iSixSigma

We Need Six Sigma in the Voting Process

Six Sigma – iSixSigma Forums Old Forums General We Need Six Sigma in the Voting Process

This topic contains 14 replies, has 13 voices, and was last updated by  Marc Richardson 18 years, 2 months ago.

Viewing 15 posts - 1 through 15 (of 15 total)
  • Author
    Posts
  • #26961

    Jim Parnella
    Participant

    It’s too late to fix the problems for THIS election, but how about fixing them PRIOR to the next big election?

    A six sigma voting process will result in not more than 3.4 defective votes for each one million cast. Since we can expect 100 million votes in the 2004 presidential election, that amounts to not more than 340 defective votes in the entire country. And if these defective votes are randomly spread among the 50 states it is highly unlikely that a recount in any state will change the outcome in any state.

    Six Sigma tools will be appropriate for this objective. What do you think? Who can lead such an effort?

    0
    #65551

    Tom Avery
    Member

    You were reading my mind Jim! I was just thinking how atrocious the process was — how utterly broken — and that it needed a good process map and measurement system.

    This is a tremendous opportunity for the world of consultants to put together a proposal for individual states. I don’t think it would be a tough sell.

    I guess it just goes to show that sometimes even statistics should be trusted. I once read an article about how many votes were needed to statistically represent the United States. I guess they didn’t get a representative sample!

    Cheers,
    Tom

    0
    #65560

    Vallejo
    Participant

    My family now lives in the United States and I cannot believe that this happens on my first vote. It is amazing. I believe it should be solved soon, but see no quick finish.

    I also cannot believe the amount of rework that has been amassed in this processes. How much has been spent? Is there ways to poka yoka this process? Error proof? Why not?

    0
    #65563

    Steve MacDougall
    Member

    I believe that bringing Six Sigma into the election process would be a fantastic approach to fixing the process! What better way to get the tools and methodology out in front of those not aware of the power of doing things in a Six Sigma manner too! I can only envision the great things that would come out of a project as large as this, but truly believe that it could be done. I’d love to be part of that effort, but see it coming from a level of consultants such as one person touched on in a previous reply. If there was a call for volunteers though, I’d surely be one to sign up.

    0
    #65564

    Bonghanoy
    Participant

    A great idea. This will take a substantial investment in new voting machine technology.
    In Florida, it appears that counties with the newer machines have fewer discarded ballots. The newer machines allow the voter to insert their punched ballots into a tester and they get instant feedback on whether or not their ballots are valid & will be processed. If there are errors, they immediately receive a new ballot.
    To ensure fairness across the U.S., the measurement system, ie the technology that counts the ballots in all the counties, will have to be proven capable. To ensure every voter is ‘heard’ we’ll have to reduce the variation in the current measurement system.

    0
    #65565

    Schultz
    Participant

    I’m sure the candidates will still find something to argue about! Even Six Sigm can’t be that black and white!

    0
    #65567

    Patrick Waddick
    Participant

    The press never reported why Florida requires a recount when the vote tally differential is within 0.5%. It must have been an arbitrary threshold, but it begs the question: Should vote counting be based on reliability studies (Gauge R & R) of the various methods of counting votes? I think it should. There should be a known threshold when the election is declared a virtual (statistical) tie and the decision goes to the House of Representatives.

    That would be a huge improvement over all the manual recounting, litigation, and protests.

    0
    #65569

    Jack Kennel
    Participant

    I heard that the statistical margin of error associated with the Florida recount is larger than the difference in votes! Is it possible to “certify” a result that has that large a margin of error?

    0
    #65570

    curious
    Participant

    That’s a good question. I don’t believe that you can trust numbers that are that close. Although they must have statisticians gainfully employed at the areas of tally, I believe that their confidence intervals are not what they should be.

    What must a confidence interval be in order for the margin of error to not have an effect on the outcome?

    Curious George

    0
    #65575

    Jim Parnella
    Participant

    Gage R&R won’t cut it for the voting process. All it will do is assess how much ERROR we have. We need Six Sigma, as it will result in not more than 3.4 defective votes per million cast. Now, wouldn’t that be great to have no more than 340 defective votes in the 100 million votes cast? And if you spread them equally over the 50 states, it would be highly unlikely that a recount would change the result of any election.

    The problem is … who will organize such an effort, and who will fund it?

    0
    #65591

    Patrick Waddick
    Participant

    I agree that Gage R&R won’t solve the problems inherent in a vote counting process that is flawed. A Six Sigma vote counting process should be the standard for every state or locality to follow. Perhaps additional funding is not the concern here, but merely that more emphasis needs to be placed on defining a reliable, certifiable, (Six Sigma) vote count.

    0
    #65649

    Dave Hauff
    Participant

    I agree Six Sigma tools should be used in fixing the current voting process, but I don’t agree that it needs to operate at a Six Sigma level. Although I have no data to back this up, I believe that getting to that level would take a significant investment of our tax dollars – money that would be better spent elsewhere. Do we really need to be *that sure*? Get the error to an “acceptable” level, have a process for when the difference is within the error, and hope we don’t have to worry about this problem more than one per century. Maybe we should vote on what “acceptable” means… :)

    0
    #65652

    Jerry Brong
    Participant

    We the people in order to form a more perfect union can take necessary actions to protect, and in fact strengthen, our system of electing leaders that will serve in our governments.

    The “WE THE PEOPLE” is to imply that the “corrective actions” require an open process in which there is a mutual sharing of our ideas about the processes of casting & counting. We also need corrective action on how we are informed about the results of votes & problems with OUR voting processes.

    Quality tools can, in fact will, be used to carry our voting processes far into the new century. As quality experts we can’t “demand” or “direct” corrective action but we can guide & suggest how We The People should continue to improve our system of democracy.

    Jerry Brong, EdD
    Institute for Improvement Through Schools

    0
    #66419

    James W. Angelo
    Participant

    My company, InWork Technologies, has developed interactive electronic procedures that have been designed to “engineer out” human error in critcal use applications. Converted to a ballot format, an interactive ballot is one that would provide feedback to the voter and would not allow the voter to proceed without an acknowledgement of voting action already taken. A touch-screen would be used for voter interaction, and only the candidates being considered for a given voting action would be listed along with a digital picture of the candidate. Before the voting action step can be completed, the interactive procedure would print out the results of the voter’s action, require that the voter validate the actions taken and confirm the results before electronically posting the result to the election board as an official ballot. Similarly, the ballot would prevent a voter from starting unless they were registered according to the local requirements. Several other features are available to bring voting into the six-sigma error range. Please contact me if there is any interest in our process. You might be interested in knowing that this technology was developed for use in assembly and disassembly of our nuclear weapon stockpile for the Department of Energy, where errors can not be tolerated.

    0
    #66421

    Marc Richardson
    Participant

    Margins of error only apply with sampling. In voting, one does not sample the number of votes to determine the outcome, one counts 100% of the votes. Sampling the votes and projecting the “winner” early that evening of 12 Nov was part of what led to sorry spectacle of suits and counter suits we saw over the ensuing months.

    Marc Richardson
    Sr. Q.A. Engineer

    0
Viewing 15 posts - 1 through 15 (of 15 total)

The forum ‘General’ is closed to new topics and replies.