On Voting Machines

by: Chris Bowers

Mon Oct 20, 2008 at 15:31

Cries of "they stole the election through voting machine fraud" are a pet peeve of mine. To show why, I conducted a study of final week polls in the 42 Governor, Senate and House elections since 2004 that were decided by 10% or less. You can see the report here, and read the methodology here. These were among the findings (more in the extended entry):
Chris Bowers :: On Voting Machines

  • Average polling error from final polls to final results: 2.0%
  • Number of states where polls favored a candidate other than the winner: 1, Wisconsin 2004 Presidential shifted from Bush 0.7% to Kerry 0.4%
  • Average partisan shift from final polls to final results: Democrats +0.5%
  • Largest pro-Republican shifts: VA-Pres (-4.2%), FL-Pres (-3.4%), AR-Pres (-3.3%), AK-Gov (-3.0%), MT-Sen (-3.0%). NC-Sen (-2.5%), MO-Pres (-2.2%). No other pro-Republican shifts greater than 2.0%.
  • Largest pro-Democratic shifts: KY-Sen (+7.0%), MD-Sen (+5.75%), MD-Gov (+5.5%), RI-Sen (+5.0%), NV-Pres (+3.7%), NV-Gov (+3.0%), WA-Pres (+2.2%). No other pro-Democratic shifts greater than 2.0%.
  • Most accurate polling: OH-Pres 2004. Yes, that's right--Ohio 2004 was the most accurate polling of these 42 averages. There was a difference of 0.0% from the final polls to the final result.

In other words, the research I have conducted on my own actually suggests a slight pro-Democratic shift from final polls to final results in close statewide elections. This doesn't exactly instill me with a lot of fear that a Republican conspiracy has used voting machines to shift elections in their favor. In fact, I don't see much of a conspiracy at all, given that two-thirds of these campaigns ended up with results that were different from final week polling by 2.0% or less. That actually instills me with a lot of confidence in final week polling averages, rather than a lack of confidence in voting machines. As such, it serves as the rationale behind my presidential forecast methodology.

Now, I am completely in favor of election reform. In fact, along with the Employee Free Choice Act, media reform, and immigration reform, I have repeatedly identified it as one of the four "progressive positive feedback loops." Those who raise concerns about election error and election tampering in order to either pass legislative reform and / or support election protection measures are engaged in solid issue advocacy and / or election activism. Good for them. I support their efforts whole heartedly.

However, those who only selectively claim fraud in cases where their candidate lost, those who offer up voting machine fraud as a singular rationale for progressive and Democratic defeat, and those who fatalistically raise these fears without offering a legislative and / or activist solution to the problem--well, I just find stuff like that annoying. I certainly understand people's concerns, what with the many behind the scene power grabs by Republicans over the years, but in this case I am pretty sure people are barking up the wrong tree. There absolutely are widespread Republican attempts to suppress the vote, but as far as pro-Republican voting machine fraud goes, I just don't see it.

Tags: , , , , , (All Tags)
Print Friendly View Send As Email

On Voting Machines | 35 comments
Ahhh, thank you (0.00 / 0)
You know how when you have an ice-cream-headache, and then it goes away, and it feels so great that it was almost worth having the headache just so you could feel the relief of having it gone?

I've got that feeling right now.

Conduct your own interview of Sarah Palin!

Many thanks for this post. (0.00 / 0)
Statistics are a bitch in instances where one would rather believe in anecdotes and conspiracy theories.  

Not to get all conspiracy theorist... (4.00 / 1)
But aren't final pre-election polls and the actual vote still completely different things?  In other words, just because the final vote came close doesn't necessarily mean that there wasn't funny business going on at the polls. Maybe the Democrat was going to out-perform the polls but the cheating brought it back to where the (incorrect) polls said they were.

Now, I know this is sort of a ridiculous, backwards argument... But, in 2000, Gore did significantly outperform the pre-election polling, and only because of some bullshit going on in FL (along with a bullshit Supreme Court decision) did Gore end up losing the election.  Couldn't it just be that the Democrats are, for whatever reason, usually over-perform the polls, but then whatever form of vote tampering has pulled them back in?

And no, I don't think that the election was necessarily stolen from Kerry from bad vote-counting or weird vote-switching, but it does seem like there was a substantial amount of voter disenfranchisement that may have cost him the election.  

voter disenfranchisement is completely different from vote-rigging (0.00 / 0)

[ Parent ]
Right... (0.00 / 0)
I'm just pointing out that comparing final pre-election polls to the actual vote may not necessarily be "proof" that tampering isn't occurring.

[ Parent ]
Sure (4.00 / 1)
But you can't prove a negative. "No fraud" is the null hypothesis.

Conduct your own interview of Sarah Palin!

[ Parent ]
Why is "no fraud" the null hypothesis? (0.00 / 0)
Has there ever been any large elction where there hasn't been fraud? Remember Nixon's ratfucking? The shenanigans of Tammany hall?

Really, has it ever been proven that any presidential election has been 100%, or even 99% fair? Is this really PLAUSIBLE? Imho it very well might be that fair elections are the exception, not the rule.

So, instead of falsely working with an unproven null hypothesis here, it would be much better to try to calculated the probable amount of vote fraud, to see if this could turn the election around. And, remember, this isn't only about machine fraud. There are unscrupuols right wingers out there who use every trick in the book, and even event new ones on the fly.

[ Parent ]
You can't ever really "prove" something is true (4.00 / 1)
only that it is false.

So no, you can't prove that any election has been 100% fair. But that's also not Chris's point. His point was that the evidence we do have indicates that the effect of whatever vote-rigging there is, is very very minor.

And I would suggest that instead of idly calculating the probable amount of vote fraud, it would be a better use of one's time who is concerned about vote-rigging, to go out and volunteer as a poll watcher for the Obama campaign.

[ Parent ]
I was answering to fwiffo (0.00 / 0)
and he brought up the "null hypothesis", which can't be used here, ok?  

[ Parent ]
really? (0.00 / 0)
I'm pretty sure I've definitively proven myself bad slam dunking.

[ Parent ]
Because you can't prove a negative (4.00 / 1)
I don't know how many times I have to repeat myself.

You can't prove a negative.

"Fraud" is "something" and "no fraud" is the absence of "something". You can't prove a negative. That's why "no fraud" is the null hypothesis. There has to be some evidence to suggest there was fraud of a scale sufficient to change the outcome of a race. You can't have evidence that there wasn't anything.

You don't just assume there are invisible pink unicorns until somebody proves they don't exist.

If you're doing a drug trial, comparing a group that takes a drug against a control group that takes a placebo, and there is no difference between the groups, you assume the drug did not work. You don't just say "hmm, the drug probably worked, but all the patients must have lied about taking their pills."

To accept the notion that there was widespread fraud (sufficient to change the outcome of a race), but we didn't notice because the polls by many different firms were all wrong in the same direction is preposterous. It requires us to accept both of the premises without evidence. You have to come up with some evidence for at least one of those assumptions before it becomes plausible.

Again, nothing here recommends not being vigilant, just being realistic about what we actually know.

Conduct your own interview of Sarah Palin!

[ Parent ]
It's just another way of manipulating the election (0.00 / 0)
Imho it's wrong to see this as an unrelated issue. All tricks are used in widespread efforts to steal the election.

[ Parent ]
Ok. (4.00 / 1)
So go out and volunteer to help stop or minimize that manipulation.

[ Parent ]
This would be inappropriate (0.00 / 0)
As a foreign citizen, I shouldn't take an active role in the election. Afaik there are even laws prohibiting this. And rightly so.

[ Parent ]
No it's not. (0.00 / 0)
It's inappropriate to vote as a foreign citizen and to donate money as a foreign citizen, and there are laws prohibiting this, and rightly so. But I know of no law stating that you can't volunteer your time to help out in some capacity. Being a poll watcher may be problematic itself since you'd have to be a voter to challenge a republican challenge. But there are still many many things you can do to help. I just don't think harping on how the election is going to be stolen in blog comments is the best use of your time if you truly want to do something about it. Unless, you don't really want to do something about it, but want someone else to do something about it, or think that nothing can be done about it.

[ Parent ]
Interesting. However... (0.00 / 0)
this is not my election, so don't expect me to invest time, and more important, money on it. But still, if this discussion results in even one participant here volunteering, I think it was worth my time.

One other thing:
I noticed you trolled every single one of my recent comments with adding the very same proposition you added here. Pls note that I will rate every single one of these comments with a troll rating , ok?

[ Parent ]
I was just trying to get you to respond once (0.00 / 0)
since you seemed so concerned about expressing concern rather than about going out and doing something about it.

[ Parent ]
That's not an argument (4.00 / 1)
That's like saying gravity is caused by pink unicorns, but you just can't prove it because the pink unicorns are invisible.

I understand that an absence of evidence isn't evidence of absence, but it's still an absence of evidence.

In Florida 2000, the shenanigans were there for all to see.

Conduct your own interview of Sarah Palin!

[ Parent ]
Hmm.. (0.00 / 0)
This is one way to look at it, though I'm not sure it's the same thing.

I'm not trying to say that there is definitely fraud or whatever, just that I'm not sure that pointing to pre-election polls really tells us all that much about the amount of fraud in the system.

On the other hand, I suppose if I don't want to accept that the pre-election polls are at least a rough indicator of actual support, I suppose there's no point in even following the polls at all. =)

[ Parent ]
That's an important question: How did 2000 influence 2004 polls? (0.00 / 0)
Did the pollsters adjust their methods with data generated from the 2000 election? If this is so, every vote manipulation that happened in 2000 has distorted the 2004 polls. And this systemic error is becoming a self fulfilling prophecy now.

[ Parent ]
Of the last 12 polls in Ohio (0.00 / 0)
10 showed a Bush lead (and one of the two showing a Kerry lead was an online poll).

The bottom line is that the machines were supposed to eliminate undervotes by making them effectively impossible.  In this sense they were superior to optical scanners.  I think a paper trial is absolutely critical, and because the machines were expensive they resulted in longer lines at the polls, and so I am not sad to see them go.

But I have never seen any evidence that they tilted GOP.  In fact, it was marginally the reverse (which you would have expected).

For a review of all state polling in 2004, go here for the excellent analysis from the national council on polling.

Paper trails only inspire false confidence (0.00 / 0)
All experiments show that real voters much too often don't care to check the recipes. And even if they do, and find a discrepance, officials tend to explain this away with voter error. Imho optical scanners are the only reasonable compromise to manual paper ballot counting. And even optical scaniing should be checked for accuracy in every election, in randomly chosen districts.  

[ Parent ]
Let me add to the thanks on this as well (0.00 / 0)
I figured this was one of Chris's hot buttons based on a reply he made to me last week about the Cleland vs Chambliss race so it is nice to see him dig in deeper into the polling.

I'm curious (0.00 / 0)
as to why you used the final polls at a reference instead of the exit polling information.  Or don't you have access to the exit poll data?

Anyways, I really like to see the US as a whole move towards the Oregon 100% vote by mail system.  Which isn't prefect, of course, but is cheaper and does seems to work just fine and have few if any problems.  Maybe an Obama Administration can "encourage" all states to adopt that system by 2012.

Don't know if this is a partial reason why (0.00 / 0)
but there were definitely problems with the exit polls too. One of the more notorious examples was how they way over-sampled african-american voters in NC.

I agree though we need real reform. I think making election day a nation holiday will be a big help.

[ Parent ]
what about transients? (0.00 / 0)
Its like addresses are given the vote instead of people. Maybe that's okay.
Also: I hate all this early voting. Its like the critics are walking out before the show is over.

[ Parent ]
However, the vast majority of vote shifting incidents by machines... (0.00 / 0)
were in the favor of the GOP. Actually, I don't remember EVER having read about ANY instance where it has been different, but, of course, it's possible. Remember, there was just another example where this happened in early voting in West Virginia. And, OF COURSE, once again the shift was from Dem to GOP votes:

There have been several demonstrations how easy it is to manipulate the machines. There are still lots of incidents where this equippement isn't locked away before the election. And the actions of the manufacturers don't inspire confidence, either -some are outright illegal.

Well, when it's possible to manipulate the vote this way, and the measures to prevent this aren't even remotely adequate, it's very farfetched imho to categorically deny that this doesn't happen.

And then, there's also the purging of voter rolls, disenfranchising voters by deliberately punitive ID reqirements, the spreading of false informations at universities, in working class and minority neighborhoods, falsified robo calls, the challenging of voters by GOP election "observers", and every other trick the rethuglicans have come up with so far. And all of this is supposed to have only a negligible impact?

Maybe. But maybe the real truth is that this is already going on for so long that it is already incorporated in the polls. So, the polls don't show the real voter intent anymore, but the voting result after all this ratfucking has come into play. How about this for an explanantion?    

there are other ways to distort the vote (4.00 / 2)
They may not have rigged the voting machines in Ohio but they sure did make it a lot harder for people to vote in more Democratic parts of the state. I remember a friend talking with his sister in college there in 2004 and they were still waiting in line to vote at midnight.

If you don't want poor folks and college students to vote, you just have to "underestimate" the number of polling places and voting machines needed in those areas, while making sure there are plenty in parts of the state who will vote for your side.

Rigging the computers is more complicated. Making it harder to vote is easier.

Yes (4.00 / 3)
And that's where our efforts need to be focused. Which is why everybody should go to their local Democratic campaign office and volunteer to help watch the polls.

Conduct your own interview of Sarah Palin!

[ Parent ]
Odds of Bush win in Ohio '04 1/9600, at best (4.00 / 2)
I thank you for broaching the topic, but my conscience compels me not to thank you for the brevity with which you are treating the topic.

According tNational Election Data Archive at uscountvotes.org (NEDA), the odds of Bush winning Ohio in '04 was no better than 1/9600, and as high as 1 in 16 million.

I dare you to contact NEDA, and run your argument past them. You might convince me if you got the mathematicians and statisticians at NEDA to agree that the poll results you mention somehow are of greater significance than exit poll data. Until you do that, you're not going to convince me that the Presidential election wasn't stolen in 2004, via Ohio.

The Gun is Smoking: 2004 Ohio Precint-level Exit Poll Data Show Virtually Irrefutable Evidence of Vote Miscount

Six Percent of Ohio's polled precints Show Virtually Impossible Vote Counts, and Over 40% Show Improbable Vote Counts, Given Their Exit Poll Results. The Patterns Of Ohio's Discrepancies Are Consisten With Outcome-Altering Vote Miscounts.

Ohio's exit poll discrepancy pattern includes three precincts with virtually impossible outcomes and an unusually high number of precincts with significant discrepancy.1
· 6% of Ohio's precincts each have virtually zero chance (less than one in 15,000) of occurring
due to sampling error, given their Kerry official vote count.2 Even if the "within precinct
discrepancy" (WPD) is adjusted for all the precincts to remove any possible effect due to Kerry
voters completing more exit polls, the probability of obtaining Ohio's exit poll discrepancies are
virtually impossible.3
· Over 40% of Ohio's polled precincts have discrepancies having less than a 5% chance of
occurring, given the official vote counts.4 The expected number of such precincts in a sample of
49 precincts would be five such precincts, not the 20 found.
· Ohio's exit poll discrepancies, when plotted against precinct exit poll share show a pattern that is
consistent with vote miscounts that benefited Bush, and
· The pattern of Ohio's exit poll discrepancies cannot be explained by random sampling error or
partisan exit poll completion rate differences.


Ohio's precinct-level exit polls over-estimated Kerry official vote by an average 5.8%. The overall
discrepancy between exit poll margin and official vote margin was double that, 11.7%.
On June 6, 2005 The Election Sciences Institute (ESI) with Mitofsky12 released a report on the Ohio
precinct level exit poll data purporting to rule out vote fraud as the cause of the discrepancies. The (ESI)
report entitled "Ohio Exit Polls: Explaining the Discrepancy" by Susan Kyle, Douglass A. Samuelson,
Fritz Scheuren, Nicole Vicinanza, Scott Dingman and Warren Mitofsky, concluded:
"...the data do not support accusations of election fraud in the Ohio Presidential election of 2004".13
ESI's premise is that if there were vote fraud, then the 2004 exit poll discrepancy would be correlated
with Bush vote share increases from the 2000 election. Finding no such correlation, ESI ruled out vote
fraud as an explanation of the exit poll discrepancies.14 ESI's method of exit poll analysis was included
on October 14, 2005, in a presentation by Warren Mitofsky to the American Statistical Association fall
conference in a talk entitled "The 2004 U.S. Exit Polls".
In an October 31st paper, NEDA mathematically proved that ESI's and Mitofsky's analyses were
incorrect because many counterexamples exist to its basic premise.15 In other words, NEDA proved
mathematically that ESI's and Mitofsky's analysis of Ohio's and national exit poll data is of no
analytical value and no conclusions about the presence or absence of vote fraud can be drawn from

The ESI report had made no attempt to explain or mathematically analyze the actual 2004 exit poll
discrepancies and the ESI report was missing key data. To date, Mitofsky and ESI have provided no
explanation for the exit poll discrepancy that is supported by data and analysis

A history of election auditing (1975 - 2008) is here: http://uscountvotes.org/ucvAna...

Also, there's enough up-to-date coverage of vote fraud issues at bradblog.com to make anybody puke....

If it makes anybody feel any better, a Princeton history professor that I was friendly with told me that traditionally, it's been the Democrats who were associated with vote fraud. Personally, this doesn't make me feel better, at all.  

435 Dem Primaries 2012
Coffee Party Usa

StealBackYourVote (4.00 / 1)
What do you think of the big StealBackYourVote(.org) push by Greg Palast and Robert Kennedy? They produced a comic book to distribute to vulnerable voters, complete with agressive fundraising, which I fell for to the tune of $20. They published it in the latest print issue of The Nation.

Its full of scary statistics about lost ballots and purged voters.

Block the Vote (0.00 / 0)
Here's a link to an article by Palast & Kennedy on the subject in the new Rolling Stone - Block the Vote:


And here's Kennedy's article in Rolling Stone from a couple of years ago - Was the 2004 Election Stolen?


I'm not thrilled with people voting 3 weeks before election day.  But until we fix the problems with not enough voting machines in Democratic leaning precincts, and hundreds of thousands of eligible voters getting illegally kicked off the voting rolls, early voting is the best we can do for now to try to offset these attempts to steal our elections.

[ Parent ]
"Analyst: GOP voter fraud scandal 'really serious'" (0.00 / 0)
The switches could impede Democratic get-out-the-vote efforts and could prevent those affected from voting in next year's Democratic primaries. Some of the voters were even switched to absentee status, meaning they could lose their vote entirely if they show up at the polls on Election Day without bringing an absentee ballot.


This begs the question: Why California? McCon can't really hope to turn that state around, or can he? Is this designed to help rethuglican congressmen keep their seats?

I agree the shrieking is not helpful, but small problem with your numbers... (0.00 / 0)
While shouting "election fraud" from the roof tops only serves to crush voter confidence and legitimize (just by adding to the noise) GOP/right-wing claims of "voter fraud," it seems like you are discounting swings of less than 2% and I don't know if that's wise.

Even a fraction of one percent can make a difference in some elections, especially combined with other voter suppression efforts (right-wing attempts to steal elections are very clearly multi-pronged attacks). In addition, these small percentages in several states in a national election can pad the pop vote numbers in a way to make people more "at ease" if the election really is stolen by any kind of behind the scenes outright fraud in any state.

I would also bring up that your findings did show a 3.4% swing in FL (and that is significant, with that state's history and the people who are in charge). AND I think the the exit polling factor is worth more exploration -- the narratives in the mainstream press that have become adopted as conventional wisdom to discredit exit polling should NOT be taken at face value.

The problem with the scare about vote-tampering/switching due to the intentional rigging of electronic voting machines or software is that it would take a conspiracy so massive to actually pull off in the way it's described that it is just plain unlikely. However, could it be be pulled off in several targeted districts or counties in just a few key states to ADD to an illegitimate victory for a state or federal candidate? I think the answer is probably yes. And could problems with voting machines swing an election one way or another, especially if those problems occur in machines in certain areas that lean one way or another? Well, look at FL-13 from 2006 and that answer is a clear yes.

Thanks for your good research. I just think that looking at states and even smaller voting jurisdictions and examining who is in charge there are important to the whole story....

On Voting Machines | 35 comments

Open Left Campaigns



Advanced Search

Powered by: SoapBlox