OK, a 28 term model from which you have dropped 4
terms leaving 8 df for error. The fact remains
that the response appears to be in the space
spanned by the 24 terms, and this has nothing to
do with the design per se. If you have no
replicates or near replicates, and if the data was
collected in a reasonable fashion, then it seems
that you have a very good prediction indeed: so
good that it is worth questioning, which is what
you are doing -- but the explanation for the
phenomenon lies in the data not the design.
[EMAIL PROTECTED] wrote:
>
> My response variable is continuous and ranges betwen 0.77 to 1.2.
> I have 31 observations, 6 factors each at 3 levels. Its not a full
> factorial, since that would require 729 runs. My full model contains
> all the main effects (which would be 6 terms), the pair wise second
> order interaction terms(which would be 15 terms) and the square terms ,
> i.e. x1*x1 x2*X2 X3*X3 X4*X4 X5*X5 and X6*X6.
> When I did the stepwise regression I was able to get rid of X1*X2 ,
> X1*X3, X2*X5 , X4*X4 and X5*X5(I apologize the error in my post, which
> says I was able to get rid of only 3 variables) So, I basically have 8
> df for my error term.
> Please let me know if I can further clarify.
> I appreciate your response.
> Thanks!
>
> In article <[EMAIL PROTECTED]>,
> Bob Wheeler <[EMAIL PROTECTED]> wrote:
> > You haven't said anything about the response. It
> > is obviously in the space spanned by the terms of
> > the model. It could be something simple as the
> > presence of exact replicates. (If it is truly
> > D-optimal and if you truly used all 27 terms, then
> > it must be a full factorial. Is it?)
> >
> > [EMAIL PROTECTED] wrote:
> > >
> > > Hi,
> > >
> > > I constructed a D-optimal design for 6 continuous variables, each at
> > > three levels. I have 31 runs. My initial model includes, all the
> main
> > > effects, all interactions and polynomial terms. I was only able to
> > > remove 3 of the terms using step wise regression, My final model
> has an
> > > R square of 0.9954, which looks very artificial. My adjusted R
> square
> > > is also very close to R Square. Does anyone have any suggestion on
> what
> > > could have gone wrong or if there is a different analysis technique
> > > that I can use?
> > > I even did a lack of fit test, and the interaction terms and square
> > > terms were significant. I'll appreciate any help in this regard.
> Thanks
> > > for your help!
> > >
> > > Sent via Deja.com http://www.deja.com/
> > > Before you buy.
> >
> > --
> > Bob Wheeler --- (Reply to: [EMAIL PROTECTED])
> > ECHIP, Inc.
> >
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
--
Bob Wheeler --- (Reply to: [EMAIL PROTECTED])
ECHIP, Inc.
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
