iSixSigma

Round robin versus ANOVA

Six Sigma – iSixSigma Forums Old Forums General Round robin versus ANOVA

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
    Posts
  • #39845

    Jeremy
    Participant

    Hello,
    I am looking at comparing the results within operators and between two locations. I have heard about round robin and ANOVA. Which test is the best to answer the following questions:
    – Is there any difference between the 2 locations? What is the difference between the 2 locations?
    – Is there any difference between the operators at each plant? What is the difference between the operators?  
    I would appreciate any exemple or website reference.
    Many thanks.
    Jeremy

    0
    #122271

    AP
    Participant

    You may use two way ANOVA and also GLM.You could get combined effects between operator and location also.Take care of ANOVA assumptions while analyzing.

    0
    #122273

    A.S.
    Participant

    My questions are what is round rabin ? and what is GLM?.
    For Operators comparison simple one way anova is adequate if you compare within a plant and other factors are constant.
    In one of my project Anova was used to compare the operators.Interesting thing is Anova done after two months has given totally different results when compared to previous one.During the investigation it was found that there was a drastic change in their family situation and attitude which has resulted in the output.
    Please be carefull on this type of noise factors while doing Anova.
     

    0
    #122299

    Robert Butler
    Participant

     Usually, ANOVA is the tool used to analyze round robin data.  I’ve been involved in a number of such tests most of which were focused on agreement between plant testing labs.  The usual procedure was to create a standard sample of some material, split this into the requisite number of sub-samples and then ship them along with other material to be tested to the resepective labs.  The material was double blinded – neither the sender nor the tester knew the identity of the test material samples and the results were recorded by operator and location.  Since operators were nested within location the analysis consisted of a nested ANOVA.  The model you could entertain was that of the effects of operators and locations.  interactions weren’t possible since none of the operators were crossed with locations. 

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

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