open-discussion
open-discussion > RE: shapeAnalysisMANCOVA and p-values
Feb 25, 2015 01:02 PM | Beatriz Paniagua
RE: shapeAnalysisMANCOVA and p-values
Hi Marco,
Great to hear that shapeAnalysisMANCOVA is working for you.
I hope that helps!
You are more than welcome.
Thank you for using SPHARM-PDM,
Beatriz
Great to hear that shapeAnalysisMANCOVA is working for you.
When I load output_meanAll_uncorrected in
ShapePopulationViewer, I can understand from DiffMagnitude whether
the difference (in mm?) is A-B or B-A, right?
Correct. Now, how can I obtain different p-values maps
for A-B and B-A?
In theory you can change the order of your input csv files, in
order the groups are assigned differently. In practice your
p-values are going to look exactly the same, since the significance
is evaluated in base to the magnitude of the difference, and it is
the same in each direction.And also, what's the difference between
output_GLM_correctedMean and output_meanAll_uncorrected? which sort
of correction is applied?
correctedMean is computed from the General Linear Model (GLM) that
is built for group comparison in MANCOVA, and though it accounts
correction for any covariates included in the GLM. uncorrected is
just the arithmetic average of both A and B means.I hope that helps!
You are more than welcome.
Thank you for using SPHARM-PDM,
Beatriz
Threaded View
| Title | Author | Date |
|---|---|---|
| Marco P | Feb 24, 2015 | |
| Beatriz Paniagua | Feb 25, 2015 | |