The only correct way to make a comparison of two groups is a 2-sample t-test. Showing significance occurs in one group, but not the other, in 2 separate tests does *not* imply that the groups are significantly different. Cheers, Jeanette On Fri, Apr 12, 2013 at 7:20 AM, SUBSCRIBE FSL Mariam Zabihi < [log in to unmask]> wrote: > Hi, > I did set my design and contrast like this, but i didn't get any > significant p values. then I decided to change the glm features! > I'm just wondering if it's possible to investigate the events' relevancy > of data separately by setting the specific contrast [1 0] for ev1 and [0 > 1]for ev2. > the out put will be two set of significant pvalues for each event. by > removing the common significant pvalues, i have 2 different set of > significant for each events which show me the event's relevancy cross data. > But I'm not sure is it a common way (or at least correct way!) of analysis > of the comparing two group effects or not! > Thank you > Mariam >