[Eeglablist] Unequal group size and STUDY use

[Bradley Voytek]

The easiest work-around is to just do stats in matlab. Most people do
t-tests, but if you want to be safe and if you have the statistics
toolbox, use the ranksum command in matlab to perform the Mann-Whitney
U test, which is a non-parametric t-test.

Create two vectors, one for each group, where each element in the
vector is the mean amplitude (or the sum of amplitudes) of a time
window for each trial, for each subject.  Use these two vectors as
your input for your ttest2 or ranksum analyses.

It's pretty easy.

If you want to do bootstrapping statistics for each time point like
EEGLAB does, it'll take a little more work for unequal groups, but
it's possible.

::bradley voytek

On Sun, Nov 30, 2008 at 3:41 PM, Hamish INNES-BROWN
<HINNES-BROWN at bionicear.org> wrote:
> Hi Micheal, yep, the same thing happens with my data.
>
> I can get ERSP and ITC statistics with unequal groups, but not ERP
> stats.
>
> At the moment I plan to work out how run the stats myself, either using
> the EEGLAB data structure or by exporting, but I haven't started that.
>
> I too would be glad of any pointers antone can give!
>
>
> Hamish Innes-Brown
> Senior Research Assistant
>
> The Bionic Ear Institute
> c/o St Vincents Hospital
> 6th Floor Daly Wing,
> 35 Victoria Pde, Fitzroy Vic 3065
> Tel: +61 3 9288 3523
> Fax: +61 3 6997 7518
>
> -----Original Message-----
> From: eeglablist-bounces at sccn.ucsd.edu
> [mailto:eeglablist-bounces at sccn.ucsd.edu] On Behalf Of Michael Stevens
> Sent: Tuesday, 25 November 2008 4:38 AM
> To: eeglablist at sccn.ucsd.edu
> Subject: [Eeglablist] Unequal group size and STUDY use
>
> Hi everyone,
>
> Sorry if this is a known issue that I just didn't see in previous
> posts... But it'd be quite helpful if someone could confirm a
> problem/limitation.  I had been beating my head against my computer
> trying to figure out why I couldn't get simple group stats to display
> properly when doing t-tests (and similar analyses) in a STUDY.  I first
> thought it was code I whipped up to create the STUDY, then I thought I
> had a bad dataset somehow in the mix, etc.  However, one of my students
> and I were working on this when she finally noticed that group
> comparison results could be visualized perfectly when there were equal
> numbers of participants in each group.  Until that time, I had been
> trying to compare one group of 26 subjects with another of 19.  Sure
> enough, when I selected exactly 19 vs 19, all the results/visualization
> plots worked fine.  Until then for instance, it'd give me ERP plots for
> one group, but blank plots for the second, etc.
>
> Is this a known issue?  Or perhaps it was just something I was doing
> wrong locally?  Obviously, it'd be more flexible to be able to use
> unequal numbers of participants in group comparisons.
>
> Thanks for anyone who can confirm this, or just point to something I'm
> doing wrong.  (FYI - I'm using the version of EEGLAB as of 10/08, the
> most recent one as far as I know - 6.03b, I think).
>
> Best,
> Mike
>
>
> Michael C. Stevens, Ph.D.
>
> Director, Child and Adolescent Research
> The Institute of Living / Hartford Hospital
>
> Director, Clinical Neuroscience & Development Laboratory Olin
> Neuropsychiatry Research Center
>
> Assistant Clinical Professor of Psychiatry Yale University School of
> Medicine
>
> Contact Information:
> 200 Retreat Avenue
> ONRC, Whitehall Building
> Hartford, CT  06106
>
> Tel: (860) 545-7552
> Fax: (860) 545-7797
>
>
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to
> eeglablist-unsubscribe at sccn.ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to
> eeglablist-request at sccn.ucsd.edu
>
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to eeglablist-unsubscribe at sccn.ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to eeglablist-request at sccn.ucsd.edu
>