[SciPy-User] stats repeated measures ANOVA

[Christopher Brown]

On Monday, April 11, 2011 10:31:18 PM Peter Tittmann wrote:
> Thanks for your reply Josef, sorry for the obscurity, I've third to clarify
> below....
> 
> On Apr 11, 2011 8:44 PM, <josef.pktd at gmail.com> wrote:
> > On Mon, Apr 11, 2011 at 10:52 PM, Peter <ptittmann at gmail.com> wrote:
> > > All,
> > > 
> > > I would like to determine the effect of  two independent variables
> > > (tree size class, and canopy percent cover) which are factor vectors
> > > (non-continuous) on the variance between two methods of measuring tree
> > > height.
> > > 
> > > I believe this is referred to as two-way repeated measures ANOVA.
> > 
> > As far as I understand ANOVA only tests for differences in means, or
> > mean effects, not differences in variances, equality of variances is
> > usually a standard assumption for ANOVA.
> 
> There's a tutorial on doing it in R here
> http://www.r-bloggers.com/r-tutorial-series-two-way-repeated-measures-anova
> / but it has a lot of array operations involved which I would preferr to do
> in python.
> 
> > But it's not clear to me what you want to test or estimate, with size,
> > cover and method you have actually 3 factors.
> > Are the two measuring methods for height used for every tree?
> > Do you want the effect of size and cover on the measurement variance
> > for each method separately?
> > ....?
> 
> One method for measurement is field based and the other uses aerial data.
> The objective is to determine if the two factors (canopy density and tree
> diameter class) explain variance between of the measurement methods (ie
> accuracy of the aerial method)

From what you describe, this does not sound like repeated measures anova to 
me. I am not sure exactly what you want, but it sounds like ancova might be 
close:
 
http://en.wikipedia.org/wiki/Analysis_of_covariance

-- 
Christopher Brown, Ph.D.
Associate Research Professor
Department of Speech and Hearing Science
Arizona State University
http://pal.asu.edu