iSixSigma

2 SD Versus 2 Sigma

Six Sigma – iSixSigma Forums Old Forums General 2 SD Versus 2 Sigma

Viewing 5 posts - 1 through 5 (of 5 total)
  • Author
    Posts
  • #53188

    Bill Bond
    Participant

    My apologies if this is a stupid question I’m familiar with how SD is calculated & what ± X sigma means. But I don’t understand why ±2 SD as a measurement uncertainty relates to 95%. Whereas ±2 sigma relates to a process yield of around 69%. I thought ±2 SD & ± 2 sigma were the same thing ?
    I’m obviously missing the point & would be grateful for clarification from someone a bit more expert than me.Thanks – William Bond

    0
    #188814

    w. g. miller
    Member

    The process sigma calculation sometimes contains a “process sigma shift” term, which is used to account for normal and expected variations in process output.  Here is a link that will explain it and let you see how it works:
    https://www.isixsigma.com/index.php?option=com_k2&view=item&id=1450:process-sigma-calculator-assumptions&Itemid=198
    Running the process sigma calculator in the advanced mode lets you have hours of fun by varying the sigma shift term.
    Be aware that the selection of the proper value for the sigma shift term is arbitrary and has been the subject of much discussion on this forum.
    W. G. Miller

    0
    #188826

    MBBinWI
    Participant

    Bill: You’ve been fooled by the dirty little “secret” of six sigma.  If you also look at the probability for being outside of 6 std dev’s (6 sigma) you won’t find 3.4 ppm, but rather less than 1 in a billion.  The difference is that back in the “old days” when it was difficult to take data and virtually all data that was taken was considered short-term, folks wanted to protect against longer term shifts/drifts, so they adjusted the data by 1.5 sigma.  This has been perpetrated through numerous summaries of defect rates and their sigma equivalents.  If you look in a Z-table or use Normsdist function in Excel, you will get the actual value.  If you subtract 1.5 from the number of std dev’s (sigma) and look that up in the table (or use Normsdist in Excel) you will get the often advertised values.

    0
    #188844

    Bill Bond
    Participant

    Thanks for such a prompt reply MB.  I think I understand what you are saying, as I’ve read about this 1.5 sigma shift before. I’ll have a look into this again though.
    Cheers, Bill

    0
    #188845

    Bill Bond
    Participant

    Hi WG,
    Thanks for that, I’ll take a look & have a play around.
    Cheers, Bill

    0
Viewing 5 posts - 1 through 5 (of 5 total)

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