open-discussion > RE: how to control for covariance
Dec 18, 2014  12:12 PM | Martin Styner
RE: how to control for covariance
Re 1) your command is nearly correct, but you actually have 2 group-types (group and gender), so 
shapeAnalysisMANCOVA ./designfile.csv --infileColumn 8 --numGroupTypes 2 --numIndependent 6 --columnGroupTypes 0,1 --columnIndependent 2,3,4,5,6,7 

Now, how does shapeAnalysisMANCOA know which one of the group columns to use for testing. By default it will use the first column (index 0), which is why you do not need to specify it in the above example. So the above call is the same as this one:
shapeAnalysisMANCOVA ./designfile.csv --infileColumn 8 --numGroupTypes 2 --numIndependent 6 --columnGroupTypes 0,1 --columnIndependent 2,3,4,5,6,7 --testColumn 0

Now let's assume you want to test for gender then you would call: 

shapeAnalysisMANCOVA ./designfile.csv --infileColumn 8 --numGroupTypes 2 --numIndependent 6 --columnGroupTypes 0,1 --columnIndependent 2,3,4,5,6,7 --testColumn 1

Threaded View

TitleAuthorDate
Chao Suo Dec 8, 2014
Beatriz Paniagua Dec 11, 2014
Olof Lindberg Feb 6, 2015
Martin Styner Feb 6, 2015
Chao Suo Dec 17, 2014
Martin Styner Dec 17, 2014
Chao Suo Dec 18, 2014
RE: how to control for covariance
Martin Styner Dec 18, 2014
Chao Suo Jan 8, 2015
Martin Styner Jan 8, 2015
Chao Suo Jan 9, 2015
Martin Styner Jan 9, 2015
Martin Styner Dec 18, 2014
Olof Lindberg Dec 11, 2014
Beatriz Paniagua Dec 16, 2014
Olof Lindberg Dec 16, 2014
Martin Styner Dec 16, 2014