[FieldTrip] one/two-tails analysis

[Sysoeva, Olga Vladimirovna]

Dear group,

I do face some weird results during my analysis.
Usually I do t-test analysis to compare 2 condions/groups and use 2-tails, as far as I do not have any preliminary hypothesis about the direction of the effect.
So my script looks something like Script 1 (see below). Important here is that I do keep the default parameters, cfg.tail = 0; cfg.clustertail = 0;
To my understanding of statistic, if you have the two-tail effect, and group 1 has bigger mean than gr2, you should have the effect with one-tail as well and even more significant. But what I discovered just today that if I change this parameter to 1, cfg.tail = 1; cfg.clustertail = 1; in order to look spesifically for the one-tail effect (gr1 is bigger than gr2), my very significant 2 tail effect just disappeared completely. Is it a bug?

Your comments are appreciated.
Best reGards,
Olga.


Script 1
cfg = [];
cfg.method = 'montecarlo';       % use the Monte Carlo Method to calculate the significance probability
cfg.statistic = 'indepsamplesT'; % use the independent samples T-statistic as a measure to evaluate
                                 % the effect at the sample level
% cfg.statistic = 'indepsamplesF'; % use the independent samples T-statistic as a measure to evaluate
% cfg.computecritval = 'yes'                                 % the effect at the sample level
cfg.correctm = 'cluster';
cfg.clusteralpha = 0.05;         % alpha level of the sample-specific test statistic that will be used for thresholding
cfg.clusterstatistic = 'maxsum'; % test statistic that will be evaluated under the permutation distribution.
cfg.minnbchan = 1;               % minimum number of neighborhood channels that is required for a selected
                                 % sample to be included in the clustering algorithm (default=0).
cfg.tail = 0;                    % -1, 1 or 0 (default = 0); one-sided or two-sided test
cfg.clustertail = 0;
cfg.alpha = 0.05;               % alpha level of the permutation test
cfg.numrandomization = 1000;      % number of draws from the permutation distribution
cfg.neighbourdist = 0.2;
%cfg.neighbours      =  cfg_n;
cfg.channel          =  {'all', '-HEOG', '-VEOG', '-EKG'};
cfg.latency          = [0. 0.5];
cfg.layout           = 'C:\Matlab_programs\Fieldtrip\My_Fieldtrip\EEG1010_modified32.lay';

subj1 = length(GROUP {nGroup1,1});
subj2 = length(GROUP {nGroup2,1});
design = zeros(1,subj1+subj2);
design(1,1:subj1)        = 1;
design(1,subj1+1:subj1+subj2) = 2;
cfg.design   = design;
cfg.ivar     = 1;
[stat] = ft_timelockstatistics(cfg, peak1_gr1 , peak1_gr2 );



____________________________________________________________________________
Olga Sysoeva,
Research Associate, PhD
Psychiatry Department,
Washington University School of Medicine
Campus Box 8134
660 South Euclid Ave
Saint Louis, MO 63110-9909


________________________________
The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20120213/3aeec61c/attachment-0001.html>