Dear List,
I constructed a second level ANOVA using the flexible factorial design and
the factors:
Side (2 levels, independence: no, variance: equal)
Condition (3 levels, independence: no, variance: equal)
Group (3 levels, independence: yes, variance: unequal)
Subject (28 subjects, independance: yes, variance: equal)
I included all 4 Main effects and no interactions into the model.
The F-contrast for the effects of Group (3 rows, each of them giving a 1 to
one of the groups and 0 to the others, in combination with the appropriate
weights for the other factors) produces a result that seems resonable.
However, when I plot the parameter estimates, the third column (giving the
third row of the F-contrast and the contrast estimate according to group 3)
seems to have no confidence interval (compare attached screenshot). Matlab
tells me:
Warning: Imaginary parts of complex X and/or Y arguments ignored
> In spm_graph at 322
looking at contrast.interval and contrast.standarderror in the workspace I
find a complex value for the third element. It looks like somewhere around
line 298 in spm_graph.m
there is a negative value appearing in the diagonal, which leads to the
complex result of the sqrt in:
SE = sqrt(diag(SPM.xCon(Ic).c'*Bcov*SPM.xCon(Ic).c));
Did I use to much model for not enough data? I read about a limitation to
three factors in SPM, but it happily accepted this model. Any other ideas?
Thank you for your help!
Thomas
|