Hi Hongjiang, Thanks so much. I also got the advice that it's not good idea to use ess* files to do random effect analysis. I'll think about what you said. Really appreciate your help. Best, Chunhong -- Chunhong Shao [log in to unmask] Research Associate Department of Psychological Sciences Purdue University Quoting Hongjian He <[log in to unmask]>: > Hi,Chunhong > > So sorry that i misunderstand your question and for my delayed reply. > But i don't think it a good idea to take random analysis using > ess*files. I'll use ttest results for next F-test if it needed. > > Best Regards. > On 8/2/07, [log in to unmask] <[log in to unmask]> wrote: > > Hi Hongjian, > > > > Thanks so much for your reply. But I'm still so confused about this. You > know, > > if we select F-test, that means I want to find the difference between > > regressors, which should be two-way difference. Whereas, If I select > t-test, > > that means I want to find the one-way difference. The resulting results are > con* > > file (if t-test) and ess* file (if F-test). If I select F-test, there is no > con* > > file in the directory. I may misunderstand your meaning. If so, could you > let me > > know more? Thanks so much! > > > > Best regards, > > -- > > Chunhong Shao > > [log in to unmask] > > Research Associate > > Department of Psychological Sciences > > Purdue University > > > > > > Quoting Hongjian He <[log in to unmask]>: > > > > > I think, if i've catch your idea, one consider whether it's a F-test > > > or T-test only when he chooses the different contrast for the output > > > result, i.e, which statistical result you want to review. However, the > > > inputs for both are the same, which means you also select the > > > con*.imgs for "random effect analysis for F-test". > > > > > > On 7/31/07, Chunhong Shao <[log in to unmask]> wrote: > > > > Dear SPM users, > > > > > > > > When I did statistical analysis for the individual subject, I selected > > > F-test > > > > above contrast window. It resulted in corresponding "ess" files. How > could > > > I do > > > > a random effect analysis for the F-test? Thanks in advance! > > > > > > > > Regards, > > > > Chunhong > > > > > > > > > > > > > -- >