Count Votes tried to show Elizabeth prior to her printing her work, how she was selectively considering some facts and disregarding others to support her hypothesis. I haven't read her paper yet, so she may've adjusted her conclusions.

It may be that all she's showed is that exit poll discrepancies were not concentrated in particular types of partisanship precincts, but are more likely to have occurred everywhere, in which case I agree.

At any rate, several of US Count Votes' statisticians and computer scientist Bruce O'Dell are working on a further development of our Analysis of Exit Poll discrepancies and a response to Elizabeth's work.

However, it is very time-consuming (much more so than you might imagine) to coordinate these collaborative papers, and I need five of me because I would rather be focusing on implementing the technical systems to allow secure volunteer login, and finish the first phase public archive for cataloging all election data docs from the over 33,000 separate election entities in America, and I'm also working to prevent Utah from purchasing DRE voting machines which takes a lot of my time. I am trying to coordinate US Count Votes' groups of volunteers, including volunteer lawyers, volunteer data gatherers, volunteer statisticians and scientists, volunteer programmers, etc. Plus I am not very adept at running the business that a nonprofit 501-3c is. We've received enough donations for two Internet servers, and other office and system administration expenses, but haven't come close yet to the 2,000 $10/month subscribers we need in order to hire the full-time programmers, system admin and an office/business manager, election office liason, and eventually statistician, we need. So, five of me would be great. (I have an extra bedroom w/ own private bath if I could find a computer or business savy intern to come work with me in Park City, UT)

Anyway, the long and the short of it is that I am aiming to release a response to Elizabeth's work, and also the public announcement of our public election data archive, by end of May. However, I was really hoping "not" to have to continue to focus on the exit polls, so that the sole focus could be development of our system to analyze the 2004 election thoroughly in order to put in place the systems whereby we can immediately analyze the 2006 election after it is over, and ensure that the right candidates are sworn into office in January 07.

All of you on this list can help by educating all the democrats in the U.S. Congress and any candidates who run for the U.S. Congress in Nov 06, NOT TO CONCEDE after the election and to ASSUME IF THEY "LOST" THAT THE VOTES WERE MISCOUNTED BY ELECTRONIC VOTE COUNTS THAT ARE NOT BEING INDEPENDENTLY AUDITED. i.e. All of you on this list can work on convincing the democratic leaders that almost all of the computer scientists, mathemticians, and statisticians in America, ooops, I mean, all us conspiracy nuts, are convinced by the data that the democrats DID WIN IN NOVEMBER 2006, AND DO NOT HAVE TO GET OUT THE VOTE OR CHANGE THEIR MESSAGE. They have to do one thing alone, NOT CONCEDE, AND FIGHT FOR RECOUNTS, INVESTIGATIONS, AND EVEN, IF NECESSARY, RE-ELECTIONS IN NOV 06. All other issues will take care of themselves, if we begin routinely independently auditing our elections and ensure that the right candidates are sworn into office.

So, fight the battle for states and counties to adopt paper ballot Op Scan voting systems rather than electronic ballot DREs also. Send letters to your state and county election offices.
http://uscountvotes.net/scripts/lettergen/ltr_pick_addressee.php

And go give public testimony at your next meeting of County Commissioners or Council Members and tell them how they can save money and easily independently audit their own election results if they purchase Op Scan rather than DRE voting machines.

All of you could also help immensely by helping us raise the money US Count Votes needs to be able to hire the half dozen or so full-time staff it needs to complete its National Election Data Archive project so that we can provide all the court-worthy evidence that is needed right after the 2006 election. We have completed a database design that will work that has close to 1,000 tables in it, 14 tables for every state plus shared tables, but more tables will be needed as we expand to be able to analyze additional issues. We have designed it such that we can split each state's own election results analysis onto separate servers as it becomes necessary.

Anyway, not to worry, US Count Votes will eventually release a response to Elizabeth's work, but collaborative papers take time. We are also working on a new mathematical computer simulator that will be able to analyze the statistical footprint of exit poll results to determine if bias or vote embezzlement or miscounts are the more likely cause.

However, we are stretched very very thin due to lack of funding and staff that is needed to implement a huge but do-able project like our National Election Data Archive, so could use your help.

Thank You.

Kathy Dopp
http://ElectionArchive.org
[e-mail]

Kathy

[TruthIsAll]

2. My math question is this. How can there be a RBR when

in Mitofsky's FINAL 13660 NEP, Bush 2000 voters comprised 43% and Gore voters only 37%?

Now we know the 43% is IMPOSSIBLE; at best it was 39.82%.

You have seen the numbers: 50.456/122.26 million =41.26%
And 1.75 million (3.5%) of those 2000 Bush voters died.

50.456-1.75 = 48.70 and 48.70/122.26 = 39.8%

So the question is moot.

Lizzie should address this issue first:

HOW DO YOU JUSTIFY A FINAL EXIT POLL WHICH
1) USES AN IMPOSSIBLE WEIGHTING WHICH
2) REPUDIATES THE RBR AND
3) MATCHES THE FINAL VOTE

LEMON = LEMON

THEY CAN'T HAVE IT BOTH WAYS.

IF THE RBR THEORY IS TRUE,
THEN THERE MUST HAVE BEEN MORE GORE VOTERS INTERVIEWED, NOT FEWER.

BUT IF THAT IS THE CASE, THEN THE WEIGHTS MUST BE CHANGED..
AND THAT WILL SHOW THAT KERRY WON THE ELECTION...

THEY ARE CAUGHT IN A TRAP FROM WHICH THERE IS NO LOGICAL EXIT.

[davidgmills]

4. USCV needs to have this pointed out

As I understand that they are now beginning to waffle on their analysis.


By the way, did you do the same analysis for all the states with the 73,000 polled?

[TruthIsAll]

8. I did that analysis in November, using the state exit polls.

I have done many permutations of the state exit poll data, looking at deviations to MOEs, Z-scores and computed the probabilties for 16 states exceeding the Moe, and 35 states with Z> 1 (in the trillions).

I have compared the regional groupings of state EP data to the regional weights of the NEP.

RBR? Spread out all across the country?
Bull.

Bandwagon effect"
Bull.

43%/37% Bush/Gore 2000 voter?
Bull.

37/37/26 Party ID?
Bull.

51% of the female vote for Kerry?
Bull.

The last 613 respondents of 13660 respondents from 51-48 Kerry to 51-48 Bush?
Bull.

Thousands of machines turning Kerry votes to Bush?
No Bull.

Blackwell fixing the recount?
No Bull.

Mitofsky not releasing the FULL EP Raw data?
No Bull.

[mgr]

5. You seem to forget

That you don't know how they got the weights in the first place.

In multiple regression equations, the values used in weighing correspond to the correlation with each eigenvector. In following Bookchin's suggestion that allometry usually is the significant factor on the first eigenvector (precincts with larger numbers), the next likely variable is going to be party identification and it will be on the second eigenvector. That is going to be the strongest weight, and would likely explain 70 to 90% of the variability observed. In other words, it is almost sufficient to explain the phenomena observed. Other significant variables are likely gender and ethnicity, that will likely get your total explanatory power to 95%-97% with re-weighting. Income may finish it off. Each time you increase your explanation closer to 100%.

If party identification is the category with the strongest explanatory power, then what makes the Gore/Bush 2000 voter selection independent of that variable, since that voter selection would be almost entirely dependent on party identification? In other words, the variable has almost no explanatory value in the patterns observed, but would be altered by the re-weighting for party identification, and would have no reason to be corrected or readjusted for. In other words, it is what we used to call a 'howler'--an unintended impossibility that is only an artifact of re-weighting.

Mike

[Name removed]

6. Deleted message

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

[TruthIsAll]

7. Her own words are very relevant...READ THE REPORT

THE CAPS ARE MINE.

“Firstly, they indicate that WPE is a CONFOUNDED DEPENDENT measure, at best introducing NOISE into any analysis, but at worst creating artifacts that suggest that bias is concentrated in precincts with particular degrees of partisanship WHERE NO SUCH CONCENTRATION MAY EXIST”

“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 EXIT POLL ERROR"

“Mathematically, the observed pattern COULD ARISE FROM WIDESPREAD FRAUD AS WELL AS FROM WIDESPREAD RESPONSE BIAS;; differential SPOILAGE rates for Kerry votes, or “BALLOT STUFFING” of Bush votes, would produce results INDISTINGUISHABLE from “reluctant Bush responders”.

[anaxarchos]

15. Not just her words but the substantial concluding line of her paper...

How does this hurt again?

Instead of reluctant responders we could have had reluctant ballots (fraud) and produced exactly the same result on the exit polls.

[Melissa G]

10. Why are we getting posts from a banned DUer again? n/t

[davidgmills]

11. I don't know if it was Rick Brady or not,, but if it was

Although he is a Republican conservative he is getting lots of priase from Mystery Pollster and others for his input on the topic of exit poll discrepancies. He is much like mgr in his views. Liddie (the author of this article) even gave Rick some credit at the end of her paper for his criticism of her work.

Rick says he is not a freeper. He understands very much that if freepers were allowed on this board it would be ruined. All he wants to do is be involved in the debate.

He emails me all the time because he thinks I have an open mind when it comes to free speech. He knows we disagree. But he and I have had some healthy debates and both he and I genuinely want to be right about what caused the exit poll discrepancies.

[Melissa G]

23. You KNOW Perfectly Well He Is Banned! Here is the evidence!

How many other Fictions are you telling? Your previous thread and Link Below...

http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=203&topic_id=348218

davidgmills (582 posts) Thu Mar-24-05 05:11 PM
Original message
Criticism of Freeman and USCountVotes



The author, who is banned from DU, asked for me to post his article. In the interest of fairness on the issue, and for critical review, I thought I should.



http://www.stonescryout.org/archives/2005/03/critical_r...

davidgmills

[tommcintyre]

12. Another approach to handling the reluctant Bush responder hypothesis

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x363706#363731

[Peace Patriot]

14. tommcintyre, correct me if i'm wrong, but isn't it the skew to Bush in...

...the exit polls vs. the official official--the "red shift" that runs east to west, and involves mainly the battleground states (the states where the election was closest, and that Bush needed to win) that makes the official result so suspicious, and not merely or solely because Kerry won the exit polls?

...US Count Votes doesn't give 10 million to one odds against Bush winning in the official tally and losing in the exit polls, but rather against THE IMPOSSIBLE WAY that he wins the official tally. If there were a bias in the exit polls, it would be more or less evenly distributed across the landscape. That bias could be discovered. But there is no possible exit poll bias that can account for the way the official tally varies from the exit polls.

Am I not right? Even if Bush had won the exit poll totals, the official tally would STILL be suspect? (--suspicion of padding his victory, in that case).

This is a very important point, it seems to me. Please enlighten me if I am wrong about it.

[tommcintyre]

17. Yes, there are many "improbabilities" such as this that "favor" Bush.

Way too many to have occurred just by chance.

Another good example is the "Dixiecrat defection" in Florida. This was dismissed way to easily; and definitely deserves another look. Once again, they (Bushco) "created doubt", and were allowed to get away with it. The percent changes in many counties (from Gore to Kerry) are just too glaring. And the "son of the South" argument just doesn't hold up since Gore didn't even win his own state.

I think we need to revisit ALL of these obvious "anomalies" and start building sound hypotheses of HOW they pulled it off (see link below for more thoughts on this).

Further, I think it's OK to speculate (brainstorm) on how it was done, but we need to support these hypotheses with carefully constructed logic and facts (as I just did with the RBR hypothesis). Admittedly, sometimes it is much easier to do than other times (as it was with the RBR hypothesis), but it must be done in each case just the same. Otherwise, our credibility suffers, and the "weak links" can be used in an attempt to taint the whole case (create doubt).

The evidence the election was stolen is overwhelming, now we just must carefully put the case together.

EDIT for link:
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x362509

[Senator]

18. A bit of speculation "support"

A saw another item of support for my (election night-cursory view of FL returns) speculation on how it was done in Bob Fitrakis's latest fraud/coverup roundup article. Since I'm not one of those who's had their head in the numbers perhaps it was rehash for most of you, but to me it was another small revelation of what I think was perhaps the most important/tangible nefarious activity.

Fitrakis mentions a ChristoFascist college with 186 registrations from the same (business) address; out of which 46 votes were tallied. This dovetails with similar "ghost voter" findings and "overtallies" in other states.

But the crucial factor for numbers crunchers is the role that fraudulent registrations (voted or not) may well play in data analysis.

My hypothesis is that they artificially pumped up registered "R voters" so that pumped up tallies would not be as noticeable. But it also makes such numbers unreliable in general.

This was also buttressed by the anonymous election worker quoted by Fitrakis who revealed that ballot boxes (with log books?) were secreted in a separate area where she feared signatures could have been forged (her suggestion).

The reason to use this method is that it would only be (systematically) detectable by labor intensive comparison of actual people with tallied totals. Something not even the GAO is likely to undertake, even at a spot-check level.

But for our purposes I think it would be of great help to establish it as far as we can, even if only anecdotally because:

• it allows us to bootstrap onto their similar (but unfounded) accusations that are sure to be higher profile in the Euphemedia (among which are still some who are well aware of the neofascist tactic of "projection")
  
  
• it personifies the fraud by simply saying what people did, and doesn't require any math
  
  
• it is more "believable" in the sense that phony registrations can be accomplished over a long period of time and can be rationalized as a "lesser crime" (shenanigans, gamesmanship, etc)

---

[tommcintyre]

22. Hmm... that would make a very interesting smokescreen

"...artificially pumped up registered "R voters" so that pumped up tallies would not be as noticeable. But it also makes such numbers unreliable in general."

I have a feeling that Florida is a virtual "hothouse" of election/vote fraud; considering it's history and who is currently Governor. A study of Florida in this regard would very likely make it easier to detect similar fraud in other parts of the country since they seem to be "the old pros" at it.

Both parties seem to have been at it for quite awhile. See "Votescam" for more info. Here's a link to the first 95 pages of the book online:
http://www.constitution.org/vote/votescam-.txt

[anaxarchos]

16. Tom, you are so right....

The reason something as thin as BRH exists at all, with very little supporting evidence, is entirely political. Since the alternative (fraud) is so "inconceivable", we begin our trip out onto ever thinning ice. T'aint science at all. Pointing it out won't win you any friends though...

[tommcintyre]

19. It's time we help break 'em through that "ever thinning ice" ;)

And, please correct me if I'm wrong, but I think there is currently NO supporting evidence of the reluctant Bush responder. I say currently, because (as the report I sited points out), it would be possible only in the future to test this hypothesis. Then and ONLY then is it possible for such evidence to exist.

Your last observation is interesting. It had crossed my mind prior to posting this; but the stakes are just too high for this to be a consideration. The RBR hypothesis is being used in an attempt to discredit the possibility that major election fraud put Bush in the White-house, and gave him his so-called "mandate". In fact, it is the linchpin to Mitofsky's argument that the exit polls do NOT indicate fraud. Pull that linchpin, and their whole argument falls apart. I believe I did just that. ;)

My purpose is to vet the idea here,and then "take it on the road" (email it to Freeman et. al, post it at mystery pollster, etc. If it doesn't get more response in the three threads I just posted it in, I'll give it it's own thread to see if I can get more feedback. We will see.

[davidgmills]

20. See my posts on the other thread

You are on to something.

I don't think it can ever be tested. How are you going to use math or statistics to test what is in the mind of an unknown person?

The arguments I got to refute yours were really far fetched.

[tommcintyre]

21. LOL! Let's spread this virus of truth and watch their heads explode ;)

OK, I'm being overly dramatic. ;) But seriously, anyone who wants to hang onto any shred that they think logically, will have to conclude the RBR argument is fallacious.

[sunshinekathy]

24. US Count Votes' Rebuttal & Plea for Support of Election Data Archive

Bruce O'Dell, VP of US Count Votes, has created a simulation model of Lizzie's algorithm. It indicates that her model requires a "participation by partisanship" profile that is -totally- inconsistent with E/M's empirical data.

Lizzie's model requires -greater- exit poll participation in High-Kerry precincts than in High-Bush precincts. E/M's data shows quite the contrary.

If you reproduce Lizzie's results, and delve down into the data, you face a fundamental contradiction with E/M's "participation rate by precinct" empirical data.

Lizzie (and E/M's) "uniform response bias" hypothesis simply does not fit the facts. Period. There is simply no way to reconcile a "uniform response bias" hypothesis with E/M's "participation rates by partisanship" data.

US Count Votes will eventually publish the raw data and my simulator source code, to allow anyone to confirm my assertions.

For Bruce O'Dell's professional qualifications to simulate mathematical models, please refer to www.digitalagility.com/Odell_home_page.htm
35 years after writing his first Basic program, Bruce has reached the peak of his profession.

He is currently engaged as the lead architect at an enterprise security project at one of the Fortune 20 ("20" is not a typo). He is in charge of the technical aspects of security at one of the twenty top public corporations in America, and for the last 15 years, I have focused a major portion of my consulting practice on formal modeling of complex systems.

US Count Votes needs funding NOW to hire full-time programmers, system administrators, liason with the 33,000 election offices in America, statiticians, etc. if it is to complete its project to audit elections to have a system in place by Nov 06 to ensure that the right people are sworn into office in Jan 07. This cannot wait. The technical systems US Count Votes is building for its National Election Data Archive project are huge and cannot be built in six months prior to the Nov 06 election.

Every other issue that the Democrats care about will be resolved as soon as we audit elections to ensure that the right candidates are sworn into office in Jan 07. US Count Votes' needs 1/2 million dollars minimum for the staff to accomplish its database systems (which require about 1000 database tables - 15 per state and some shared tables). Please help out by donating or helping us to raise funds:

http://electionarchive.org/fairelection/donate.html

This project is the Democrats' best hope for restoring Democratic control of Congress in Jan. 07.

Best,

Kathy Dopp
http://electionarchive.org

[davidgmills]

25. Hope you get the funding Kathy

As a lawyer, I am considering doing what Land Shark is doing, taking on the voting machine manufacturers in court. But the cost in time, and especially money, is very worrisome. I am a solo practitioner and taking on an army of lawyers whose clients have bottomless pocketbooks could quicky bury me. That is how they usually play the game and I am sure this will be no exception. Plus, election law and contract law are not my fields; I am a personal injury lawyer.

I look forward to USCV's next paper. I gave your last paper to one of the judges here. Seeing his eyebrows rise when I told him its implications was something to see.