Hi Barbara
>In the SPM2 basic models, what is the difference between simple
>regression(correlation)and multiple regression using one covariate?
>
>
should be similar :-\
>
>Using the same design matrices and input parameters, I am getting
>significantly different results using "simple regression" and "multiple
>regression". Which method is to be preferred?
>
>
did you check the last column?
one explanation is that in one case you had a constant and not in the
other case.. (multiple reg with/without cst)
but in general I would say 1 cov = simple regression.
Best
Cyril
|