iSixSigma

Interpret DOE results

Six Sigma – iSixSigma Forums Old Forums General Interpret DOE results

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
    Posts
  • #29788

    Tim
    Member

    Need help in interpreting a DOE results. Below is the result from doing Multiple Regression Analysis.
    What is this telling me:
    Rsq = 0.9913,   Adj. Rsq =  0.9905,  Std Error = 4.1
    F =  1361,    Sig F = 0
    What is the significant of  TOL and P(2 Tail)?
    Many thanks.

    Y-hat Model

    Factor

    Name

    Coeff

    P(2 Tail)

    Tol

    Active

    Const

     

    54.450

    0.0000

     

     

    B

    Time

    19.650

    0.0000

    1

    X

    C

    Distance

    -35.400

    0.0000

    1

    X

    BC

     

    -9.00000

    0.0000

    1

    X

     

     

     

     

     

     

    Rsq

    0.9913

     

     

     

     

    Adj Rsq

    0.9905

     

     

     

     

    Std Error

    4.1049

     

     

     

     

    F

    1361.2522

     

     

     

     

    Sig F

    0.0000

     

     

     

     

     

     

     

     

     

     

    Source

    SS

    df

    MS

     

     

    Regression

    68811.3

    3

    22937.1

     

     

    Error

    606.6

    36

    16.9

     

     

    Total

    69417.9

    39

     

     

     

    0
    #76916

    SmartGuy
    Member

    That you need to get Design & Analysis of Experiments training before attempting to design or analyze an experiment.

    0
    #76940

    Ø6 Sigma BB Coordinator
    Participant

    Wow……, you use DOE KISS, my favotite software.
    Rsq = 0.9913,   Adj. Rsq =  0.9905, 
    R sqaure and adjusted R squre show the model you get can explain your Y about 99%.
    Both can be used for this interpretation.
    Adj. Rsq is modified Rsq, being less when many terms included in the model.
    Std Error = 4.1
    From the model you can predict Y. The standard deviation of the actual result around your Y predicted is = 4.1.
    F =  1361,    Sig F = 0
     F value ( a statistical term is = 1361). You have to know about F distributtion and your degree of freedom to interprete this.
    For easier way, I use Sig F (significant of F).
    This case your Sig F = 0.000 somthing. Surely it is not 0 exactly.
    This means the proability that your overall model is not significant is near to 0 or you can have confidence nearly to 100% that your model is significant (your X factors really affect your Y)
    What is the significant of  TOL and P(2 Tail)?
    Tol, tolelance, shows degreee of dependence between your Xs.
    Tol = 1 means , there are absolutely no dependence between your Xs.
     ROT — Tol should > 0.5.
    P(2 Tail) is the probability that the terms may not be significant.
    ROT — P(2 Tail) <= 0.1 keep the terms.
    Hope this hellps.
    Six Sigma Black Belt Coordinator

    0
    #77087

    Tim
    Member

    Thanks SSBBC,
    What happen if not all the Tol =1. Would the design still consider as good?
    Why some DOE yield no S hat?

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

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