Slashdot Mirror
Berkeley Researchers Analyze Florida Voting Patterns
empraptor writes "Researchers at UC Berkeley have crunched numbers and determined that 130,000-260,000 excess votes went to Bush in Florida. They have held a conference and posted their findings online. You can find articles on their research from CNet, Wired News, and many other sources. While the research used statistical analysis based on past elections and demographics, how else do you verify that a paperless voting system is working properly?"
A. They neglect to factor in the "Hurricane effect." The President's visits and aid raised him popularity in the area.
B. They performed the same study on Ohio and found no irregularities.
I really do think that Florida went to Bush.
The question is Ohio. It has been a stuanch Democrat state. It lost 10's of thousands of jobs under Bush. And it voted for him in a close election? So why are these researchers looking at Florida?
I prefer the "u" in honour as it seems to be missing these days.
Not sayin' it is so... but HAD the election been accidentally given to Bush, now that Kerry has conceded, what would the legal recourse be??
Luck favors the prepared, darling.
You're nuts. Paper and pencil are NOT more reliable than computers. Haven't you ever heard the term "ballot stuffing?" Physical media such as paper are also fraught with security concerns. They boil down to the same thing as computers: do you trust the election officials running them? Who has physical access to the vote once it has been cast? Etc. I'm not saying they are the same, but c'mon, all of a sudden the old paper method is the gold standard? No way.
The trouble with voting security is that it requires authentication, anonymity and ability to verify later. The verification necessarily must be done by the voter himself, or else somebody else will know how you voted.
Here's my idea: after you vote, you get a random ID and password associated with your vote. Later, you can log onto a website and verify that your vote is as you cast it, without divulging your identity. Make the process for getting votes from the machine to the central data repository open-sourced, open, open open, totally so that we know exactly what is happening.
Hey, it's a start. But I'm in favor of these voting machines. Let's not throw the baby out with the bathwater.
Currently hooked on AMP
And who is going to watch to make sure the same exact code is going to go on the machine?
I sent this letter to the editor of the washington post a few days ago on the evoting topic (wasn't published)...
l es/A556 91-2004Nov16.html
re: In ATMs, Not Votes, We Trust
http://www.washingtonpost.com/wp-dyn/artic
I'm a programmer a major corporate bank in Manhattan.
Anne Applebaum's analogy of e-voting to ATM and credit card banking was misleading and uninformed.
Users receive regular bank statements, with each ATM transaction itemized.
Cross-checks of all transactions can be validated by the user through this method or
at any time with a phone call or with web access.
This is a paper trail.
For a credit card, it's the same deal, of course.
Again, there is a paper trail.
Increasingly e-voting machines have no paper trail requirement.
This is highly troubling.
Anne seemed to label this as "conspiracy", but it is no such thing.
To say so is irresponsible.
There is no way for the individual to verify that their vote was counted as they registered it, as you
can at an ATM, with or without a receipt. Do you find this troubling? I do.
This is just one short-coming in the system, among many.
As a computer programmer and security expert, I know how easily computers can be manipulated.
It is a fact that the coding on these machines could literally do anything.
We're irresponsibly putting our votes into a black box, and don't even have an audit trail.
This issue has nothing to do with whether fraud occurred in this particular election or not.
Glitches frequently occur due to human and machine errors.
An audit trail is a minimum necessary requirement -
And this is just the beginning of the problems with e-voting as currently implemented.
I'm surprised that the Washington Post allowed such a flimsy analysis to be published.
a quick scan of the paper reveals that they're saying that there were 130,000 abberant (for lack of a better word) votes.
If you want to think those votes are ghost votes (perhaps they would have gone for Nader) then subtract 130,000 from Bush. If you want to think those votes should have gone to Kerry than subtract 130,000 from Bush and add 130,000 to Kerry.
If you don't buy into their statistical modeling, then don't do anything. But isn't it curious that the largest disparity between expected and actual e-voting results occurred in heavily democratic counties?
It breaks my pluginses, my precious!
Uh, run a test? Before the election, vote for Kerry 50 times. Vote for Bush 50 times. Tally the results. If it's not 50 and 50, something is jacked up. It doesn't seem to be rocket science to me.
YOU DONT LEAVE WITH THE RECEIPT.
In a "paper trail" situation, the receipt is the ballot! The only purpose of the machine is to give the people who can't punch a hole properly a chance to have their vote count, and maybe you can plug all the machines in at the end and get a quick count, but in the end if something smells fishy, you pull those paper ballots out.
If the machine recorded a vote for A and printed out a vote for B, then this would be caught either when A wins by an unexpected landslide, or by chance by random sampling.
If I have been able to see further than others, it is because I bought a pair of binoculars.
Well said. I wish I had mod points.
In many respects I think the bottom line here is accountability. The problem therein is that you can't please all the people all of the time. You also not create a fraud free system, as we know with the hacker culture, rules/systems/processes/et al. are meant to be bent and sometimes broken, and anyone who has the desire to attack the system can do so with enough effort.
I do find it very ironic that we have two distinct crowds, largely both in the Democrat leaning arena which desire to challenge the election results. There are those who want to challenge the electronic voting and those who want to challenge the paper voting. Each group implies the other system is the better system. You can't have it both ways.
No one likes to loose, and I'm not trying to rub a nose is someone's defeat. There are a many people out there who supported Kerry and who have just as deep convictions for his agenda as I do for Bush's. An old coach of mine used to say loosing builds muscle (because after a loss he'd basically kill us with PT) and character (because it should teach you to have dignity in your loss and to make sure you work hard enough not to loose the next time). I sincerely hope for the Democratic party that this election loss does both.
What I don't like about paper audit trails in electronic voting machines is that everyone thinks they should be printed out in real time, like a cash register receipt at the grocery store as each item (voter) goes past. That makes it rather simple to match up voters to their votes if someone wished, and remove all the protections of the secret ballot process. Are you concerned?
And I do find it curious that voting machines are only being questioned in states that Republicans have won. Don't you?
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
The way I would think you could come close is.
1. Put the software on bootable CDs.
2. Ship an excessive number of CDs to each polling location.
3. Boot the machines using CDs choosen by random voters.
4. Allow voters to take home and verify the excess CDs
Appearently an election watcher went to collect official poll tapes from one florida county and found the staff throwing them into the trash. They compared some of them on live tv with the ones certified by the state and found they were different. This should get very interesting....
That's pretty good, but it still doesn't guarantee that the machine's firmware isn't tampering with the code after it's in the machine.
No no no....The main advantage to an e-vote is that it gets counted automatically. When you have an e-vote with a paper receipt, you can vote electronically, the number gets beamed to a central DB, AND you get a paper ballot you can visually check and put in a box. Sure there won't be any faster recounts, but once the technology has been around for a while, people will trust the data. The whole state of Nevada was like this during the last election, everything got counted fast, and if there is any questions about the outcome, it's easy to double check by counting the receipts.
The REAL question is why are there electronic voting machines that DON'T have a paper trail?
"If it sucks without butter, it still sucks with butter, only creamier." - AC
> That's pretty good, but it still doesn't guarantee that the machine's firmware isn't
> tampering with the code after it's in the machine.
You could allow each of the N parties to supply Required/(N+1) PCs to run the voting software, and the electoral commision would supply Required/N+1 PCs of its own. The software records which machine was used and a firmware hash of some kind on the voter's receipt.
Then statistical analysis could be used to determine if one or both parties are cheating; it will be easy to detect and VERY hard to do.
You could also hash the running total on the voter's receipts, along with timestamps. That might also prove to be very interesting.
Do daemons dream of electric sleep()?
Who cares about the code. It is data that should be signed and countersigned at every step and travel by at least 2 or more parallel pathways which are crossverified as well as the signatures. I am sorry, but 5th world country like Bulgaria has been doing this for nearly 10 years now. India has done it in the last election. It is time for the US to actually get a clue and learn how to run an election or import the Bulgarian or Indians who designed the election data flow (note the architecture, not the code) for a short H1B stint.
Baker's Law: Misery no longer loves company. Nowadays it insists on it
http://www.sigsegv.cx/
* Unless those trails are voter verified, nobody has any way to determine if the trail matches the actual votes cast by the voters.
This is the core problem with electronic voting.
We either need to put the actual vote on paper, or make sure the machine printed votes match voter intent, or the election cannot be trusted.*
you use a machine to vote. the machine spits out a paper with your vote number into a transparent casing. you verify (if you're not lazy) that the number is the same that you chose on the machine and extract the paper with the correct number(s) and then take it into a normal ballot box. the result is computer readable and later verifyable by hand if necessary ballot(computer readable because it was computer printed).
either way, the voter should be able to verify that his vote has the right markings before going into the ballot box.
bad motives for a totally paperless ballot are way too bad to accept(too many what if's even by ACCIDENT not to mention the situations possible by intent).
world was created 5 seconds before this post as it is.
I disagree. At least as far as 'important' only applies to my lifetime. While the prospect of voter fraud and election buying is certainly a dreadful prospect, the possibility of multiple Bush appointed Supreme Justices ruling on my civil rights for a decade or two seems to be something which will have a far greater impact over the next 60 expected years of my life.
No, no, no. Sure, some people think that the different system is better, but most people who have problems with the system do not. Even people who have a problem with electronic voting machines do not want to get rid of them, they just want to make sure that they *work*, instead of giving their vote to someone else. Paper ballots aren't too good either. When you have a candidate *surprised* about how many votes he got in a country, and many people say they accidentally voted for that candidate, you can't say they're just trying to change their vote, there really *was* a problem.
Hurricane Ivan: A 17th century prison collapsed. All of the inmates escaped.
A statistical discrepency can have multiple causes.
1) There was something fishy about the election results -- this seems to be the assumption behind the investigation, the researchers wanted to find something and they found it.
2) The statistical model is invalid -- There may be other factors that weren't considered, this is where peer review would help.
3) The data is poor -- The data in this case being previous election results and results from non-electonic voting precincts. So it is possible (though probably not likely) that it could be the other counties where something isn't quite right.
I think the main legit advantage of electronic voting is accessibility. See Electronic Voting - Overview and Issues.
A Quote:
This is exactly how we do things in Canada, by the way, and it works wonders. Of course, TV-wise, it is much less interesting: we have a projected winner one hour after polls close and have the final and definitive results for all counties two hours after that. A party can ask for a recount without being accused of "hurting the country" and said recount happens in days. On the up-side, we do save money on exit polling. When the electoral system works, who needs exit polling?
Here's what he found:
Objection: assumes facts not in evidence.
Bush won.
Not yet.
(I didn't vote for Bush.) Get over it.
Read the latest news regarding Florida machinations.
You are being MICROattacked, from various angles, in a SOFT manner.
Uh... well, do the math (if these figures are right).
Bush wins by 400,000
subtract (possibly) 260,000
Now he's only 140,000 ahead...
Add those 260,000 votes to Kerry's tally...
and you have Kerry winning the state by 120,000 votes.
So... if this analysis holds up under scrutiny (which I doubt it will), it definitely could have affected the election.
Another possible cause is that the Democrats were cheating in 1996 and 2000 and the use of electronic voting machines eliminated the cheating. The existence of a correlation doesn't elucidate the cause of a correlation.
Anon..
Yes, I know you are joking, but how does it feel to be a robot? I've read several articles that predicted jokers like you would come out and make light of the fact that the study came out of Berkeley. If you are not aware, so did BSD. Is BSD some flakey project built by partisan nutcases? NO, and your perpetuation of a stereotype does nothing but muddle the truth.
LS
There is a fine line between being a cultivated citizen and being someone else's crop. - A. J. Patrick Liszkie
I have been a scrutineer at the past two Canadian federal elections, and the parent is correct. There were 2-4 people counting votes from each of six ballot boxes. We had everything counted, recounted, and called in 20 minutes after the polls closed. Technology is great, but it does not need to replace everything.
In 2002 the governer had a huge lead in the polls and was defeated. Also, a popular Vietnam vet incumbant (who left most of his limbs in Vietnam) was defeated for Senate, by a draft-dodging empty suit. Again, polls showed a commanding lead for the incumbant.
Another thing which should be considered: If there is a certain flaw in the voting system (e.g. puncher working statistically worse on one side of the ballot), the ones in charge of the voting process can utilise this fact to favour their candidate. The machines don't even have to be rigged, just be imperfect!
I don't know if Diebold has changed their system since, but I can't believe that Diebold has been claiming that the software used for voting machines is proprietary and secure. Without MS access, it makes absolutely no use.
Got democracy?
What's to stop them from changing the code on enough of the machines to win? We'd never know what happens after we inspect the code. In the right area they COULD possibly win with only a handful of doctored machines.
It is already certain that vote fraud occured in an alarming number of isolated cases. The only question now is if it occured and went undetected in enough places to actually swing the election. Here are a few of the things we already know for certain:
In several districts, electronic voting machines were preloaded with thousands of votes for Bush before the election started. Where it was discovered, the machines were reset and did not effect the outcome. The question is, in how many districts did this go undetected because voter protection advocates were not there to check the machines.
In at least one case, a location in which only about 600 people voted recorded over 4000 votes for Bush. No explanation has been given for this, though it is likely another example of 'pre-loaded' machines.
In at least one local election, a manual recount of the ballots swung the vote total by a large amount compared to what the electronic vote machines had reported, enough to move the winner from the republican candidate to the democrat.
But the biggest smoking gun is in Florida's Volusia county where election offitials were caught red handed throwing out the official signed poll tapes from Nov 2nd. When these tapes were compared to the reported vote numbers, they showed that votes had been added to Bush's total IN EVERY SINGLE PRECINCT EXAMINED. If this was done in many more Florida precincts, it could explain the eight point swing between the exit polls showing Kerry winning and the official tally showing a Bush win. We must at least acknowledge the possibility, and insist on a full audit of the Florida results... not just a recount done by the same Florida partisans, but full, impartial audit.
The Bolachek Journals
They won the presidency, they emasculated the democrats (won key senate seats, tom daschle, etc).
I happen to live and vote in South Dakota, and while I don't like the Diebold machines any more than most people on Slashdot, you cannot blame Daschle's loss to them. (well you could, but you'd be wrong) Guess what we used for voting? Good old number 2 pencils and paper. They showed the counting machines on the news the night of the elections and they're essentially the same type of machines that ACT uses to score results on their tests. The precincts send their paper ballots in to the central counting location (in my case the county courthouse), the workers put the ballots in the counter, and voila! As for Daschle losing, I can't explain that one to you. You'd have to ask the other voters...
echo $SIG
That's kind of interesting.
I copy-pasted the numbers from that page and ran a simple query against them. In most counties, the turnout was less than the number of registered voters, but there are 30 accounting for 97,489 mystery voters -- more votes were counted in those precincts than there are registered voters.
This isn't just a matter of absentee ballots being put in the wrong category, nor is it minor double-counting of ballots. Here are a few excerpts (check 'em yourself!)
HIGHLAND HILLS VIL has 760 registered voters, yet counted 8822 votes, for an overage of 8062 votes, or a 1161% voter turnout rate.
WOODMERE VIL has 558 registered voters, yet counted 8854 votes, for an overage of 8296 votes, or a 1587% voter turnout rate.
Note that I'm only looking at the cases where num_votes > num_voters. If you plot the voter turnout percentages, MANY more precincts show an abnormally high turnout rate, just less than 100%.
Now I'm off to see if Colorado posts the same raw numbers online.
--
Conspiracy crap? A good percentage of liberals I know are very uneasy about the choice of companies that created these voting machines.
Here is a test. Next 4 years, we can choose our companies to build the machines and to count the numbers. Michael Moore, and George Sourros will head the companies. Does that make you feel comfortable? Don't complain if somehow Barbara Streisand wins California, You just have to Move On.
Oh. and just because you can site an example where the Republicans didn't win, when they've had a great showing of blithering failures (oh, the economy, pollution, the rising cost of healthcare + anything else I'd bother to mention), does not mean that they didn't try to cheat.
The Libertarian you mention may actually be pushing the same NeoCon agenda that has worked so well for Mexico. I don't want to get into that debate, but having been a Libertarian and a Republican for I while, I had to leave because their economic concepts were not sustainable, and the Dems looked the least evil by a smidgen.
But I also live in Georgia, which is the Belt Buckle of the Bible Belt, so no amount of self interest or reality will outweigh a good rhetorical moralizer. And the ignorance of people listening to Neal Bortz and nodding to his ideas of a Value Added Tax are making me want to retch.
By the way, some months ago, the president of DieBold publicly stated that he would do everything in his power to see that President Bush was re-elected.
Can you not admit, that a system where elected officials approve the budgets for private corporations who control who gets elected IS a system that is bound to be corrupted? What are we paying for these boxes anyway? About $100k a piece? Doesn't that mean that most of the expense is for "services rendered".
And note, that in 2000, the Florida Government payed the people who conducted the voting about 10 times as much as 4 years before. The number of rejected voters went from about 8,000 to over 90,000. It has now been verified, that many of the people who were rejected was unwarranted (and of course, mostly from Democratic voters). I could point to a number of articles discussing this, but you would not be convinced.
Why are people so dead set against an idea of a "conspiracy." It is damn well profitable to have a president give taxpayer money to corporations. It is worth Billions. And we have many examples of overpaid contracts to look at. There are all sorts of conspiracies. But it seems that anyone pointing it out is automatically a nut. So what does anyone do about a conspiracy? Hand the crooks the keys and hope they run over a school bus full of kids on prime time news so that we can be sure they are the bad guys?
I'll say it. I think the Bush administration is a bunch of crooks. They behave like crooks. They act like crooks. They want everything secret and they punish anyone who criticizes them. They were conveniently incompetent on 9/11 and it has done nothing but give them a green light to push through their agenda. They have pandered to just about every corporate supporter, in historically cynical ways. They have lied and said Iraq was an immanent threat. Oops. Now we must forgive them because it is a tough job. Meanwhile, Billions of dollars of taxpayer money are going to companies owned by the Carlyle group, which has financial dealings with almost all of the Bush administration (Halliburton ain't half of it). And we are supposed to shrug that off because it's only coincidence that it's their pockets the money lands in "hey, it could happen to anyone".
Wow, the energy bill even indemnifies oil companies from lawsuits they might incur over gasoline additives. OK. The future looks bright. King George will start the "No two-headed baby left behind" program. Retraining as a circus freak can help a large portion of the genetically damaged. Good thing they can't sue.
And all 5 of the electronic voting companies have been major donators to the Reelect Bush fund.
This statement; f
>>"ad space available -- low rates!!!"
Has anyone done a comparison of the battleground states won by Bush and whether they used electronic voting machines or not?
"Anyone that has ever gotten an idea based on any of my work and done something better with it-good for you."--J.Carmack
Actually, up until the 2000 election, exit polls were remarkably accurate.
They are still remarkably accurate -- if done for elections held anywhere but in the U.S. Magic.
And if exit polls no longer work, statistically the variant outcomes should scatter for Bush and Kerry roughly equally. They do not. They all skewed way, in some cases REALLY WAY, over to Bush.
And something is definitely wrong. Check Bev Harris's work these past few days. In Florida, she was issued unsigned audit tapes in response to her requests for evidence after the election, rather than the signed and verified ones.
After being denied the originals, she actually found them in the TRASH. Police were called to stop her, but she got the tapes.
Kids, they compared the unsigned results to the actual, disposed-of results from the dumpster.
The copies she was given do not match the originals. The vote was way, way adjusted for Bush. In. Every. Case.
She won't make the conclusion outright, but it's obvious. Where Jeb could cheat, he did. Mygod, how could he NOT cheat??
Long Answer: The purpose of electronic voting machines is not to provide an inexpensive election - paper ballots counted by hand are the cheapest way to run a secret election - nor is it to provide a guaranteed accurate election - paper ballots with check marks are the gold standard for proof of who voted for whom - but to allow undetectable election fraud. Any election without a real-time unchangeable audit trail - which means a paper log of every vote generated at the same time as voting is done - should be presumed to be intentionally fraudulent.
Back in 1966, Robert A. Heinlein gave the exact formula in The Moon is a Harsh Mistress for stealing an election without the public realizing they'd been robbed: have all the votes collected by computer where there is no audit trail and no way to prove the validity of the actual vote versus what is recorded.
And what do electronic voting machines give us? A voting system collected by computer where there is no audit trail and no way to prove the validity of the actual vote versus what is recorded. Why should it surprise anyone that the voting machines are inaccurate; they're intended to efficiently steal elections in a concealed fashion, not necessarily to efficiently count them.
Let's not forget that the head of one of the companies that sell electronic voting machines said that they intended to make sure they got Ohio for a specific candidate in the 2004 election. (No candidate has ever won the presidency without Ohio.)
Has anyone here considered that since it takes 270 electors to win, all that one needs to get elected President is 11 states?
- California.... 55
- Texas......... 34
- New York...... 31
- Florida....... 27
- Illinois...... 21
- Pennsylvania.. 21
- Ohio.......... 20
- Michigan...... 17
- New Jersey.... 15
- North Carolina 15
- Georgia....... 15
- -------------
- Total........ 271
Get (or steal) these 11 states and you can forget the other 39.The lessons of history teach us - if they teach us anything - that nobody learns the lessons that history teaches us.
I am blessed (cursed?) with a background in econometrics from school.
.5 for all of them, which means there's quite a bit to do before their models are actually believable and worth using as "evidence" of voter fraud.
.5 really isn't terrible for social science stuff. However, this is the only statistic they posted regarding the correctness of the models, and I'd like to see some more numbers in this regard.
The figures look nice, until your eyes stray to the R-square (goodness of fit) results for their regressions - it's about
More formally, R-square is the percentage of sample variation in the dependent variable that is explained by the independent variables. So, in this study, they can generally explain about 50% of the variation - which is not exactly what I'd want to take to court.
In fairness to the researchers, R-square is not the end-all, be-all of proof, and
I find the willingness of people to take this as "proof" of vote fraud is disturbing. This is evidence that places that had electronic voting had more votes for Bush. This evidence is of correlation rather than causation. Maybe Bush supporters were more likely to come out in places with electronic voting?
In any case, I would also direct people to read page 4, where it points out that electronic voting in Ohio didn't cause any change in percent voting for Bush (using model 1, I believe).
-Erwos
Plausible conjecture should not be misrepresented as proof positive.