That should be Rush Holt (D-NJ). He's a great guy, who knows physics and everything!
Dammit pjs, you stole my nitpick.
Thanks for the spelling correction. Yeah, apparently one of his bumperstickers is "My Congressman Is A Rocket Scientist".
Has there been any discussion of an investigation, congressional or otherwise?
So let's say I were writing an email to Carolyn Maloney. I'd be asking her to support the Electoral Fairness Act of 2006 (H.R. 4989), right?
Also, while I don't think I have a problem with issuing proof of registration cards to all registered voters, surely this will lead to some anecdotes of people who are registered voters but don't remember or care to carry such cards being denied access to the vote because of poll worker confusion as to what they mean.
The voting machines used in Mississippi have a "verifiable paper trail."
Sort of.
When you go through the process of electronic voting, the machine produces what looks like a credit card slip visible in a little window to the lower right of the machine. *If* you think to study it carefully and *if* you read it to match against your vote, there is that piece of paper that is auditable. It is not convenient or easy to check for the limited number of people who might think to check it. I'm hugely dubious about it as a solution. Though it is better than nothing.
I absolutely fail to understand why the scanning method we previously used in this county did not work well-- you just shadded in the appropriate cell, a machine counted it, and the machine count could be verified by hand. But guess what? Republicans in Congress and the state legislature have been pushing for these new machines.
Under the relatively low pressure environment of voting here, the new machines did seem to work in the sense that people voted wtihout delays. I view as almost unknowable whether they'll produce accurate counts in a tight election.
I am friends with the Assistant Secretary of State in charge of elections in Mississippi, and, when the decision to buy these machines had not yet been made 18 or so months ago, I called him and started to explain the problems I'd heard about. I got no reassurances other than that there'd be a paper trail.
Yeah, there were stories this election about places where the 'paper trail' on that sort of machine jammed, and they just went on voting regardless, without fixing it. I want the printout to be the ballot -- the computer can provide accessibility, and can keep a running unofficial tally, but I want a human (and computer -- optical scan works) readable paper ballot to be the official vote.
I know I am not entitled to query this, being a foreigner with precisely zero knowledge of US voting procedures and the relevant law, but the "but, if" question shouts for attention: If 18 000 lost votes have "decided" a seat, then is that election not void, thus demanding a new poll?
Or do I miss an "obvious" truth elsewhere?
Put another way... that story is just plain surreal, and frightening. Big Time.
I believe that question is being decided by the Florida courts now.
What is the evidence that the votes were lost due to machine malfunction rather than bad ballot design?
I think the US has a much more conservative view of elections than a lot of other democracies, maybe because our current goverment has been around the longest (biggest recent change being direct election of senators in 1913) and so made the early mistakes, or at least didn't have other examples around to learn from. The government doesn't get to call elections when it wants, first-past-the-post, no votes-of-no-confidence, etc.
So the standard for "proof" is pretty high, probably higher than it should be. A bunch of ballots that were largely Democratic but with no vote for representative is exceedingly suspicious, but hey, there's no proof. Yay for local control of elections, or federalism, or something.
Thirteen percent of the voters missed one of the highest profile races in the district. While bad ballot design might result in people not remembering to vote for dog-catcher, most people walk into the voting booth with a formed intention to vote for someone for Congress. That bad ballot design could explain that number of missing votes is literally incredible.
Of course, there's no evidence of what happened beyond its incredibility. That's the beautiful thing about not having a paper trail. There's no evidence of anything.
I think that an organized, multi-prong stealth campaign to discourage voters is the explanation of a lot of stuff. The SecState in MN here (a squeaky-clean state) was involved in a sneaky trick to discourage native American voting.
I'm sure that it's all justified with pious shit about Discouraging Fraud, but Republicans know that on the average, low turnout and difficult voting help them. If every Republican official everywhere knows that, every judgement call goes the same way automatically, and there will be various svattered voter-restriction initiatives. (In other words, you don't need the coordinated conspiracy the anti-conspiracy-theorists joke about).
Yeah, at least here in MN there's an actual, paper ballot for every voter, with pretty tight controls on the transportation and disposition of the ballots.
Rush Holt is definitely a right-on guy. I doubt there'll be much dissension from either the outgoing (Sabo) or incoming (Ellison) representatives in my district, but it sounds like a decent initiative. Having said that, I still don't see why we can't mandate same-day registration nationwide. The Democrats could pick up like 10 million votes in one fell swoop. Also, there ought to be motor-voter laws in every state (and DC), and there should be mandatory prime time PSAs with detailed, location specific instructions on voting every election season on radio and TV. And there ought to be some kind of mechanism by which precincts which historically have long lines of people waiting to vote get split up, or add staff or something.
Right. Clearly, there was either fraud or a bug in the voting system software. My money is on the latter. Reading the articles though, I'm not sure a paper trail would have helped. If the machine isn't telling you that you voted for Jennings, even though you wanted to (which is what the complaints in the newspaper articles seemed to be saying), and you shrug and say "hey, ok, maybe it actually worked, maybe it didn't, but I have to go home so I won't worry about it"... you're screwed regardless.
11:Yes, and I generally approve of this American rigour. It is in part a conservatism about elections, but it is also a sense of the law. The example here is that, IIRC, the voting has to be done on the first Tuesday in November. We cannot constitutionally have a "redo." It should go to the House, where the results must be validated. As the 2000 election should have gone, to the Florida or US House, instead of the Courts.
There is an opportunity in this rigour, the fact that no matter what is done on Election day, at midnite the votes are tabulated and the game over, subject only to appeal of, I think, the newly elected house.
Voting fraud & force might get you sent to jail, but in certain scenarios it would be worth it.
Incidentally, if the election is validated by the "old" house instead of the "new", then Democrats have not won any house seats...yet.
12 Via Big Tent Democrat , the ballot . I think it would be fairly easy to overlook the Congressional race on the second page. Whether this plausibly explains a 13% undervote is another matter but I do not find the idea incredible on its face.
17, 18: and warning - eye-bleedingly poor ballot design. I wonder what kind of checkmark appeared when a vote was made.
14: if one of the biggest perceived advantages to a change in voter registration law is that it will add 10 million Democratic voters in one fell swoop, it shouldn't be a surprise that it hasn't happened.
Couple of things of interest here:
1) a TPM link explaining why there are likely missing votes and that they are overwhelmingly Democratic votes in that Florida race.
2) Another TPM link explaining the "audit" they're conducting this week.
It's a joke. The "audit" will consist of test voting on some backup machines (they won't be using the actual machines, which are off-limits during the recount), to be performed by partisan officials who had conducted tthe election.
Oh, great. And no effort to see if there might be something in, say, the software or design of the machine that could have produced a miscount.
Congress has the option of refusing to seat the wrongly-elected Congressperson, and they should exercise this option.
It's a weird system where election rules are purely local, but seating someone is a Congressional prerogative.
20 According to a local paper on 17846 ballots with no vote for Congress, Democrats led in the other races by 3-7%. This is not what I would call "overwhelmingly Democratic" although it might be enough to change the result.
Well, according to the Orlando Sentinel, it's enough to swing the election if the D candidate for Representative did as well among the undervotes as any other Democratic candidate, even the least popular, by a margin of over a hundred votes.
All right, a little o-earnestness never hurt anyone. I sent a letter to the editor.
If anybody wants to crib from it, e-mail me for the text. (I don't think that highly of my writing -- I just know how handy a form letter can be.)
That's odd. E-mail should have shown up as macdanceatgmail.
You want conservative approaches to elections? In Britain we still vote by making an "X" against a name on a printed paper which we then stuff in a tin box. Counting (and recounting) is manual, by teams of municipal employees on overtime, and 95% of divisions are called before dawn the next morning. So it has been, AFAIK, since secret ballots were introduced in the middle of the 19th century, and it fucking WORKS.
What the hell is it with you people and half baked technology?
OFE: Some parts of the US use that method, but it isn't convenient for the TV networks. "Dawn the next morning?" Ha!
John - Then American TV producers have got to be twats as well. Watch the coverage of a British election some time - the best media circus ever! Tightrope walkers - faded politicos trying to spin their party's fortunes on the hoof, Clowns/psephologists, Lion tamers - a couple of journos with no manners who aren't afraid to embarras the leaders, custard pie fights in the studio, plus the fact that the results come out all through the night gives you plenty of time for drinking games. There's no party like it.
Also, it ends with a fair and incontestable result (not that the TV people care about that).
Then American TV producers have got to be twats as well.
That's a bit of British understatement there. American TV people are the most hideous people on the face of the earth.
The particular issue is that they want to package the election in a predictable chunk in order to sell advertising, but without impinging on other things they can also sell advertising to.
Yep, one of the few things the Brits and the bits of continental Europe I know have in common is a ballot system (even if the representation systems are different) and a glassy-eyed daze over stories of american voting machinations.
But hey, you're big. So you get to define
democracy. Whose going to tell the Swiss though?
30 gets it so right.
Jesus Christ. Yes, with all the money we would save from not instituting those stupid-ass machines and butterfly-ballots and paying for recounts and lawsuits, the government could hire a KICKASS circus for each goddamned TV network. We could call it the "Orgy of Democracy!".
Well, machine counts are not just faster but more accurate than hand counts, if the machine actually works. For this reason it's good to have a machine-readable ballot that can also be read, if needed, without the benefit of the machine.
And using a single voting system for over 100 years doesn't necessarily strike me as a particularly reliable way to avoid fraud, but a way to refine and perfect it. One advantage of changing systems is that it requires any cheaters to change their method of cheating as well, or else give up on it, both of which are easier to detect than consistently cheating the same way for 20 elections in a row.
Technology does have a lot of good things to offer to elections. It's a pity that the people on the forefront of implementing it seem intent on confirming people's worst fears about voting machines.
Well, machine counts are not just faster but more accurate than hand counts, if the machine actually works.
Got some basis for that? Seriously, my understanding is that the gold standard for checking whether any machine for reading a ballot that is also human readable is to check it against a hand count.
35,36: It's also a pretty bald assertion to assume that introducing technology makes cheating harder. It could work as a just-so story, but I'm skeptical that it's true. Particularly with this latest round (diebold etc.) it's pretty clear that the technology has been designed with an eye to allowing fraud, or at the very least no serious effort has been made to counter it.
I don't know how it's done elsewhere, but in the UK ballots are sorted by candidate and then counted by a different person, providing a level of double checking at the outset. Also, accredited representatives of the candidates are present and can intervene whenever they want if they suspect funny business.
Technology should have already made elections obsolete. Time travel, people -- when the fuck are we getting our time machines? Once we can look into the future and check on who was elected, the election itself is just a silly formality.
That's because a hand count is the only way to prove that the machine is counting correctly, not because a hand count is more accurate. (This was of particular importance, of course, in Florida in 2000, when we all learned that the machines are very bad at counting punch-card ballots correctly. We also learned there, that invariably, if a county does a hand count twice, it comes up with two different numbers.)
Although this isn't the source that I learned it from, this paper (PDF) analyzed 415 elections in New Hampshire which had recounts. It found that the discrepancy between the initial count and the recount was just under 1% for hand counts (over approx. 50 years) and 0.5% for optical scan (only in 2002). This is not the same as error rate, it's just what sort of variance you get with both counting methods.
It's also worth pointing out that the error rate for a human hand count changes dramatically based on human factors, such as how many ballots each person has to count, how quickly, how much of a break they get, etc. I suppose there are similar factors for machines based on how well-kept they are, but the point is, human error in counting is a major factor and is worth eliminating.
(Human error in casting a ballot which is correctly machine-read is a separate factor, although I think I've seen other studies that show optical-scan ballots have a lower spoilage rate as well, which is a good way to estimate the unmeasurable quantity of how many voters had their vote counted in a way they did not intend.)
37: I didn't say introducing technology made cheating harder and I vigorously dispute the assertion. But changing the counting method makes cheating harder for established cheaters because they have to devise a new method of cheating which is consistent with their earlier method.
(The counterpoint is that introducing a new voting system is an excellent way to begin cheating where you weren't before. But I assume we all already knew that.)
I have not read that there were systematic problems in the old methods, and the problems that there were were a function of corruption and machine politics. Voting machines are a solution without a problem, as far as I can tell. Getting the results in time for TV is not a good enough reason. There has to be a better way to help bling and handicapped people.
As people have pointed out, ATM machines are virtually error-free. If voting machines aren't, it's very fishy. This isn't rocket science.
It's really a political breakdown -- the makers of voting machines have been allowed to make them closed-source, so that no one but the patentholder can know what's going on. No one should have consented to that.
Also, pretend that I emphasized for established cheaters in the end of 40.
42 -- bling people are fine, they don't need any help from us.
But changing the counting method makes cheating harder for established cheaters because they have to devise a new method of cheating which is consistent with their earlier method.
And it makes cheating easier for new, entrepreneurial cheaters. What is this -- creative destruction and the circulation of cheating elites?
I don't think that anyone here is dogmatically against any particular method, we're just pretty irate at what has actually happened.
And, ATM machines have been pretty error-ridden over the past little while -- didn't you notice all those Kittys of Doom? We'll see how the new server holds up -- doing pretty well so far -- but "virtually error-free" I think is an overstatement.
(Human error in casting a ballot which is correctly machine-read is a separate factor, although I think I've seen other studies that show optical-scan ballots have a lower spoilage rate as well, which is a good way to estimate the unmeasurable quantity of how many voters had their vote counted in a way they did not intend.)
You know, if a person successfully marks a ballot in a manner that's unambiguously readable to the human eye, but the machine's limitations keep it from reading it correctly, there's something wrong with eliminating that from your calculation as to whether hand-reading of ballots is more accurate than machine reading. Whether the error should be blamed on the voter or the machine, the point is that a person can accurately read many votes that a machine can't.
neil: I don't think anyone would disagree that machine counts are potentially more accurate. Hand counts will have an error, but if done with care it will be small. More importantly, there are practical methods to counteract fraud that work well.
Machines, on the other hand, are easier to game, and I think it's a pretty strong assumption to presume that these methods are actually being applied with best intentions. There is a long and colorful history of electoral fraud, and no reason to believe that has changed.
*Particularly* when you see mechanisms like the diebold machines. Here the only possible conclusions are that the makers are stupid or they are crooked. Which is hardly heartening.
The only sensible way to approach this sort of thing is assume that people are going to try an cheat, and try to make this very difficult. This is clearly not done in a lot of cases.
It's really a political breakdown -- the makers of voting machines have been allowed to make them closed-source, so that no one but the patentholder can know what's going on. No one should have consented to that.
This too -- next to a voter-verified paper ballot, open source software is the biggest thing I want out of any voting machine.
40: Ok, I'll buy that it creates a nuisance as well as an opportunity; I don't think the nuisance is a big deal for them and the opportunity is sometimes very large.
47: I was just clarifying that that factor lies outside the statistics I cited.
Although I also think it is worth separating the factors when discussing, in the abstract, the advantages of a hand-counting system versus those of a mechanical-counting system. The human error involved in counting is impossible to eliminate and can only be mitigated to a certain point. On the other hand, 'unambiguously' human-countable ballots being machine-uncountable is a design issue, and one that I'm not sure is unsolvable. (If nothing else you could have those ballots hand-counted.) (Note that most misvotes are ambiguous. If you circle a candidate's name when you were supposed to put an X next to it, that's ambiguous in that a human counter could easily miss that vote if he's only looking at the line where the X is supposed to go. The 'stray mark' and 'hanging chad' and 'wrong pen'-type flaws are what I'm thinking about when I say solvable design issues.)
Also: The ATM comparison is fatally flawed because an ATM is intended to record the entire transaction. ATMs have redundant data storage out the wazoo because the people at both ends of the transaction want to be able to track down any inconsistency. Conversely, a voting machine is unacceptably flawed if it provides any way of reconstructing the transaction (i.e. connecting a voter's identity to her ballot). It's common sense that this requires a tradeoff of accuracy.
51: Sure, the ATM comparison is flawed, but the current voting machine implementations don't even get the basics right. Some of the implementation details of ATM networks would potentially be useful to address that.
no that I'm a fan of the ATM comparison... I don't find it that helpful.
My conspiratorial side thinks, if you're an established electoral cheater and you want to keep technology from making your work harder, what do you do? Perhaps you promote a fatally flawed electronic voting machine in order to build support for the status quo. FL-16 seems to provide the perfect test case: the result of the election was obviously reversed, but there is just as obviously no fair way to fix it. Better just to go to those old punch cards... why, Katherine Harris's successor kept all the old readers in storage just in case this were to happen...
..uh, my conspiracy theorizing side, i guess.
To me the ATM comparison is completely valid. People care whether they work, and they do work. If they didn't work, they wouldn't be out there.
Voting-machine promotion seems to have consisted of a series of one-off deals between entrepreneurs and faceless, small-time politicos who might have been bribed.
56: I think there are two ATM comparisons. On a technical/network level, they don't really compare so well. On the level of `box we want to rely on' level, it works.
Neil was objecting to the former I think.
The only thing I object to as a comparison is that because they are superficially similar, it can be misleading to think of voting machines as being `like' ATMs.
Personally, I'm in favor of any analogy that will keep people saying "ATM" once or twice per comment.
(We don't really see enough of that here ATM.)
One big difference between ATMs and voting machines is that ATMs are in use at full capacity 24/7, and are attended to by people who have that as their full-time job, while voting machines only get used on a big scale a couple of times every two years or so, and are attended to primarily by volunteers who have received a few hours training. So problems with ATMs tend to get found and fixed quickly, while large-scale problems with voting machines only show up when the stakes are huge. If we used our voting setup regularly to vote for American Idol winners or something, we'd probably shake the bugs out a lot quicker.
37: I generally tend to believe in the maxim "Never attribute to malice that which is adequately explained by incompetence," and I think that applies here. As a software developer, most of the Diebold, etc., misfeatures seem to me like classic symptoms of a company trying to put a limited amount of effort into modifying an existing product quickly to serve a new market, rather than designing a brand-new product from scratch. They are just getting a lot more visibility because of the stakes involved. Now if those misfeatures are allowed to persist, I would expect to see a growing amount of fraud over time exploiting them. I just don't think that fraud was part of the plan from the start, as opposed to "grabbing market share in the electronic voting market."
That's the great thing about this race. There's no particular reason to suspect fraud rather than accident, and nothing huge riding on the outcome, which means that we can focus on the fact that it looks very much as if the voting machines have produced an incorrect outcome and are unacceptable for that reason alone, regardless of whether anyone did it on purpose.
60: `misfeatures' is too easy. I'm somewhat willing to go along with an argument of incompetence rather than fraudulent intent. I'm not convinced, mostly because there is a long history of electoral fraud.
From a development point of view, sure, you can understand they could have been cutting corners. But somebody had to design the thing, and it's as if nobody who really understood elections or electoral fraud or computer security ever looked at these things. So at that level we are stuck choosing between malfeasance and gross incompetence, regardless of the poor schmucks who were asked to bang something out the door quickly.
There has also been congressional testamony to the effect that in at least one case, fraud was part of the plan from the start.
I generally tend to believe in the maxim "Never attribute to malice that which is adequately explained by incompetence"
That maxim has been downgraded to a cliche and should be used only with the utmost caution.
The new version of that cliche is: Sufficiently advanced incompetence is indistinguishable from malice.