 |
 |
Hi Réka Papers certainly still get published with uncorrected thresholds, although among these papers, I think p < .001 is much more common than p < .05. However, I think the general consensus is that, for a whole brain analysis, the number of statistical tests performed makes inferences drawn from any kind of uncorrected threshold somewhat dubious. That being said, a great deal depends on your study and the specific hypotheses you are testing, and you may well get differing opinions from everyone on the list (or from each of your reviewers). Jonathan On Tue, Oct 28, 2008 at 12:35 PM, Réka Horváth <[log in to unmask]> wrote: > Dear List Users, > > I have a question regarding the first level analysis in the SPM5. Is it > necessary to choose p<0.001(uncorr.) value for thresholding ? > Can I use or is it appropriate to use a less strict threshold like p<0.05? > > Thank you for your help > > Reka Horvath > >