iSixSigma

IMR / XBar-R usage

Six Sigma – iSixSigma Forums Old Forums General IMR / XBar-R usage

Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
    Posts
  • #52280

    Lee
    Participant

    I have been told that most processes can be monitored with an IMR chart.  The assertion is that if increased accuracy is desired, then take 5 samples and then enter the average into the IMR chart.  That sounds like a $3 bill to me.
    My thought is that the XBar-R chart was created for a reason — because it is frequently needed.
    I work in a food processing field, and 99% of our monitoring is of meat characteristics (such as slice thicknesses, package weights, etc.) wherein we take a sample of 5 packages each hour at the end of the process and chart using an XBar-R chart.  Therefore, from my perspective the norm is to use XBar-R charts.
    Am I correct that putting the data into an IMR chart is poor practice and that XBar-R charts have a high frequency of application?  I would expect that there is a wide variation across industries, and process characteristics should dictate the type of chart used.  In responses please indicate the type of industry that corresponds to your reply.
    Ah.  No, this is not for a research paper.  I’m just trying to determine the frequency range of use of the XBar-R chart and then have a firmer basis to assert that our processes should not abandon them for the IMR type.
    Thanks for your kind replies.
     

    0
    #183761

    GB
    Participant

    I prefer I/MR for greater discrimination.   Your “5 sample” comment sounds off though.    Running control charts is so easy these days, so you might as well include as much data as is available for either type of chart.
    XBAR/R is great for averages and high-level views.
    my 2 pennies

    0
    #183766

    Darth
    Participant

    Shewhart developed the control chart to examine both within sample variation and between sample variation. He used the within sample variation for his calculations of the between sample variation. The R/s chart can be thought of as the noise while the Xbar chart can be thought of as the signal. It is the signal or change over time that you wish to monitor. The I/MR does not permit any valid measure of within sample variation. I/MR is a last resort when you don’t have an option for rational subgrouping. If the sample of 5 that you are taking is a rational subgroup then I would stick with what you have. Taking the average of the 5 and putting it on an I/MR is silly. Given today’s computing power, you can use an Xbar/s if you wish.

    0
    #183789

    hbgb2
    Participant

    Darth, I agree that I/MR for the avg of 5 would be cuckoo…My stated pref for I/MR is couched in what you stated concerning rational subgrouping.  At my shop, it is common to be approached with data up the wazoo and puzzled looks by the bearers.    Every time I’ve attempted a discussion of sampling, or Rational Subgrouping, their pupils start to dilate and they get that 1000 yard stare.    That’s when I take their data, smile and hunker down for the deep dive.   What can I say? I like Individuals charts.
     

    0
    #183793

    Darth
    Participant

    I am not judging you. You also like Mikel and Stan but I am still your friend and respect you. Hope things are good.

    0
    #183795

    GB
    Participant

    Things are OK, given the turbulent times we are in…
    Hope you are enjoying your grandbaby.
    Stan iis good people…Mikel, or Michael?   There are big differences!
    The Michael’s are good peeps too!

    0
    #183796

    Lee
    Participant

    Thanks for the replies. 

    0
Viewing 7 posts - 1 through 7 (of 7 total)

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