iSixSigma

Special Causes

Six Sigma – iSixSigma Forums Old Forums General Special Causes

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
    Posts
  • #38688

    Aman
    Participant

    Hi Guys,
    I have read quite a bit information on SPC but still am not very clear as the approach to handle special causes once detected… Should the control charts be recreated after removing the subgroup containing special causes ( and investigating these special causes later) or the control charts should only be created after special causes are removed and measurements then made for SPC..Any thought welcome!!

    0
    #116183

    JM
    Participant

    Remove the data point and then re do the Control Chart

    0
    #116184

    Anonymous
    Guest

    Aman,
    A Shewhart Chart is really a test of homogeneity. When a few points fall outside a set of control limits you either have a special cause of variation, or the control limits are too tight.
    One cause of tight tolerance limits is autocorrelation – another is irrational subgrouping. You should check all of these before assuming that the process is ‘out-of-control.’
    Regards,
    Andy

    0
    #116186

    Aman
    Participant

    removal means action to be taken or ignoring special cause

    0
    #116617

    Jonathon Andell
    Participant

    Well stated. You need to step away from statistical analysis, and understand the underlying process instead. The sampling scheme may or may not contribute to your problem.Minitab has a utility that allows you to leave the point in question on the display, but to eliminate it from the computation of control limits. You can use that to see what the limits might become. Whether that will enhance your understanding of the process is unclear.The key concept here is “Assignable Cause” (another term for Special Cause). If you cannot trace the point in question to an event that made it happen, then eliminating the point from the display does nothing to ensure that it cannot happen again.

    0
    #116643

    melvin
    Participant

    Aman
    Yes – the control chart is designed to be used to identify special causes that should be acted upon.  After they have been addressed (removed) you can update the control chart – revise the control limits.
    Prior to development of the control chart, you may want to analyse the data to insure you have a stable process.  Failing that – you may want to starrt eliminating the causes of instability right away.
    Bob 
     

    0
Viewing 6 posts - 1 through 6 (of 6 total)

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