iSixSigma

DOE with an Attribute Response

Six Sigma – iSixSigma › Forums › Old Forums › General › DOE with an Attribute Response

Viewing 6 posts - 1 through 6 (of 6 total)
  • Author
    Posts
  • April 1, 2009 at 7:10 pm #52151

    newbie
    ★ 10 Years ★
    Participant

    I am studying basic factoral designs and I have not seen any examples that use attribute data as the response variable….How is this done, generally speaking?  I am wondering power and sample size for the number of replicates needed and then how one would analyze the findinngs…i am thinking it would include a form of logistical regression…Any references would be greatly appreciated.  Thanks!

    0
    April 1, 2009 at 10:05 pm #183062

    Bower Chiel
    ★ 10 Years ★
    Participant

    Hi NewbieThe book by Ellis Ott et al entitled Process Quality Control – Troubleshooting and Interpretation of Data is well worth a look. It has a chapter on Troubleshooting with Attributes Data which includes examples of factorial experiments with attribute responses. The book uses Analysis of Means (ANOM) rather than Analysis of Variance (ANOVA)and comes with a CD that has software for ANOM.Best Wishes
    Bower Chiel

    0
    April 1, 2009 at 10:38 pm #183064

    Sigmordial
    ★ 10 Years ★
    Member

    Couple of possibilities and all involve repeated measures within each run (not to be confused with replication). 

    If your response is binary, you can determine the proportion for the event of interest from the repeated measures and then use a variance stabilizing transformation such as the Freeman-Tukey transformation
    Another approach is to design your experiment with repeated measures and analyze the results using logistic regression. Applied Logistic Regression by Hosmer and Lemeshow is a great reference, and it will provide good insight into sample size considerations
     

    0
    April 2, 2009 at 2:45 am #183068

    Eric Maass
    ★ 10 Years ★
    Participant

    Newbie,
    I recently was asked a similar question, and gave a presentation on DOE with an attribute response to some Motorola engineers in Malaysia.If you’d like a copy of that presentation, please email me at [email protected] and I’ll email you the pdf file.
    Best regards,Eric Maass, PhDMaster Black Belt

    0
    April 2, 2009 at 7:25 am #183073

    riyan
    ★ 10 Years ★
    Member

    if we have predictor as variable and response of attribute, we can use logistic regression instead. Logistic regression use logit formula in calculation of hipotesis test.

    0
    April 2, 2009 at 2:02 pm #183088

    newbie
    ★ 10 Years ★
    Participant

    Thanks everyone!  I will follow up on some of the references mentioned.  I appreciate the help!

    0
  • Author
    Posts
Viewing 6 posts - 1 through 6 (of 6 total)

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