IM IN UR VOTING MCHINS STEALIN UR PRIMARIES

January 11, 2008 10:29 AM

Kucinich Asks for New Hampshire Recount in the Interest of Election Integrity.A little earlier in the election cycle this time around, many people are worried about serious discrepancies between pre-election opinion polling and exit polls, which both had Obama winning by a substantial percentage, and the official results. Obama also appears to have won in hand-counted precincts while Clinton seems to have dominated in precincts which counted the votes with the Diebold Accuvote TSx optical-scan machines, which have been shown to be susceptible to a memory card hack.

Note: there is no reason for a vote-counting machine to ever run arbitrary executable code from a memory card.

Voting activist Bev Harris thinks Kucinich is walking into a trap due to a lack of ballot chain-of-custody oversight.

posted by dinsdale (99 comments total)

It is so ridiculous that in 2008 this can even be a question. We have the technology to be mathematically certain of the voter intent.
posted by DU at 10:33 AM on January 11, 2008

Once again, the Black Box site has extremely interesting info ... that's also extremely depressing. Voters need to demand transparency.
posted by mrgrimm at 10:37 AM on January 11, 2008

CNN's New Hampshire exit polls.

They don't seem horribly inaccurate to me. There's a question where 97% of those polled chose one option, and Obama and Clinton both got 39% of that part, with Edwards getting the other 17%. Final results were 39/39/17, and a 2% error on an exit poll is not a conspiracy. It's actually pretty good.

That they don't agree with the pre-election polls is almost surely a problem with the pre-election polls, not the exit polls.
posted by smackfu at 10:40 AM on January 11, 2008

dinsdale, thanks for that second link, which really is the money. It makes what would have been simple newsfilter into a far stronger post.
posted by George_Spiggott at 10:44 AM on January 11, 2008

And another discussion of the Diebold / handcount thing.
posted by smackfu at 10:44 AM on January 11, 2008

(I mean the last, link, sorry: walking into a trap.)
posted by George_Spiggott at 10:45 AM on January 11, 2008

It is so ridiculous that in 2008 this can even be a question. We have the technology to be mathematically certain of the voter intent.

How so? Certainly improvements can be made, but there will always be a theoretical possibility to steal an election. All you would have to do would be swap out the entire voting machines before the election.

On the other hand, clearly we should be able to keep the machines from being hacked at the retail level, by a voter or poll worker.

I seriously doubt Hillary actually stole the election. Obama got all the votes he was expected too, it's just that the undecideds decided for Hillary. In fact, Zogby claims their overnight tracking poll picked up a big Hillary swing the night before, but because they use a 3-day average it didn't affect anything.
posted by delmoi at 10:48 AM on January 11, 2008

This post is missing the fact that in at least one precinct in NH, Ron Paul showed up as having zero votes, despite one entire family saying they voted for him. On recount, votes mysteriously showed up. They were "left off the tally sheet" and it was human error.

I second the notion that Black Box Voting is both amazing and scary and depressing.
posted by inigo2 at 10:48 AM on January 11, 2008 [1 favorite has favorites]

Also, isn't it true that Hillary and Obama both get the same number of delegates coming out of NH? Or is NH a winner take all state?
posted by delmoi at 10:49 AM on January 11, 2008

Why are people so quick to assume the actual election results were doctored in some elaborate fraud, but no one questions the exit poll results, which could be easily doctored.

More to the point, given that in the current climate, whenever a discrepancy arises between poll results and the actual results people assume the results are wrong and the poll is right, all that is required to compromise an election and get people to question the outcome is to fake the poll results.

That's easier to do, because the exit poll taking and tabulation process is not well known buy the public, nor are the identities of the people doing them. Furthermore, there is no trail whatsoever in exit polls - exit polls are conducted verbally as people leave the polling places.
posted by Pastabagel at 10:50 AM on January 11, 2008 [1 favorite has favorites]

So, considering 2000, why the hell is the ballot chain-of-custody so friggen vulnerable?
posted by edgeways at 10:50 AM on January 11, 2008

This also raises another question--sorry to go back to Chris Matthews, who was busy claiming the polls were wrong only because New Hampshire voters had lied and were racist--said they were going to vote for a black man and then did not. Would the same people not also lie to people on an exit poll? Am I missing something?
posted by etaoin at 10:52 AM on January 11, 2008

And, yeah I think the election was counted fairly. Still, it seems like with well established data retention protocols you can avoid any hint of impropriety and sooth any future questions.
posted by edgeways at 10:54 AM on January 11, 2008

Why can't we just have actual thugs at the polling stations, like the rest of the world?
posted by not_on_display at 10:54 AM on January 11, 2008 [6 favorites has favorites]

This conspiracy was demolished on Daily Kos, of all places -- a center of Hillary hating. Give it up. http://www.dailykos.com/storyonly/2008/1/10/02623/2264/85/434176

Key points:
1) there is a paper trail everywhere the machines are used (paper ballots are kept). It would be idiotic to hack it, cause it can be easily proven.
2) fewer than half the towns in N.H. even use machines.
3) results on a town by town basis are perfectly consistent with georgraphical and demographic results, IE Obama skews upper class, Hillary lower; Hillary got Kerry voters from 2004, etc.
4) If this is so true, why aren't Obama &co saying anything?
posted by msalt at 10:55 AM on January 11, 2008 [1 favorite has favorites]

Unless I'm making a basic mathematical error, the CNN exit poll for New Hampshire predicted 39% for Clinton and 37% for Obama.
The actual results were 39% for Clinton and 37% for Obama.

Pick your battles, folks.
posted by East Manitoba Regional Junior Kabaddi Champion '94 at 10:57 AM on January 11, 2008

sorry to go back to Chris Matthews, who was busy claiming the polls were wrong only because New Hampshire voters had lied and were racist

He's an idiot though. The big difference with pre-election polls is that you take the raw data and try to turn that into results for the people are actually going to vote ("likely voters"). Even worse since New Hampshire has the whole "Independents can swing both ways" complication. Contrast to exit polls, where at least you know these people voted and didn't blow off the election because Obama had a huge lead OMG.
posted by smackfu at 10:58 AM on January 11, 2008

By the way, your post is wrong there -- New Hampshire does NOT use the Diebold Accuvote TS. As the HinMI wrote on KOS:

" The Diebold Accuvote-TS has been shown to be a piece of crap. The Diebold Accuvote-OS, the machine used in New Hampshire, has much of the same hardware and runs much of the same tabulation software, so these machines could conceivably be hacked. However, the incentive for hacking them is not very great, because unlike with the paperless voting, again, there's the paper trail. So if there were ever a recount—and there was after the 2004 election, when a survey of New Hampshire voting districts chosen by the Nader campaign showed there was virtually no difference between the scanned tabulation and the hand recount—the malfeasance would be easily discovered. "
posted by msalt at 11:00 AM on January 11, 2008

delmoi: How so? Certainly improvements can be made, but there will always be a theoretical possibility to steal an election. All you would have to do would be swap out the entire voting machines before the election.

Pencil and paper.
Pencil and paper.
Pencil and paper.

Why do people think it has to be so difficult? Surely the man-hours invested in counting votes by hand with the Canadian system would be far cheaper and more valuable than the development hours spent working on these machines. As has been said many times before, being absolutely certain that the vote is correctly counted is fundamental to people believing in this whole democracy thing.

Remember, democracy is still experimental.
posted by lostburner at 11:02 AM on January 11, 2008 [11 favorites has favorites]

Oh, apparently exit polls are skewed after the election to fit the actual results. My mistake.
posted by East Manitoba Regional Junior Kabaddi Champion '94 at 11:03 AM on January 11, 2008 [1 favorite has favorites]

Yeah, I tend to think major voter fraud is saved for the general election, not to get 7 delegates instead of 6. The polls *were* accurate, it's just that the media reports them as if they were scores. They do not mention margin of error, except in a small print byline. They do not discuss the fact that something like 15% of NH was undecided going into the polls. The polls are not wrong, they are being used wrong by lazy and/or ignorant pundits. (I will not call them journalists.)
posted by absalom at 11:04 AM on January 11, 2008 [1 favorite has favorites]

msalt, do you have citations for those points, as opposed to a blogger board which you characterize as monolithically "Hillary hating", as if that had any bearing on the facts?

1. Do you have a response to the chain of custody problems raised by Bev Harris?
2. "Fewer than half the towns" is extremely misleading. The larger population centers used voting machines, so as a matter of population, i.e. actual count of votes, far more votes were machine counted
3. Is this supposed to be proof of something?
4. This fallacy isn't worth touching.

I'm not drawing any conclusions about the facts of the matter but I'm very skeptical of anyone who says "don't look at this, because I'm telling you there's nothing there", and backs it up with such weak stuff.
posted by George_Spiggott at 11:05 AM on January 11, 2008 [4 favorites has favorites]

Another post on Drunkard's Lamppost shows that the exit polls actually gave Hillary the victory. If this is true then why is everyone saying the exit polls showed Obama won?
posted by spaceviking at 11:09 AM on January 11, 2008

The president of Pew Research had an illuminating Op-Ed in the Times yesterday, detailing

another possible explanation [which] cannot be ignored — the longstanding pattern of pre-election polls overstating support for black candidates among white voters, particularly white voters who are poor.

posted by coolhappysteve at 11:11 AM on January 11, 2008

There appear to be four possible reasons....

1. White Democrats are racist, and voted for Hillary even though they told the pollsters that they were voting for Obama.

2. The pollsters screwed up.

3. There was a last minute break of many voters for Hillary, obviously moved to vote for her by her hurt feelings on Saturday and her tears on Monday.

4. There was a broad vote fraud consipracy in Hillary's favor that was so effective that no one noticed it at the time.

Yawn....
posted by Durwood at 11:13 AM on January 11, 2008

Yeah, so, can anyone find some exit poll data that definately hasn't been weighted to fit the actual results? Otherwise they're pretty useless.
posted by East Manitoba Regional Junior Kabaddi Champion '94 at 11:13 AM on January 11, 2008

So, considering 2000, why the hell is the ballot chain-of-custody so friggen vulnerable?

Because it works! *wink*
posted by agregoli at 11:15 AM on January 11, 2008

Completely ridiculous, and a cheap (financially and morally) way for Kucinich to keep his name in the papers. As the Kos link above noted, New Hampshire actually has one of the most verifiable paper trails for its elections in the country.

There were no irregularities or voting anomalies in the actual vote, like the infamous "Jews for Buchanan" fiasco of 2000 or the struck voter rolls in 2002... there was just a difference between certain tracking polls and the actual vote outcome.

Of all the places to whine about voting irregularity, a Democratic primary with record turnout across all demographics is a laughable pick. If you want to address voter fraud issues, focus on the GOP attempts to find new and unique ways to prevent black, elderly, and poor people from getting to the voting booth in the first place.
posted by XQUZYPHYR at 11:16 AM on January 11, 2008 [1 favorite has favorites]

spaceviking writes "Another post on Drunkard's Lamppost shows that the exit polls actually gave Hillary the victory. If this is true then why is everyone saying the exit polls showed Obama won?"

Are those the same exit pollsters who so accurately predicted the victory of President Kerry?
posted by mullingitover at 11:17 AM on January 11, 2008

Paper and pencil, or robust optical scanners would work well as long as (and this is a major part of the linked story folks) there is credible handling and storage of the ballots, used and unused. A paper trail is useless for recounts if ballot retention and accounting is compromised. You can steal an election, ANY election, if people don't treat those ballots correctly. RTA
(still think NH was, most likely, accurate)
posted by edgeways at 11:17 AM on January 11, 2008

There are some interesting possible ways to make electronic voting safer and more transparent. For example, how about publishing all of the votes as salted hashes? That way people could check their vote has been recorded correctly (by putting their details into a simple app then searching the database for that hash) without any privacy problems (providing the system's not designed by Diebold).
posted by malevolent at 11:26 AM on January 11, 2008

East Manitoba, Here's another link explaining how the exit polls are "corrected". I guess we'll never know what they were really seeing on election night. But here's the best I've found so far...
posted by spaceviking at 11:27 AM on January 11, 2008

Exit polls have implicit errors, and shouldn't be used to call tight races until a great deal of real data is available. The tighter the race, the more real results you need. The problem with the past errors in calling races is that the networks forgot that. Notice how cautious they were with the NH Primary.
posted by smackfu at 11:27 AM on January 11, 2008

Ron Paul was the true winner of the primary, but everyone stole his votes. Ron Paul tried to save the day, but he could not go Super Saiyan because his blimp was sabotaged by the mainstream media.
posted by Krrrlson at 11:28 AM on January 11, 2008 [3 favorites has favorites]

The other big problem with exit polls is that you can't really "correct" them like we should, since we don't have demographic data for the real votes. Like the exit poll says 57% of voters were females and 43% were male, but there is no way to independently validate that. If that is skewed versus the real results, there's no way to tell.
posted by smackfu at 11:29 AM on January 11, 2008

Are those the same exit pollsters who so accurately predicted the victory of President Kerry?

Good point. The fact is that I don't know what happened in Ohio, and neither do you.
posted by George_Spiggott at 11:30 AM on January 11, 2008

IM IN UR VOTING MCHINS STEALIN UR PRIMARIES

Clearly, there's a conspiracy.
posted by LordSludge at 11:30 AM on January 11, 2008

America just needs to copy how Canada does their elections already.
posted by chunking express at 11:39 AM on January 11, 2008

Yanno, an exit poll is just some dude with a clipboard saying 'So, who did you vote for?'. This being New Hampshire, I wouldn't discount quite a few people not saying anything, and a ton of people saying 'Oh, I voted for X', when they actually voted for Y.
posted by Happy Dave at 11:43 AM on January 11, 2008

It is a non-issue that was explained shortly after the election, and Dennis, stating that he was doing it for all and not for himself, thus makes himself seem foolish for having not made himself aware of the explanation of the results.
posted by Postroad at 11:44 AM on January 11, 2008

Are those the same exit pollsters who so accurately predicted the victory of President Kerry?

In countries with clean elections, exit polls rarely have more than a 2-3% error margin. And there's more than one poll, often commissioned by competing sides, and they're rarely off from each other by more than 1 point.
posted by signal at 11:44 AM on January 11, 2008 [3 favorites has favorites]

msalt, do you have citations for those points, as opposed to a blogger board which you characterize as monolithically "Hillary hating", as if that had any bearing on the facts?

No kidding. And the use of "Give it up," teamed with an inability to make a hyperlink, and a copy/paste from elsewhere without a cite make me think that this msalt fellow is exactly the kind of new user this place could do without.
posted by Kwantsar at 11:46 AM on January 11, 2008

And of course, legitimate exit polls' results are made public before election results, so there's no post-election skewing.
posted by signal at 11:53 AM on January 11, 2008

Pencil and paper.
Pencil and paper.
Pencil and paper.

Yep, no problems there.
posted by DevilsAdvocate at 11:55 AM on January 11, 2008

America just needs to copy how Canada does their elections already.

I've read there are constitutional reasons the U.S. can't have a single organization running elections (like in Canada).

But really each state should have a single elections body that buys equipment in bulk and that has specific rules and procedures for ballots and recounts. Instead things seem to be handled at the county level so every election is amateur hour with different ballots and counting methods and lineups in places and inconsistency between wealthy and poor counties.
posted by bobo123 at 12:03 PM on January 11, 2008

I thought it was extremely odd that both Obama's and Clinton's internal polls showed Obama with a > 10 point lead going into the primary, and he ended up with a two point deficit.
posted by goethean at 12:23 PM on January 11, 2008

And of course, legitimate exit polls' results are made public before election results,

The CNN ones were made available at around 8:30/9 PM. Before most of the election results, at least.
posted by smackfu at 12:24 PM on January 11, 2008

Completely ridiculous, and a cheap (financially and morally) way for Kucinich to keep his name in the papers.

Dude, have you seen how hot his wife is?
posted by humannaire at 12:24 PM on January 11, 2008

Ugh. This makes the people investigating and trying to bring to light the very real problems with the paperless machines look like hysterical nutjobs.

Either that, or they always were hysterical nutjobs and this just demonstrates it. You see how hard it is to tell? Thanks Dennis, you elf-eared twit.
posted by rusty at 12:25 PM on January 11, 2008

I've read there are constitutional reasons the U.S. can't have a single organization running elections (like in Canada).

Strange if true, though I guess the worry there is that it would be corrupt and evil. The good thing about election Canada is that it's not tied to any of the political parties. Why you would let people with a stake in the election results manage elections is beyond me.
posted by chunking express at 12:27 PM on January 11, 2008

malevolent writes "There are some interesting possible ways to make electronic voting safer and more transparent."

Yes, but it's not really transparent. Not unless you have read all the code and know the engineering of the machines. That's why they call it "black box" voting.
posted by krinklyfig at 12:28 PM on January 11, 2008

DevilsAdvocate writes "Yep, no problems there."

Every voting system has its flaws. None are impervious. Pen and paper are the simplest, most reliable tools available, which are understood inherently by all the voting population.
posted by krinklyfig at 12:31 PM on January 11, 2008

Ugh.

It's amazing to me how biased so many of you are. When this was Bush & Co, you guys weren't saying " I really doubt Bush would steal an election" and making tons and tons of excuses for it. Did everyone forget that Rupert Murdoch and the Clintons are friends? Same hands, different puppet. I think it's VERY likely Ohio freaked them out, so they pulled some strings, pushed some buttons, and... there ya go.

These same things were going on in 2004 and everyone here freaked out.
Now that it's a different name underneath all the allegations, most of you are calm and collected. Sounds to me like it's working perfectly as planned.
posted by Espoo2 at 12:33 PM on January 11, 2008 [1 favorite has favorites]

Yeah, in listening to the coverage last night on NPR, I was amazed at how the top story kept being "How did the pollsters get it so wrong? Are NH residents racist?" when the data I'd seen all had a fairly significant margin of error, into which the final results fell.

But man, Kucinich sure won't be president now, right?
posted by klangklangston at 12:40 PM on January 11, 2008

These same things were going on in 2004 and everyone here freaked out.

Apparently you have a different definition of "everyone" than I do.

That some MeFites "freaked out" about 2004 and some other MeFites are not freaking out now does not a contradiction make. MetaFilter is not a hive mind, despite the popular notion. Different users have different opinions, and unless you can identify particular individuals with contradictory opinions, you have not demonstrated hypocrisy.
posted by DevilsAdvocate at 12:42 PM on January 11, 2008

I'm reading over the CNN exit poll, and I can't see the malfeasance of the election other than still using Diebold machines.

Look at the tally by population. Large cities, 43-34 Clinton. Those places are more likely to be using machines instead of pen and paper. It also has a larger concentration of lower class folk without little or no college -- a group Hillary did well with.

It really does appear that the two factors the experts are gravitating towards, support of Boomer women and independents swayed by the debate, were what caused the radical shift in the closing days. That, and I bet that most pollsters forgot to factor in the independents.

I go back to what I said in an earlier thread: Fast, accurate, and cheap -- pick two.
posted by dw at 12:43 PM on January 11, 2008

"Did everyone forget that Rupert Murdoch and the Clintons are friends? Same hands, different puppet. I think it's VERY likely Ohio freaked them out, so they pulled some strings, pushed some buttons, and... there ya go."

Is that more likely than the alternate scenario, where the votes weren't stolen and you're an idiot? I can't help but think that's the simplest explanation.
posted by klangklangston at 12:44 PM on January 11, 2008 [2 favorites has favorites]

I'm skeptical to these claims not because they're allegations against a Democrat but because Obama and Clinton would still end up with the same number of delegates. Hilary gets a media boost now, and a new campaign narrative, but that's all (and a close 2nd place in the primary could have been spun to much the same effect). Besides, up until Obama won Iowa, Hilary was leading in NH in almost every poll.
If there was fraud....worst stolen election ever.
posted by Bromius at 12:49 PM on January 11, 2008

Pen and paper are the simplest, most reliable tools available,

"Most reliable?" By what criterion?
posted by DevilsAdvocate at 12:49 PM on January 11, 2008

Stanford prof Jon A. Krosnick had a different theory: Clinton got at least 3 percent more votes than Obama simply because she was listed close to the top.

Pencil and paper.
Pencil and paper.
Pencil and paper.

I know it's obsolete technology so the origin of the metaphor is now lost in the mists of time, but there actually was once a fraud technique called ballot-box stuffing. I remember when Harold Washington was elected Mayor of Chicago, one of the most notorious machine aldermen took something like three hours to get from the polling place to the board of elections with the "locked" suitcase ... and this in an era when TV cameras were watching.
posted by dhartung at 12:54 PM on January 11, 2008

dhartung writes "I know it's obsolete technology so the origin of the metaphor is now lost in the mists of time, but there actually was once a fraud technique called ballot-box stuffing. I remember when Harold Washington was elected Mayor of Chicago, one of the most notorious machine aldermen took something like three hours to get from the polling place to the board of elections with the 'locked' suitcase ... and this in an era when TV cameras were watching."

Well, those are all potentially open to scrutiny. Additionally, paper ballots provide a system which is transparent to the voters. Ballot stuffing has always been with us, as long as there have been elections. Computerized voting machines make the process much easier to thwart without any sort of record or trace. You don't even need a "notorious alderman" to do your dirty work, as you could accomplish the task without anyone seeing it happen or having any sort of reliable recount mechanism.
posted by krinklyfig at 1:06 PM on January 11, 2008

DevilsAdvocate writes "'Most reliable?' By what criterion?"

You should do some research of your own, if you really don't know. Start with some of the stuff Bruce Schneier has written.
posted by krinklyfig at 1:13 PM on January 11, 2008

It's amazing to me how biased so many of you are.

Indeed. I, too, often bemoan just how aggravatingly in the tank most of MetaFilter is for Hillary Clinton. (Bong hit)
posted by XQUZYPHYR at 1:14 PM on January 11, 2008

Maybe we could invite the UN to overlook our elections this time.
posted by Blazecock Pileon at 1:20 PM on January 11, 2008

Pencil and Paper? Meet eraser.
posted by ogre at 1:21 PM on January 11, 2008

All of this controversy is kind of silly. I appreciate Kucinich for doing what he's doing, because he's going to a) use his rather unique position of having the power to force a recount and b) pay for it himself. The worst that can happen, as the NH guy in charge of the elections said, is that the results will be confirmed.

Here's a summary of reasons why I initially dismissed the idea of fraud:

a) polls are statistical extrapolations based on certain assumptions that can be right or wrong.
b) one of those assumptions was about the likely number of women to turn out. "likely voters" are who count in those polls and that means people who voted last time. if new or lapsed voters are driven to vote, they are going to skew the math in a way polls did not account for.
c) NH has an open primary, meaning Is get to pick an R or D ballot at the polling place. That means that if they are polled separately about R and D tickets, they might choose both Obama and McCain for the purposes of those polls, but on election day, they only got to pick one or the other. For whatever reason, a large number of Is who were counted for Obama in the polls ended up voting for McCain. Possible reasons: wanting to vote against Romney, believing Obama had a dominating lead (due to stupid polls).
d) NH has day-of registration, meaning voter rolls (upon which some of the aforementioned assumptions are based) do not mean much. In other states, you already know who can even unlikely-ly vote, because they are already registered. Often the lists are pared down even further to likely voters, but at least you can do math that includes the outlying, unlikely voters. In NH, you cannot account for an unknown crowd of day-of registerers.

whoo. see, conspiracy unneccessary.

But here's the thing that's come to my attention since then...

We do note, however, the following rather remarkably anomalous result which was reported late this afternoon by analysts from the Election Defense Alliance (EDA). As noted by one of the researchers, IT Consultant Bruce O'Dell:

"Analysts at the Election Defense Alliance (EDA) have confirmed that based on the official results on the New Hampshire Secretary of state web site, there is a remarkable relationship between Obama and Clinton votes, when you look at votes tabulated by op-scan v. votes tabulated by hand:

Clinton Optical scan 91,717 52.95%
Obama Optical scan 81,495 47.05%

Clinton Hand-counted 20,889 47.05%
Obama Hand-counted 23,509 52.95%

The percentages appear to be swapped. That seems highly unusual, to say the least."

( http://www.bradblog.com/?p=5544 )

So... doesn't that look a little suspicious? You don't really have to dig too deep for that to look like something's wrong. I'm not thinking fraud, I'm thinking error. All it would take is for something to have been swapped somewhere in the tabulation process for that to happen. I don't think there's a conspiracy here. But I do think there is a chance of error. And I think that in that case, it is a very good thing that Kucinich is doing what he's doing.

Other things I've read here that I'd like to comment on:
a) delegates are not as important as media perception, and a NH win for Clinton may have saved her hide
b) Recounts have such a bad rep after being done half-assedly or with no change that it amkes sense that Obama might hesitate to call for one while Kucinich would have little to lose by doing so. I'm certainly torn as to whether this is a way to keep Kucinich in the race, or the kind of thing that is his entire reason for running in the first place, to be in a position to do this. One is ignoble, the other is noble, and who knows which is right?
posted by Embryo at 1:21 PM on January 11, 2008 [3 favorites has favorites]

Pencil and Paper? Meet eraser.

Meet purple finger.

If America is bold enough to pay for democracy for its colonies, it can have democracy at home.
posted by Blazecock Pileon at 1:30 PM on January 11, 2008

So... doesn't that look a little suspicious?

No, not really. What I'd rather know is which precincts used the optical scan machines. I'm guessing it's primarily urban precincts with optical scanners, primarily paper and hand counting in rural precincts.

In a case like that, then you have to check for any additional factors that might have created the divide.

You don't really have to dig too deep for that to look like something's wrong. I'm not thinking fraud, I'm thinking error. All it would take is for something to have been swapped somewhere in the tabulation process for that to happen.

Well, that's usually not the case. Most of the errors, as those of us who lived through the Washington state governor's race four years ago learned, come from the problem of the optical scanners not being 100% accurate. They're, on a clear day, 99.99% accurate. When you have a million votes, then that means 100 votes will be mistabulated, at the least. And then you factor in all the times when a voter doesn't do a good enough job on making their intent known (not filling in the bubble, e.g.) and pretty soon you have 500 votes per million being off.

In most elections, this doesn't matter. In Bush v. Gore and Gregoire v. Rossi, it did matter because of the margin. Thus, the recounts. And even the recounts could introduce errors.

I don't think there's malfeasance or substantial error here. I wish people would stop looking for shadows. I also wish they'd stop with the "Diebold is inaccurate" meme. The problem is not the accuracy of the machines. The problem is the lack of transparency with the machines and the code that runs it.
posted by dw at 1:45 PM on January 11, 2008 [2 favorites has favorites]

I think we should have the recount because i f we have we will no longer need to speculate. Its too bad Dennis Kucinich lack the cachet to actually make this a reality.
posted by Rubbstone at 2:12 PM on January 11, 2008

These are the rules for delegate assignments in NH:

14 district delegates are to be allocated proportionally to presidential contenders based on the primary results in each of the State's 2 congressional districts (each congressional district being assigned 7 National Convention delegates). In addition, 5 at-large National Convention delegates plus 3 Pledged PLEOs are to be allocated to presidential contenders based on the primary vote statewide. A mandatory 15 percent threshold is required in order for a presidential contender to be allocated National Convention delegates at either the congressional district or statewide level.

By these measures , Hillary and Obama actually receive the same number of delegates, even if the percentages quoted above are in fact the opposite of what they should be. So, where is the conspiracy here?

I agree with the above posters who contend that the Independents swung the vote here. Nearly half of the voters in NH are registered Independents that can decide at the time of voting which party they'll vote with. I'm also convinced that much of the error in pre-election polls came from polling fatigue. I think that fewer people answered their phones as election day got closer - I know that I refused to answer any call with unfamiliar caller ID in the last couple of days before the election. And exit polls? That's the final insult. No way was I going to stop and answer one more time.

If anyone is interested in a more detailed breakdown of the NH vote, look here.
posted by Flakypastry at 2:15 PM on January 11, 2008

Meh. Here in Muncie, we don't need any damn technology to screw-up vote counts. We do it the old-fashioned way...by hand.
posted by Thorzdad at 2:23 PM on January 11, 2008

Anyone who thinks a second place finish would have sunk Clinton is a god damned moron.

Embryo: Is it fair to compare two sample sizes where N(1) is something like twice as much as N(2)? Not to mention, as quoted above, Optical Scanners and Hand Ballots are not exactly randomly distributed.

Espoo2: Biased? I thought I was being reasonable.
posted by absalom at 2:34 PM on January 11, 2008

Clinton Optical scan 91,717 52.95%
Obama Optical scan 81,495 47.05%

Clinton Hand-counted 20,889 47.05%
Obama Hand-counted 23,509 52.95%

The percentages appear to be swapped. That seems highly unusual, to say the least.

Even more unusual, when you total the votes for each candidate, you find that they don't quite match the official results!

And the claim on the EDA front page doesn't seem to match their own table.

Ignoring this, let's suppose the optical scans just switched votes, as seems to be the implication (nevermind that it would be remarkable by itself if the hand count and machine count agreed with each other to 1/1000 of 1%). Then Obama really got 40% instead of 36%, and Clinton got 36% instead of 39%. There's still a very big discrepancy between the votes for Clinton the and pre-election predictions.

Anyhow, it's pretty clear to me that the numbers above are completely made up to produce the desired effect.
posted by dsword at 2:34 PM on January 11, 2008

So... doesn't that look a little suspicious?

No, because it's statistical irrelevance that's padded on to make it seem even more super spooky. What, exactly, is Bradblog arguing because of the numbers matching? There's no case presented as to what the coincidence actually means, save for a suggestion that the vote totals may have been switched... which I guess could be relevant if there was any precinct in the state of New Hampshire where Clinton and Obama were actually the only two candidates, like they are in this fake statistical model.

All the other candidates like Edwards and Richardson are removed from the totals to make it a comparison of two numbers that have to equal 100. Since 100% is a constant, tacking on their inverses would obviously yield a mutually equal result. It's like saying that if I bought a pack of gum for 25 cents and you a candy bar for 75, it's "strange" that I got 75 cents change and you 25.

In reality, nowhere were votes counted based on Hillary Clinton and Barack Obama being the only two candidates. So why tabulate them that way now? Because doing so creates a mathematical coincidence to favor suspicion.

The argument that this suggests the totals were "reversed" is nonsensical, since in the real tabulation those percentages would have factored against the other candidates, thus making the percentages of total real vote unequal. The only reason these match, suggesting a "switch," is because the data was cherry-picked. It's fabricated math.
posted by XQUZYPHYR at 2:39 PM on January 11, 2008 [1 favorite has favorites]

DevilsAdvocate writes "'Most reliable?' By what criterion?"

You should do some research of your own, if you really don't know. Start with some of the stuff Bruce Schneier has written.

Schneier himself advocates electronic voting, just as long as there is a voter-verified paper trail. See here, for example.

"Sounds like an impossible task, but in reality, the solution is surprisingly easy. The trick is to use electronic voting machines as ballot-generating machines. Vote by whatever automatic touch-screen system you want: a machine that keeps no records or tallies of how people voted, but only generates a paper ballot. The voter can check it for accuracy, then process it with an optical-scan machine. The second machine provides the quick initial tally, while the paper ballot provides for recounts when necessary."

If there's somethine else relevant he's written which promotes a different view, feel free to point me to it directly, rather than just at his website.

Please note: I am not disputing that a pencil-and-paper, hand-counted ballot would be better than some forms of electronic voting currently in use. I am questioning your assertion that it is the "most reliable" form of voting, which would mean it is more reliable than all forms of electronic voting, including those with a voter-verified paper trail.
posted by DevilsAdvocate at 2:39 PM on January 11, 2008

DU: "It is so ridiculous that in 2008 this can even be a question. We have the technology to be mathematically certain of the voter intent."

Oh that's so cute!

Repeat after me: There is no such thing as an unhackable program.

If an election process goes digital, it can be tampered with, and even further, it can be tampered with in ways that leave no evidence to said tampering. At least when you have hand ballots, there's something to track. And again, I reiterate I have no intentions of voting in this presidential election. THIS. RIGHT HERE. IS WHY. Our votes aren't really counted. The idea that the common man still has any voice in American politics is a fallacy. It's a lie. I for one no longer wish to participate. This is no longer a democratic republic. I see no reason why anyone should pretend to vote in a corporate oligarchy, unless they're being paid.
posted by ZachsMind at 2:56 PM on January 11, 2008 [1 favorite has favorites]

Clinton Aide: Obama Is For People Who Want "Imaginary Hip Black Friend"
posted by homunculus at 3:49 PM on January 11, 2008

DevilsAdvocate writes "Please note: I am not disputing that a pencil-and-paper, hand-counted ballot would be better than some forms of electronic voting currently in use. I am questioning your assertion that it is the 'most reliable' form of voting, which would mean it is more reliable than all forms of electronic voting, including those with a voter-verified paper trail."

Tell me this: what is the need for the computer, if you have the paper trail? Is it just an elaborate device to make a mark? Why bother with the expense, the complexity, and the inherent flaws in the idea that local governments are capable of dealing with the issues involved in using these machines? ANYONE can count paper, and read it. Nothing to break down. Nothing to turn on. Nothing to call the IT guys about because it won't connect to the network.

But, even if you could control all that and minimize it, the biggest problem with having computers as the I/O and counting device is that the process of counting is opaque to anyone except those with access to the code, and the knowledge necessary to understand how it works.

Your question was how do I know that pencil and paper are the most reliable. I pointed to Scheier, not because of his opinion that electronic voting may be possible to do securely, but because of his research on the matter in general. All things considered, I'd rather not have local governments dealing with complex issues like network security. The companies making the machines have not proven themselves to have an interest in serving the public. The cost of paper, the simplicity, the utility, the transparency, all point to it being ideal. You can get machines to count the ballots, and even mark them, but you need to have paper as the ballot, and people have to be able to see it being marked, if they don't do it with their own hand. We could control the variables far better in a paper ballot system through a chain of custody. People may still abuse it, but physical evidence is still possible to obtain, and there is nothing to learn to understand how it works. The system we use has to be trustworthy in the eyes of the public, or the whole point of doing this is out the window.
posted by krinklyfig at 4:27 PM on January 11, 2008

Tell me this: what is the need for the computer, if you have the paper trail?

It seems in your desire to prevent deliberate, malicious tampering, you are too quick to discount the possibility of accidental, unmotivated error. A voting system needs to minimize the possibility of both.

ANYONE can count paper, and read it

I think you severely overestimate the ability of humans to do tedious, repetitive tasks without error.

You can get machines to count the ballots, and even mark them, but you need to have paper as the ballot, and people have to be able to see it being marked, if they don't do it with their own hand.

I agree with this completely. It does, however, seem to contradict your earlier insistence that a hand-counted, pencil-and-paper ballot is preferable to any electronic system, including one which would mark a paper ballot, present it to the voter for verification, and count the ballots.
posted by DevilsAdvocate at 5:29 PM on January 11, 2008

P.S. Also note, the NH township where Ron Paul was initially reported to have received 0 votes, linked by inigo2, was one where ballots were hand-counted. Does that bolster your faith in hand counts?
posted by DevilsAdvocate at 5:47 PM on January 11, 2008

I've been a pretty staunch defender of Kucinich in the past, and I guess I'd have to say that on principle, it's never a bad thing when an election process gets closer scrutiny, but I've got to agree with the naysayers on this one--intentional fraud seems pretty unlikely here.

Still, what rubs me the wrong way a little bit is all the spin surrounding the NH result, how the prevailing storyline about Clinton II being another "come-back kid" just like Clinton I ignores the reality of how much better than expected Obama actually ended up doing in NH, which was for the majority of the pre-primary run-up considered a lock for Clinton ("firewall" was the term her own campaign used, as I recall).

The real story here is that Obama did much better than anyone expected in NH. I even half suspect that Clinton's own polling numbers toward the end intentionally showed Obama with a significant lead, so that the final result, in which Clinton barely won a state she was expected to win handily, ended up looking like some kind of surprise victory for Clinton, rather than what it actually was: A surprisingly strong finish for Obama.

Some nice perception manipulation possibly being done here, sure, but I doubt there was any outright fraud.
posted by saulgoodman at 6:45 PM on January 11, 2008 [2 favorites has favorites]

votescam
posted by hortense at 7:03 PM on January 11, 2008

First of all, I'd like to make it clear that I don't mean to suggest that I believe that the NH primary was hacked - that's why I didn't link to BradBlog or any of the other "alarmist" sources that are covering this issue. The most common come-back to these sorts of assertions is that "there's no proof" that the vote was manipulated. True enough - but neither is there any proof that it wasn't manipulated - there's no proof either way. But what the hell good is an election if the results can't be proven? An election is complicated, especially American elections that have lots of races and propositions on the same ballot, but it's not that complicated!

Why do we have elections? To try to create an orderly transfer of power, a situation in which the losers admit defeat and go along with the results.

There's a little parable that most likely apocryphal, but illustrates the point - it goes something like this: "democratic" elections were invented by a group of soldiers stranded in some foreign land after a battle in which their commanders were killed. Unable to decide what to do, some argued that they should push on in hopes of finding an undefended village to plunder - others were of the opinion that they should head for home. The discussion became heated and the rowdy mob, all armed to the teeth, were on the point of unsheathing their weapons and having at it, when someone decided to ask for a show of hands, to get an indication of the relative strengths of the opposing camps. When the minority group saw that they were badly outnumbered, and would most likely be all hacked to pieces, they backed off and went along with the majority. Politics really is war by other means.

The key to the effectiveness of the whole process is transparency. The losers will not go along with the results if they think that they were cheated, so the whole process has to be done in a manner that is as above-board as possible, with every step clearly and provably fair and honest.

Here's another parable, somewhat closer to our own time and place, which is from a first hand description that an acquaintance of mine related of an election which the UN facilitated in Africa in the early 1990s. The country had seen a colonial government hand over power to a local "strongman", who only relinquished power after a long and devastating civil war. Finally it was agreed that democratic elections - the first ever in the country, would be held. Because of this history, there was no "culture of democracy" in the country, other than the sort of "white of black marble in pot" elections held traditionally in the villages to choose local leaders. In particular, due to the many years of warring factions, all seen as outsiders, periodically marauding through the villages, there was absolutely no trust of central institutions, and a bunch of white people flying around in helicopters didn't inspire much trust either. So here's how the ran election: in each village an open area was created, and the villagers and others living nearby all gathered on the edge of it. In the center, a simple box was built in plain view of everyone, so that everyone could verify that it was empty at the beginning of the election. Then, one by one, again in plain view of everyone, the voters approached the box, filled out their ballot, and walked away from the box. No one else was allowed to go anywhere near the box until everyone had voted. But everyone stayed there and watched, to verify that no hanky-panky was afoot. And when they finished voting, the box was opened and ballots removed one at a time, held aloft, and shown to everyone present, along with the announcement, "one vote for [...]". (African ballots tend to be very large and colorful, with the logo and colors of the various parties prominently displayed beside a very large oval, so everyone could see how the ballots were marked.) After each vote, a mark was made on a large board, under the banner of the respective party. This went on for hours, until the last ballot was removed from the box, and the final count was tallied - and during this whole process, everyone stayed to watch and verify that the result was accurate. Many people wrote down these results so that they could look up the results from their village when the final election results were published, and verify that they were identical numbers.

The point is, because these people were deeply distrustful of the whole process, they went to extreme lengths to make the process as simple and tamper-proof as humanly possible, and came up with something that I would consider to be a paragon of transparency, a benchmark if you will. Ironically, because Americans in general actually trust the process and assume that in "the world's greatest democracy" people couldn't possibly steal elections, the system has been allowed to become so byzantine and kafka-esque that there is effectively no transparency to the process, and hence lots of room for suspicions to form and fester.

So my point in posting about this was not to weigh in on one side or another in a partisan political debate, or to engage in finger-pointing towards any particular faction, but to point out that, in 2008, Americans are finally reaching the point where their skepticism is beginning to overcome their natural inclination to trust the process. This is a good thing, not because it means the system can't be trusted (that's the bad part) but that it is only through a lack of trust in the current system that they will become interested in the actual processes and begin to demand accountability and transparency at every level.

In the international elections community, it is generally accepted that the American electoral system is hands-down the worst in the developed world - for example, nowhere is it considered appropriate to have the campaign manager of one of the candidates also be the person in charge of running the election (*cough* Katherine Harris, Kenneth Blackwell) - that's considered an obvious conflict of interest, and it would be laughed at anywhere else. Not that anyone's laughing.

In America, exit polls diverging by double digits from the count is considered evidence of an unknown "bias" in the polling process. In the Ukraine, the very same symptom is held up as evidence of "massive fraud" - by the same people who won't accept the possibility of fraud in an American election. That, my friends, is a mind-fuck of world-historical proportions. And again, I'm not saying that fraud has been proven, only that the system appears to have been designed to leave many doors open, and it won't be fixed until people take a genuine interest in the entire process from start to finish and accept nothing less than complete transparency.
posted by dinsdale at 7:06 PM on January 11, 2008 [4 favorites has favorites]

No, not really. What I'd rather know is which precincts used the optical scan machines. I'm guessing it's primarily urban precincts with optical scanners, primarily paper and hand counting in rural precincts.

This page stratifies the information by town size. For both small and medium towns the Clinton advantage and Obama disadvantage between optical scanned and handcounted is very, very clear. For large towns, Obama has an advantage. The thing is, NH isn't like other states; it doesn't really have any "cities" by most states' standards. "Large towns" is a much better term for them (even if technically cities are by population, not development type.) There is really no "urban" part of NH. The large town advantage doesn't make a lot of sense. The first two towns to vote didn't register a single vote for Clinton, and they are among the state's most rural. I know that doesn't prove anything, but the urban/rural dichotomy between progressive/establishment or black/white voters doesn't make sense in NH. I'm just glad that Kucinich will get a count so that this discussion on metafilter isn't the end of this question. And, if anyone can explain the below data, please do. Anyway, peep this:

http://checkthevotes.com/index.php?party=DEMOCRATS

Small Town Machine vs Hand Counts (less than 750 votes)

Candidate  	Machine VS Hand in Small Towns
Clinton 	  3.733%
Edwards 	  1.080%
Gravel 	          0.014%
Kucinich 	 -0.720%
Obama 	         -4.550%
Richardson 	  0.107%
Other 		  0.336%

Medium towns:

Medium Town Machine vs Hand Counts(between 750 and 1,500 votes)

Candidate 	Machine VS Hand in Medium Towns
Clinton 	  5.572%
Edwards 	  0.549%
Gravel 	77 	  0.005%
Kucinich         -1.037%
Obama 	         -4.257%
Richardson 	 -1.284%
Other 	          0.453%


Large Town Machine vs Hand Counts (more than 1,500 votes)

Candidate 	Machine VS Hand in Large Towns
Clinton 	 -3.988%
Edwards 	 -1.815%
Gravel 	          0.046%
Kucinich 	  0.496%
Obama 	          4.454%
Richardson 	  0.678%
Other 	          0.128%

posted by Embryo at 7:19 PM on January 11, 2008

damn, that <> block was not double-spaced in preview.
posted by Embryo at 7:19 PM on January 11, 2008

The argument that this suggests the totals were "reversed" is nonsensical, since in the real tabulation those percentages would have factored against the other candidates, thus making the percentages of total real vote unequal. The only reason these match, suggesting a "switch," is because the data was cherry-picked. It's fabricated math.

See, the thing is, that is just silly. Of course that's not the actual OVERALL relative totals, percentagewise. But that's not the point. For those two numbers to match exactly, taking the other (non-contender) candidates out of the mix, to 2 decimal places, is just way beyond math you could "fabricate." It could be an eerie coincidence, and we may or may not find out. But to just say, there's most certainly nothing to those numbers, is silly to me.
posted by Embryo at 7:25 PM on January 11, 2008

intentional fraud seems pretty unlikely here.

So? All the better. If we really think that tampering was unlikely, we now have the potential to see an open and public audit on different voting systems' performance in actual use. There's been endless speculation about whether electronic systems are more or less reliable than hand counted paper ballots, and how touch screen or optical scan measure up. So now we can find out. NH is a small state, and with the relative homogeneity described above it could probably be one of the best places to do this.
posted by dilettante at 8:27 PM on January 11, 2008

I heard about the Hillary Clinton + Martin Luther King gaffe on NPR yesterday. Miscounts or not, her campaign is desperate.
posted by VulcanMike at 9:33 PM on January 11, 2008

why can't we simply get paid for our votes the good ol fashioned way?
posted by dougiedd at 11:49 PM on January 11, 2008

And the use of "Give it up," teamed with an inability to make a hyperlink, and a copy/paste from elsewhere without a cite make me think that this msalt fellow is exactly the kind of new user this place could do without.

That ain't the patented Metafilter snark^TM that I joined this site for, Kwantsar! More like some junior high girl trying to imitate a stuffy London private club member.

I guess ad hominem attacks on a newb and his incorrect link to an article are easier than responding to the 75 other mefites who share my skepticism. Or what I said, for that matter.
posted by msalt at 11:50 PM on January 11, 2008

msalt, do you have citations for those points, as opposed to a blogger board ...
Do you have a response to the chain of custody problems raised by Bev Harris?

So I read the Bev Harris post, which is (wait for it) a "blogger board". Of course.
Heaven forfend we slum among those dread blogger boards!

It's funny that you're so coy about my source. I think people here have heard of Daily Kos and can figure out how much they trust it.

As for chain of custody; sure, it could be an issue. In New Hampshire, each town handles its elections separately. If "they" got to each of hundreds of townships separately, stole and swapped the ballots with differently marked ones that -- by your count -- are 3-4% different than the real ones, and did all of this without anyone noticing, then yes, flawed chain of custody could have allowed "them" to substitute tens of thousands of paper ballots that perfectly matched slightly hacked vote totals that don't change the delegate count.

You are absolutely right; it is mathematically possible.
posted by msalt at 12:12 AM on January 12, 2008

Dennis Kucinich talks about his UFO sighting. Seems he is interested in mysteries and conspiracies.
posted by iviken at 8:54 AM on January 12, 2008

Here's a pretty reasonable write-up from Salon.
posted by oddman at 10:42 AM on January 12, 2008

Nice link, oddman. Hard to argue with Manjoo's arguments for random audits of every election's ballot counting, not having partisan secretaries of state in charge of voting, and general transparency. These shouldn't even be debatable.

I would love it if Kucinich paid for a recount in New Hampshire. Unfortunately, it looks like he may have been bluffing. As of Friday evening, he had not formally requested a recount despite his public announcements.
posted by msalt at 12:32 PM on January 13, 2008

Judge says MSNBC must let Kucinich debate
posted by homunculus at 11:15 PM on January 14, 2008

Glenn Greenwald: The Kucinich court decision and "judicial activism"
posted by homunculus at 9:35 AM on January 15, 2008

Recount update: Kucinich's campaign paid $25,000 to start a recount. That's enough to count the votes in Hillsborough and Rockingham counties (about 100,000 votes), but a statewide recount would cost $42,600 more. They also asked for the right to cancel the recount at any point -- and get the rest of their money back -- if there are no irregularities. Republican Albert Howard put down $56,000 for a full recount, a day after the deadline, but NH will honor his request anyway as soon as the Democratic recount is done.

Counting began Wednesday. Only the Kucinich and Clinton campaigns sent observers; they were joined by several freelance watchdogs with video cameras. Official raw results. No one sees significant problems yet, not even Kucinich observer Manny Krasner. Except of course Brad Blog: "Scores of Votes Mistallied for Every Democratic Candidate..." After 35,000 votes counted, Kucinich is down 1 vote. The most significant issue was in Manchester Ward 5, where vice-presidential votes were added into presidential totals. Every major candidate lost votes when corrected: Clinton 683 to 619, Eddwards 255 to 217, Obama 404 to 365. This was a human, not machine error.

Finally, chain of custody: "The last area of concern some have raised is the security of the ballots. Town clerks hold the ballots until a state trooper arrives to deliver them to Concord, and the ballots are under lock and key until the recount is over."
posted by msalt at 4:56 PM on January 18, 2008 [1 favorite has favorites]

The recount is finished; after 40% of votes were hand-counted, Kucinich's payment was used up and he wouldn't pay for the rest of the state. Official results: Hillsborough County (complete); Rockingham County (partial).

AP: no significant changes, errors were all human. Clinton dropped 25 votes, Obama dropped 7. Daily Kos diarist mspicata -- "There is no evidence of hacking of the Diebold scanners; there are no issues with chain of custody". BradBlog: "New Hampshire Sec. of State Appears to Lie to Media About Kucinich Hand Count - Bill Gardner Says 'Kucinich Satisfied With Recount' When Letter from Congressman to SoS Says Otherwise...

Meanwhile, the recount of Republican votes, which has been paid for in full, has begun.
posted by msalt at 2:51 PM on January 24, 2008

Kucinich causes House floor fuss in announcing Bush impeachment drive
posted by homunculus at 10:03 AM on January 25, 2008