Skip to main content

View Diary: "Election Reform": Bringing Back The Big Ideas (139 comments)

Comment Preferences

  •  about paper receipts (none)
    Paper receipts can be made to work pretty well.  The term "voter verified paper trail" is the key.  Here's how it works:

    1.  You cast your vote.
    2.  A paper receipt shows up covered under glass for you to double check it's accuracy.
    3.  If you indicate that it is not correct the receipt is shredded and you may alter or redo your vote.
    4.  If you indicate that it is correct, the receipt is dropped into a lockbox and the vote is tallied electronically.
    5.  Spot audits are done on random (real random, not "Ohio random") precincts throughout the nation to verify, through manual recount, that the receipts match the electronic record.  Discrepencies trigger further investigation and recount.

    The fifth step is the key to the whole thing working.  No matter what you do there will always be small scale, localized, fraud, but the spot audit will prevent anything on a large scale and make even small scale stuff very risky to attempt.  Even if the whole thing was done manually without any computers, as you suggest, the spot audits are still essential to a secure voting system.
    •  Yes, even with (none)
      a manual system, audits should be used, and I did not comment that no auditing would be necessary, nor did I comment that all fraud would be detected or prevented. [I'm an internal auditor].

      However, your scheme regarding the printing of receipts still does not mitigate the weakness that the program can print something on a piece of paper while it tabulates/sums/transmits the data differently.  And for some tabulation procedures described by some diarists--where the file is transmitted to a central location for adding--you have the added weakness of unverified processing or software at the adding/tabulation level [or a different level where the count may be tampered with].

      If you use the printed receipts to verify the final total, it would still be possible for the program to move small amounts from one candidate's count to the other candidate.  Over many precincts, the small amounts shifted could add up enough to change the outcome, but the discrepancy between the receipts and the count could still be small enough at the audited level [I'm assuming precinct level here] such that the auditors could conclude no problem occurred.  [It is my understanding that audits are done on a very small number of precincts, and I don't recall what error rates are in use now--I'm thinking I saw around 5% somewhere.]

      Maybe the problem could be mitigated by accepting only a very small error rate; larger discrepancies would kick in additional audits.  

      Then again, maybe I'm just full of it.

       

      •  print versus electonic (none)
        If the machine is printing and tallying electronically all automatically, the numbers should match 100% in the audit.  Anything else indicates a problem.
        •  I agree with you (none)
          but currently, I think for recounts, there were error rates that are allowed [where the recount could differ from the original count by a certain % w/o kicking in more recounts].  I think the error rates are set by the state laws.

Subscribe or Donate to support Daily Kos.

  • Recommended (141)
  • Community (67)
  • 2016 (55)
  • Environment (44)
  • Elections (42)
  • Republicans (38)
  • Media (37)
  • Hillary Clinton (35)
  • Iraq (32)
  • Law (30)
  • Jeb Bush (30)
  • Barack Obama (30)
  • Trans-Pacific Partnership (29)
  • Climate Change (29)
  • Culture (29)
  • Civil Rights (27)
  • Labor (24)
  • Economy (23)
  • White House (18)
  • Science (18)
  • Click here for the mobile view of the site