Breaking news in Exit Poll Analysis & Ohio Analysis - coming soon.

Several new discoveries have been made which will be released publicly in two papers Two new papers are coming from the National Election Data Archive very soon:


1. I've derived a new WPD function for analyzing exit poll discrepancy caused by (all of) exit poll response bias, vote miscounts, and random sampling error. This will permit automatic analysis of exit poll discrepancy data to determine if vote miscounts were likely, and even enable us to evaluate if such miscounts were likely to have changed the outcome of an election

2. The Ohio exit poll discrepancy analysis has almost been completed but we need atleast a week to work to do on it (see if it needs expansion, double and triple check it, and automate the analysis for various data sets and run them). In the future, such analysis can now be automated as long as we can obtain the exit poll data to do it. Our conclusions from conservatively analyzing the Ohio exit poll discrepancies show

virtually irrefutable evidence of vote miscounts and a pattern consistent, not only with vote miscounts, but with "outcome-altering" vote miscounts in Ohio. In other words, it looks like, if votes had been accurately counted, Kerry would have won Ohio's electoral votes, even despite all the voter disenfranchisement. (Now if we could only get candidates who have the guts not to concede until after helping us to obtain and analyze their election data - which even the DNC did not make any attempt to do in their sham Ohio analysis led by Donna Brazill.)

3. The Liddle/Lindeman Exit Poll Response Bias Measures that Mitofsky presented at the spring AAPOR conference in May and the fall ASA conference in October, and which the pollster Mark Blumenthal used to discredit the exit poll evidence, completely Distort Discrepancy Patterns caused by random sampling error and are thus useless for analyzing exit poll discrepancies. Mitofsky, Blumenthal et al were all grossly misled by this useless "bias measure". And I would still like to know who was responsible for inventing that invalid illogical hypothesis that ESI and Mitofsky used from June to October. This makes three wrong for three theories proposed by Mitofsky to explain the exit poll discrepancies by something other than vote miscounts.

Liddle and Lindeman, along with the Election Science Institute and Mitofsky have done American democracy a great disservice by their misuse of shoddy hypotheses and analyses to mislead Americans concerning the exit poll discrepancies in the 2004 presidential election. Even if their mistakes were innocently made, U.S. democracy deserves a great deal more care and double checking by those who release such invalid analyses before they publicly release them next time.

Please sign up to receive the update email as soon as the Ohio exit poll analysis and the exit poll analysis method paper are released by emailing:

election-subscribe@uscountvotes.org and replying to the confirmation email.

The upcoming Ohio exit poll discrepancy analysis will be scientifically sound and unimpeachible.

Congratulations on all the other great stuff going on in this Discussion Group on Elections.

Kathy Dopp
http://electionarchive.org

Just stay off small planes for the next few weeks, OK?

:kick:

from the Globe.

The media needs to be dragged kicking and screaming into this area, vote fraud, that they are so afraid of.

:-)

It's the difference between blaming those who voted and those who counted their votes.

Vote fraud is manipulating votes

Voter fraud is voters voting illegally. Research shows that voter fraud is not a prevalent problem and yet the NASED is focused on it and many state election officials are making a big push to detect and prevent voter fraud. They seem to agree that the election systems needs to be protected from voters, rather than from insiders - not very logical as the biggest threat to any system is always from insiders.

of my work, and I expect Lindeman will too.

Elizabeth Liddle.


On edit, to clarify, as this thing already has seven votes: while clearly it is impossible for me to assess the utility of the new measure proposed here, as it is apparently still under wraps, I would certainly defend my own, and do not agree that it is "useless for analyzing exit poll discrepancies". Indeed I would argue that it is very much more useful than the "WPE" or "WPD".

I find it somewhat curious that Ms Dopp would characterise my own proposed measure in such terms, as in previous papers she and her fellow authors have insisted that the measure that they used was no different to mine, and indeed, have implied that I merely derived it from their own work (or, alternatively, that I derived it independently but in a way that "obfuscated" it).

I would agree with Ms Dopp that it is distorted by sampling error at extremes of partisanship, where sample sizes are small; because of this, Lindeman and I have, over the past few months, developed an approach that minimises the distortion. However, if Ms Dopp's measure proves to be superior to the one proposed by Lindeman and myself, I shall be the first to give her full credit, and I am sure that Lindeman will join me. However, I certainly deny the charge of "shoddy" work. I also simply reject Ms Dopp's claim that the hypotheses she refers to as "illogical" are, in fact illogical, and I do not believe that she has demonstrated that they are.

However, I would be extremely suspicious of the validity of any analysis based on the exit poll data for a single state that claimed to show "virtually irrefutable evidence of vote miscounts and a pattern consistent, not only with vote miscounts, but with "outcome-altering" vote miscounts in Ohio". I simply do not believe that the 49 precinct data points are adequate to demonstrate any such thing.

But until the paper is published we cannot tell. And until it is, it is scarcely "breaking news".

that I myself pointed out the problem Kathy correctly* reports with my proposed measure, by email to her on Christmas Eve. I also explained that we had spent several months working on a development of the measure (dubbed "tau prime") that substantially compensates for the problem. Indeed all analyses conducted using the measure have been "double checked" using "tau prime".

Kathy's allegations seem inappropriate in these circumstances.



*up to a point

that The Nation, Salon, and Mother Jones do not officially abide by evidence of multiple election thefts. They won't touch it. Miller said he thought it was because that to address it would open up the immense damage to the country that it has caused. Denial is safer.

Malloy was dumbfounded.

Malloy told MCM that he should try very hard to get onto Al Franken's show, because Franken is another one who poo-poos the whole idea of stolen elections. Hope he will.

"The Most Valuable Progressives of 2005," published in The Nation this week.

Apparently The Nation isn't staying away from the subject, though they may not be editorializing on it.

Hope he gets with the program very soon.

Malloy kept asking Mark to stick around -- until after the break-- its clear that Malloy gives this issue the proper place it desreves-- you go MIKE

Is there a mathematical denotation to "virtually irrefutable"? I've only heard the expression in advertising circles:

dispense with the hype (potatoes)-- give us the meat-- please

It isn't my favorite phrase, but one which Ron Baiman and Bob Klauber who helps edit our papers, both like.


To me, nowadays, it connotes computer "virtual reality".

That figures. Here we have virtual breaking news of two virtual papers, containing "virtually irrefutable" evidence that Kerry won Ohio.

I've never been that keen on CGI. I'd find The Stolen Election so much more convincing without the SFXit poll stuff.

I am looking forward to the new studies.

Kathy Dopp can't find the time actually to support any of her assertions, but does find the time to head-butt the people she is disagreeing with. (Is anyone reminded of -- oh, never mind.) There is nothing here I could rebut, since there is nothing here.

(Nevertheless, I would be most grateful if DU members would find the time to besiege my dean with complaints that I am undermining American democracy by critiquing arguments that Kerry won the 2004 election. Outside ER, it is hard to depict this work as relevant and edgy. Let me state clearly that I try only to critique the bad arguments.)

Maybe I could mark time by rebutting some other stuff that I haven't rebutted yet. Here's one, footnote 57 in the timeline paper: "If ESI's analysis had been valid, then candidates would always win second terms unless there were vote miscounts." Oh, my. Do I have to rebut that? I'm tempted to say that Dopp just made that up, but I am sure she thinks it has some foundation in ESI's work. It's just that... you guessed it, she didn't find the time to cite anything in the ESI report that would support that inference. So she might as well just have made it up.

Hmm, let me try one: "If Dopp's critique had been valid, then the National Election Data Archive would be useless, because it would be impossible to use data from previous elections to analyze the current election." Actually, I don't see why mine isn't true.

My response to Dopp's mathematical proof remains available at http://inside.bard.edu/~lindeman/doppresponse.pdf , and I welcome serious comments, questions, and criticisms (need I add, from people who have read both papers). As far as I know, I am the only person in the world who has bothered to write up a critique of Dopp's mathematical proof. Sadly for Dopp, that isn't because everyone else agrees with it. Some people will agree with Dopp reflexively because they like her conclusions. Heck, I sorta like some of Dopp's conclusions myself. But that isn't how science is supposed to work.

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

"Personal attacks", Kathy, are attacks against the person, not the argument.

The examples you give are attacks on your arguments.

Attacks on a person mean attacks on their personal motivations and attributes, such as accusing someone of deliberately lying or misleading.

They are called "ad hominem" attacks because they are against the person not against the argument.

I do not accuse you of lying or misleading. I do consider your arguments faulty. If I did the first, I would be making a personal attack. But attacking arguments is the stuff of legitimate debate.

If you want to engage in academic debate, do not indulge in the first, and be prepared for the second.

I can't really read much of this, so I encourage anyone who has questions about it to direct them to me.

Actually, I was not aware that Dopp had posted her response to my critique. I thought she had accepted the advice of her colleagues not to, and I haven't seen an announcement on the Dopp website. Again, if anyone has questions about Dopp's response, I encourage them to direct them to me.

I welcome serious discussion on the very serious topic of the integrity of the 2004 election and other elections, past and future. Dopp's conviction that she knows the "facts" and the "true shape" of things does not in itself materially advance the discussion.

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

"Truth twisting and using mathematics wrongly and carelessly to do it, is highly unethical", whether done deliberately or through lack of diligence.

However, neither your original Logic Proof, nor your new critique persuades me that anyone involved in the analyses you claim to demonstrate to be illogical is misusing mathematics.

The ESI study, according to their website, is undergoing peer-review. I suggest you submit your Logic Proof to peer-review also. This is the way academic disputes are resolved.

Publicly accusing fellow analysts of incompetence, lack of diligence, or dishonesty is not.




(edited for typo)

Sorry to be the one (well, among the few) to break it to you, but your colleagues are letting you down.

If you really want to play Credentials Smackdown -- if you really want to argue that your MA in mathematics trumps my Ph.D. in political science when we are talking about polling data -- then I would suggest that Fritz Scheuren has to be the walkover winner, with his Ph.D. in statistics (dissertation on "Topics in Multivariate Finite Population Sampling and Analysis") and everything he has done since. You should be worried about this. You should worry a bit less about whether my mind is open enough, and a bit more about whether yours is, IMHO. I can certainly imagine the president of the American Statistical Association being wrong, but it seems like a Big Stretch to suppose that he isn't qualified to do data analysis.

Yes, you argue that the analysis is moot because of faulty logic. Do you really want to argue that ESI's argument is bunk because "showing one valid counterexample... is sufficient to disprove any inference"? OK, then I am here to tell you that smoking doesn't cause cancer, 'coz just look at all the people who haven't died from smoking! My counterexamples aren't even hypothetical! Will it be clearer if I add some symbolic logic notation?

eomer actually had a serious discussion with Febble and me about substance, and I worked to incorporate his insights into my response to you. I think you should read it again, and then get back to the actual debate.

You are losing this argument, messily. Here is part of the mess, from your latest PDF response:

OK, let's fact-check this, briefly. Was there ever a thread entitled "The ESI/Mitofsky analysis is Illogical Bunk"? Does anyone here think that I or anyone else got a thread frozen on the grounds that I/we dislike the words "bunk" and "illogical"? Does anyone suppose that I "admitted" such a thing? (I do "admit" that I have alerted on some of your posts.) Does anyone think that your posts here are limited to the purpose of addressing inaccuracies posted by O'Dell, Lindeman and Liddle? (It's a matter of record, by the way, that you joined DU in December 2004, Febble joined in April 2005, and I joined in May.)

Kathy may have "discovered" on Christmas day that the measure I proposed ("tau") "reduces the size of random sampling error in precincts which are 50/50, and increases it at the end points".

In fact I emailed Kathy on Christmas Eve, in response to a question from her about the relationship of the "tau" to precinct partisanship as measured by vote count, and, inter alia, wrote:



I have simply had enough of Kathy Dopp's repeated attacks on my moral and intellectual integrity. If she herself does not have the integrity to at least minimally credit me for pointing out a property of my measure that she then claims renders it worthless (it certainly does not) without producing so much as an equation in support of her assertion, then, frankly, she has no business impugning mine.

Sub-thread removed by moderator. Click here to review the message board rules.

Finally!!! The "new WPD function" that'll make the sElection 2004 theft even more obvious than the sElection 2k USSC 5 to 4 decision. :sarcasm: Ever heard of "analysis paralysis"?

A single WPD function that handles:

1. exit poll response bias

2. vote miscounts

3. sampling error

and any combination of the above.

Its use is really to study all the patterns and shapes caused by the above factors so that we can see what exactly is and is not consistent with existing exit poll data.

I worked very hard to come up with it, but it will seem very easy when I finish writing it up with the help and advice of other members of the National Election Data Archive. I've received some suggestions on improvements to the random sampling error section of it I need to evaluate and possibly implement tomorrow, plus write up the derivation in a way that is logically easiest to follow, which I haven't done yet.

It is exciting and good news. We'll repair our democracy yet.

Kathy Dopp
http://electionarchive.org

I'm looking forward to checking this out in a week or 2-- thanks.

I was going to cut and paste the fla here, but the equation stuff doesn't paste across.

The derivation of the new WPD function was so easy (although you wouldn't know it from the multitudinous hours it took me) that it is going to be embarassing that we didn't derive this function a LONG time ago.

Right now we're having discussions because Josh thinks that we need to use a different function to produce random sampling error that gives smaller amounts of sampling error than the one I proposed. So my current proposal is more conservative, and Josh hasn't convinced me yet, but I'm openminded to the fact that he may have something we should use instead - but have not understood how it is logically correct to do so yet.

Also, we're waiting for the two statisticians who offered to help automate the analysis using the WPD function to find time to do so in their busy schedules, so that may take a week, and then I've got a TON of work to do to finish writing the "Lessons Learned - How Not to Analyze Exit Poll Discrepancy" section and other sections of both papers, then we've got to use whatever tools the two guys create to recreate all the charts in our papers, then double and triple check our input data. Phew.

I can't wait to get this over with! Bigtime. I never wanted to analyze exit poll data at all, but felt been pushed into it everytime the other side released another bunk invalid analysis because I can't stand to see Americans misled by all the wrongful analyses and invalid conclusions that have been proferred and reported on - such misinformation needed to be countered by scientific fact.

cough it up.

Just put into a pdf and post it on your site. Or a jpg.

If you are going to trash our measure, the least you can do is let us see yours. I want to use it. I've been trying to find this thing for about 11 months. I think ours works pretty well, but if yours is better - bring it on.

Let's see whose Fancy Function works the best. If it's yours, I might even donate to NEDA.

in a fair election/count
http://www.flcv.com/ohiosum.html

more fraud in the non-recount:
http://www.flcv.com/greenrc.html

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

Looking forward to it. . . .

While it is ok to criticize a poster's public works, the moderators feel that this thread has gotten too personal.