Hard evidence of electoral fraud claimed!

[Kevin_Lowe]

The journalists at www.blackboxvoting.org are claiming that they have hard evidence of electoral fraud using electronic voting machines, not just hints like exit poll discrepancies.

They are calling for US volunteers to chase further evidence, donations to fund FOIA requests, and for computer security experts willing to go on the record about flaws in the machines' security.

As I said elsewhere, these are not Democrat partisans. They are journalists who have been investigating Diebold for some time, who know their stuff and who do not seem to me to be the kind of people who would cry wolf.

When these people say they have hard evidence, I think it's time to sit up and pay very close attention. The real question now is the scope of the fraud, and whether the people involved can bury the evidence faster than blackboxvoting and its volunteers can dig it up.

American JREFers: This is an issue where your skills may be highly relevant, and this is a lot more important than spoonbending or homeopathy. Please, please, keep an eye on this if you live in a relevant area and you have any free time to help out.

[a_unique_person]

I would be very suspicious of anything like this. Without a paper trail, independent audit, etc, the appearance of certainty is gone. Even if no fraud is committed, the perception that it could be and not detected weakens democracy.

[DaveW]

Any kind of conspiracy theory immediately makes my woo sense twinge. It is possible that it happened, but it is more likely that, in fact, the results are genuine (or that any errors are, in fact, just that: errors, and not intentional). I do agree that some kind of backup audit system would be preferable to none, though.

[Matabiri]

Originally posted by DaveW
Any kind of conspiracy theory immediately makes my woo sense twinge. It is possible that it happened, but it is more likely that, in fact, the results are genuine (or that any errors are, in fact, just that: errors, and not intentional). I do agree that some kind of backup audit system would be preferable to none, though.

I would have thought that a backup audit system should have been required by the terms of the bid for the contract, to eliminate this to start with. But there seems to have been a rush to get the machines in, regardless of any objections.

The population need to trust elections if they are to have any meaning. That means votes should be private, and auditable.

[Marquis de Carabas]

Interesting that they say they have hard evidence, when they're asking for donations so they can find hard evidence. Don't get me wrong, I hope they're right. I could be saved from 4 more years of theocracy. It just seems a little suspicious to me.

[BPSCG]

Originally posted by DaveW
Any kind of conspiracy theory immediately makes my woo sense twinge. It is possible that it happened, but it is more likely that, in fact, the results are genuine (or that any errors are, in fact, just that: errors, and not intentional). I do agree that some kind of backup audit system would be preferable to none, though. What I've noticed over the last four years is all these sophisticated systems designed to make voting idiot-proof and fraud-proof.

The problem is, the more sophisticated and complex a system - any system - is, the more prone it is to imperfection, error, or outright failure.

In this country, when someone is accused of a crime, there is a "chain of custody" for the evidence. Evidence is collected and secured so nobody can tamper with it. Whenever it is removed from its secure location, the person removing it has to sign for it, so at trial, prosecutors can show that the evidence was safe from the moment it was collected to the moment it was presented at the trial.

Defense lawyers love it when the chain has been broken, because they can then ask, "Where was this evidence during the unaccounted-for twenty minutes? How do you know nobody tampered with it?" In fact, a break in the chain can lead to the prosecutor's dropping the charges - even if the break was entirely inadvertent and innocent.

So now we are supposed to have printed voting machine logs that are entirely free of errors, discrepancies, and glitches. But the system is increasingly complex, and as such, increasingly prone to imperfection and error. And we have thus created a trial lawyer's delight, where what is in all probability an innocent, immaterial error becomes fit matter for a lawsuit and for challenging the entire election.

One other thing: The implication of the website at the link is that the Republicans committed voter fraud. Do they seriously believe only Republicans are sophisticated enough and dishonest enough to try to screw around with the ballot boxes in a tight race?

These people are evidently intent on seeing any discrepancy - and as the voting process becomes more complex, there will increasing cases of discrepancies - as evidence of fraud. What they're inviting is a future of elections followed by lawsuit after lawsuit.

[Ladewig]

Originally posted by Marquis de Carabas
Interesting that they say they have hard evidence, when they're asking for donations so they can find hard evidence.


That was my reaction as well. If they have hard evidence, then can't they call in federal or state law enforcement officials?

[Savagemutt]

Bev Harris, who apparently runs the Black Box Voting site certainly doesn't seem to be non-partisan.

http://en.wikipedia.org/wiki/Bev_Harris

In fact, she seems a bit woo herself.

Nonetheless, I do agree that there are serious problems with Diebold in particular, and paper trail-less machines in general.

BTW, whats up with the split between the "org" and "com" sites?

[Ladewig]

Originally posted by BPSCG

But the system is increasingly complex, and as such, increasingly prone to imperfection and error.

Also consider that with this system, the end users (poll workers) operate the system at most once per year and have little experience in handling unexpected errors.

[Matabiri]

Originally posted by BPSCG
What I've noticed over the last four years is all these sophisticated systems designed to make voting idiot-proof and fraud-proof.

"The trouble with making things completely fool-proof, is that it underestimates the ingenuity of the complete fool."
- Douglas Adams

[corplinx]

Sorry but I am a skeptic. These people have made the serious claim that they have _hard_ evidence of _fraud_. Yet I see none.

Sounds like wishful thinking and meme creation. Bush will once again be touted as the illegitimate and unelected president.

[corplinx]

It appears that such an attack may already have taken place, in a primary election 6 weeks ago in King County, Washington -- a large jurisdiction with over one million registered voters. Documents, including internal audit logs for the central vote-counting computer, along with modem “trouble slips� consistent with hacker activity, show that the system may have been hacked on Sept. 14, 2004. Three hours is now missing from the vote-counting computer's "audit log," an automatically generated record, similar to the black box in an airplane, which registers certain kinds of events.

Unfortunately, their "evidence" is a discrepency in the logs on a machine in the state of Washington. That is still soft evidence and could be explained by software errors or other more mundane explanations.

I guess Kerry's win in Washington state wasn't legit afterall.

[Jocko]

I refuse to believe any of this until I hear if from Michael Moore.

[Rob Lister]

Originally posted by corplinx
Sorry but I am a skeptic. These people have made the serious claim that they have _hard_ evidence of _fraud_. Yet I see none.

Sounds like wishful thinking and meme creation. Bush will once again be touted as the illegitimate and unelected president.

Yes and no. The people who are asking for donations are going to get lots and lots of them. I'm sure some percentage of those donations will indeed be used to fund FOIA requests. The rest will be expended on...er...administrative costs. All legit, legally speaking. Salaries, expenses, etc.

[jj]

Originally posted by Rob Lister
Yes and no. The people who are asking for donations are going to get lots and lots of them. I'm sure some percentage of those donations will indeed be used to fund FOIA requests. The rest will be expended on...er...administrative costs. All legit, legally speaking. Salaries, expenses, etc.

Why now instead of before the election, one might ask.

[Leif Roar]

It looks rather iffy. While I quite agree with the dangers in using black-box voting machines, I find it very unlikely that anyone would manage to find "hard evidence of fraud" so shortly after the election.

If they had talked about "irregularites" or even "voting errors" I could have accepted it, but that they so quickly and easily suggest there was fraud in a national election in a modern democracy clearly marks them as less than serious.

[Kevin_Lowe]

Originally posted by BPSCG
What I've noticed over the last four years is all these sophisticated systems designed to make voting idiot-proof and fraud-proof.

The problem is, the more sophisticated and complex a system - any system - is, the more prone it is to imperfection, error, or outright failure.


This is true, but it's also a deceptive simplification. Word processors are subject to imperfection, error and outright failure but they are still superior to a pad and pencil in important ways. Computerised voting should be as safe and auditable as paper voting, but quicker and more accurate. It doesn't have to be a field day for fraud.


So now we are supposed to have printed voting machine logs that are entirely free of errors, discrepancies, and glitches. But the system is increasingly complex, and as such, increasingly prone to imperfection and error. And we have thus created a trial lawyer's delight, where what is in all probability an innocent, immaterial error becomes fit matter for a lawsuit and for challenging the entire election.


Well, we'll see if these errors are innocent and immaterial won't we? I think it's a bit soon to be organising a pity party for the poor misunderstood Diebold company.


One other thing: The implication of the website at the link is that the Republicans committed voter fraud. Do they seriously believe only Republicans are sophisticated enough and dishonest enough to try to screw around with the ballot boxes in a tight race?


I don't see where they said that at all. Could you point that bit out to us?


These people are evidently intent on seeing any discrepancy - and as the voting process becomes more complex, there will increasing cases of discrepancies - as evidence of fraud. What they're inviting is a future of elections followed by lawsuit after lawsuit.

Can you clarify this? You seem to actually be suggesting that it's a bad thing for people to be scrutinising elections.

Under what circumstances do you think it's okay to scrutinise election procedures?

[Kevin_Lowe]

Originally posted by Leif Roar
It looks rather iffy. While I quite agree with the dangers in using black-box voting machines, I find it very unlikely that anyone would manage to find "hard evidence of fraud" so shortly after the election.


Well, we've already had AP reporting on an individual voting machine that gave Bush 4000-odd votes from a county with only 600-odd voters. If they have evidence of a discrepancy something like that, combined with a modem log that shows someone accessing the machine when they shouldn't, that would do it. Just as an example I made up.

I admit I was surprised they made such a strong claim so quickly, but if I was a betting man I'd lay money that they will be able to back it up.


If they had talked about "irregularites" or even "voting errors" I could have accepted it, but that they so quickly and easily suggest there was fraud in a national election in a modern democracy clearly marks them as less than serious.

I think you are jumping to conclusions that are unwarranted.

[gethane]

You know, it doesn't have to be a conspiracy. If just 31 machines in Ohio mistakenly granted Bush an extra 4K votes (like one machine did) then Kerry actually won Ohio. If they don't know what caused it in one machine, how in the world can you be certain it didn't happen in 2? Or 10? or 31? or 100?

What if it were true? What if it were? Wouldn't that matter? Shouldn't someone look into it? FOI requests aren't asking people to make stuff up. They are just asking to see the data. What's so scary about that? If Bush really won, if the machines are really accurate, if there's no conspiracy, then what's the problem with making the results transparent to everyone so everyone can see there was no problems?

But its much more fun I'm sure just to accuse people of being woo woo isn't it. It isn't like election fraud has EVER happened here in the U.S. right?

[Leif Roar]

Originally posted by Kevin_Lowe
Well, we've already had AP reporting on an individual voting machine that gave Bush 4000-odd votes from a county with only 600-odd voters.

In a completely separate case, and which, as far as I've noticed, nobody has suggested had anything to do with fraud.

If they have evidence of a discrepancy something like that, combined with a modem log that shows someone accessing the machine when they shouldn't, that would do it. Just as an example I made up.

Sorry, but no. While that would have suggested fraud, it would not be "hard proof" on its own - weird stuff happens, particularly around computer systems, and particularly around new computer systems.

I admit I was surprised they made such a strong claim so quickly, but if I was a betting man I'd lay money that they will be able to back it up.

I think you are jumping to conclusions that are unwarranted.

My conclusion is only that they do not seem to be serious, for wich reason I will disregard what they say until they present something solid.

[a_unique_person]

Originally posted by Jocko
I refuse to believe any of this until I hear if from Michael Moore.

I repeat, even if there was no fraud, (and it is entirely possible there wasn't), the inability to provide an independently verifiable paper trail weakens democracy.

[Jocko]

Originally posted by a_unique_person
I repeat, even if there was no fraud, (and it is entirely possible there wasn't), the inability to provide an independently verifiable paper trail weakens democracy.

Forget about Bush's re-election - my agreeing with you is a far more concrete sign the apocalypse in near.

[corplinx]

Originally posted by Kevin_Lowe
Well, we've already had AP reporting on an individual voting machine that gave Bush 4000-odd votes from a county with only 600-odd voters. I

Exactly, because the votes in every country were put through an anal exam before Kerry conceded. The major parties are constantly reviewing census results, voter registrations, etc to make sure the other party isn't cheating. Did you notice how quickly this anomaly was found?

I think the people jumping to conclusions are the ones who think the obsessed insiders of the two big parties haven't already gone over every county in Ohio with fine tooth comb looking for even hints of fraud to accuse the other party of.


Once again, occam's razor is a great skeptical tool.

[gnome]

The only part that worries me is, how come a majory of people expected and want a paper trail for electronic voting, but it still didn't happen? Who decided not to include that obvious feature and WHY didn't they want it?

If I were planning on tampering with an electric election (tm) ... I would allow it to count the results legitimately for a few years, until people were used to them and trusted them.

[zakur]

Originally posted by corplinx
Exactly, because the votes in every country were put through an anal exam before Kerry conceded. The major parties are constantly reviewing census results, voter registrations, etc to make sure the other party isn't cheating. Did you notice how quickly this anomaly was found? This anomaly was found by regular citizens noticing the errant number on the county's election results web page, not by some in-depth audit (or "anal exam," as you put it).

[corplinx]

Originally posted by zakur
This anomaly was found by regular citizens noticing the errant number on the county's election results web page, not by some in-depth audit (or "anal exam," as you put it).

Sorry, I know that from reading it but foobared my post. Still the point is that there are many obsessed people (DU, cough) looking for evidence of real fraud. Its naive to think that there is fraud out there and many resourceful people who hate Bush and have and axe to grind haven't found it.

[NickW]

It is true, that whenever technology is used to cast votes, there is always a security risk to flaws in the systems. However, there are of course many security devices in place to keep the votes secure. The benifits of efficiency in using the different touch-screen voting systems highly outweigh the possible security issue. There are no problems with holes from punch card balots still attached or voters not knowing which hole they were supposed to punch out for each canidate. It is of course essential that people such as the FBI's Cyber Crimes unit and the different security agencies within the State Department continue to develop new security hardware and software to combat new advances in hacker technology, but with this, touch-screen voting could make the US voting system a lot more efficient.

[Kevin_Lowe]

Originally posted by corplinx
Sorry, I know that from reading it but foobared my post. Still the point is that there are many obsessed people (DU, cough) looking for evidence of real fraud. Its naive to think that there is fraud out there and many resourceful people who hate Bush and have and axe to grind haven't found it.

As I might have mentioned earlier, I know a little about polling and I've yet to see an explanation for the exit poll discrepancies in the swing states that holds water. So I'm reasonably sure something is up. It's naive to think that there was a problem with the polls but many of the resourceful people who hate Kerry and have an axe to grind haven't found it, right?

I think I'll christen this the "if it was true, it would logically have come out yesterday" argument. You can use it on any day for any point of view and it works equally well. :)

Given that computer fraud can be carried out by one individual given a few minutes alone with a machine, or alone with a USB key, or alone with a modem, and that detecting computer fraud requires people to audit electronic logs in detail, I would not be surprised if people are still digging up dirt in weeks to come. Assuming there is dirt to find, of course.

If a week or two goes by from today and no one has anything, I'll admit it looks like a closed case. If in a couple of week's time we have myseriously vanishing computer logs, officials stonewalling FOIA requests, further individual incidents of thousands of Bush votes materialising out of nowhere and no plausible explanation for the exit poll discrepancies then... well, we'll see what everyone thinks then.

[The Central Scrutinizer]

Originally posted by Kevin_Lowe
The journalists at www.blackboxvoting.org are claiming that they have hard evidence of electoral fraud using electronic voting machines, not just hints like exit poll discrepancies.

They are calling for US volunteers to chase further evidence, donations to fund FOIA requests, and for computer security experts willing to go on the record about flaws in the machines' security.

As I said elsewhere, these are not Democrat partisans. They are journalists who have been investigating Diebold for some time, who know their stuff and who do not seem to me to be the kind of people who would cry wolf.

When these people say they have hard evidence, I think it's time to sit up and pay very close attention. The real question now is the scope of the fraud, and whether the people involved can bury the evidence faster than blackboxvoting and its volunteers can dig it up.

American JREFers: This is an issue where your skills may be highly relevant, and this is a lot more important than spoonbending or homeopathy. Please, please, keep an eye on this if you live in a relevant area and you have any free time to help out.

I'll pass. It's my day to sort my underwear drawer.

[zakur]

Originally posted by corplinx
Sorry, I know that from reading it but foobared my post. Still the point is that there are many obsessed people (DU, cough) looking for evidence of real fraud. Its naive to think that there is fraud out there and many resourceful people who hate Bush and have and axe to grind haven't found it. Agreed.

[BPSCG]

Originally posted by Kevin_Lowe
Under what circumstances do you think it's okay to scrutinise election procedures? I have a weekend to get underway here, so I'm not going to give this the attention I might otherwise.

The 2000 election, as we all "know" resulted in mass disenfranchisement of voters who spoiled their ballots.

The idea of all these electronic voting machines is to make sure that it's impossible to spoil your ballot, no matter how ignorant, uneducated, or just plain stupid you may be.

What we're doing is replacing a simple system that allows the ignorant, uneducated, or just plain stupid to screw up, with a complex one that doesn't.

Neither system is perfect, but the latter one costs us untold millions of dollars in making sure the ignorant, uneducated, or just plain stupid don't screw things up. And what are we getting for that money? Are we getting more accurate results? You seem to be suggesting that we aren't. Or are we getting a system with a lot more places for things to go subtly or egregiously wrong, a lot more places where lawyers can find a cause for a lawsuit?

The BPSCG solution: Keep the voting mechanism and the ballot as simple as possible. I like the optical reader type that scans your ballot the same way the nationwide college board exams are scanned, without incident, every year in this country. And if the ignorant, uneducated, or just plain stupid find even that too complex, then that's just one of the costs of functional illiteracy.

[NoZed Avenger]

Originally posted by Kevin_Lowe
As I might have mentioned earlier, I know a little about polling and I've yet to see an explanation for the exit poll discrepancies in the swing states that holds water.

In the other thread on this, you also talk about the general reliability of polling in general, but I am not sure how that applies to exit polling.

(1) Exit polling is not designed to try and predict the voting outcome; that was not its purpose.

(2) Exit polling has not been particularly reliable in predicting state outcomes -- data from the 2000 and 2002 elections show problems in a number of states (Colorado, Florida among them) with swings of 10% or more compared to the real vote. So this discrepancy is hardly a new phenomenon.

(3) The information that has been released regarding the early polling cause some doubts regarding the methodology: why does the initial sample contain 59% women, for example?

(4) Who did the exit polling? Where was it done (urban settings would obviously skew the results)? What questions were asked? How many declined to answer? Who sponsored the polling? Are there also vulnerabilities in the exit polling method to fraud (when comparing to the actualk vote)? In the absence of information regarding the methodology, I am not sure why the sanctity of the exit polls is assumed and suspicion placed solely on the actual votes cast.

(5) Less importantly, because most of the problems polling do not apply to exit polling: Many pollsters commented on the difficulty in polling this particular race because of a number of factors that distinguished it from previous ones.


In short, I do not say that you might not be correct and I welcome anyone that wants to take a serious look at the process and investigate, but I do question any conclusions about the reliability of the exit polls at this point.

Knowing about polling in general is fine -- but what do we know about this poll. Stating that "polling is usually done professionally, so we can safely assume that this one was done professionally" [gross and overdone paraphrase] seems a bit dodgy.

N/A

[NoZed Avenger]

I also took a quick look for indormation reagrding Bev Harris, the owner/register of the black box site listed above. As far as I can tell, she is the "journalists" mentioned in the first post.

I have read a fair nuimber of articles and I tend to agree about having a paper trail that could be looked at by the voter before leaving the site. I think she advanced some good arguments for having such a system.

However, saying that she is not partisan seems not quite right, as well:

Bev Harris is a left wing political activist who has led the charge against Diebold electronic voting systems. She is a frequent guest on political commentary sites BuzzFlash and WorkingForChange. She is also the co-owner of Talion, Inc., a public relations firm for several prominent authors of books that claim that President George W. Bush had advanced knowledge of the September 11 attacks.

Harris has offered her own plan for election reform: she believes each vote should be hand-counted in the presence of observers. Her feeling is that this could be accomplished in a reasonable amount of time through the use of high-speed scanners.

That's from wikipedia ( http://en.wikipedia.org/wiki/Bev_Harris ), so take it with a shaker of salt, but the Talion reference seems correct. I am unconvinced that a publisher that handles books stating that President Bush knew about the 9/11 attacks before they happened is nonpartisan. She may well be, but it raises doubts.

[Kevin_Lowe]

Originally posted by NoZed Avenger
In the other thread on this, you also talk about the general reliability of polling in general, but I am not sure how that applies to exit polling.


My understanding is that exit polling is actually much more accurate than regular opinion polls, barring electoral fraud. Most opinion polls have a sample size of a thousand or so, exit polls can be ten times that. Regular polls sample the population who are at home with the phone, exit polls sample the population who actually got out and voted.

The exit polls that showed Kerry winning comfortably were co-sponsored by CNN, C-SPAN, AP and Fox. They should have forecast the election results highly accurately.

In fact, the reason the networks sit on the exit polls until late in the day is that they are so accurate that typically there's no point in watching election coverage or voting against the flow once the exit polls show a clear trend.


(1) Exit polling is not designed to try and predict the voting outcome; that was not its purpose.


I'm startled by that claim. Evidence?


(2) Exit polling has not been particularly reliable in predicting state outcomes -- data from the 2000 and 2002 elections show problems in a number of states (Colorado, Florida among them) with swings of 10% or more compared to the real vote. So this discrepancy is hardly a new phenomenon.


Again, evidence? This contradicts my current understanding which is that the single biggest margin of error in the major 2000 exit poll was a 2.6% swing in Alaska. Were these oddities confined to the (partially electronic :D) 2002 election?

(A niggling memory that the 2002 Colorado poll was somehow suspect is plagueing me, but I can't pin it down. Was that the race where the Republican candidate who also owned the company responsible for all the voting computers won a surprise landslide? I must check).


(3) The information that has been released regarding the early polling cause some doubts regarding the methodology: why does the initial sample contain 59% women, for example?


I think you'll find that's a highly accurate historical figure for the proportion of the voting public which is female.


(4) Who did the exit polling? Where was it done (urban settings would obviously skew the results)? What questions were asked? How many declined to answer? Who sponsored the polling? Are there also vulnerabilities in the exit polling method to fraud (when comparing to the actualk vote)? In the absence of information regarding the methodology, I am not sure why the sanctity of the exit polls is assumed and suspicion placed solely on the actual votes cast.


Because these polls are not done by amateurs, they are sponsored by the major news networks, they are conducted by professionals, and before electronic voting came along it was news if they were off by more than one or two percent.

There's a reason why exit polls are the gold standard, and routinely used (partially) as a check for fraud around the world. In fact, they are a vital part of the machinery of democracy.


(5) Less importantly, because most of the problems polling do not apply to exit polling: Many pollsters commented on the difficulty in polling this particular race because of a number of factors that distinguished it from previous ones.


Actually that's the sole source of any possible explanation for the 2004 exit poll anomaly. ;) If a genuinely disproportionate chunk of the electorate stayed home or got out and voted who have not done so in the past that can cause errors. It would have to be a huge and novel change to result in an 8% swing though, because an 8% swing in a competently conducted exit poll is practically unheard of.


In short, I do not say that you might not be correct and I welcome anyone that wants to take a serious look at the process and investigate, but I do question any conclusions about the reliability of the exit polls at this point.


As I said, I know a bit about this from the stats end and from the coal face end. I know there is cause to sit up and take notice. I don't know if hard evidence of fraud is going to be dug up.


Knowing about polling in general is fine -- but what do we know about this poll. Stating that "polling is usually done professionally, so we can safely assume that this one was done professionally" [gross and overdone paraphrase] seems a bit dodgy.

I think it's safe to say that criticising the exit poll's pedigree would require extraordinary evidence.

Edit to add: Having C-SPAN in the list above was an error. My bad. The actual full list of people paying for it seems to be ABC, AP, CBS, CNN, Fox, and NBC.

[Kopji]

The Franklin Ohio problem was the only incident I've heard of so far.
Linky (http://www.commondreams.org/headlines04/1105-23.htm)

The article mentions a second one in San Francisco, but that was a different kind of error, with compilation.

The issue of electronic ballots being "discarded" because of a memory problem is very serious, and may represent deeper problems with the systems.

I'm skeptical of any theories involving conspiracy though. We should expect to see a small random distribution of voting problems. Something to be suspicious of would be consistent errors favoring a particular candidate or position.

I think there IS something wrong with exit polling. My guess would be that a characteristic of the person doing the polling is somehow biasing the results. The pollers are not blindfolded, and can choose their sources of input by personal preference, even an unconscious preference.

Some kinds of voters do not like responding to polls. (I can attest to that).

[DaveW]

Originally posted by Kevin_Lowe
I think it's safe to say that criticising the exit poll's pedigree would require extraordinary evidence.

I think putting the exit polls' results above the election's actual results would require even more extraordinary evidence. It would be much easier to skew the exit polls' results...

[NoZed Avenger]

Originally posted by Kevin_Lowe
My understanding is that exit polling is actually much more accurate than regular opinion polls, barring electoral fraud.

Am I the only one supposed to do research, or -- since the original claim was that these exit polls should be accurate for predicting the election -- may we see some support for the initial proposition?

But warnings about exit polls (an polling in general) and who is doing them can be located on the National Council on Public Polls wesite, for example -- where they set out a number of questions on methodology. A few of those questions I highlighted, but I have not seen anyone that has put out the answers (at least on-line). Do you have information on the methodology used (as well as the polling company? Mitofsky International IIRC) The timing? The location? The refusal rate? The explanation for 59% women?

We have little or no information regarding the refusal rate (at least that I have seen -- perhaps you have a link?). Refusals are one of the main sources of error in all polls, and in fact can be more serious than sampling error because it is difficult or impossible to judge their effect the way you can with sampling error.



I'm startled by that claim. Evidence?

Exit polls were created as a tools to explain voter motivation, decision-making and ideology -- not as a predictive tool. They are commonly used for that now, but praxctically every pollster or editor I have seen quoted on the subject includes a caveat the exit poll was and is not designed to be a predictive tools. Indeed, I am equally startled at your apparent surprise.


Again, evidence? This contradicts my current understanding which is that the single biggest margin of error in the major 2000 exit poll was a 2.6% swing in Alaska. Were these oddities confined to the (partially electronic :D) 2002 election?

Ever hear of the Voter News Service? Created in the 60's, started with exit polls in the 90's, disbanded after 2000 and 2002 -- paritally because of the results not matching up in 2000 and 2002, and the fact that the organization declared that it could not function in 2002 in any realistic sense.

See also:
http://www.cpsa-acsp.ca/papers-2004/Docherty%20and%20others.pdf

The problem with exit polls resurfaced in the 2000 U.S. presidential election. By the 1990’s the large US media outlets decided to pool resources and conduct one omnibus exit poll. Without competing polls, there could be no immediate check on accuracy. The American example, particularly in 2000, illustrates that, as with any poll, election day exit polls can generate faulty results, and can be abused if their users are more concerned with getting questions answered quickly rather than accurately.

There are dozens and dozens of news arctiles from 2000 and 2002 on this -- check the CNN archives, for example. The Colorado mistake

In this case, we are not even dealing with a completed exit poll, but expectations driven by polling results released llong before the polls even closed. Do you have anything showing the historical reliability of half an exit poll?

And the early exit polling has had a number of criticisms leveled at it already. Typical example:

http://www.al.com/opinion/mobileregister/index.ssf?/base/opinion/109964989371420.xml

As it happened, the early poll samples included too high a proportion of women, too few Westerners and far too few Republicans.

So lets hear how and why these gold-standard polls included the above over- and underrepresentations?


I think you'll find that's a highly accurate historical figure [added for context: 59% N/A] for the proportion of the voting public which is female.

Well, as long as we're asking for evidence -- do you have something showing that this is an appropriate percentage for the exit polling done early Nov 2nd? I say that because I have seen references to this very fact -- by the pollsters themselves, even -- as a reason that their results were skewed. Since traditionally the number of women voters in US elections has been closer to 52%, I'd love to see something showing that the real number should be 59%.

In fact, the Alabama Reg article listed above explicitly states that women were over-sampled -- what shows 59% tto be accurate?

You have also made several appeals to your knowledge/experience in the area of polls. As that is part of the pitch, what is your background related to polling?



Because these polls are not done by amateurs, they are sponsored by the major news networks, they are conducted by professionals, and before electronic voting came along it was news if they were off by more than one or two percent.

Then lets have the pollsters present their methodology and have everyone take a look. But your post appears to paint exit polling as all-but-infallible -- leaving the impression that the safeguards on exit polling make any doubts require exptraordinary proof of mistake or fraud, while holding out a much friendlier standard for anyone questioning the actual election.

Exit polls have a number of vulnerabilities beyond those already listed (and despite your assertion that there is only one source of possible error): ever hear of "slamming" exit polls?

We also don't know -- and probably can't -- if the early reporting of the exit polls affected the final vote tally by convincing people or galvanizing people to vote.

In short -- I still want people to look into the election. But at the same time, I want them to be looking into the exit polling. Crossing it off as a source of error before the investigation begins seems a bit short-sighted.

The Alabama article states that:

Variously, they [the exit polls] showed huge margins for Sen. John Kerry in three states that actually were quite close -- Minnesota, Pennsylvania and New Hampshire -- and wins for the senator in six states he actually lost: Florida, Ohio, New Mexico, Colorado, Nevada and (probably) Iowa.

Since electronic voting machines were not used in all of those venues, what explanation do you have for the chronic problem that seems to have skewed results in the same way across at least 8-9 different states?

8-9 or more different, independent problems that skewed the results in the same direction despite different people (and parties) being in control of the voting machinery in each state, or maybe. . . . the one constant factor, the polling, has a problem with its methodology?

But feel free to show how punch card ballots, electronic ballots, and optically scanned ballots can all be tied to the new black boxes.

While we are at it, do you still maintain blackbox.org is run by non-partisan journalists?

N/A

[Zep]

Originally posted by Jocko
I refuse to believe any of this until I hear if from Michael Moore. Jocko, you are a useless twerp.

Sgd. Michael Moore

Happy now?

[NoZed Avenger]

Originally posted by NoZed Avenger
Exit polls were created as a tools to explain voter motivation, decision-making and ideology -- not as a predictive tool. They are commonly used for that now, but praxctically every pollster or editor I have seen quoted on the subject includes a caveat the exit poll was and is not designed to be a predictive tools. Indeed, I am equally startled at your apparent surprise.


Though I don't have a ton of time, I meant to come back to this, but forgot. Slate had a recent article that mentioned this bit about exit polls.

http://slate.msn.com/id/2109134/

"Somebody should reassess exit polling," Carlson said. "It's useless." When Juan Williams threw a similar fit on Fox News Channel about the worthlessness of the exit polls, William Kristol attempted to calm him by noting the obvious: Exit polls are polls, and all polls contain a margin of error.

To that Kristol could have added that no poll is better than the methodology behind it; that exit polls weren't designed to predict winners; and that no state was given to either candidate due to them.

The good thing about today's uproar is that it's accelerating the much-needed demystification of exit polls that Kinsley commenced.

The article links to an online chat on the subject from the Washington Post's Steven Coll (reg. required) that talks about the methodology. I hope he takes the points that are very generally discussed and does an investigative piece on it.

http://www.washingtonpost.com/wp-dyn/articles/A13590-2004Oct31.html

This doesn't even mention the fact that the people that created and ran the poll have admitted to potential problems:

http://slate.msn.com/id/2109310/

Joe Lenski of Edison Media Research and Warren Mitofsky of Mitofsky International admit in a report they wrote that some of their exit pollsters failed to get close enough to voting places to gather adequate samples of voter opinion, and that the exit polls may have overstated Kerry support because his supporters were more willing to be surveyed than Bush supporters.

Their report, covered in today's New York Times ("Report Says Problems Led to Skewed Surveying Data," Nov. 5), "details systematic glitches that skewed the data in ways of which several news organizations, who paid tens of thousands of dollars for the service, were not aware[.]"

While I have doubts about some of the reasons being given by the pollsters, it does not look they are insisting that their data must be correct absent extraordinary proof to the contrary. (It looks like the same excuse they gave for blowing Montant by more than 5% in 1996, as well as Florida in 2000.

Even the pollsters do not appear to claim inerrancy.

[Kevin_Lowe]

Originally posted by NoZed Avenger
Am I the only one supposed to do research, or -- since the original claim was that these exit polls should be accurate for predicting the election -- may we see some support for the initial proposition?


Trust me on this one. The recent inaccuracy is fairly unprecedented, and their use in detecting fraud is well established.


But warnings about exit polls (an polling in general) and who is doing them can be located on the National Council on Public Polls wesite, for example -- where they set out a number of questions on methodology. A few of those questions I highlighted, but I have not seen anyone that has put out the answers (at least on-line). Do you have information on the methodology used (as well as the polling company? Mitofsky International IIRC) The timing? The location? The refusal rate? The explanation for 59% women?


They were always accurate before. What changed in the swing states with the GEMS tabulators? As I said earlier, you need an explanation specific to the places and times showing anomalies.

To put it another way do you have any story for why refusal rates would have spiked just this election, just amongst Rep voters, and just in a few states? That's the story some people are floating, and it's a floater.


We have little or no information regarding the refusal rate (at least that I have seen -- perhaps you have a link?). Refusals are one of the main sources of error in all polls, and in fact can be more serious than sampling error because it is difficult or impossible to judge their effect the way you can with sampling error.


Okay, why would refusals spike so specifically? If you're in the business of asking me to prove that kind of negative, I'll admit up front that I can't prove it didn't happen. I can't prove elite squads of Democrat poll-loaders didn't home in on the pollsters in swing states to swing the poll in the morning, and then go home in the evening. I can't prove that hordes of fundy christians who consider exit polls to be tools of Satan didn't show up in the afternoon. I'm just saying that I consider these scenarios less likely than cheating using the Republican-owned voting and tabulating machines.


Exit polls were created as a tools to explain voter motivation, decision-making and ideology -- not as a predictive tool. They are commonly used for that now, but praxctically every pollster or editor I have seen quoted on the subject includes a caveat the exit poll was and is not designed to be a predictive tools. Indeed, I am equally startled at your apparent surprise.


If you're getting your info from the big news sources, it's fair to point out that they are not touching the fraud angle despite the healthy amount of evidence that the Republican party has been laying the groundwork for electronic election-rigging well in advance. Reps own Sequioa, ES&S and Diebold and they've blocked attempts to insitute mandatory electoral paper trails.

Currently the big news services seem to be mostly running Just So stories to explain the exit poll discrepancies in any way that avoids mentioning the elephant in the room.


Ever hear of the Voter News Service? Created in the 60's, started with exit polls in the 90's, disbanded after 2000 and 2002 -- paritally because of the results not matching up in 2000 and 2002, and the fact that the organization declared that it could not function in 2002 in any realistic sense.


I'm unfamiliar with it, although I vaguely remember some kind of controversy. Could you elaborate further on its significance?


See also:
http://www.cpsa-acsp.ca/papers-2004/Docherty%20and%20others.pdf


Hmm. Canadian academics reinvent the wheel.

It would be a good paper for introducing beginners to the difficulties involved in exit polling. The thing is, these difficulties (refusal rates amongst different age groups and so forth) are either factored into or eliminated by the quotas the pollsters use, based on the data from previous years. Assuming they are competent of course, and I think they are in this case.

To use a skeptical analogy, we're currently at the "how did eyes evolve, huh?" level of argument from ignorance.


In this case, we are not even dealing with a completed exit poll, but expectations driven by polling results released llong before the polls even closed. Do you have anything showing the historical reliability of half an exit poll?


I haven't got anything in my bookmarks folder that would keep you happy but my guess would be that it's close to that of a full poll, as you'd expect.


And the early exit polling has had a number of criticisms leveled at it already. Typical example:

http://www.al.com/opinion/mobileregister/index.ssf?/base/opinion/109964989371420.xml


This guy thinks we should get rid of exit polls. Who benefits?


So lets hear how and why these gold-standard polls included the above over- and underrepresentations?


Which in particular? Were they in the WP article that required registration or what? Sorry to look like I'm ducking the question, but I'm not sure what specific incidents you want me to take a look at.


Well, as long as we're asking for evidence -- do you have something showing that this is an appropriate percentage for the exit polling done early Nov 2nd? I say that because I have seen references to this very fact -- by the pollsters themselves, even -- as a reason that their results were skewed. Since traditionally the number of women voters in US elections has been closer to 52%, I'd love to see something showing that the real number should be 59%.


Having looked into that one further, it does appear that there was a combination of polling sloppiness and bad reporting that would have produced a slight slant to Kerry in the early exit polls. Slightly more women took the poll in the morning than men, and the unweighted data got sent on to the networks.

That's not nearly enough of an effect to produce the anomalous results we have seen, though, based on the final results I can find.


You have also made several appeals to your knowledge/experience in the area of polls. As that is part of the pitch, what is your background related to polling?


I spent half a year working as a phone pollster for Roy Morgan Research, the Morgan-Gallup poll mob, and a briefer period working for another polling organisation who shall remain nameless because theiy were revoltingly unethical. RMR were the good guys. This other bunch were the slime you called when you needed a push-poll, or needed disgruntled ex-customers "randomly polled" at arm's length. I bailed after a month. It was not a happy place.


Then lets have the pollsters present their methodology and have everyone take a look. But your post appears to paint exit polling as all-but-infallible -- leaving the impression that the safeguards on exit polling make any doubts require exptraordinary proof of mistake or fraud, while holding out a much friendlier standard for anyone questioning the actual election.


The thing is that the exit polls are actually a lot more transparent than the vote registration and tabulation process, and the means and motive for fraud far less clear.

On one hand you need either professionals who are prepared to be incompetent in public, or a nationwide band of keenos numbering in the thousands poised to swamp the pollsters. On the other you need insecure machines (proven) with no paper trail (thanks to Republican party obstruction) that partisans have access to (Reps own Sequoia, Diebold and ES&S).

On one hand you have professionals who won't get hired again, on the other you have career politicians furthering their careers and profiting handsomely when their companies get contracts to install more machines. (Help America Vote indeed).

I wouldn't be calling the election fraudulent just on the basis of the polls, if it wasn't already very clear that the democratic cookie jar was in party-owned hands and hidden behind a curtain to boot. If you think that coming down on the side of fraud is a long shot in the circumstances we differ.


Exit polls have a number of vulnerabilities beyond those already listed (and despite your assertion that there is only one source of possible error): ever hear of "slamming" exit polls?


It can be done and it has been done, but lacking evidence that it happened in the vital states I don't think it's a particularly likely explanation, although as I said earlier it's nigh impossible to prove such things didn't happen after the event.


We also don't know -- and probably can't -- if the early reporting of the exit polls affected the final vote tally by convincing people or galvanizing people to vote.

In short -- I still want people to look into the election. But at the same time, I want them to be looking into the exit polling. Crossing it off as a source of error before the investigation begins seems a bit short-sighted.


Sure. Check everything.


Since electronic voting machines were not used in all of those venues, what explanation do you have for the chronic problem that seems to have skewed results in the same way across at least 8-9 different states?

8-9 or more different, independent problems that skewed the results in the same direction despite different people (and parties) being in control of the voting machinery in each state, or maybe. . . . the one constant factor, the polling, has a problem with its methodology?


This would be the "magic bullet" methodological problem that hits states with Republican-owned tabulation machines? Like I said before, you need some state-specific explanations.

If the one constant factor was, say, Diebold's GEMS software counting each vote as it came back to the central server, I think that's got to be the #1 candidate.


But feel free to show how punch card ballots, electronic ballots, and optically scanned ballots can all be tied to the new black boxes.


Are you aware that GEMS by specific design allows you to maintain two sets of books for counting votes, and alter their relationship by entering an undocumented two-keypress code?

All the cards and ballots get counted and then the count is relayed back to computers programmed by a convicted embezzler, who was employed by a Republican campaign leader, that store the votes in an unsecured MS Access database. Typically these servers are hooked up to modems which, while not net-accessible, are remotely accessible by anyone who knows the correct phone number(s) to call.

I couldn't make this stuff up if I tried.


While we are at it, do you still maintain blackbox.org is run by non-partisan journalists?


What's the evidence they are actually partisans, as opposed to being investigative journalists with left-wing opinions? I hold left-wing views, but that doesn't make me a Democrat (or Labour/Democrat/Green in Aussie terms) partisan. I'd have to actually make furthering the goals of those political parties my goal to qualify as a partisan. The available evidence is that they're doing legitimate journalism at BBV that should by rights be wholeheartedly supported by people of all political stripes.

Ensuring free, fair and open elections should not be a partisan issue. In my opinion, anyway.

[glsunder]

BPSCG has the right idea: keep it simple, stupid. My county used the exact same method he mentioned: I filled in ovals just like I did in highschool taking the SAT and ACT tests. I'm sufficiently confident that my vote could be recounted if necessary, unless it was lost. The only good evote system I've heard people describe is one where the machine basically fills in the ovals for the voter, who can then verify the paper ballot.

Slashdot has had numerous stories (http://slashdot.org/search.pl?query=diebold) on diebold. They're either incompetent or corrupt, either of which is enough reason to not use their equipment in elections.

If it's so easy to commit voter fraud, it doesnt really matter if it occured, it needs fixed and it needs fixed now. And it doesn't matter if it's happened all along. If it takes lawsuits to do it, so be it. I don't care if it makes a few people uncomfortable because it challenges their belief that the USA can do no wrong and that we live in a perfect black and white world.

[corplinx]

Originally posted by glsunder

Slashdot has had numerous stories (http://slashdot.org/search.pl?query=diebold) on diebold. They're either incompetent or corrupt, either of which is enough reason to not use their equipment in elections.


There is an even simpler explanation. Government is incompetent and corrupt. Diebold makes these machines according to the whims of these county and state voting officials who want these things to where the blue haired poll workers can work them.

Having to build according to government spec sucks.

[corplinx]

Originally posted by Kevin_Lowe

What's the evidence they are actually partisans, as opposed to being investigative journalists with left-wing opinions? I hold left-wing views, but that doesn't make me a Democrat (or Labour/Democrat/Green in Aussie terms) partisan. I'd have to actually make furthering the goals of those political parties my goal to qualify as a partisan. The available evidence is that they're doing legitimate journalism at BBV that should by rights be wholeheartedly supported by people of all political stripes.


Posted today on democrat underground:

Posted by BevHarris
Added to homepage Mon Nov 08th 2004, 07:39 AM ET

I was tipped off by a person very high up in TV that the news has been locked down tight, and there will be no TV coverage of the real problems with voting on Nov. 2. Even the journalists are pretty horrified. My source said they've also been forbidden to talk about it even on their own time, and he was calling from somewhere else. He was trying to figure out how to get the real news out on vote fraud.

Definitely not a partisan at all being posted on democratundergound.

[NoZed Avenger]

Originally posted by Kevin_Lowe
Trust me on this one.

As they used to say: Trust, but verify. So you're comments are appreciated, but some linked support would also be nice -- especially as a couple of your earlier assertions (say the percentage of women voting historically, for example) were incorrect.


They were always accurate before. What changed in the swing states with the GEMS tabulators? As I said earlier, you need an explanation specific to the places and times showing anomalies.

Always?

Then Pat Buchanon won at least 1 primary ans was possibly the Republican nominee for President in 1992.

In 1996, Seantor Robert Smith of New Hampshire was surprisingly defeated, as shown by the exit polls. Someone should tell him.

May I also present Prime Minister Neil Kinnok, who defeated John Major in 1992 and threw the Conservatives out. We all remember that, right, as the BBC announced the "Labour landslide"? "As every politician knows, the pollsters got it wrong in 1992, so why should we trust them now?" Bill Bush, The BBC News General Election Guide, (1997).


Or the US in 2002 again, as the trend in 2000 and 2002 seems likely important for 2004.

http://www.buzzle.com/editorials/11-5-2002-29700.asp

America's television networks faced a rerun of election night embarrassment yesterday - two years after their spectacularly wrong calls of the 2000 presidential contest - when the main pollster withdrew its exit polls because it could not guarantee their accuracy.

There are some offline sources, as well regarding weighting errors in the 1992 national elections in the US that caused quite a bit of concern.

Interview with Warren Mitofsky (that name again) "The Polls and the 1992 Elections: Problems in Exit Polling," The Public
Perspective, January/February, 1993 (weighting mistakes)

The Public Perspective, January/February 1993 (weighting procedure re: original over estimation of Clinton vote based on preliminary raw data).



To put it another way do you have any story for why refusal rates would have spiked just this election, just amongst Rep voters, and just in a few states? That's the story some people are floating, and it's a floater.

I have listed quite a bit more than that for potential problems with the polling. We know, for example, that despite all the professional safeguards, the proportion of women polled was overstated. You initially stated that the 59% number was historically accurate. Now that number has been admitted to be a mistake, and you indicate that the percentage is not enough to explain the whole problem.

While I have not worked the mathematics myself, I think it unlikely that the 52-59% disparity explains everything -- still, it would be nice for everyone to show their work on an admitted flaw before moving blythely on.

More importantly, if the pollsters made that error -- a pretty obvious one, IMO -- then the entire "trust them not to make mistakes, their professional" line of argument seems a bit weak, does it not?

Additionally, Republican support in the swing states comes from rural areas -- while pollsters economically (generally) rely more heavily on urban settings.

And while we're at it -- what happened to the note that the early results that shows this disparity were based on exit polls performed before the polls closed -- the timing of the polling (to that point) may well have an additional impact.


Okay, why would refusals spike so specifically? If you're in the business of asking me to prove that kind of negative, I'll admit up front that I can't prove it didn't happen.

Of course, I asked for affirmative information regarding what the actual refusal rates were -- the rest is simply a rewording and reformation of my request until it bears little resemblence to my point. We don't have to guess about the refusal rates -- what are they?

But on the subject, surely you've heard of differential refusal? Social factors or traditional differences in response rate between the parties (they do already exist) could easily impact the numbers here. A perceived unpopularity of Bush could also add to it. This phenomenon is hardly unknown; it has happened in the past and cannot be ruled out, even tentatively, until we have some actual data about the real refusal rates. This should not be guesswork on whether the rates were higher or not -- professional pollsters are supposed to keep that data, as well. What was the refusal rate?


If you're getting your info from the big news sources, it's fair to point out that they are not touching the fraud angle despite the healthy amount of evidence that the Republican party has been laying the groundwork for electronic election-rigging well in advance. Reps own Sequioa, ES&S and Diebold and they've blocked attempts to insitute mandatory electoral paper trails.

I am assuming that the above is not the "healthy amount of evidence" that you refer to.

Got some?

At the moment, it looks like deciding between exit polling and the actual voting is merely a matter of speculation regarding what happened. Which is fine, as far as it goes -- but selling fraud as the far more likely outcome without any real data to back it up is over the top.


I'm unfamiliar with it, although I vaguely remember some kind of controversy. Could you elaborate further on its significance?

If you do not see the signifigance of the exit polling problems in 2000 and 2002 (when in that year the entire results were thrown out as untrustworthy), I doubt that anything I can write will help. It ties in to the "exit polls have always been accurate" assertion that you repeat.

Here is a brief site on exit polling in 2002, though it does not explain the total breakdown which caused the pollsters to withdraw the exit poll information entirely as too unreliable to use:

http://politics.fandm.edu/November192002.htm

Exit polls do have some problems. One of these is that they differ from other scientific polling--largely because dispersed polling places preclude exit pollsters from using normal sampling methods. Without wading too far into the technical aspects of exit polling, it is safe to say they should always be taken with the proverbial grain of salt--not because they are necessarily wrong, but because they are less likely to be right than other scientific polls.

The other point about exit polls is that they are sometimes misused. Calling elections on election night with exit polls in order to boost viewer ratings is a flagrant misuse of polling. The public criticism of this practice is right on point. In fact, many pollsters condemn it.

The misuse of exit polls is rooted in how they are paid for. In effect, a Faustian bargain has been made with national mass media news outlets that finance them. We get the poll results, but they get to decide how they are first used. And that first use--to call elections and build ratings--can be a misuse.

For free, it has a bit on these polls not being intended to call races, as well.




Hmm. Canadian academics reinvent the wheel.

It would be a good paper for introducing beginners to the difficulties involved in exit polling. The thing is, these difficulties (refusal rates amongst different age groups and so forth) are either factored into or eliminated by the quotas the pollsters use, based on the data from previous years. Assuming they are competent of course, and I think they are in this case. [/b]

I know that you are confident in that assumption. I'd like to see some support than shores up why anyone else should be.

For example, the exit polls were quite a bit different than the daily polls conducted right up to the eve of the election -- and those daily polls were much closer to the actual results.

Do you factor in the other polls from a dozen different professional sources when trying to locate the source of the error?

As far as this election being different, there are some rather glaring differences between this election and others (with a possible relationship to 2000, which had its own problems).

(1) Iraq changed a lot and almost necessarily affected assumptions and weighting of poll answers. (2) Women broke heavier for Bush than in the past -- affecting the model and doubly affecting polls using too many women. (3) Record turnout by a wide margin. By itself, it might skew results, but on top of that, the turnout (3a) simultaneously did not break for the challenger (as expected) nor (3b) assist the Democrats (as expected), two huge differences from previous elections. (4) More crossover voters making the weighting assumptions about party affiliation suspect. (5) Mail-in ballots and early voting were very heavily used this year -- making another big difference that hits exit polling. As voters vote early, it has a real potential to skew exit polls by physically removing voters from voting booths; this also compromises the ability of pollsters to project election outcomes.

Why wouldn't those trends make a difference in this election? It wasn't like the others used to model for it.

[snip]



If the one constant factor was, say, Diebold's GEMS software counting each vote as it came back to the central server, I think that's got to be the #1 candidate.

If?

Well, that was one of my questions. Shouldn't we know this before deciding that fraud is "likely" and that problems with the exit polls are nigh impossible?

I have read that Ohio's ballots are 70% punch card, for example. The physical ballots are still there and would not appear to use this mysterious black box, etc. Doesn't that impact the fraud claim?

We have 8 states where the early exit polling failed to match up with the outcome (1 of them, Colorado, where Kerry actually did better than the polls predicted, but not enough to win). Of those states, which used this system, and how widespread in the states was its use?

Isn't that a basic, basic question that needs to be asked before we start assigning probabilities to the likelihood that voter fraud exists?



Are you aware that GEMS by specific design allows you to maintain two sets of books for counting votes, and alter their relationship by entering an undocumented two-keypress code?

All the cards and ballots get counted and then the count is relayed back to computers programmed by a convicted embezzler, who was employed by a Republican campaign leader, that store the votes in an unsecured MS Access database. Typically these servers are hooked up to modems which, while not net-accessible, are remotely accessible by anyone who knows the correct phone number(s) to call.

I couldn't make this stuff up if I tried.

Possibly not, but why not show the link, anyway?



What's the evidence they are actually partisans, as opposed to being investigative journalists with left-wing opinions? I hold left-wing views, but that doesn't make me a Democrat (or Labour/Democrat/Green in Aussie terms) partisan. I'd have to actually make furthering the goals of those political parties my goal to qualify as a partisan. The available evidence is that they're doing legitimate journalism at BBV that should by rights be wholeheartedly supported by people of all political stripes.

Um. You made the claim that this website was non-partisan. Your assertion. You felt confident enough to state that affirmatively. I don't even have information that the owner(s) is/are journalists at the moment; I did not spend a lot of time looking at her/their background -- I was hoping you already had it.

I was asking if you were sticking to that assertion because of the owner's ownership of a publishing house which has apparently put out books tying Bush and the CIA to a 9/11 conspiracy. An implied question was what made you assert so strongly that the journalists who ran the site were so definitely nonpartisan.

What "available evidence" are you referring to? Do you have anything other than what they (she) say (says) on her web site?


Ensuring free, fair and open elections should not be a partisan issue. In my opinion, anyway.

Is this a method of insinuating that anyone that doesn't accept these conclusions unreservedly must feel that ensuring free, fair and open elections is a partisan issue -- or indeed, that they must oppose free and fair elections?

N/A

[Kevin_Lowe]

I'm putting together a longer reply to NoZed's very reasonable requests for cites and links. But in the process I ran across a wikipedia article on this very topic that was, frankly, staggeringly well put together given the scope of the issue and the timeframe involved. It has all sorts of maps and things in pretty colours.

Cue speculation about Democrat conspiracies.

In any case, it's at:

http://en.wikipedia.org/wiki/2004_U.S._Election_controversies_and_irregularitie s

...and it presents most of the relevant issues better than I have.

[Kevin_Lowe]

Originally posted by NoZed Avenger
Always?


Well, maybe not always. :)

I've been unable to find a suitable link via google, so I'm in the process of looking at the academic literature in search of something we could agree on as suitably definitive. I'll get back to this later, and I'm sure you'll remind me if I do not.


Or the US in 2002 again, as the trend in 2000 and 2002 seems likely important for 2004.

http://www.buzzle.com/editorials/11-5-2002-29700.asp


Exit polls went amiss in 2002, when electronic voting machines were beginning to be rolled out? I think you can see how that could be taken either way.


I have listed quite a bit more than that for potential problems with the polling. We know, for example, that despite all the professional safeguards, the proportion of women polled was overstated. You initially stated that the 59% number was historically accurate. Now that number has been admitted to be a mistake, and you indicate that the percentage is not enough to explain the whole problem.


Actually what I said was that I thought you'd find it was an accurate estimate of women voters, and I said that because I misunderstood your earlier claim. I thought you were claiming that the full poll was 59% female, and if that were the case it would be because that was a good estimate of the proportion of people voting who were female.

The fact turned out to be that by halfway throgh the poll they'd polled more women than men. They were aiming for the correct number, they just got ahead on one demographic. That happens.


While I have not worked the mathematics myself, I think it unlikely that the 52-59% disparity explains everything -- still, it would be nice for everyone to show their work on an admitted flaw before moving blythely on.


I ran the (fudged) "final exit poll numbers" from CNN, and if they were accurate and consistent then a 59/41 split would have given Kerry a spurious boost in the polls of about 1%. That's bad but within the usual margin of error in a good exit poll. Errors of five to fifteen percent are too large to be explained by the given disparity in respondent's sex.

The wikipedia people seem very organised and very cautious about putting out anything other than sourced facts, and they seem to think it's a solid claim that the big exit poll discrepancies are all in the swing states, and all in favour of GWB.


More importantly, if the pollsters made that error -- a pretty obvious one, IMO -- then the entire "trust them not to make mistakes, their professional" line of argument seems a bit weak, does it not?


In my experience with big polls, it's almost inevitable that some quotas fill up faster than others. The last hour or so of a poll is usually a hunt for the last few 25-39 year old male registered voters or whatever. (That last hour sucked because I got paid by the poll, and obviously your hit rate plummets in the closing hours).

You could say they erred by giving the unfinished data to the media at all, but with suitable caveats it should have been tolerably valid. Within one percent, as I said.


Additionally, Republican support in the swing states comes from rural areas -- while pollsters economically (generally) rely more heavily on urban settings.


If they haven't compensated for that I'll agree on the spot that they are incompetent and we should bin their results.


And while we're at it -- what happened to the note that the early results that shows this disparity were based on exit polls performed before the polls closed -- the timing of the polling (to that point) may well have an additional impact.


These effects cannot be the mystery effect we are hunting for unless they are very hefty effects indeed, and confined to the swing states where the poll discrepancies occurred. That fits neither bill.


Of course, I asked for affirmative information regarding what the actual refusal rates were -- the rest is simply a rewording and reformation of my request until it bears little resemblence to my point. We don't have to guess about the refusal rates -- what are they?


Actually, refusal rates are something most polling companies are very coy about indeed. Mostly because they've dropped from 35% refusal decades ago to a current low of 70%+ refusal and any Tom, Dick or Harry can see that's suboptimal for accuracy purposes. Don't run too far with that though: for it to be the explanation we want, their would have to be a historically massive shift in refusal rates amonst specific voters in specific places and not elsewhere.


But on the subject, surely you've heard of differential refusal? Social factors or traditional differences in response rate between the parties (they do already exist) could easily impact the numbers here. A perceived unpopularity of Bush could also add to it. This phenomenon is hardly unknown; it has happened in the past and cannot be ruled out, even tentatively, until we have some actual data about the real refusal rates. This should not be guesswork on whether the rates were higher or not -- professional pollsters are supposed to keep that data, as well. What was the refusal rate?


No one seems to know, and it's quite possible that info has not been released.


I am assuming that the above is not the "healthy amount of evidence" that you refer to.

Got some?


Let's see.

This covers a lot of the ground.

http://www.motherjones.com/commentary/columns/2004/03/03_200.html

This covers the GEMS functions facilitating election fraud in more detail, and shows the extent of the planning that has gone into GEMS system.

http://www.ejfi.org/Voting/Voting-30.htm


At the moment, it looks like deciding between exit polling and the actual voting is merely a matter of speculation regarding what happened. Which is fine, as far as it goes -- but selling fraud as the far more likely outcome without any real data to back it up is over the top.


Once you've read the above, and verified any details you find questionable, I think you'll be less likely to dismiss the possibility of fraud. It looks a whole lot like certain people in the Republican party have been working deliberately towards the goal of controlling your elections for some time now.

I wouldn't say it if I didn't think the evidence supported the claim.


If you do not see the signifigance of the exit polling problems in 2000 and 2002 (when in that year the entire results were thrown out as untrustworthy), I doubt that anything I can write will help. It ties in to the "exit polls have always been accurate" assertion that you repeat.


The 2000 and 2002 problems look superficially like they might tie back to the same source: rigged elections. You'd like think that in the Land of the Free that would be impossible to get away with, but I'm not of the opinion that is the case.


Here is a brief site on exit polling in 2002, though it does not explain the total breakdown which caused the pollsters to withdraw the exit poll information entirely as too unreliable to use:

http://politics.fandm.edu/November192002.htm


The VNS seems to have a murky history. I'll flag them for further research.

snippage of long list of factors

Why wouldn't those trends make a difference in this election? It wasn't like the others used to model for it.


No, but those factors are not specific to the swing states with the exit poll discrepancies. Despite all those reasons to expect the exit polls to be a bit off this year, they nailed it pretty accurately everywhere except in a handful of vital electorates.


I have read that Ohio's ballots are 70% punch card, for example. The physical ballots are still there and would not appear to use this mysterious black box, etc. Doesn't that impact the fraud claim?


Read the links above. The voting machines are quite likely to be biased, given the US history of voting machines. But the tabulation by ES&S and Diebold is where the real threat lies. Those paper ballots are still there, but unless they are manually recounted they do no good because the final tallying takes place inside a proprietary black box.


We have 8 states where the early exit polling failed to match up with the outcome (1 of them, Colorado, where Kerry actually did better than the polls predicted, but not enough to win). Of those states, which used this system, and how widespread in the states was its use?


That's the thing. That data has been impossible for me to track down, except for isolated claims that 80% or so of all votes are tabulated by ES&S or Diebold computers. The details of voting machines are everywhere, but no one I've found has listings of how votes in each area were tabulated.


Possibly not, but why not show the link, anyway?


A fair request, and one fulfilled above.


Um. You made the claim that this website was non-partisan. Your assertion. You felt confident enough to state that affirmatively. I don't even have information that the owner(s) is/are journalists at the moment; I did not spend a lot of time looking at her/their background -- I was hoping you already had it.

I was asking if you were sticking to that assertion because of the owner's ownership of a publishing house which has apparently put out books tying Bush and the CIA to a 9/11 conspiracy. An implied question was what made you assert so strongly that the journalists who ran the site were so definitely nonpartisan.


I think you confused publicity with publication. She co-owns a publicity firm that represents the authors of said books. Without having read the books myself I can't comment on them, and the wiki doesn't link to anything concrete.


What "available evidence" are you referring to? Do you have anything other than what they (she) say (says) on her web site?


That and the lack of anything saying otherwise, yes. What standards of proof for non-partisan status are you after?


Is this a method of insinuating that anyone that doesn't accept these conclusions unreservedly must feel that ensuring free, fair and open elections is a partisan issue -- or indeed, that they must oppose free and fair elections?

No. Even the freepers get it.

http://www.freerepublic.com/focus/f-news/1205221/posts

I'm just saying that focusing on eliminating every possible source of error in the exit polls should be a priority but far from the highest priority. The highest priority should be auditing the voting system, because even circumstancial evidence of large-scale electoral fraud should be treated as a national crisis, and if the problem does lie in the electoral machinery then the fastest way to expose it is to do a thorough audit.

I'm not saying you should storm the White House, I'm saying your should be demanding an immediate and thorough recount and investigation, because that's what the available evidence warrants.

[DaveW]

Originally posted by Kevin_Lowe

The wikipedia people seem very organised and very cautious about putting out anything other than sourced facts, and they seem to think it's a solid claim that the big exit poll discrepancies are all in the swing states, and all in favour of GWB.

Yup, which is probably why they have a nice yellow box with notes stating that the information in the article is under dispute, and that the accuracy is not considered verified as of yet.

[NoZed Avenger]

Originally posted by Kevin_Lowe

Exit polls went amiss in 2002, when electronic voting machines were beginning to be rolled out? I think you can see how that could be taken either way.

Not unless you have something that ties the inaccuracy to the places where the voting machines were or a showing that the outcomes in 2002 are somehow suspect, no. I haven't seen anyone make that case.



Let's see.

This covers a lot of the ground.

http://www.motherjones.com/commentary/columns/2004/03/03_200.html

This covers the GEMS functions facilitating election fraud in more detail, and shows the extent of the planning that has gone into GEMS system.

http://www.ejfi.org/Voting/Voting-30.htm



Mother Jones?

I'll try to look at anything they cite, but I have seen previous predictions of dire right-wing conspiracies from Wasserman before, so I'll admit up front I have some doubts about your source here. IIRC, Wasserman predicted that the end of democracy in America some 4-5 times now.

A review of a fan on his book on US History, for example:

". . . a superb account of the truth of America's past, with all the lies and evils of capitalism and the genocide and war and corruption necessary to make obscene amounts of money for the filthy rich. Emphasis on the word filthy."




Once you've read the above, and verified any details you find questionable, I think you'll be less likely to dismiss the possibility of fraud.

I wasn't aware I had dismissed the possibility. I have questioned what I felt and feel is a premature conclusion regarding the likelihood of fraud.


t looks a whole lot like certain people in the Republican party have been working deliberately towards the goal of controlling your elections for some time now.

I wouldn't say it if I didn't think the evidence supported the claim.

AS I have said, I welcome people that want to look into it -- but declaring it likely at this point still looks like you've come to conclusions based primarily on speculation and wishes.



The 2000 and 2002 problems look superficially like they might tie back to the same source: rigged elections. You'd like think that in the Land of the Free that would be impossible to get away with, but I'm not of the opinion that is the case.

I'd prefer not to cry a national conspiracy for rigged elections based on what anything superficially looks like. Maybe I'm overly cautious.

[snip]


That's the thing. That data has been impossible for me to track down, except for isolated claims that 80% or so of all votes are tabulated by ES&S or Diebold computers. The details of voting machines are everywhere, but no one I've found has listings of how votes in each area were tabulated.

I've had a short look and come up empty. I'll try to make some time to do a more thorough search.



[
That and the lack of anything saying otherwise, yes. What standards of proof for non-partisan status are you after?

Personally, I'd want more than that before I assured everyone of their nonpartisan status. It is probably my mistake for inferring too much from your statement, but I thought you were affirmatively vouching for their nonpartisan status -- which seems to be based on nothing except her web site.

As for the rest, I am happy for recounts to be made and welcome any investigation. I'd like to see some scrutiny of the exit polling at the same time, however.