 |
 |
See below. On Wed, Oct 21, 2015 at 10:39 AM, André Schmidt <[log in to unmask]> wrote: > Hi Donald, > > 1) No my covariates match the number of scans for each group (G1:13 and > G2:23). I just added two covariates. But maybe I have to use the multiple > covariates option? > >> Where in the GUI did you enter the covariates? > > 2) Is there a way to compare the correlations? > Yes. See http://core.ecu.edu/psyc/wuenschk/docs30/CompareCorrCoeff.pdf > > Thank you > André > ------------------------------ > *Von:* MCLAREN, Donald [[log in to unmask]] > *Gesendet:* Mittwoch, 21. Oktober 2015 16:35 > *An:* André Schmidt > *Cc:* SPM > *Betreff:* Re: [SPM] comparing correlations across groups > > Andre, > > (1) Your covariate length is different than the number of scans. How did > you enter the covariate? > (2) Your analysis is comparing the slopes between groups, not the > correlation. > > Best Regards, > Donald McLaren, PhD > > > On Wed, Oct 21, 2015 at 10:18 AM, André Schmidt <[log in to unmask]> > wrote: > >> Dear experts, >> >> I want to test whether the correlation between brain activity and a >> behavioral score is different between two groups. I tried to do this with a >> 2-sample t test and adding the two behavioral scores as covariates for each >> group separately and then using a contrast 1 -1 over the covariates. >> However, I got the this error message: >> >> Running job #2 >> ------------------------------------------------------------------------ >> Running 'Factorial design specification' >> Failed 'Factorial design specification' >> Error using spm_meanby (line 31) >> row dimension mismatch between inputs >> In file "/Users/andre/Documents/MATLAB/spm12/spm_meanby.m" (v1143), >> function "spm_meanby" at line 31. >> In file >> "/Users/andre/Documents/MATLAB/spm12/config/spm_run_factorial_design.m" >> (v6219), function "spm_run_factorial_design" at line 623. >> >> The following modules did not run: >> Failed: Factorial design specification >> >> Can anybody guide how the specify the batch for this. >> >> Thank you very much >> André >> > >