iSixSigma

Sample size

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
    Posts
  • #53401

    LIM
    Participant

    Hi need some advice from the experts in here. I would like to reduce the cycle time on one of the machines and worry that by reducing the cycle time, its might increase the defect rate on the products. Basically, the defect rate has to be the same or less, after the change. The baseline of the defect rate is 500ppm. what sample size is needed to monitor the change? can i take 10 days data after the change and use 2 proportional test to check whether there is any significance difference before and after the change? Please advice. Thanks

    0
    #189986

    Severino
    Participant

    I would use a control chart for this not a hypothesis test.

    0
    #189996

    Mike Carnell
    Participant

    There is no straight forward answer to you question. If you choose to use a hypothesis test you need to know how much of a shift you are trying to detect and your confidence level.

    Minitab will calculate it for you and the help menu explains it pretty well.

    0
    #190204

    the dude
    Member

    Why not run the project with CT as your Project Y and capture the quality metric as the secondary and track it via Control Chart (per JSVEV). If you wanted to use a 2 proportion test on a before/after yield measure, I suppose you could and easily quantify the CI/CL.

    0
Viewing 4 posts - 1 through 4 (of 4 total)

You must be logged in to reply to this topic.