iSixSigma

Sample Size for Second Level Pareto

Six Sigma – iSixSigma Forums General Forums Tools & Templates Sample Size for Second Level Pareto

This topic contains 5 replies, has 4 voices, and was last updated by  manuela aumick 9 months, 2 weeks ago.

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
    Posts
  • #55943

    MKB
    Participant

    On a production line working a around 6000 pcs per hour, all three shift production, the average rejection is around 5% ( data from 30 days production) . The automatic detection system gives the first level pareto for the cause of these rejects. Cause 1 is around 40% and cause 2 is around 25%. For the next level analysis each part is to be opened to know the reasons for the reject under a broader category. How to calculate the sample size and period of analysis for the next level reject analysis? That is if we want to be sure that Cause 1 sub categorized into x1,x2,x3,x4, to make a right pareto how many samples should be opened and what frequency. ( with respect to margin of errors and statistical significance)

    #202289

    Ahiru-san
    Participant

    For a control chart, we like to have 100 samples before calculating control limits. But you don’t always need to go that far. There is a concept I’ve heard Jay Arthur call The Dirty Thirty. If you can go back to your original 30 days worth of data, that would be great. If you can get 30 shifts worth of data (10 days, 3 shifts) that would be pretty good too.

    #202295

    Chris Seider
    Participant

    Ummmm, why not address the top defect. Don’t complicate it.

    My two cents.

    #202331

    MKB
    Participant

    Hi Chris @cseider, thanks for a quick response. Trying to address the top defect only but there are 6~7 possible causes for this top defect and would like to prioritize on which cause/s to attack first. Practically it would require deeper investigation on each failed item ( time and resources on one hand) but s required to be sure on the most significant ones and also to know if there are differences between shifts ( multi-vari). So the question is the sample size for this next level pareto.

    #202332

    Chris Seider
    Participant

    My approach is simple…work on the top 2nd level pareto cause and confirm if defect rate drops after you’ve solved the root cause. I’d less worry about whether you have enough data and use what you have. I’m excited you know to get to a 2nd level pareto to find where to tackle initially.

    #202335

    manuela aumick
    Participant

    Get secondl level of pareto

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

You must be logged in to reply to this topic.