iSixSigma

Use of R&R Study Results in Analyze Phase

Six Sigma – iSixSigma Forums General Forums Methodology Use of R&R Study Results in Analyze Phase

Viewing 11 posts - 1 through 11 (of 11 total)
  • Author
    Posts
  • #54194

    Marc van der Peet
    Participant

    Dear all,

    I’m doing a DMAIC study at the moment and would appreciate some feedback. I concluded an R&R study in my research and this shows that both repeat ability and reproduce ability are far from desirable. I assume that this, together some other small factors accounts for the large amount of process variance in the process.

    So I go on in the analyze phase now with:

    – The factors causing the process variance
    – The factors which are causing the R&R variance (fe: applying same methods etc…)

    Right approach?

    0
    #194153

    Nick
    Participant

    Please quantify “far from desirable”. A marginally adequate measurement system might allow a little different approach. However, without seeing the results of your R&R I would suggest that you improve the measurement system before you spend any time trying to diagnose/improve things with the process. If you cannot trust your measurement system then you cannot really be sure that there is any “problem” with the process nor will you be able to tell if your proposed improvements actually changed the process. Efforts to fix the process with a bad measurement system may cause more harm than good.

    0
    #194156

    Chris Seider
    Participant

    @marcvanderpeet12 Fix your gauge R&R first, as Nick M. offered. If it’s so bad, you will have a tough time finding root cause to the process variability because your regressions, DOE’s, etc. won’t find enough through the noise.

    0
    #194158

    I can agree with the comments above. Seems like your measurement system is not reliable. I suggest to do first a gage RnR on repeatability and see if your appraisors has discrepancy. If they do, fix your measurement system. Maybe your way of measuring isnot the correct one? If you are sure of your measurement system, do an other gage RnR on repeatability. If your p/t is still above 30 you need to take a look at your process, change it and improve. Then do the gage RnR again. By the way, are you working with variable or attribute data?

    0
    #194163

    Marc van der Peet
    Participant

    Thanks a lot for the feedback guys! Highly appreciated.

    To give some more background data about the project. I working on a LSS project for reducing the cycle time of the underwriting process within my firm. I used an R&R study to find out that both repeatability (0,47) as reproducability (0,4) numbers are quite low. This is because the different underwriter measure in a different way whether certain cased need extra judgement or not.

    So if I understand you guys right we first have to make sure all underwriters use the same methods to assess whether cased need further judgement before I can go to the analyze phase where we going to look at the different x’s influencing the y (cycle time).

    Dear Regards,

    Marc

    0
    #194165

    Marc van der Peet
    Participant

    By the way: does anybody have an example of an R&R study performed in the service industry. Tried to google some but couldn’t find any.

    0
    #194169

    Nick
    Participant

    Marc – you are correct in your understanding. Fix the measurement system before you do further analysis. One way to improve the measurement system might be to ensure that all underwriters are using the same method to assess the case. There are potentially other ways to improve the measurement system but I think would be a good place to start.

    0
    #194201

    Hari VS
    Participant

    Marc, as you told it is difficult to find R&R study performed in the service industry. What I suggest is that (Don’t know whether you are following the same method) you can create a set of Yes/No/NA type of questions (with the help of SMEs’)relevant to your Underwriting topic and ask the underwriters to answer accordingly.

    0
    #194209

    The set of questions Hari is mentioning could work. Perform the questionaire among the selected people wich could be the reproducebility. A week or two later perform the same test with the same people and you have yourepeatability. Just dont know if doing a R&R in the service sector is worth doing. You will always have x’s wich are unstable like: stress factor, the mood where a personisin,social situation,etc..

    0
    #194210

    MBBinWI
    Participant

    So, Dirk, it’s acceptable to you to have different outcomes for the same situation?

    It is because of the possibility of different assessments that training, standardization, and conversion of everything that can be standardized is necessary.

    0
    #194221

    Chris Seider
    Participant

    I was just thinking how else to help you.

    You can use the DMAIC methodology to fix your gage, truly. Your metric is the gage s.d. I’ve had many teams solve gage R&R’s.

    Let us know or send a file to look at your gage study.

    0
Viewing 11 posts - 1 through 11 (of 11 total)

You must be logged in to reply to this topic.