-------- Original Message -------- Subject: Re: Regression and DFA/CVA in MorphoJ Date: Wed, 5 Oct 2011 15:37:17 -0400 From: Chris Klingenberg <[email protected]> Reply-To: [email protected] Organization: University of Manchester To: [email protected] Dear Eric I have found the bug and fixed it -- this will be in the next update. Frankly, I would not do the analysis you are describing. A simpler way to correct for both age and size is to include both as two independent variables in a regression. I presume you have age in your dataset as a covariate. When it comes to select the independent variables, select covariates *and* centroid size as the data matrices; in the variables list, select age and centroid size. The resulting analysis will take into account both independent variables simultaneously. (You could also do this for more than two variables.) Best wishes, Chris On 10/5/2011 6:51 PM, morphmet wrote:
-------- Original Message -------- Subject: Regression and DFA/CVA in MorphoJ Date: Wed, 5 Oct 2011 12:36:55 -0400 From: [email protected] To: [email protected] Hi, I find that when I do the following, MorphoJ does not perform DFA or CVA. I often regress Procrustes coordinates against an independent variable such as age, and perform a second regression of those residuals (not the regression scores) against centroid size. MorphoJ will perform PCA on these new residuals, but it will not compute a DF (though it will, of course, if the procrustes coordinates are regressed one time against an independent variable). I've seen this with several different datasets, and was wondering if anyone might comment on this. Thanks. Eric.
-- *************************************************************** Christian Peter Klingenberg Faculty of Life Sciences The University of Manchester Michael Smith Building Oxford Road Manchester M13 9PT United Kingdom Telephone: +44 161 275 3899 Fax: +44 161 275 5082 E-mail: [email protected] Web: http://www.flywings.org.uk Skype: chris_klingenberg ***************************************************************
