Back to previous post: Open Thread 99

Go to Making Light's front page.

Forward to next post: Ars Technica on recounting New Hampshire

Subscribe (via RSS) to this post's comment thread. (What does this mean? Here's a quick introduction.)

January 11, 2008

Recounting New Hampshire
Posted by Jim Macdonald at 11:59 PM * 50 comments

What’s up with that?

The primary is over. Hillary won. So why is Dennis Kucinich demanding a recount? He polled less than two percent. Does he think that he’s going to find another 112,000 votes somewhere and be declared the winner?

Kucinich is many things, but delusional isn’t one of them.

Instead, hardly had the last balloon dropped in the Hillary headquarters when the usual wingnuts and Hillary-haters started saying that she must have cheated and stolen the election. The allegations against her range from using Diebold sleight-of-software to busing in fake voters from out of state.

Here’s a fairly typical example.

This all ignores that New Hampshire uses paper ballots, not touch screens. There’s a complete, verifiable, paper trail. The primary results matched the exit polls. But, since the pollsters didn’t take into account the fact that the turnout would be around 33% above the previous record turnout, their list of “likely voters” that the samples were based on didn’t match actual voters on the day.

Here’s what I think is happening: In order to prove that Hillary won fair and square, there has to be a full hand recount (totally possible since we do have the original pieces of paper). But how to get one? Hillary can hardly request a recount—she won. Neither Edwards nor Obama can ask for a recount without looking like sore losers, and perhaps hurting themselves in later primaries. This race isn’t over yet, and neither of them is willing to fall on his sword just yet.

So who is it who has the standing to request a recount, and who is it who has nothing to lose by doing so? Enter Kucinich.

The Ohio congressman cited “serious and credible reports, allegations and rumors” about the integrity of Tuesday results.

He can quash this right-wing talking point right now, before it gets started. He’s still got money in his campaign account, and he can use it for the good of the Democratic Party. Go, Dennis!

This is going to be the first state-wide recount since 1980.

Technorati icon
Comments on Recounting New Hampshire:
#1 ::: Linkmeister ::: (view all by) ::: January 12, 2008, 01:46 AM:

But wouldn't the state resist spending a pile of money to recount what is essentially a now-unimportant election relative to the upcoming primaries, particularly Super Duper Tuesday on Feb. 5? I would if I were Secretary of State or whatever officer is in charge of elections.

#2 ::: Stefan Jones ::: (view all by) ::: January 12, 2008, 01:59 AM:

After reading that web page, I get a creepy feeling that the Paulistas will probably dismiss any recounts too. Every election will be suspect, and the crazier followers will feel entitled to take action.

#3 ::: Mary Frances ::: (view all by) ::: January 12, 2008, 02:00 AM:

My God, Jim. I bet you're right. In which case, Go, Dennis, indeed.

Linkmeister, if I'm reading the article Jim linked to correctly, New Hampshire isn't going to pay for the recount. Kucinich is.

#4 ::: Linkmeister ::: (view all by) ::: January 12, 2008, 02:10 AM:

Mary Frances, thanks. I agree with you. I clicked the second link to read the Ron Paul argument, but not the first linking to the Kucinich story. There's a lesson for me there.

#5 ::: Jon H ::: (view all by) ::: January 12, 2008, 02:34 AM:

"But wouldn't the state resist spending a pile of money"

Kucinich would pay for it.

From what I read, if the requesting candidate lost by small margin, the recount costs only $2,000. If the loss is by a larger margin, like Kucinich's, the candidate pays the whole cost.

#6 ::: Bruce Adelsohn ::: (view all by) ::: January 12, 2008, 03:11 AM:

One analysis I read suggested that the reported deviations between poll results and actual results were only in locations that used electronic scanners to count the votes. The votes themselves were paper, but were counted by machine. If that's so, then at least there will be an opportunity to verify any differences NOW, rather than waiting until the general. Also, if that's the case, then there should be doubled and redoubled scrutiny of any similar polls countrywide.

But, as long as we're talking about R*n P**l and Dennis Kucinich, together again at last, I should at least point to yesterday's Bob the Angry Flower.

#7 ::: . ::: (view all by) ::: January 12, 2008, 03:17 AM:

I want to tell you a thing, but
The world keeps threatening to fall.
So join me, O well-arméd wingnut
In spamming to vote for RuPaul!

[posted from]

#8 ::: Lindra ::: (view all by) ::: January 12, 2008, 04:27 AM:

#7 -- Drive-by?

I want to know what 'The United American Freedom Foundation' has to do with a television network.

(I'm not about to go and see if it actually is a televison network. Introductory splash page with auto-loading Flash? Frames on the main page? Back button!)

#9 ::: abi ::: (view all by) ::: January 12, 2008, 05:35 AM:

Lindra @8:
I opened the link in a suitably clean environment.

Wingnuts, emphasis nut. Citizen militias, forcible deportation of illegal immigrants, disband the Department of Education, yadda yadda. Three guesses whom they endorse for president, and the first two don't count.

I have translated and summarized for your convenience.

#10 ::: Zander ::: (view all by) ::: January 12, 2008, 06:37 AM:

Stefan, I can understand that. That's how I've felt about American elections since 2000. Once a system has been proved (or is perceived as having been proved) to be hackable, anything run on that same unmodified system is going to be suspect. I imagine the Republicans are more worried about whether the Democrats will beat them at vote-fiddling than they are about losing popular support.

#11 ::: dr.hypercube ::: (view all by) ::: January 12, 2008, 06:44 AM:

Apparently, there will be a recount on the R side, too: NH to Recount Ballots in Light of Controversy. Dunno who Albert Howard is ("Mr. Howard, I know Vermin Supreme, and you sir, are no Vermin Supreme.") - just wanted to toss the factoid in for the sake of completeness.

The whole thing on the D side is moderately insane - it's been debunked multiple times - but 2 things to give thanks for:
- paper trails. Any voting machine exec that tries to foist a system on us that does not produce a verified (you look at the paper and confirm that's your vote) audit trail should be taken out into the parking lot and...
- the fact that I (NH resident) am not going to have to pay for this.

#12 ::: Giacomo ::: (view all by) ::: January 12, 2008, 07:24 AM:

I'm no Paulista (yet another hijacked word, sigh), but the 31-disappeared-votes is a trust-shattering farce as much as the Hain funding row in Britain ("sorry, i forgot to register £100,000+, human error!"). Seven years after the Gore-Bush shambles, one would think that people involved in the US electoral system had got a bit wiser.

Mind, Europe is no paradise... my parents worked at poll stations more than once, and they remember situations like this: one night while counting, the available table-space was all occupied by different piles of votes (Italy has dozens of parties). Suddenly, the "announcer" pulled out from the box a (extremely rare, in that area) vote for M.S.I, the ultra-right fascism-nostalgia party. The "counting director" suggested that "we'll really have to put this pile on the floor (wink wink(, in this dark corner (wink wink), and hopefully we won't forget to eventually count them (wink wink)".

Unfortunately, times have changed (and the MSI guys even ended up in government for quite a few years recently).

#13 ::: Connie H. ::: (view all by) ::: January 12, 2008, 07:30 AM:

This makes me like Kucinich all over again, truly.

One thing that makes me shake my head over the accusation of "busing in voters from out of state" is that by and large, poll workers are actual members of their communities who know a great many of their neighbors by name or by sight. It's one of those byproducts of small-town small-d democracy -- neighbors greet each other in line at the voting place, you get to sneak a glance at the ledger upside down and see which of your neighbors voted, et cetera.

#14 ::: Matt McIrvin ::: (view all by) ::: January 12, 2008, 07:32 AM:

My thought was that Kucinich sees himself to some degree as a surrogate for the frustrations of marginalized, vocal left-of-center folk on the Internet, and if some of them are claiming that Clinton stole the election from Obama (which they are), he has an opportunity to at least look into the matter.

#15 ::: James D. Macdonald ::: (view all by) ::: January 12, 2008, 07:34 AM:

Albert Howard is a chauffeur from Ann Arbor, Michigan, who feels that he was robbed. He got 44 votes, but one of the early reports said that he had fifty.

An anonymous donor is paying the cost of the recount on his behalf.

#16 ::: Earl Cooley III ::: (view all by) ::: January 12, 2008, 07:56 AM:

Won't a recount affect issues like who qualifies for Secret Service protection and federal matching funds?

#17 ::: P J Evans ::: (view all by) ::: January 12, 2008, 09:19 AM:

Earl @ 16

Probably not.
The FEC is currently not exactly running, due to Shrub having another hissy fit because one of his minions has been disapproved. (Long story. Shorter version is, he or Cheney have been dicking with the FEC to give it majority GOP membership when it's supposed to be 50/50; they were trying to run nominees as an up-or-down set of four, claiming that's how it's always been done, when all it needs is a pair at a time to be voted on. One nominee was firmly squashed before the vote; hissy fit ensued.)

#18 ::: roman_eagle ::: (view all by) ::: January 12, 2008, 09:26 AM:

How are we going to guarantee the integrity of the voting data? where is it now? can they be in the process of altering it? Are any independent. people gaurding the data now?

#19 ::: dr.hypercube ::: (view all by) ::: January 12, 2008, 09:31 AM:

@18 roman_eagle - if you have that little faith in the process, why do you care at all?

#20 ::: Xopher ::: (view all by) ::: January 12, 2008, 09:33 AM:

roman_eagle, welcome to Making Light. I doubt the ballots are in open boxes in someone's basement; I've seen the seals on some of these boxes (in NJ, not NH) and they're difficult to fake.

Note that it's not "data" that needs to be guarded from alterations. The ballots will be recounted to generate the data all over again. Altering hundreds of thousands of ballots is much harder than hacking a Diebold no-paper-trail machine. I wouldn't worry about it.

#21 ::: Caroline ::: (view all by) ::: January 12, 2008, 10:15 AM:

Now all we need is a nationwide paper ballot system, and we could settle this everywhere.

I look forward to finding out that Clinton won NH fair and square.

I still shudder to remember the rickety, funky electronic voting machines I used in the 2000 election. Now, my county went blue. But I still did not like just trusting the thing. At least we have optical scan of paper ballots in my current county.

#22 ::: fidelio ::: (view all by) ::: January 12, 2008, 10:16 AM:

Yes, Xopher, it takes really iron-clad control and organization to mess with paper ballots to a wide enough extent to really mess up a large election--James Squires' Secrets of the Hopewell Box gives an idea of what's involved in doing that, especially the number of people who have to be co-opted into the tampering process.
This is one reason why machines liked to manipulate the registration rolls instead--it's easier and safer, especially because it's harder to monitor who's registered, as registration does not take place all at once, on one day, in a few locations.

Clearly, the process of messing with electrons is far simpler, if you have the technology (and the technology workers), which is why there is such a need for a paper trail. If I can get a receipt at my ATM, why not at the polling place? If I can get a receipt for an electronic transaction (debit/credit card) from a cash register, or from my computer when I make an on-line payment, or even, for godsakes, from a gas pump, why not at the voting booth?
As one of my uncles would have said, "There just ain't no excuse for such sorry doin's, and the whole world knows it, even if some won't admit it."

and Go! Kucinich! I was at a loss when I heard this, and although Jim's reasoning makes perfect sense, I am awestruck to realize that the Democrats have reached such a level of internal cohesion that they'll get each others' backs this well, because all too often, the statement "I am not a member of an organzised politcal party; I'm a Democrat," seems to apply.

#23 ::: Michael Bloom ::: (view all by) ::: January 12, 2008, 10:19 AM:

Count me among those who think the "debunkers" didn't actually debunk anything. Many of them have tried to confuse the issue with the known problem of touch screen machines, and said that just because there are actual paper ballots involved, the counts from the scanning machines must be accurate. Sorry, the scanning machines are themselves computers that can be hacked in much the same way, and because voters never see their workings, we're not even aware of their anomalies (unlike the anecdotes of the voters who said they selected Gore and the screen came up "Bush"). To say that nobody would dare cheat because the ballots could always be counted by hand is insufficient until somebody actually counts them by hand, which is why Kucinich is stepping up to the plate.

As far as I can tell, Bradblog is one of the best aggregators of the available information. Apparently there is a persistent discrepancy of about 7% between the polling and the results, and apparently this discrepancy is most evident in the (larger, more urban, more establishment) communities that use optical scanners-- the smaller polities that count by hand reported results closer to what the polls predicted.

I agree with Teresa that nowadays everything makes me feel like a nutbar conspiracy theorist, but I don't think the Clinton forces fixed it, I assume if anyone did it was the Republicans, as usual-- they read the same polls as the rest of us, and they see that their chances of eking out another 50%-plus-one election are better against Clinton than against Obama or Edwards. But I'm with Kucinich on this, I want a recount.

#24 ::: Dave Bell ::: (view all by) ::: January 12, 2008, 10:34 AM:

And there I was thinking that Tammany Hall was a porn star.

#25 ::: albatross ::: (view all by) ::: January 12, 2008, 10:45 AM:

Xopher #20:

The tamper seals are somewhat hard to defeat in large numbers, but only if they're used properly. (If you want to do large-scale fraud, you will get economies of scale, but I think it's going to be a labor-intensive, pain-in-the-ass sort of process.) My reasonably well informed picture of this is that the procedures followed vary widely from place to place--some jurisdictions are really careful about stuff like writing down the numbers on the seals and then verifying them when the seals are opened, others aren't. If nobody writes down the numbers on the seals or checks them, or if it's the same guy doing both, then the tamper seals have the same property as paper trails that nobody ever looks at--they add hassle and cost, but not much security, to the election.

#26 ::: albatross ::: (view all by) ::: January 12, 2008, 10:50 AM:

Connie H:

Of course a recount would have no use in catching "bussing in voters from out of state." There is a push in many places to demand picture ID to vote, to prevent such attacks. One of the more pleasant ways to re-enforce my deep cynicism about people has been to watch this discussion.

Five or six years ago:

Activists, mostly liberal: We need paper trails on voting machines, because God knows what the software in those black boxes is doing.

Establishment types, mostly conservative, but including nearly all election officials: This is silly, nobody's ever proven an attack on DREs, these machines are just like Fort Knox, moving to paper will inconvenience the blind, etc.


Acrtivists, mostly conservative: We need picture IDs to prevent bogus voters being counted.

Establishment types, mostly liberal: This is silly, there's no proof, etc....

#27 ::: Connie H. ::: (view all by) ::: January 12, 2008, 11:21 AM:

albatross -- at least in my neck of the woods (MA), the poll workers check off addresses in a log book that is also put under lock and key. It wouldn't be that tough to randomly select a log book in one of these suspect precincts and go door to door in the neighborhood, ascertaining that all votes were in fact voted.

#28 ::: Sebastian ::: (view all by) ::: January 12, 2008, 11:28 AM:

fidelio @ 22: The reason you can't get a receipt for your vote, as I understand it, is that it would enable vote-buying: If there's no receipt, then someone can pay me/threaten me/etc. to vote for Candidate A, but there's nothing stopping me from actually voting Candidate B once I get inside the voting booth. However, if I get a receipt for my vote, Candidate A's thugs can come around to my house after the election and demand to see the receipt.

#29 ::: sburnap ::: (view all by) ::: January 12, 2008, 11:49 AM:

This obsession with who "won" is annoying. This isn't a winner-take-all vote like the presidential election. Clinton didn't "win" anything concrete other than the ability to get the media to label her the "winner". In truth, Clinton and Obama both got nine delegates.

I find it fascinating how the media seems to avoid talking about the actual delegate count. For instance, how many people realize that Mitt Romney, who has not won a primary, is the current Republican frontrunner? You wouldn't know it from the media.

On the actual subject at hand: recounts are never a bad idea, even if vote fraud allegations seem silly. If recounts happen at the drop of a hat, you make it harder to commit fraud, and hopeful dissuade it in the first place. But I bring up the real nature of these elections, because while the alleged fraud might effect who "won", it is unlikely to change the delegate count.

#30 ::: sburnap ::: (view all by) ::: January 12, 2008, 11:59 AM:

addendum: Note that while Kucinich would need 112,000 more votes to be declared the "winner", he'd only need around 8,000 more to get a delegate.

Delegates are important even if final victory is unlikely. Kucinich certainly knows that if neither Obama nor Clinton can get a majority of delegates, candidates like him, with small blocks of delegates, get power to shape the debate. If you are wondering why people like him (or Ron Paul) are running seemingly quixotic campaigns, this is why. Kucinich isn't in it to win. He's in it to force the ultimate winner to talk about the issues he cares about.

#31 ::: albatross ::: (view all by) ::: January 12, 2008, 01:21 PM:

Connie H #27: Do you know if anyone ever does that kind of audit? The kind I've heard about using pollbooks is the count where you make sure there aren't more ballots cast than voters who signed in. (Though as with most after-the-election audits, it's not 100% clear what to do when you find problems, since it's hugely expensive to rerun an election, and it also changes the whole process.)

Sebastian #28: There are some very clever proposed voting schemes (still in the "interesting research" phase, IMO) which give voters a receipt which in principle lets them verify that their vote was counted as they intended, while not letting them prove how they voted to anyone. Google for ThreeBallot, which is a conceptually very simple scheme for doing that. (Most of the proposed schemes involve some fairly hairy crypto protocols, but that one doesn't. ThreeBallot has some security problems that other similar schemes don't have, but the simplicity of the scheme makes it a good one to start with.) The inventor of that scheme has come up with a couple other useful ideas in his life, too.

#32 ::: Sebastian ::: (view all by) ::: January 12, 2008, 02:21 PM:

albatross @ 31:

I've heard in passing of various schemes to allow people to verify their own vote without enabling vote-buying; as you point out, though, most of them are entangled with Serious Math and have usability problems, security problems, or both. The point I was going for is that there is a very good reason fidelio's ATM receipt analogy doesn't apply.

#33 ::: sburnap ::: (view all by) ::: January 12, 2008, 04:20 PM:

A verifiable secret ballot is actually fairly easy. The voting machine assigns a random number to every vote. It prints the number and the vote that it is tied to on a receipt. You must then put this printed number, and the vote, in a big box at the polling station. You can then take this number, enter it in a website and see what the vote was counted as any time after the vote is cast. A machine with access to the numbers and votes (but not names) for *all* voters is in the polling booth, making it trivially easy to write down a number for anything you want to claim to have voted for.

This means that you, the voter, can be sure that your vote equated to exactly what you intended, but makes it impossible to run vote buying schemes or coerced voting, because while you know for sure what your vote is, it is trivially easy for you to lie to a third party about what you voted for and thus makes it impossible for you to prove to any third party that you, in particular, voted a certain way. It also makes recounts verifiable, because the polling station has a big box with paper receipts representing all votes cast.

#34 ::: Mia ::: (view all by) ::: January 12, 2008, 05:08 PM:

Albatross @26
I think it's very important to take note of the problem of "inconveniencing the blind" being used as a defense of not using paper ballots only. "The blind" and people with other disabilities have a right to cast a secret ballot as much as every other citizen does, but without the recent technological advancements, have not been able to exercise that right. Rights are not "conveniences".

I know, because I am affected by this, and quite frankly, entrusting that some friend or stranger is accurately recording my vote for me, and is not judging me for it is massively discomfitting.

This is why it is exceedingly important that there be ways of making sure that the electronic voting systems are secure, and that there is a way of verifying one's vote (like the random number generation technique cited above.) Then I can both vote securely, (those machines have speech output listened to through headphones, and other accessible interfaces), AND check that my vote was recorded correctly (since I can use a scanner and OCR to read my randomly numbered receipt, and log on as everyone else does.)

#35 ::: Jon H ::: (view all by) ::: January 12, 2008, 05:15 PM:

"The reason you can't get a receipt for your vote, as I understand it, is that it would enable vote-buying"

This might be getting deeper into the technological weeds, but I imagine a receipt could carry an encrypted vote written as a barcodey thing, which could be read and verified using a machine after voting. Being well-encrypted, nobody outside the voting location could find out who the voter voted for.

#36 ::: Lydy Nickerson ::: (view all by) ::: January 12, 2008, 05:32 PM:

If I was doing vote buying and the receipt thing was in place, I'd just demand the receipt from the person I thought I'd bought the vote from, enter the information, and find out how they voted. I don't see any way to ensure that only you can enter your own receipt number in the system.

#37 ::: P J Evans ::: (view all by) ::: January 12, 2008, 06:02 PM:

"Secret ballot" is actually a recent idea. I've voted where everyone was handed a ballot and pointed at a table with pencils for marking it. (Think of all those tests in school, where you learned (presumably) not to look at the other test papers.)

What I want is honest counting, and if that requires putting elections in charge of, say, the Nevada Gaming Commission, then that's what we need to do.
(The other thing I want is two boxes after each name, one 'for' and one 'against'. Whichever candidate has the least-negative total after all the ballots are counted is the winner.)

#38 ::: Sten Thaning ::: (view all by) ::: January 12, 2008, 08:06 PM:

Would it work to create one receipt for each party/candidate, but print the actual voting receipt first? As the voter would know which piece of paper he got first, he can verify the vote, but since the papers are identical there is no way to prove it to someone else.

For electronic voting boots, is there anything stopping the voter from taking a picture of the screen? I'm not very clear on exactly how these machines work...

#39 ::: James D. Macdonald ::: (view all by) ::: January 12, 2008, 08:12 PM:

Here's what would work:

Print a paper receipt showing the actual vote. The voter gets that, verifies it -- and deposits it in a ballot box on the way out of the polling place.

If there's a recount, those printed paper receipts are what's counted.

#40 ::: Nicole J. LeBoeuf-Little ::: (view all by) ::: January 12, 2008, 09:13 PM:

Boulder County does something like what you suggest, Jim - and also makes independent voting for the blind possible, which, as Mia points out, is important.

I think the 2004 primaries where when we started having one voting machine per precinct. This voting machine has very simple controls - pretty much a selector wheel and a button. It has headphones so that the blind can navigate it via screen reader, and it has a "suck and puff" tube for those without use of their hands. Most of the people who chose to use it (everyone was given a choice between that or the usual Scantron paper ballots, and we poll workers were forbidden to express preference between the two; I really chewed out my fellow poll worker who, when asked, "what's the difference?" answered, "depends on whether you care about fair ballots" - whatever our personal beliefs, the choice is the voters' and not ours to guide) did so based on the Oooh! Shiny! factor.

This machine requires that you confirm your vote three times. First time you confirm that you're done, and it prints your choices on the screen for you to review. Then you confirm again, and it prints up your choices on a paper in a scroll-type printer whose output is locked up. (We poll workers aren't allowed to open it. If we run out of paper, we have to call headquarters to send a tech out along with a witness. It's got a seal on it whose number is recorded at the beginning of the day and checked at the end.) Then you confirm for a third time, and the screen waves a cute American flag animation while the printer scrolls your paper vote away out of view. You're done.

The votes are, however, counted in the Scantron fashion. Even so, it takes awhile; there was a bit of a flak in '04 when by the weekend the vote still wasn't all done being counted. I was very disappointed with my fellow Boulder County citizens, at least those quoted in the newspaper story, for being so well trained by the media as to prize Fast over Accurate.

#41 ::: Don Fitch ::: (view all by) ::: January 12, 2008, 09:21 PM:

Macdonald #39:

That's close to what was done in my precinct in the recent local election. Voter is given a printed-card ballot, inserts this into a frame, marks with a rubber-stamp pen. Removes ballot, checks accuracy (if desired), inserts in Big Black Box counting machine which verifies readability & tabulates, then spits it out. Voter can check again for accuracy before depositing it in the ballot-box. (I disremember whether there was a receipt tab for the voter to remove and keep.)

The dependability of this system hinges on the state's provision that one out of every hundred precincts will be randomly selected for manual-counting verification. I'm a bit uncomfortable about the "randomness" criteria, and know nothing about the security of the manual counting, but I think something like this _could_ work fairly.

#42 ::: Lindra ::: (view all by) ::: January 12, 2008, 11:29 PM:

abi @ 9:

Many belated thanks.

#43 ::: CHip ::: (view all by) ::: January 13, 2008, 01:07 PM:

albatross@26: that's a crude and partial summary of the arguments; you omit, for instance, the commonly-cited fact that all the ID proposals amount to a poll tax.

#44 ::: albatross ::: (view all by) ::: January 13, 2008, 02:38 PM:

Chip #43: Yep. I also left out much detail from the anti-paper arguments, like adding a lot of operating cost and failure points to elections (all election officials who've used paper have horror stories about it, as far as I can tell), some evidence that almost nobody looks at the paper trail in a VVPAT machine anyway, a long history of documented attacks on the paper part of elections, the legal and moral problems of having a whole verification path that only sighted peopel who can read English get to use[1], etc. And I'm not sure voter ID requirements make sense, though it sure seems like they would, if we're actually worried about election fraud.

The interesting thing was the broad pattern of argument being the same in both cases. I listened to an NPR piece on this awhile back, and it was amazing how perfectly the talking points of the opponent and advocate of voter ID requirements were pretty much just repetitions of the VVPAT argument, but swapped around. I was especially taken with the lovely argument that when there's no evidence of an attack you're doing nothing to detect, this is evidence that it's not happening, and so nobody should take any measures to detect it.

#45 ::: Aidan Thornton ::: (view all by) ::: January 13, 2008, 03:46 PM:

From what I've read, people were complaining that the exit polls didn't match the primaries. Be careful where you get your election poll data from - I think that some of the providers fix it up to match the actual results. (This is based on the assumption that the actual results are accurate - they can then get further statistics out of them that they can't from the official results.)

The other detail that you're missing is that there are a number of known ways of exploiting this model of optical scan machine with a moderate amount of access and some technical skills. The usual people were complaining about this before the election, but were ignored.

(For example, if the information available is accurate, they're using the paper tabulator printouts for the vote totals. These are generated by a program - written in Basic, of all things - on the memory card containing the election data. Anyone who can modify the data on that card can reprogram how the totals are calculated. Another method is to preload the card with votes adding up to zero. There's a "zero report" that can be used to check for this but it's produced by - you've guessed it, a Basic program on the same card. Yes, this is a really stupid design.)

Sure, they're only tabulators and there are paper ballots, but unless the paper ballots are actually counted that doesn't matter.

#46 ::: James D. Macdonald ::: (view all by) ::: January 13, 2008, 09:09 PM:

the legal and moral problems of having a whole verification path that only sighted peopel who can read English get to use[

Say what?

That's like talking about the legal and moral problems of using symphony conductors who can read music.

#47 ::: albatross ::: (view all by) ::: January 16, 2008, 08:44 AM:

James #46: Maybe so, but it sure is an objection that gets raised to moving from DREs to DRE with VVPAT. Telling blind voters that the total number of audio ballots isn't enough to swing an election in most cases, so they don't need to be verified, may be a true statement, but it won't make you any friends.

#48 ::: James D. Macdonald ::: (view all by) ::: January 16, 2008, 12:52 PM:

We aren't talking about blind voters. We're talking about the verification process. Is there some reason why a sighted person can't verify an audio ballot?

There are numerous reasons why a blind verifier can't verify a pencil-and-paper ballot.

#49 ::: James D. Macdonald ::: (view all by) ::: January 16, 2008, 12:54 PM:

Incidentally, today's the day the people challenging the ballot need to cough up the money or the recount won't happen. $0.24 per ballot verified. Around $65K for the Democrats and $55K for the Republicans.

#50 ::: James D. Macdonald ::: (view all by) ::: March 01, 2008, 11:21 AM:

Here are the results for the Democratic recount. The recount was stopped on 23JAN08 when the funds ran out.

Here are the results for the Republican recount.

Welcome to Making Light's comment section. The moderators are Avram Grumer, Teresa & Patrick Nielsen Hayden, and Abi Sutherland. Abi is the moderator most frequently onsite. She's also the kindest. Teresa is the theoretician. Are you feeling lucky?

Comments containing more than seven URLs will be held for approval. If you want to comment on a thread that's been closed, please post to the most recent "Open Thread" discussion.

You can subscribe (via RSS) to this particular comment thread. (If this option is baffling, here's a quick introduction.)

Post a comment.
(Real e-mail addresses and URLs only, please.)

HTML Tags:
<strong>Strong</strong> = Strong
<em>Emphasized</em> = Emphasized
<a href="">Linked text</a> = Linked text

Spelling reference:
Tolkien. Minuscule. Gandhi. Millennium. Delany. Embarrassment. Publishers Weekly. Occurrence. Asimov. Weird. Connoisseur. Accommodate. Hierarchy. Deity. Etiquette. Pharaoh. Teresa. Its. Macdonald. Nielsen Hayden. It's. Fluorosphere. Barack. More here.

(You must preview before posting.)

Dire legal notice
Making Light copyright 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017 by Patrick & Teresa Nielsen Hayden. All rights reserved.