The Digit or the Ballot


New at Reason: Julian Sanchez votes on the Diebold controversy.

NEXT: When Private Isn't So Private

Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Report abuses.

  1. I would add one thing to the touch screen machines, have them print out your ballot when you’re done and slip it into a regular ballot box. That way people can verify their own votes and there is a paper trial so that a recount can be made (data corruption is always a possibility). Heck, OCR the paper ballots for the recount.

  2. I agree on the need for a paper backup. One could even have the paper ballots placed into a machine that reads them (similar to how the fill-in-the-oval-like-on-the-SAT ballots are counted). Keep two counts: One on the touchscreen machine that the voter used to input his votes and generate the paper ballot, and one with the machine that reads the generated paper ballots. Have the machines made by different companies, or at least have different election officials in charge of each set of machines. If the two counts agree then we know we’re in good shape.

  3. thoreau,
    and if the count is off by a few, which number do we use? Is the second count a recount?

  4. Here is an article on a recent electronic voting screw up – . It’s not just theoretical that there can be problems.

  5. amr-

    We can expect the second count to be slightly lower in each category (i.e. fewer votes per candidate), if people for one reason or another fail to submit their paper ballots. However, since the second count (the one of paper ballots) is the most easily verified and most difficult to tamper with (although I hasten to add that nothing is 100% immune to tampering) that should ordinarily be used. And with two counts, any significant discrepancy that can’t be explained by a handful of forgetful people not submitting their paper ballots to the second machine would trigger a fraud investigation.

    Come to think of it, though, the fill-in-the-oval-with-a-pen ballots are so easy to use that I don’t see the need for touch-screen machines in the first place. One could make an argument that regular handling might disturb a punch-card ballot, but the ones you write on are much harder to disturb.

    So I’m not such a big fan of touch-screen machines. I think the fill-in-the-oval-with-a-pen ballots are good enough.

  6. thoreau,
    I’ve always been a fan of the ScanTron ballots. A variant I’ve seen is the complete the arrow with the black marker. It takes a fraction of a second to run and can be recounted easily. It’s fast, easy and leaves a paper trail. Plus after years of standardized testing, everyone can do a ScanTron. Why use punch cards instead?

  7. Mo-

    I agree. Scantron is the way to go.

  8. Of course, once you have paper ballots as the “official” ballot for legal purposes, there really is no point at all in having touch-screen voting, is there? The paper ballots can do everything the touch-screen does, and more!

  9. I like the idea of a paper trail as a check against digital fraud. But it seems like it would be redundant and costly. Why have digital voting at all?

    I’ve never used a scantron type voting machine; it’s always been punchcards here. But the punchcards seem as idiot-resistent as anything… though I did once puncture my finger playing around with the little punch-needle device:) As for chad issues, all you have to do is look at the damned card after you’ve punched it.

  10. >A variant I’ve seen is the complete the arrow with the black marker.

    That’s the type used in San Francisco. Leaves a paper trail, but the count is fast and electronic. IMO one of the few things the SF elections department does right.

    An added bonus vs. touchscreens [which I used in Virginia till the mid-90s] is no waiting in lines for the machines. Any number of people can mark their ballots on their own time simultaneously. Then the scan into the machine takes only seconds.

    Back in Virginia, the lines at peak periods for the 2 touch screen machines in our precinct would be maddening.

  11. Since I write software, I think I can make an edumacated guess as to a) what’s going on, and b) what’s the best way to do things.

    What’s going on is (I have no knowledge of this, just speculumating) Diebold decided to create the Mother of All EVoting Systems. I’ve seen things like this before, where things are way overde$igned. That’s just speculation, of course. I have no knowledge other than the fact that all their data requires 6 or so CDs.

    The best way to do things is a) keep it as small and simple as possible given the constraints, and b) keep the source code publicly available.

    Other than the code required to run the various hardware, which may or may not be complicated and custom, I don’t see all that much software required here. All it needs to do is read the cards, add up the totals, and then transmit or display those totals.

    As I and many others have said before, you need a paper trail. For instance, when you vote you’re given a print out with your votes. You’re required to compare that printout with the way you voted, and then you’re required to put the printout in a box.

    After the electronic voting is over, the results are announced subject to review. That review consists of a sampling of the printouts. If a significant enough discrepancy is found, then all the printouts are counted.

    In close races, the printouts would have to be counted just like paper ballots are now. However, there would be no chad problems. In not so close races, all of the printouts wouldn’t have to be counted, saving time and money.

  12. “the ‘s part.”

    Typo. Paragraph 9, line 5.

  13. “the ‘s part”

    hmm, whitespace? Think, n-dimensional. Arbitrary place-holder maybe?

    But they just have to use the touchscreen to condition us to accept [shudder] on-line voting. Any public official advocating that particular monster should be taken out and shot as an idiot or a criminal.

  14. There will always be ways to screw up elections. Making the source of a voting system public does not ensure that the public code is what actually gets used on the machine. Leaving a paper trail doesn’t really help, because paper ballots are subject to tampering just as they always have been. There is no perfect way to ensure that an election is accurate, and there never will be.

  15. You all miss another reason the ‘ScanTron’ ballots are superior to any other method of voting I’ve seen (or I missed someone saying it). A co-worker of mine, who works at the polls at election day, clued me into the fact that if a ballot is spoiled, i.e. arrows going the wrong way, incomplete lines, otherwise messed up, the machine will reject the ballot on the spot, in essence spitting it back out at the person who put it into the machine. And since the poll workers here are extremely scrupulous about making YOU put your own ballot into the machine, there is no question about spoiled ballots. If it is rejected by the machine, people are given the opportunity to correct their ballot, or given a new one. So to recap just some of the benefits of the optical readers:

    1) Original paper ballot, not printed by the computer, can be easily reread and counted if a machine error happens
    2) Instant confirmation of good ballots
    3) Only 1 machine needed per voting location / precinct.
    4) Extraordinarily high capacity / throughput since the time taken by the machine is very low. More felt tip pens can increase the number of concurrent voters to whatever capacity is necessary.

    I just can’t understand why people have been suckered into preferring this touch-screen system over the optical readers. I can understand the ‘gee-whiz’ factor, but do we really want to be the country that’s distracted by the bright and shiny object?

  16. Its not just a “bright and shiny object.” Its something that can be programmed to speak in multiple languages, also accommodating the blind.

    OptiScan is the most reliable as it stands now, but if DRE were open source, we wouldn’t need to worry

  17. diebold’s actions should make it clear to everyone (even average people) that they cannot be trusted.

    just like laws are open, so must the software be. if they want to retain copyright, that’s fine. just don’t put any restrictions on people for reading it.

    of course, i think the whole thing is a bad idea. my brother told me, where he voted (in fairfax county, va) the touch screens had no privacy. that, of course, means voting is no longer anonymous. that opens voters to coercion(sp).

    he also said it took some people 10min just to enter their vote. it usually takes half that.

  18. Software for a vote counting machine should be simple addition. No algorythms, no other data manipulation. Nothing. Any extra features are simply invitations for corruption

  19. EMAIL:
    DATE: 12/10/2003 06:51:55
    In his errors a man is true to type. Observe the errors and you will know the man.

  20. EMAIL:

    DATE: 12/10/2003 02:20:59
    Hello, this is a nice site you have

  21. EMAIL:
    DATE: 12/20/2003 06:33:55
    Have no friends not equal to yourself.

  22. EMAIL:
    DATE: 01/09/2004 06:34:42
    During the Samuel Johnson days they had big men enjoying small talk; today we have small men enjoying big talk.

Please to post comments

Comments are closed.