six-sigma, a special definition?

Six Sigma – iSixSigma Forums Old Forums General six-sigma, a special definition?

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
  • #27708


    I reached an uneducated conclusion that “six-sigma” is a special use of statistical terms which amounts to one of three things:

    it indicates some inverse relationship of variance between sample-mean and desired mean as well as sample standard deviation.
    (my choice) it is merely a counting process of outcomes and the proportion of good outcomes to total plotted against the probability or area represented by the central region a standardized normal curve.
    some other thing unknown to me.
    My assumption is of course that n+1 Sigma is “better” than n Sigma.
     Is “six-sigma” then a justified statistical term, or is it a redefined term borrowing liberally from statistics?
    Does anybody have a formula to show how “process sigma” is determined? 



    You are correct in stating that the quantitative value of six sigma is determined through tried and true statistical calculations.  To determine a “sigma level” for a process is no more difficult than calculating the z scores for that process.  If you get a Z score of 6, that is, quantitatively at least, defined to be a six sigma process.  If you get a 3, that is a 3 sigma process and so on.  The difficulty arises when you start discussing that pesky 1.5 standard deviation shift in the mean that is unique to SIX SIGMA the methodology.  I think I’ve confused enough people on my response to a previous post on that one, so I will refrain here. 
    Teresa Tobias ([email protected])



    does “6-sigma” use bernoulli trials?
    The 1.5 sigma drift sounds a lot like black-magic to me at this point, and I keep getting this whif of dogma.
    I hope it only smells like bunkum!



    Matthew –
     I’m afraid your hope is in vain.  While good and sufficient reasons have been cited for why Motorola chose1.5 sigma to estimate the shift which had occurred over time BEFORE they had adopted 6 Sigma as a goal and good-to-excellent statictical control methods throughout their organization, there is (in my not-very-humble, experience-based opinion) no reason to believe that this estimate is valid for other firms with moderately to extremely different collections of products, processes and process control strategies.  There’s even some externally-available evidence that it’s no longer valid for Motorola!  Elsewhere in this forum, the “1.5 Sigma Shift” has been referred to as a weed in the 6 Sigma garden, and suggested that the time to pull this weed has come.  I heartilly concur.

Viewing 4 posts - 1 through 4 (of 4 total)

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