1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.
Dismiss Notice
You must be a registered member in order to post messages and view/download attached files in this forum.
Click here to register.

Capability analyses with nonnormal data

Discussion in 'Capability - Process, Machine, Gage …' started by essegn, May 18, 2018.

  1. essegn

    essegn Member

    Joined:
    Feb 5, 2016
    Messages:
    47
    Likes Received:
    6
    Trophy Points:
    7
    Hi All,

    I would like to perform a capability analysis in Minitab - I ran the normality test and got the result, that my data are not normal distributed.

    I believe that the measurement data are correct and there is no reason to improve the process.
    I would like to have my data transformed, but it doesnt work.

    I have run an Individual Distribution Identification with all distributions and transformations, but any of them increased the p-value above 0,05.

    May I run the Nonnormal Capability Analysis?

    Attached are my data with a Subgroup size of 2.

    Regards
    P.
     

    Attached File(s): 1. Scan for viruses before using. 2. Report any 'bad' files by reporting this post. 3. Use at your own Risk.:

    essegn, May 18, 2018
    #1
  2. Miner

    Miner Moderator Staff Member

    Joined:
    Jul 30, 2015
    Messages:
    613
    Likes Received:
    532
    Trophy Points:
    92
    Location:
    Greater Milwaukee USA
    Your process is stable (see control chart), and the histogram is reasonably normal in appearance (see histogram). The probability plot shows the issue. The data lie reasonable straight, but the data are chunky. This chunkiness is what causes the normality test to fail. The number one cause of chunky data is the measurement system. Have you performed an R&R study? What were the results? A low ndc or high %Study Variation will result in chunky data.
     

    Attached File(s): 1. Scan for viruses before using. 2. Report any 'bad' files by reporting this post. 3. Use at your own Risk.:

    Last edited: May 18, 2018
    Miner, May 18, 2018
    #2
    essegn and Andy Nichols like this.
  3. essegn

    essegn Member

    Joined:
    Feb 5, 2016
    Messages:
    47
    Likes Received:
    6
    Trophy Points:
    7
    Thank you Miner! To be honest, the Gage R&R was performed years ago. But this could be a good argugment to perform Gage R&R Studies on a regular base. I am not sure, but it seems to be that decimals are being cut and this could be reason of these chunky data.
    Which Capability Analysis should i choose? For normal or for nonnormal data?
     
    Last edited: May 18, 2018
    essegn, May 18, 2018
    #3
  4. Miner

    Miner Moderator Staff Member

    Joined:
    Jul 30, 2015
    Messages:
    613
    Likes Received:
    532
    Trophy Points:
    92
    Location:
    Greater Milwaukee USA
    That could definitely do it.

    The data are normal despite the chunkiness, so use the normal capability analysis.
     
    Miner, May 18, 2018
    #4
    essegn likes this.