Ed Brayton is a journalist, commentator and speaker. He is the co-founder and president of 
With much controversy over allegations that some touch screen voting machines in West Virginia and elsewhere are switching votes from one candidate to another, a county clerk tries to demonstrate how well the system works and ends up showing that even after recalibrating the machines, they still don't work right:
Touch screen technology really isn't that difficult. It's been around forever. We've all used ATM machines with touch screens a thousand times and likely never seen a machine switch a $20 withdrawal to a $100 withdrawal. No bank would tolerate a touch screen that didn't register accurate results; our voting machines should be at least that good.
Politics
Comments
Here we go again. Voting and ATM's machines are very different. Touchscreen ATM machines are a small percentage of the overall number of machines out there and they frequently have buttons as backups. Secondly ATM machines have a very basic purpose with very few options that don't change that often. Thirdly - the number of ATM machines is much smaller than the number of voting machines used on election day. Therefor the quality is different to save costs. Fourthly - ATM machines are setup by trained technicians - not 80 year old blue haired ladies. Lastly and more importantly - ATM machines are a disaster from technical security perspective. The risks are mitigated by the fact that they are basically vaults with built in location devices and cameras connected to a central mainframe. Please tell me how that's the same as a voting machine?
The bottom line is that the voting process is not the same as dispensing cash and people need to stop thinking like it is. There is no over lap. Voting is a infrequent event that has to start and finish on one day for all states but one. Rules are different from state to state and each election is unique. It needs to be anonymous. Its needs to verifiable by all parties (not just your bank) which means a paper ballot.
The state of Minnesota uses pencil, paper, and optical scanners. That is the best tools to use for voting and we have one of the highest voter turnouts in the country. This is not a complicated issue so lets stop making it one.
(don't get me started on requiring IDs for voting)
Posted by: yoshi | October 30, 2008 10:02 AM
There should always be a paper trail. Paper, pencil and optical scanner are pretty good, but not perfect. The optical scanner should print a receipt that can be compared to the ballot.
Video voting machines should print a receipt that clearly shows how you voted.
Posted by: calladus | October 30, 2008 10:12 AM
Btw - according to the AP - Maryland and Virginia scrap their touchscreen machines and will return to paper ballots after this election.
Posted by: yoshi | October 30, 2008 10:19 AM
@calladus
I think you are missing the point - complicated tools like video voting machines don't provide the reliability needed for an election. Its the wrong tool for the job.
Posted by: yoshi | October 30, 2008 10:22 AM
I don't understand what's wrong with the old fashioned paper-pencil system.
Many countries -- even modern democracies -- have used it successfully, with voter turnouts that can add up to more voters than you get in most US states.
It's also pretty secure -- the only way to do voter fraud with paper ballots in a box is to have a lot of poll workers on the take. While that is possible, it's certainly less so than it was back in the day when many states and precincts were one-party states (think Southern Democrats) and there was no oversight. These days it would be harder, with poll workers from both campaigns, to get away with that with enough votes to make a huge difference.
Screw the touch screens. Give me a paper, pencil and a tick box. If it's good enough for France, Britain, Germany and Italy, to name a few, it's good enough for us.
Posted by: Jesse | October 30, 2008 10:36 AM
The second "error" does not appear to be a fault of the machine, particularly a calibration error as the man points out.
Here's what I see:
* After calibration, no candidates are selected.
* At 1:33, he taps the 2nd party choice (apparently straight Dem.) This automatically selects the 2nd pres (apparently Obama)
* At 1:37, he hits the 4th pres (apparently Nader) to continue demonstrating that calibration is fine. This keeps the straight party ticket he had previously selected, but overrides the pres choice.
* At 1:42, he hits the 1st party choise (apparently straight Rep.) This takes him to a screen which asks him to verify something. He selects the green choice.
* Returning to the main screen, the straight Rep. choice is selected, and the 4th pres choice (apparently Nader) is *still* selected.
I can't read it, but I bet the screen at 1:42 asked him whether he wanted keep his previous pres choice or switch with the straight party choice. Remember, he didn't get this screen when he hit straight Dem, but there were no previous choices at that point.
This is not to say that TS technology is better or worse than any other method (I prefer optical scan), but this case appears to be user error.
Posted by: Mr. Upright | October 30, 2008 10:41 AM
Yoshi, Im sorry, but I have to question some of your assertions. (1) Touchscreen ATMs are probably a minority of all ATMs, but I would take issue with the claim of "a small percentage." Actually, in my area, they seem to be a slight majority, but I would expect that machines in more rural locations would reduce that percentage. (2) Ill buy this. I think it could be stated that Voting machines have a very basic purpose too, but you are correct that the changing of the options on voting machines is a significant issue. (3) This does not seem to correspond to my experience. Most towns in this area use between 6 and 12 voting machines. Most towns in this area have at least 4 bank branches with 2 ATMs each. So the numbers seem comparable. (My previous voting station used 8 machines itself, but had 20 ATMs within a 2 block raidus, but I dont think NYC is representative.) (4) Voting Machines, atleast in this area, are not set up by blue haired old ladies. They are set up by trained technicians. Blue haired old ladies check your name off the voting rolls, and you report problems to the same BHOLs, but they call trained technicians to solve any problems. I know some of those technicians, theyre not the always the sharpest sticks in the bunch, but neither are those setting up the ATMs, and I know some of those as well. Your last point, I agree with. ATMs are a disaster from a technical security perspective. (I know of one major ATM switch that uses wireless regularly in their ATM installations.) I would suggest however, that the risks are primarily mitigated by the owners and operators of the ATMs having a vested interest in preventing fraud (they lose money) and a judicial system that takes a very dim view of bank robbery.
Posted by: Dave | October 30, 2008 10:41 AM
There's no question that certain varieties of electronic voting machines are unreliable. The question for me, however, is: is there a pattern to the errors? Do they consistently favor one party or candidate over others? The fact that Diebold machines' errors have rather consistently favored Republican candidates seems to me to be rather blatant evidence of tampering, which electronic machines are always especially vulnerable to.
Posted by: gary l. day | October 30, 2008 10:59 AM
Yoshi, to list just two examples, both the iPhone and Nintendo DS use touch-screen technology that's accurate with a slew of different software setups (and on a far smaller screen with tighter tolerances), and are produced in quantities far exceeding voting machines. Maybe ATMs aren't entirely comparable to voting machines, but I still can't see any valid reason why modern voting machines should be this unreliable.
That said, I agree that it is technological overkill. All voting should be done with a pencil and paper, or at a minimum a device that leaves a clear paper trail.
Posted by: H.H. | October 30, 2008 11:02 AM
Yoshi, please stop saying "ATM machine". It's giving me a headache.
Posted by: Ryan | October 30, 2008 11:12 AM
I love our voting "machines" here in wisconsin. You draw a line with a black marker on a piece of thin cardboard/paper, which is placed in a machine that collects and opticaly scans the votes. It's simple, fast and has a paper trail.
Posted by: Caliban | October 30, 2008 11:15 AM
Because I was curious, I exercised a little google-fu, and it appears there are approx 400,000 ATMs in the US (1.7 Million globally) compared with 180,000 voting precincts in the US.
Oh and Ryan, http://ars.userfriendly.org/cartoons/?id=20080614
Posted by: Dave | October 30, 2008 11:21 AM
Attention: I am a computer programmer with over 10 years experience programming touch screen interfaces.
Yoshi, I'm not sure what you are trying to say.
The part of the ATM or voting machine that is out of calibration is the touch screen.
The job of the touch screen is to say "This point, designated by an x, y coordinate pair, was touched." It is up to the software running on the system to map that x, y coordinate pair to a button on the screen, and then to a database entry corresponding to a candidate or a position on a Yes/No ballot measure.
The fact that ATM interfaces are relatively static, and that voting machines are not, is an issue of the database programming and interface software, not of the touch screen. Recalibration does NOT change the database or the software -- just the conversion of the physical touch to x, y coordinates.
Thus, miscalibration is no more of an issue for voting machines than it is for ATMs, and should rarely be an issue in the first place, UNLESS, in fact, the voting machines are made with cheaper screens, a supposition for which I would like to see some evidence.
So, Yoshi, while you may say that ATMs and voting machines should not be compared, there is no reason why that would be the case with regard to the specific issue of touch screen calibration, which is the issue in question.
Posted by: xebecs | October 30, 2008 11:44 AM
There's no paper trail for the machines in my district, but other than that they're very reassuring. You push a real button, which makes a satisfying "click", and a big green X appears next to your guy's name. It's a toggle, so you can just unclick if you want. When you're done, you press the button at the bottom, and it goes "ka-chunk!" and registers your vote. Presumably.
Posted by: Dave M | October 30, 2008 11:47 AM
yoshi,
I agree that video voting machines don't provide reliability. What they provide is accessibility, which may be considered important to some people. But unless the voting machine prints a receipt ballot, I think it would be suspect.
I even have a misgiving with the optical scanner and paper ballot. As an electrical engineer, I can't help but wonder how easy it would be to rig the optical scanner? I want the optical scanner to print a receipt that I can use to compare with my ballot.
And I want all receipts to have a unique serial number, so that in the case of a recount, officials can ask for and receive help from the voters.
But I'll admit that I've never studied or even read about voting practices and what works and what doesn't work. So my opinion is uninformed.
Posted by: Calladus | October 30, 2008 11:51 AM
the paper trail in use with these machines needs to consist of more than just a paper receipt handed to the voter. there should be 2 copies of that receipt - one for the voter (so they can see what the machine has supposedly recorded as their vote) and one for the precinct (so that should the machine tally be in question, they can go back to those paper receipts and compare). without the precinct retaining a copy that has been approved by the voter, there is no means to verify the machine's tally, as the machine can record one vote, but spit out a receipt that shows something entirely different.
Posted by: arin | October 30, 2008 12:10 PM
No.
The results that optical scanners forward can be, and often are (Ohio 2004!), faked; and if they aren't close (or if the Supreme Court stages a coup -- Florida 2000), the ballots aren't recounted, or a nonrandom sample of ballots is recounted (New Hampshire 2004).
Count the ballots by hand in the polling station. That's how First World countries do it.
Poll workers from each party, to be precise, because -- except in the USA -- observers from all parties (I didn't say "both", mind; I said "all") are allowed to be present in each polling station.
Here's an explanation of the election laws of Italy. You'll like it. In fact, you'll join me in being flabbergasted why elections are conducted in such a dilettantic manner in the USA, despite 250 years of experience!
They did in 2004. Almost all "irregularities" nationwide, out of tens of thousands, favored Bush.
And even if not, what does it help if all of the votes of the state are first sent to Kenneth "Katherine" Blackwell's desk and then counted on the computer there??? Because that's what happened in 2004.
Posted by: David Marjanović | October 30, 2008 12:17 PM
Everybody read Mr. Upright's comment. The machine WAS in calibration after he recalibrated it. He had already selected Nadar, and it kept that selection because Nadar is from a third party.
Posted by: 386sx | October 30, 2008 12:17 PM
Calladus said:
"I even have a misgiving with the optical scanner and paper ballot. As an electrical engineer, I can't help but wonder how easy it would be to rig the optical scanner"
I'm with you. IIRC, an analysis of voting anomalies in Ohio indicated that it was a particular model of either Diebold or Ess optical scanner that was producing funny results in black precincts.
And ANY electronic device is susceptible to vote fraud not only on its own hardware/software, but in addition they are usually tied up to networks, which use compilers. This is another place where attacks seem to have occurred.
I agree with Greg Palast, who says it makes no sense whatsoever to have our nations votes being counted by any private company. It really is outrageous.
I am really really sick of seeing Republicans win elections with results of 51% to 49% when all polling showed them to be expected to lose.
Voting on a piece of paper is used almost everywhere else in the world, and the results are available the next day in England. There is absolutely no reason whatsoever that I can see to justify our system.
Funny, I think, that Republicans LOVE these voting machines, and pushed through the law to mandate them. Based on an argument that they - get a load of this - were needed to help the handicap vote!
Posted by: Gingerbaker | October 30, 2008 12:18 PM
ATM-style voting machines are in the class of redundant technology: technology that can do the job it's designed for, but not in a way that's demonstrably superior to more low-tech methods. Since low-tech is inherently more relible than high-tech, it should be adopted unless high-tech provides other clear advantages. (But, hey, I'm still fuming about the fact that you can't buy cars with manual windows anymore. I fixed the window crank on my old truck, but can't on my new car.)
Posted by: James Hanley | October 30, 2008 12:23 PM
Here's the full video, by the way:
http://www.youtube.com/watch?v=sc9Gd5g3DFY
About 27:40 is where he votes after it's recalibrated.
Posted by: 386sx | October 30, 2008 12:41 PM
We had a very complicated voting machine at the last election I voted it.
It was a high tech graphite carbon deposition device, that was powered by oxidative phosphorolation. This was then stored in a long term, physical memory before being passed to an optical reading system.
Posted by: Donalbain | October 30, 2008 1:14 PM
The county clerk implies that if the machine (actually the touchscreen) is "out of calibration", the choice always goes down, not up from Obama to McCain.
There is no reason that the calibration couldn't be off in the other direction.
And how come the error is so far off? It looked like a good fourth of the screen height in the video. I would expect a touchscreen which had been previously calibrated would develop a error only a few pixels off, if any, unless there is a problem with the hardware or driver. Otherwise, I would start to be suspicious.
Posted by: Gary J. Bivin | October 30, 2008 2:10 PM
PS. Note that with recites at the voting booth - or with a small digital camera (in your cell) - you have a record you can provide to the people who bought your vote.
Posted by: llewelly | October 30, 2008 2:27 PM
I disagree; having a receipt that the voter keeps endangers the secret ballot guarantee, as he can be accosted and his vote checked. The voter should get a single receipt, which he checks and then drops into a ballot box. The voter is able to check his vote, and the hard copies of teh ballots can be used to spot check for machine errors.
Posted by: Shygetz | October 30, 2008 3:03 PM
"Save costs" on voting machines? Are you kidding me? If there's one thing on which everyone can agree their taxes should be used, it is VOTING MACHINES. I don't care how much they cost. The need to work.
Please tell me why the voting machine shouldn't be the same way.
Posted by: FishyFred | October 30, 2008 3:10 PM
Alternative: The voter can see the paper trail through plexiglass.
Posted by: FishyFred | October 30, 2008 3:11 PM
Touchscreens suck. Why can't they just use buttons?
Posted by: bullet | October 30, 2008 3:33 PM
"...having a receipt that the voter keeps endangers the secret ballot guarantee, as he can be accosted and his vote checked. The voter should get a single receipt..."
Man, where are you voting! :D
If it has gotten to the point where people are being accosted and their votes checked, worrying about how many receipts we get is kinda irrelevant.
Granted, the idea that we even think to have a secondary receipt in case the machine records are "lost" or look suspicious says all we need to know about electronic voting.
Posted by: Gingerbaker | October 30, 2008 4:49 PM
Since most people seem to agree with the need for a papertrail, then why the hell are we worried about computer systems and about 30 different printouts to verify votes were cast properly? Just go back to the damn paper/pencil/scanner systems which seem to work just fine. Jimminy.
Posted by: TomJoe | October 30, 2008 4:50 PM
Hack a voting machine in 7 minutes:
http://blogs.spectrum.ieee.org/riskfactor/2008/10/want_to_hack_an_evoting_machin.html
Posted by: KeithB | October 30, 2008 5:17 PM
In my neighborhood I used an electronic machine that prints a receipt that it retains which shows the selections you made. I liked the idea of the Diebold machines with no paper trail, and the president of the company saying in a fund raising letter that he was committed to delivering Ohio's 20 electoral votes to George Bush.
Posted by: soboco | October 30, 2008 5:33 PM
"Here we go again. Voting and ATM's machines are very different. Touchscreen ATM machines are a small percentage of the overall number of machines out there and they frequently have buttons as backups. Secondly ATM machines have a very basic purpose with very few options that don't change that often. Fourthly - ATM machines are setup by trained technicians - not 80 year old blue haired ladies. Lastly and more importantly - ATM machines are a disaster from technical security perspective.
The bottom line is that the voting process is not the same as dispensing cash and people need to stop thinking like it is. There is no over lap."
interesting. being married to a bank auditor i have a slightly different perspective. touch ATMs are not necessarily a small percentage, depending on your location and bank. ATMs can easily have as many options as voting machines, or more, and often have to interface with far more subsystems. further voting machines are very often setup by a tech from the company that builds them. at the very least that is who trains the people who set them up.
lastly, if there is no overlap between voting and dispensing cash (and i think this is important) why are we letting the same companies build the machines for both?
Posted by: shen | October 30, 2008 5:49 PM
Speaking as an Australian, the U.S. voting system continues to baffle me. Here we have compulsory voting, so 100% turnout, pencil and paper voting method, and my whole life I have never waited more than 5 -10 minutes in a line to vote. I can't recall a single issue of alleged voter fraud either, and because of compulsory voting, we have none of the efforts to discourage potential voters.
It all seems to easy to fix from our end..
Posted by: Jeremy | October 30, 2008 6:02 PM
Sorry, but this whole issue just seems absurd to me. If there is ANY legitimate reason to think that votes are not being tallied or counted accurately in ANY way for ANY reason by these machines, then they should not be used.
Period.
The video above clearly shows that the machines can go out of calibration and misrecord a vote. As such, they should be entirely removed from service until they can be remodeled so that this simply cannot happen. If something makes this impossible with touch-screen voting machines, then touch-screen voting machines should simply not be used.
It is one of the underlying, fundamental principles of democracy that each voter has the right to vote for who they choose. If ANYTHING obstructs this, then, if America is to call itself a democracy, then this obstruction should be removed immediately - even if it requires going back to having the voters physically writing an X in a box on a ballot card, and having those cards physically counted.
Posted by: Smidgy | October 30, 2008 6:17 PM
Here's in WA we have the paper + optical scan method. It's like taking the SAT or some other standardized test so the "feel" of it is familiar to most folks and the scan technology is well tested and there is a paper trail.
There is NO NEED for the touch screen machines.
Posted by: Ian | October 30, 2008 6:33 PM
Complete agreement with the support for paper ballots/optical scanning. Thanks to No Child Left Behind, we're raising an entire generation of children who are very good at filling in little bubbles with #2 pencils. We might as well take advantage of those skills when picking voting technologies, right?
For people who want to read more, the Caltech/MIT Voting Technology Project has an excellent selection of reports. They've found that paper ballots, followed by optical scanners have the lowest rates of uncounted votes [Full report].
Posted by: Malta | October 30, 2008 7:17 PM
As noted up thread here in Minnesota we use paper ballots that are scanned. In addition there is a mandatory audit of some number of randomly selected precincts whereby all the precinct ballots are hand counted and compared to the scanned count. IIRC the hand count audit precincts are selected on election day.
LM Wanderer
Posted by: LM Wanderer | October 30, 2008 9:29 PM
That's the system the voting machine in the video; Nevada uses the same machine, but with slightly different software (i.e., pressing a candidate locks it in and you have to press it again to change the vote).
Posted by: Turcano | October 30, 2008 10:05 PM
"Man, where are you voting! :D
If it has gotten to the point where people are being accosted and their votes checked, worrying about how many receipts we get is kinda irrelevant."
Not entirely. Secret balloting not only prevents intimidation of voters, it also prevents the buying and selling of votes.
Posted by: konrad_arflane | October 31, 2008 9:19 AM
Hmmm...it might not be a calibration issue. North Carolina uses similar voting machines (if not exactly the same, at least from the looks of them) and voting a straight ticket does NOT include the office of President by design. IOW, you have to make a choice for President separately from voting a straight ticket, so selecting straight ticket shouldn't alter a previous choice for President. If you look at the video, you can see that immediately prior to choosing the "straight Republican" ticket, he had selected Ralph Nader for president. When the screen resets after his straight ticket choice, it still has Nader as the choice for President. For the machines we have here, that would be exactly how they are supposed to operate.
Posted by: Bill Snedden | October 31, 2008 12:55 PM