iSixSigma

Control Chart – Why 3 StdDev?

Six Sigma – iSixSigma Forums Old Forums General Control Chart – Why 3 StdDev?

This topic contains 12 replies, has 9 voices, and was last updated by  Obiwan 10 years, 7 months ago.

Viewing 13 posts - 1 through 13 (of 13 total)
  • Author
    Posts
  • #51943

    QualityArch
    Participant

    Hello, I am trying to shift our focus away from creative story development and fussing around every time the color on our balanced scorecard changes color. The colors are based on totally seat-of-the-pants ‘what we think would be a reasonable range’ numbers. I have taken our office utilization rate (percentage) for the last 3 years and created a control chart. All points fall well within 3 StdDev upper and lower control limits.
    Can someone help me with a ‘laymans’ explanation of why 3 StdDev is used for the control limits? (or provide a link or reference) I have searched around but everything I find requires a statistics background – which I don’t have. Thanks

    0
    #181872

    lin
    Participant

    You are wrong. 
    3 sigma is only used by that crazy Shewhart – Deming – Wheeler crowd.
    In the highly acclaimed and incredibly brilliant Six Sigma, we only use control charts with 6 sigma limits.  The psychologist Mikel Harry has proven it to be true.

    0
    #181873

    Mikel
    Member

    No one uses 6 sigma limits.To the original question, this is a call made by Shewhart back in the
    late 20’s or early 30’s based on what he felt to be rational. Everyone has done the same since because it seems to serve them
    well. Do a google on Walter Shewhart or read his book – Economic
    Control of Quality of Manufactured Product – ISBN-10: 0873890760.

    0
    #181879

    Darth
    Participant

    Echoing Stan’s comments, the determination of 3 s.d was empirical with Shewhart trying to balance over reacting and under reacting to special cause. The book is entitled “The Economic Control…..” indicating that it was an empirical and economic issue rather than probability.

    0
    #181883

    J.P.
    Participant

    >>No one uses 6 sigma limits.
    I agree … 6 sigma is dead.
    Long live Lean.

    0
    #181884

    Dr.J
    Participant

    It is 3 Stdev to give way to the previous study that when you collect data, 99.73% will fall within 3 Stdev from the mean. The process won’t anymore become exciting for change if we increase the stdev to 4 or more which will capture almost or all data within the limits.

    0
    #181902

    Obiwan
    Participant

    Dr. J
    Not true…it is not because 3 SD captures 99.73% of the data.  It is very much what Stan and Darth said…it is an economic balance point between reacting and overreacting to out of control conditions.
    Obiwan

    0
    #181904

    Mikel
    Member

    Dr. J,You should stick to shooting ice cubes into Dr. Pepper.You really don’t appear to know much.

    0
    #181910

    Jenn
    Participant

    I trained a guy a few years ago who told me he changed his control limits to 6 sigmas.  He said he didn’t want any false positives.  I wonder if his process went out of control before his business closed…..

    0
    #181911

    Mikel
    Member

    What is a false positive on a control chart?

    0
    #181914

    QualityArch
    Participant

    Thanks Dr.J (an others).
    Please clarify for me, when you refer to 3 Stdev – does that mean 1 1/2 above the median and 1 1/2 below (3 total) or does that mean 3 above and 3 below (6 total)?
    Sorry if that is real obvious to everyone else, I appreciate the help.

    0
    #181916

    Taylor
    Participant

    Its + 3 Sigma or STD Dev from center. Your Upper and Lower Control limits are calculated using + 3 Sigma and -3 Sigma where in your Cp=(UCL-LCL)/(6*Sigma st) st=short term pooled standard deviation.

    0
    #181930

    Obiwan
    Participant

    And it is not above and below the median (typically)…it is above and below the mean.
    Obiwan

    0
Viewing 13 posts - 1 through 13 (of 13 total)

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