iSixSigma

Statistical Significant X – But How Much?

Six Sigma – iSixSigma › Forums › Old Forums › General › Statistical Significant X – But How Much?

Viewing 2 posts - 1 through 2 (of 2 total)
  • Author
    Posts
  • October 8, 2008 at 3:29 pm #51091

    RK
    ★ 10 Years ★
    Member

    Hi,
    My metric is reducing inflow of duplicate payments on a weekly basis. If the result of my Anova/Hov ( i have a normal data) tell me that there are 5 statistically significant Xs ( p value < 0.05):
    1. Processor, 2. Business Division 3. Day of the week 4. Type of Invoice 5. Customer

    How do I find out what is the degree of significance of these Significant Xs? Because there are some which would be having more influence on my Y; some quite less. And since we need to work on the key Xs only to make significant impovements, this identification is very important.
    Could anyone help me out with this?
    thanks..RK

    0
    October 9, 2008 at 2:46 pm #176584

    Remi
    ★ 10 Years ★
    Participant

    Grumble; first reply crashed so try #2 with the lightbulb tip
    Hai RK
    Look at the R-sq. Analyse data with 1 X less. The drop in R-sq is how much worse your model is compared to the one with 5X. So it is an indication of the importance of that 1X.
    Warning: if strong interaction X1X2 than Rsq can drop 60% when X1 outside analysis and 66% when X2 outside analysis and 68% when both outside analysis.
    Assumption: the data you analyse is the data you model. So it has to be representative of reality to be usefull for improvements.
    Hope this helps

    0
  • Author
    Posts
Viewing 2 posts - 1 through 2 (of 2 total)

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