A very late update on voting technology

[Kevin_Lowe]

Way back in 2004, there was a certain amount of discussion about the knowable legitimacy of the presidential race, given the exit poll anomalies and the various demonstrated means by which vote tabulating machines could have been compromised.

The best single thread was probably this one, and it's the most relevant to the new news:

http://forums.randi.org/showthread.p...0&pagenumber=1

Anyway, immediately after the election a certain amount of attention was focused on Diebold-made optical vote scanning machines because what looked like a large and inexplicable swing to Bush in part of Florida correlated tightly with the presence of optical scanning machines. It turned out, though, that there was a perfectly innocent explanation with the legs to cover the ground. Type of voting machine and voting patterns both correlated with geographical area.

Also, as you would see if you slogged through the thread above, after some in-depth discussion with a very valuable contributor who worked with the machines in question, we thought we had ruled out just about all of the plausible means by which Diebold's optical scanning precincts could have been rigged at the grass-roots level (assuming competent and honest officials).

It turns out that our mistake was that we hadn't considered sufficiently implausible avenues for naughtiness.

I would have been shovelled on relentlessly at the time if I had asked "How do we know that the "ballot box", the memory card that stores the votes, does not have undocumented live code on board that can control the scanning machine and make it do whatever it wants?". However it turns out that is precisely how the Accu-Vote machine works. Switch a card, or tamper with the machine that programs the cards (both apparently doable by amateurs) and you own the machine that records the votes.

BBV discussion thread, with link to a formal report in pdf format, here:

http://www.bbvforums.org/cgi-bin/for...1954/6805.html

The sheer degree of vulnerability and sloppiness in Diebold's work amazes me. Apparently branches of the US government are still buying the damn things.

[davefoc]

California doesn't seem to be too happy with Diebold.

http://www.fcw.com/article89758-08-01-05-Web

[The Central Scrutinizer]

Quote:

Originally posted by Kevin_Lowe
Way back in 2004, there was a certain amount of discussion about the knowable legitimacy of the presidential race, given the exit poll anomalies and the various demonstrated means by which vote tabulating machines could have been compromised.

The best single thread was probably this one, and it's the most relevant to the new news:

http://forums.randi.org/showthread.p...0&pagenumber=1

Anyway, immediately after the election a certain amount of attention was focused on Diebold-made optical vote scanning machines because what looked like a large and inexplicable swing to Bush in part of Florida correlated tightly with the presence of optical scanning machines. It turned out, though, that there was a perfectly innocent explanation with the legs to cover the ground. Type of voting machine and voting patterns both correlated with geographical area.

Also, as you would see if you slogged through the thread above, after some in-depth discussion with a very valuable contributor who worked with the machines in question, we thought we had ruled out just about all of the plausible means by which Diebold's optical scanning precincts could have been rigged at the grass-roots level (assuming competent and honest officials).

It turns out that our mistake was that we hadn't considered sufficiently implausible avenues for naughtiness.

I would have been shovelled on relentlessly at the time if I had asked "How do we know that the "ballot box", the memory card that stores the votes, does not have undocumented live code on board that can control the scanning machine and make it do whatever it wants?". However it turns out that is precisely how the Accu-Vote machine works. Switch a card, or tamper with the machine that programs the cards (both apparently doable by amateurs) and you own the machine that records the votes.

BBV discussion thread, with link to a formal report in pdf format, here:

http://www.bbvforums.org/cgi-bin/for...1954/6805.html

The sheer degree of vulnerability and sloppiness in Diebold's work amazes me. Apparently branches of the US government are still buying the damn things.

Bush won. There was no fraud. Get over it.

[corplinx]

And you keep linking the BBV website despite the flagrant woo they showed during that time. Any good points BBV might have about voting systems will inevitably get overshadowed by their conspiracy theories and woo.

[Leif Roar]

Quote:

Originally posted by The Central Scrutinizer
Bush won. There was no fraud.

True. True. However, there are still serious problems of trustability when it comes to electronic voting machines. The important question isn't if such machines has been used to commit voting fraud, but whether they can be.

This issue isn't about Bush.

[Kevin_Lowe]

Quote:

Originally posted by corplinx
Any documented evidence BBV might have about voting systems will inevitably be ignored by Corplinx in favour of empty chatter about conspiracy theories and woo.

I fixed your post.

As Leif said, the real issue here is whether the current electoral machinery of the USA is wide open to tampering.

[Ed]

Quote:

Originally posted by Kevin_Lowe
I fixed your post.

As Leif said, the real issue here is whether the current electoral machinery of the USA is wide open to tampering.

I suspect that with 25,000 voting places (or whatever the number) the question is not whether or not tampering will occur, it probably dies and will no matter what.

The real issue is the nature and extent of the tampering. That is to say that barring a vast conspiracy, is the tampering basically random, that is to say, does it wash out?

A very big set of issues that sort of come up when the TV is short on sound bites and then vanishes again are the questions surrounding absentee ballots and old farts. It seems that when a person is incopetant (which is a broad definition) the individual caregiver often "votes" for that person. God knows how many votes are cast with such "help" and what the nature of them is.

I think that the best we can hope for is self correcting flailing.

[Kevin_Lowe]

Quote:

Originally posted by Ed
I suspect that with 25,000 voting places (or whatever the number) the question is not whether or not tampering will occur, it probably dies and will no matter what.

The real issue is the nature and extent of the tampering. That is to say that barring a vast conspiracy, is the tampering basically random, that is to say, does it wash out?

Two things are going on there.

Firstly, it doesn't take a vast conspiracy when your votes are counted by centrally produced and programmed black boxes. There's no reason why the head programmer of ES&S couldn't decide to write and hide an altered file on their boxes to give Yogi Bear 90 billion votes for president. It would not be that much harder to arrange for a candidate to get a 51/49 victory if they would otherwise lose.

Secondly, the big hitters in the current crop of black box companies, Diebold and ES&S, have tight links with each other and the Republican party. They have also done very well out of the Republican-passed HAVA. So it's not unreasonable to be concerned that the Republican party has a head start on fudging.

Quote:

A very big set of issues that sort of come up when the TV is short on sound bites and then vanishes again are the questions surrounding absentee ballots and old farts. It seems that when a person is incopetant (which is a broad definition) the individual caregiver often "votes" for that person. God knows how many votes are cast with such "help" and what the nature of them is.

I think that the best we can hope for is self correcting flailing.

Back when vote rigging had to be done by hand, the old-fashioned way, I don't think it was too unreasonable to assume that it would mostly come out in the wash. Nowadays it is technically possible for individuals or small groups to rig elections automatically, remotely and on a vast scale, so I think we have to be specifically vigilant to make sure nobody tries it.

As I recall there's strong evidence it has already been done in Venezuela, although in Venezuela they took the exit polls seriously and fixed the problem. Bear in mind that a raft of top US statisticians published a paper stating that the exit poll anomalies in 2004 were incredibly unlikely to be chance and that none of the proposed (innocent) explanations were credible.

I still say that paper ballots, hand-counted by many eyes at the polling station, with results immediately published at the polling station level, is a system that is far simpler, cheaper and harder to corrupt than any electronic system.

The JREF challenge is an excellent model for a voting system. Absolutely no scope for shenanigans, nothing going on behind closed doors or inside proprietary boxes.

[Ed]

Quote:

Originally posted by Kevin_Lowe
I still say that paper ballots, hand-counted by many eyes at the polling station, with results immediately published at the polling station level, is a system that is far simpler, cheaper and harder to corrupt than any electronic system.

I agree. We still use those mechanical things where you pull a lever. The folks at the polling place told me that they never had a problem but we don't really have that many voters. I mean, I personally vote 7 or 8 times but the number of unduplicated voters is small, if you get my drift.

I get nervous about centralized anything since one f*uck up that would screw up the votes in a tiny town could now do the same for a vast swath of the electorate.

[varwoche]

Quote:

Originally posted by The Central Scrutinizer
Bush won. There was no fraud. Get over it.

BORING.

[The Central Scrutinizer]

Quote:

Originally posted by varwoche
BORING.

Reality is often boring. That is why we have movies!

[Beth]

Quote:

Originally posted by Kevin_Lowe
Bear in mind that a raft of top US statisticians published a paper stating that the exit poll anomalies in 2004 were incredibly unlikely to be chance and that none of the proposed (innocent) explanations were credible.

Do you have a link or cite for this? I'd like to read the paper.

Beth

[Kevin_Lowe]

Quote:

Originally posted by Beth
Do you have a link or cite for this? I'd like to read the paper.

Press release:

http://uscountvotes.org/index.php?op...d=85&Itemid=41

Executive summary by Dr Mitteldorf of Temple University Statistics Department:

http://electionarchive.org/ucvAnalys...s_summary.pdf.

I strongly encourage everyone to read the executive summary, it's short, professionally presented and totally unambiguous.

Particularly striking is the table on page five, showing the mean discrepancy from the exit poll results for each kind of voting apparatus. In precincts where people write their vote down themselves on a piece of paper, to be counted by hand by other human beings, the exit polls were highly accurate. As soon as machines of any kind get in the mix though, particularly in Republican strongholds, funny and incredibly unlikely things start to happen.

Things that are incredibly unlikely to be chance start to happen, anyway.

Corplinx, please do read it and tell us what you think. It's only five pages so I'm sure you have the attention span. It has pretty pictures and it uses very few hard words.

An interesting aspect of their numbers is that if cheating by both major parties were equally prevalent and effective you would expect to see improbable outcomes favouring Kerry in Democrat strongholds, but the noticeable effects all went in favour of Bush. It would seem based on the exit polls that if the Democrats are cheating too, it's to a much lesser extent or much less effective.

[Zep]

That last reference with the URL fixed to work...

Quote:

Originally posted by Kevin_Lowe
http://electionarchive.org/ucvAnalys...ls_summary.pdf

[Cylinder]

Quote:

Originally posted by Kevin_Lowe
Apparently branches of the US government are still buying the damn things.

Why is the US government buying voting machines? The counties in the particular states (and even precincts in very populous counties) are in charge of hardware purchasing for the most part.

[Rob Lister]

Quote:

Originally posted by Kevin_Lowe
Press release:

http://uscountvotes.org/index.php?op...d=85&Itemid=41

Executive summary by Dr Mitteldorf of Temple University Statistics Department:

http://electionarchive.org/ucvAnalys...s_summary.pdf.

I strongly encourage everyone to read the executive summary, it's short, professionally presented and totally unambiguous.

Would you trust a paper reaching the opposite conclusion if it were written by Karl Rove?

Not to disparage Dr Mitteldorf but the only thing that report proves is that he has an opinion. But is it an unbiased opinion? If he were cherry picking his data (intentionally or not) or out right pulling the statistical wool over your eyes would you even know it? Does he have a personal stake in his opinion? An emotional stake? Is he so incredibly biased as to fit perfectly the ultimate and quentisential liberal stereotype.


Here's Dr Mitteldorf's bio, in his own words. You be the judge?

Quote:

I'm a peace activist and an environmentalist, a long-time board treasurer at Pennsylvania's Clean Air Council before a dispute over financial integrity ended that relationship. I fight the deception, bribery, political and legal manipulation that keeps Big Tobacco profitable, as a past president of the Coalition for a Tobacco-Free Pennsylvania and an active member of TEACH.

Since the 2004 presidential election, I've been part of a net-organized coalition of statisticians who analyze voting patterns for evidence of errors or fraud. The National Election Data Archive has recently released my analysis of the Edison/Mitofsky pollsters' disavowal of their own exit poll, and In These Times carried a more popular analysis.

For over twenty years, I've been teaching a weekly yoga class through Mount Airy Learning Tree, an education coop named for my neighborhood in Philadelphia. This is my own brand of public health activism, as well as a way to keep my own spiritual practice alive.

I am active in Philadelphia's amateur music community, playing piano in chamber ensemble, French horn and oboe in the Chestnut Hill Community Orchestra and the Olney Symphony. Summers, my family and I enjoy the music camp in Quebec operated by CAMMAC and a Pennsylvania music camp called Summertrios. I'm an avid member of the Amateur Chamber Musician Players . Through Project Gutenberg, I am helping to make classical sheet music freely available on line in a form that can be edited and printed.

http://mathforum.org/~josh/

[Beth]

Quote:

Originally posted by Rob Lister
Would you trust a paper reaching the opposite conclusion if it were written by Karl Rove?
Not to disparage Dr Mitteldorf but the only thing that report proves is that he has an opinion. But is it an unbiased opinion? If he were cherry picking his data (intentionally or not) or out right pulling the statistical wool over your eyes would you even know it? Does he have a personal stake in his opinion? An emotional stake? Is he so incredibly biased as to fit perfectly the ultimate and quentisential liberal stereotype.

Doesn't matter what his background is. Likewise if Karl Rove produced a report reaching the opposite conclusion. That's the point of doing an analysis. If you disagree with the conclusion, then give a reason why the analysis is flawed. The background of the researcher may point to bias, but it doesn't invalidate the analysis or the conclusion.

BTW, this isn't the first election that the electronic voting machines have been implicated in suspicious results. Personally, I hope US COUNT VOTES is successful in their endeavor to ensure the accuracy of future election results.

Beth

[Rob Lister]

Quote:

Originally posted by Beth
Doesn't matter what his background is. Likewise if Karl Rove produced a report reaching the opposite conclusion. That's the point of doing an analysis. If you disagree with the conclusion, then give a reason why the analysis is flawed.

My reasons were expressed as questions. The answers to all of which are "I have no way of knowing".

IOW, I can neither agree nor disagree with his "analysis". His work is not peer reviewed, his data set is not available, his methods are not annotated, his work contridicts the findings of the original 'in place' organization, and he demonstrates incredible bias.

I'd say those are all good reasons to give his analysis little or no weight.

[kalen]

On the first page, Mitteldorf says that Edison/Mitofsky "have not released the raw data that would facilitate independent corroboration of their analysis." But seeing as how the poll was "commissioned by major TV networks and print news services," don't they (the papers and networks) own the data and have a say if it can be released or not? Why are these data under lock and key?

The code for the programs running on the electronic voting machines was a big secret, too. Before the election, one could make a case for security via obscurity, but now? The election's over.

[Luke T.]

Will somebody please just tell me how to beat the electronic slot machines? Then I just might go to the next TAM.

[Luke T.]

Quote:

Originally posted by Rob Lister
My reasons were expressed as questions. The answers to all of which are "I have no way of knowing".

IOW, I can neither agree nor disagree with his "analysis". His work is not peer reviewed, his data set is not available, his methods are not annotated, his work contridicts the findings of the original 'in place' organization, and he demonstrates incredible bias.

I'd say those are all good reasons to give his analysis little or no weight.

One of the co-author's of the report was Steven F. Freeman.

A critical review of the report

Quote:

Dr. Steven F. Freeman, visiting University of Pennsylvania (UPenn) professor is not an “expert� on exit polls or the 2004 Presidential exit poll discrepancies as suggested by this UPenn press release. In fact, his paper, The Unexplained Exit Poll Discrepancy, is highly flawed. His argument that “in general, exit poll data are sound� fails having suppressed evidence and the conclusion that “it is impossible that the discrepancies between predicted and actual vote counts in� Ohio, Florida, and Pennsylvania was not substantiated statistically. Nevertheless, Dr. Freeman is right in concluding that explanations of the discrepancy to date are inadequate and Edison/Mitofsky should address the concerns of US Count Votes in subsequent analysis of their data.

Dr. Freeman wrote a book based on his research that is due out in a couple of months and has a couple of working papers in progress. If The Unexplained Exit Poll Discrepancy is any indicator of the quality of research included in these forthcoming works, I suggest that his publishers take a closer look at the manuscripts.

Some interesting reading in the critique.

[Beth]

Quote:

Originally posted by Rob Lister
My reasons were expressed as questions. The answers to all of which are "I have no way of knowing".

IOW, I can neither agree nor disagree with his "analysis". His work is not peer reviewed, his data set is not available, his methods are not annotated, his work contridicts the findings of the original 'in place' organization, and he demonstrates incredible bias.

I'd say those are all good reasons to give his analysis little or no weight.

It's too soon to see peer-reviewed papers on this data. The summary gives some general information about what data is being used and how it was analyzed. I can't say whether or not the full paper contains the annotations you're looking for, but it should either provide the data or a cite for the data (he may not have permission to publish it).

What's published in the summary certainly shows that bias occurred and that the bias was in Bush's favor, but that's already well-established. It's whether the bias is from the exit poll results or the official results that's in question.

I'd like to see the full paper, but I didn't see a link for that anywhere. I've actually been thinking of doing an analysis for myself on the OHIO results, but haven't yet acquired all the data I'll need. I hoping I can get my University library to purchase it, but if someone else has already done the analysis I'm thinking of trying, I'd just as soon read about their results. It would sure save me a lot of time.

Beth

[Kevin_Lowe]

Quote:

Originally posted by Luke T.
One of the co-author's of the report was Steven F. Freeman.

A critical review of the report

Some interesting reading in the critique.

Luke, you have referred to this as "a critical review of the report", which is bound to mislead some people. The report critiqued in the link you give is not the same report I linked to. It has an author in common and not much else. People are bound to read your post and get the impression that an interesting critique has been made of the article I linked to, and that is just false.

I can't download the critiquing pdf for some reason, so I can't say whether its arguments are sound, but I'll try again later.

To Rob: What you have going there is a combination of an ad hominem argument (and a weak one too, given there is nothing discreditable in his posted biography) and an argument from ignorance.

To Kalen: The official reason for keeping the data secret is commercial confidentiality. Yes, you would think that the major news networks would release the data. No, they have not done so, despite at least one request from an interested US Congressperson (Conyers). No, I have no idea why although I have read all kinds of more or less exited speculation.

To Cylinder: I am unfamiliar with the US system, but I still call local government part of the US government. Is that not correct for some reason? In any case, even if only the US federal government is properly referred to as "the US government" in the USA, a large chunk of the money to buy these things with comes from the Help America Vote Act, which was a piece of federal pork mandating (if I recall correctly) a complete changeover to electronic voting by 2006. Take a wild guess why I think the current government thinks this is a good use of federal funds.

Edited to close tag

[WildCat]

Quote:

Originally posted by Kevin_Lowe
Take a wild guess why I think the current government thinks this is a good use of federal funds.

So they can STEAL THE ELECTION!!! Did I win?

[WildCat]

Kevin, the move to electronic voting was in response to Dems crying about the punch cards in the 2000 election. Oh, wait, that's all part of the Vast Right-Wing ConspiracyTM , isn't it?

[Rob Lister]

Quote:

Originally posted by Kevin_Lowe
To Rob: What you have going there is a combination of an ad hominem argument (and a weak one too, given there is nothing discreditable in his posted biography) and an argument from ignorance.

Kevin, I was being as gracious as I could. There are times, albeit few, when an inferred ad hom is impossible to avoid, even if it is not implied.

Points with which I think we can agree:

The paper is not peer reviewed. That cannot be disputed.

The author does not have the complete data set. That cannot be disputed.

The author has not made his available data set available. That cannot be disputed.

The author has not made his methods available. That cannot be disputed.

The author's conclusions contradict the stated conclusions of those that had the full data set and were in a much better position to know. That cannot be disputed.

The author shows a past and present (very) strong bias for the implied side he is presenting. That cannot be disputed.

You tell me, Kevin, how much weight should I lend this 'analysis'. Currently, I lend it exactly none.

[Kevin_Lowe]

Quote:

Originally posted by WildCat
Kevin, the move to electronic voting was in response to Dems crying about the punch cards in the 2000 election. Oh, wait, that's all part of the Vast Right-Wing ConspiracyTM , isn't it?

Do you think you could unpack that argument a bit for us, to show us what you think follows from the claim that HAVA was "in response to" the 2000 punch-card issues?

[Rob Lister]

Quote:

Originally posted by Beth
It's too soon to see peer-reviewed papers on this data...

If you think that then all I can surmise is that you suffer from the beaten wife syndrome. Lots of excuses, few results. The paper was...in layman's terms, pathetic. No better than a homeopathic explanation of dilute solutions.

You're a skeptic. Act like one.

That doesn't mean changing parties. That means picking the voices in the party of your choice that you find worthy of amplifying. You grow your party that way.

nuff said.

[WildCat]

Quote:

Originally posted by Kevin_Lowe
Do you think you could unpack that argument a bit for us, to show us what you think follows from the claim that HAVA was "in response to" the 2000 punch-card issues?

It was sponsored by Rep. John Conyers, a Michigan Democrat, and co-sponsored by many other House Democrats:

Quote:

Status: Introduced (By Rep. John Conyers [D-MI])
Introduced: Feb 2, 2005
Last Action: Feb 2, 2005: Referred to House Judiciary
Sponsor: Rep. John Conyers [D-MI]
Cosponsors: hide cosponsors
Rep. Corrine Brown [D-FL]
Rep. Sherrod Brown [D-OH]
Rep. George Butterfield [D-NC]
Rep. Lois Capps [D-CA]
Rep. Julia Carson [D-IN]
Rep. Ed Case [D-HI]
Rep. William Clay [D-MO]
Rep. Elijah Cummings [D-MD]
Rep. Rosa DeLauro [D-CT]
Rep. Lloyd Doggett [D-TX]
Rep. Chaka Fattah [D-PA]
Rep. Barney Frank [D-MA]
Rep. Raul Grijalva [D-AZ]
Rep. Maurice Hinchey [D-NY]
Rep. Jesse Jackson [D-IL]
Rep. Sheila Jackson-Lee [D-TX]
Rep. Stephanie Jones [D-OH]
Rep. Marcy Kaptur [D-OH]
Rep. Carolyn Kilpatrick [D-MI]
Rep. Dennis Kucinich [D-OH]
Rep. Barbara Lee [D-CA]
Rep. Zoe Lofgren [D-CA]
Rep. Betty McCollum [D-MN]
Rep. James McDermott [D-WA]
Rep. James McGovern [D-MA]
Rep. George Miller [D-CA]
Rep. James Moran [D-VA]
Rep. Eleanor Norton [D-DC]
Rep. James Oberstar [D-MN]
Rep. John Olver [D-MA]
Rep. Major Owens [D-NY]
Rep. Frank Pallone [D-NJ]
Rep. Donald Payne [D-NJ]
Rep. Charles Rangel [D-NY]
Rep. Janice Schakowsky [D-IL]
Rep. Jose Serrano [D-NY]
Rep. Hilda Solis [D-CA]
Rep. Fortney Stark [D-CA]
Rep. Bennie Thompson [D-MS]
Rep. John Tierney [D-MA]
Rep. Edolphus Towns [D-NY]
Rep. Christopher Van Hollen [D-MD]
Rep. Maxine Waters [D-CA]
Rep. Henry Waxman [D-CA]
Rep. Anthony Weiner [D-NY]
Rep. Lynn Woolsey [D-CA]

Yep, looks like a Republican plot to me!

[WildCat]

Also:

Quote:

To establish a program to provide funds to States to replace punch card voting systems

[WildCat]

BTW, that's the complete list of sponsors and co-sponsors. Notice how the Republicans tricked the naive Democrats right into their trap! "No!" said br'er rabbit. "Don't throw me into the briar patch!".

[Kevin_Lowe]

Quote:

Originally posted by Rob Lister
Kevin, I was being as gracious as I could. There are times, albeit few, when an inferred ad hom is impossible to avoid, even if it is not implied.

Points with which I think we can agree:

The paper is not peer reviewed. That cannot be disputed.

We'll have to do our own review then.

Quote:

The author does not have the complete data set. That cannot be disputed.

Certainly, but if the available data yields sufficiently strong results then it would be very surprising indeed if the complete data set gave us substantially different results.

Quote:

The author has not made his available data set available. That cannot be disputed.

As far as I am aware, no data set that (competently examined) does not show that the 2004 election poll discrepancy is incredibly unlikely to be chance has ever been presented. I'd like to see the full report, of course, but I have seen to evidence that any reasonably comprehensive data set is going to yield startlingly different results.

Quote:

The author has not made his methods available. That cannot be disputed.

As above.

Quote:

The author's conclusions contradict the stated conclusions of those that had the full data set and were in a much better position to know. That cannot be disputed.

If you've read the Mitofsky report, they simply do not support their stated conclusions worth a plugged nickel. As has been pointed out before, the evidence for their proposed mechanism to explain the discrepancy does not exist (the opposite in fact). Their paper wasn't peer reviewed either, nor was their data fully published, so you cannot give it a great deal of weight if you are applying your own standards consistently.

Quote:

The author shows a past and present (very) strong bias for the implied side he is presenting. That cannot be disputed.

I am sorry, you have to excuse me. How do you justify that statement? How does being a musician, a yoga teacher or an activist against tobacco indicate a very strong bias in favour of claiming that the 2004 exit poll discrepancies are highly unlikely to be chance? What is your evidence that Mitofsky are unbiased compared to Mitteldorf, Freeman et. al.?

Quote:

You tell me, Kevin, how much weight should I lend this 'analysis'. Currently, I lend it exactly none.

You cannot pick your conclusion beforehand and then decide that non-peer-reviewed papers without full disclosure of data are just great when they support the conclusion you want, but have no weight at all when they support (much better) a conclusion you don't want.

Since you asked, I think that you should take it to be provisionally true that exit polls consistently underestimated the tabulated Bush vote in Bush strongholds where voting machines were used, and hit the nail on the head wherever hand-counted votes were used. Now the field is open for hypotheses as to why this might be the case, and if you can find a good one other than mischief I think we would all be interested to see it.

[Kevin_Lowe]

Quote:

Originally posted by WildCat
BTW, that's the complete list of sponsors and co-sponsors. Notice how the Republicans tricked the naive Democrats right into their trap! "No!" said br'er rabbit. "Don't throw me into the briar patch!".

Thanks for explaining in more detail where you were going.

It seems to me that your argument, although not clearly stated, is something like

1. HAVA is Democrat-flavoured.
2. It is ridiculous to suggest that a Democrat-flavoured act could harm Democrat interests.
3. Democrats are interested in fair elections.
4. HAVA mandates a changeover to electronic voting.
5. It follows from the above that it is ridiculous to suggest that a changeover to electronic voting could threaten fair elections. That would be tantamount to saying that the Democrats shot themselves in the foot!

Or something like that.

Is that about right, or is there some other chain of logic we should be trying to follow?

[WildCat]

Quote:

Originally posted by Kevin_Lowe
Thanks for explaining in more detail where you were going.

It seems to me that your argument, although not clearly stated, is something like

1. HAVA is Democrat-flavoured.
2. It is ridiculous to suggest that a Democrat-flavoured act could harm Democrat interests.
3. Democrats are interested in fair elections.
4. HAVA mandates a changeover to electronic voting.
5. It follows from the above that it is ridiculous to suggest that a changeover to electronic voting could threaten fair elections. That would be tantamount to saying that the Democrats shot themselves in the foot!

Or something like that.

Is that about right, or is there some other chain of logic we should be trying to follow?

Uhhh, not really. You said:

Quote:

In any case, even if only the US federal government is properly referred to as "the US government" in the USA, a large chunk of the money to buy these things with comes from the Help America Vote Act, which was a piece of federal pork mandating (if I recall correctly) a complete changeover to electronic voting by 2006. Take a wild guess why I think the current government thinks this is a good use of federal funds.

I pointed out that HAVA was a Democrat idea from top to bottom. There wasn't a single Republican (what I think you mean when you cite the "current government") sponsor! If HAVA had been voted down by Republicans, it's proof that they're trying to steal the election. But it passed, and that is even more proof that the Republicans are stealing elections!

I don't know what more to say here, it just seems so loopy to me. The 2004 election was watched closely by Democrats and the details were gone through w/ a fine toothed comb, yet the conspiracy theories abound. Not through any actual evidence, but by a discrepancy w/ exit polls!

[Kevin_Lowe]

Quote:

Originally posted by WildCat
I pointed out that HAVA was a Democrat idea from top to bottom. There wasn't a single Republican (what I think you mean when you cite the "current government") sponsor! If HAVA had been voted down by Republicans, it's proof that they're trying to steal the election. But it passed, and that is even more proof that the Republicans are stealing elections!

You were doing okay until you veered off into mind reading and straw man victimisation.

I agree with you that the fact that HAVA was a Democrat job debunks the idea that voting machines and everything to do with them are an Evil Republican Plot from top to bottom. But it's not remotely correct to characterise my argument as being that because the Republicans passed HAVA it is proof they are in on an evil plot.

About all you can say is that the Republican party did not find HAVA as a whole objectionable. Seeing as the machines HAVA mandates have a funny correlation with exit poll discrepancies in their favour Republican strongholds, you can kind of see why. Assuming anyone in the Republican party knew or had an inkling in advance that was going to be the case anyway.

Quote:

I don't know what more to say here, it just seems so loopy to me. The 2004 election was watched closely by Democrats and the details were gone through w/ a fine toothed comb, yet the conspiracy theories abound.

So despite the fact that a variety of suspicious circumstances have been uncovered when the details were gone through with a fine toothed comb (as we are doing now), none of those circumstances are suspicious, because we found them by going through things with a fine toothed comb? That's the most amazing argument yet.

Quote:

Not through any actual evidence, but by a discrepancy w/ exit polls!

You say that like it's a sensible criticism. It's a genuine problem that needs to be addressed. A highly sensitive measure of how people thought they had voted fails to match with how the votes total up, by a margin which is incredibly unlikely to be chance. Then a group of statistics professors go over the numbers and put it in writing, with their names on it, that these discrepancies (in addition to being of inexplicable scope overall) also cluster in Bush strongholds, and are totally absent from precincts where humans count the votes.

Despite this, we still have the odd poster like yourself ridiculing the idea that this mysterious effect might be something to do with a miscounting of the votes facilitated by the voting machines. What's your alternative hypothesis?

[WildCat]

Quote:

Originally posted by Kevin_Lowe
You say that like it's a sensible criticism. It's a genuine problem that needs to be addressed. A highly sensitive measure of how people thought they had voted fails to match with how the votes total up, by a margin which is incredibly unlikely to be chance. Then a group of statistics professors go over the numbers and put it in writing, with their names on it, that these discrepancies (in addition to being of inexplicable scope overall) also cluster in Bush strongholds, and are totally absent from precincts where humans count the votes.

Despite this, we still have the odd poster like yourself ridiculing the idea that this mysterious effect might be something to do with a miscounting of the votes facilitated by the voting machines. What's your alternative hypothesis?

What are the numbers? Where, exactly, do humans count the votes? Where is the raw exit poll data (time, place, numbers, etc)? How does it compare (exit poll vs. actual) to previous exit poll data from those areas? Has any exit poll data been ignored? If so, why? Where is there even a shred of evidence that voting machines were rigged?

Do you not see how you're making a lot of unjustified assumptions, and then drawing conclusions based on those questionable assumptions?

[Beth]

Quote:

Originally posted by Rob Lister
If you think that then all I can surmise is that you suffer from the beaten wife syndrome.

I don't understand the connection to the beaten wife syndrome. A legimate peer review takes months to complete - after the paper has been writen which also takes time. Statewide precinct level data for the actual votes from the one state I've been looking into wasn't released until April and was revised in May. Assuming someone is working on such an analysis for a peer reviewed publication, I wouldn't expect to see it published before 2006.

Quote:

You're a skeptic. Act like one.

I'm sorry, but I don't know what you mean by this. Could you be more explicit on how you think skeptics ought to behave with regard to suspicious election results?

Quote:

That doesn't mean changing parties. That means picking the voices in the party of your choice that you find worthy of amplifying. You grow your party that way.

nuff said.

I sorry, but are you thinking I'm a Democrat? I'm not.

[WildCat]

Also, Kevin, IIRC the final results closely matched the latest telephone polling data. If anything, the exit polls were an anomaly.

[Rob Lister]

Quote:

Originally posted by Beth
[b]I don't understand the connection to the beaten wife syndrome.

A legimate peer review takes months to complete - after the paper has been writen which also takes time. Statewide precinct level data for the actual votes from the one state I've been looking into wasn't released until April and was revised in May. Assuming someone is working on such an analysis for a peer reviewed publication, I wouldn't expect to see it published before 2006.

I'm sorry, but I don't know what you mean by this. Could you be more explicit on how you think skeptics ought to behave with regard to suspicious election results?

I sorry, but are you thinking I'm a Democrat? I'm not.

para1: YOu are being abused and you make excuses for your abuser.

Para2: I'm commenting on a paper written in March of this year, not some future paper that hasn't been published.

Para 3: I'm not commenting on how you should behave with regard to suspicious election results, I'm commenting on how you should behave with regard to a suspicious 'analysis' of said election results.

para4: You assume facts not in evidence.

[Kevin_Lowe]

Quote:

Originally posted by WildCat
Also, Kevin, IIRC the final results closely matched the latest telephone polling data. If anything, the exit polls were an anomaly.

To respond to your second post first, what you have to understand is the relative reliability of these polls.

You are probably aware that positive results for homeopathy and whatnot notoriously arise in small and poorly controlled trials. Bigger, better studies show no effect.

A telephone poll asks maybe one thousand people, randomly selected from the voting-age population, who they say they will vote for. The M-E exit poll asked seventy thousand people, randomly selected from the people who actually voted in chosen "bellwether" precincts, who they actually did just vote for. There really is no comparison in the expected accuracy.