iSixSigma

Gage R and R on historical data – how to?

Six Sigma – iSixSigma Forums Old Forums General Gage R and R on historical data – how to?

Viewing 5 posts - 1 through 5 (of 5 total)
  • Author
    Posts
  • #53193

    The Kid
    Member

    I have with me last 1 years data recorded in various reporting systems. It is clear that I will not get a opportunity to collect live data and will have to do all my analysis on this historical data.
    How do I perform a Gage R and R in this situation which will give me confidence that data is of good quality?
    Any inputs/references to textbooks would be highly appreciated.
    Cheers,
    The kid

    0
    #188849

    Remi
    Participant

    Hai The Kid,
    You can’t. A gage r&R study is an experiment you do to investigates the quality of your data-collection/measurement. Unless you can perform that experiment you cannot do a gage r&R.
    This doesn’t mean that your old data is worthless. only that you can’t proof that it can be trusted.
    So the only way forward is: Write down in your report that a gage r&R is not done and that all conclusions from any analysis on these data has to be further verified by an experiment (for which a gage r&R has proven that the data from the experiment can be trusted).
    Remi

    0
    #188858

    BMWSAND
    Participant

    You can do the gage R&R by using the following formula:
    %GRR=100(5.15* Std dev)/tolerance
    The number you will get will tell you if the gage is good or not but it will not tell you if the variation id due to the process or operator or the equipment.
    If your process is under control and you can prove it with data then the GRR number you get using the formula gives you total gage variation. Going forward you can run a true GRR and see what kind of variation you get from operator or from equipment.
    BMWSAND

    0
    #188912

    Remi
    Participant

    hi BMWSAND,
    your formula assumes that the data only has variation caused by the measuring process.
    Remi

    0
    #188913

    BMWSAND
    Participant

    yes…while not perfect (you still have to do the real GRR) it will give you an idea if you should even try running a GRR or focus your efforts on controlling the process (and I assume you have some empirical data that you can use and see if it operates in control).
    Remember, a statistical number/solution has to make practical sense also.

    0
Viewing 5 posts - 1 through 5 (of 5 total)

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