Back to the main page.
Bug 783 - montecarlo p value should never be zero
| Status | CLOSED FIXED |
| Reported | 2011-06-28 21:47:00 +0200 |
| Modified | 2019-08-10 11:46:17 +0200 |
| Product: | FieldTrip |
| Component: | core |
| Version: | unspecified |
| Hardware: | All |
| Operating System: | All |
| Importance: | P1 major |
| Assigned to: | Saskia Haegens |
| URL: | |
| Tags: | |
| Depends on: | |
| Blocks: | |
| See also: |
Robert Oostenveld - 2011-06-28 21:47:49 +0200
instead it should include the statobs in the randomization distribution, which results in a minimum p-value of 1% if you have 100 permutations. Or more efficiently (sum(statrand>statobs) + 1) / (nrand + 1)
Saskia Haegens - 2012-02-01 18:22:55 +0100
this has now been implemented, would appreciate if eelke can fully test this! in case of cfg.numrandomization = 'all', behaviour is as before (i.e., smallest possible p=0) in all other cases, smallest possible p value is now 1/(Nrand+1)
Saskia Haegens - 2012-02-01 18:23:26 +0100
ps: robert heeft hier nog testscripts voor!
Boris Reuderink - 2012-02-03 21:26:04 +0100
Dear Saskia, could you indicate in which revision you fixed this problem? And did eelke already test it?
Saskia Haegens - 2012-02-14 10:34:44 +0100
revision 5225