[Eeglablist] study: permutation-based condition statistics; multiple comparisons

[Jaan]

dear list,
 
I'm a rookie, but I have enjoyed working with EEGLAB a lot. praise to the people, who developed it, but now also some questions:
 
if I run the study and get the condition statistics, then how could I get the exact significance of my results? it shows me "p<0.002" (and that's really kind of the program), but how significant is it exactly? (somehow the program does not give lower values for the permutation-based significance than 0.002 - if I move it to 0.0019 all my effects are gone /and I refuse to accept that all the effects on all the channels are exactly at 0.002 and not lower/). I used the 'naccu' of 20 000 and so far I understood, it should be able to compute lower values than 0.002. 
 
any help of how I could move the permutation based statistics bar lower than 0.002 or get my exact results, would be really appreciated.
 
I'm not satisfied with 0.002, because when comparing the ERPS of two conditions I of course perform many tests at the same time. how many ? = number of times out (200) x number of frequencies (25)?? what to you think would be the appropriate p level? (not Bonferroni, which is too conservative and well ... I can't get lower than 0.002 as I told You)
In Delorme & Makeig, 2004 one reads that the correction for multiple comparisons feature will be available in near term future. is there any hope that the developers get to it someday? 
 
thanks a lot for taking Your time and helping me.
 
cheers,
Jaan
 
 
_________________________________________________________________
News, entertainment and everything you care about at Live.com. Get it now!
http://www.live.com/getstarted.aspx
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20080508/00a863a1/attachment.html>