Hi, The simplest thing is to break down your ANOVA into the t-tests of interest and run those. For example, if you want to compare stim1 to stim2, you would run a level 2 (fixed effects) analysis to calculate the paired difference (subject-specific models) and then run a 1-sample t-test on this in a group model. This simplifies the repeated measures aspect of the data. Hope that helps, Jeanette On Thu, Sep 27, 2012 at 3:44 PM, Cara Damiano <[log in to unmask]>wrote: > We are interested in doing a group (control, patient) x stimulus type (4 > levels) ANOVA in FSL but are not sure how to do this. We have tried to set > up F tests for the within-subjects factor of stimulus type and then feed > these fstats to higher-level analyses but this doesn't seem to work. Any > suggestions? Thanks! > > Cara > >