[Eeglablist] erp analyses with unequal groups

[Arnaud Delorme]

Rob Anderson wrote:
> Hi. I am currently conducting erp group comparisons of oddball and 
> gonogo data for groups that have been defined based on a cluster 
> analysis of personality trait variables. I would like to compare the 
> erps for these different clusters (4 of them), but have only been able 
> to compare groups of equal size (i.e., taking a subsample of the 
> clusters). I do not have the Matlab Statistical Toolbox installed, so 
> the statistical analyses are defaulting to the nonparametric 
> permutation tests.

These tests are usually more rigorous as they do not rely on the 
probability distribution of the data.


> If I install the statistical toolbox, will that enable me to make erp 
> comparisons for unequal groups?

Yes, it will. Note that you can still set a threshold p-value using the 
parametric statistics in EEGLAB. The only thing EEGLAB will not do is to 
compute the exact p-value itself as this is the part that requires the 
statistics toolbox (fcdf() and tcdf() functions).

> If not, is there some other way to get statistics based on unequal 
> groups? If not, is there a relatively easy way to do a bootstrap sort 
> of analyses where the smaller groups are formed with different 
> membership and then tested statistically to see if there is a 
> difference within a given group when there are different members in 
> that group?

I am not sure I understand clearly what you want to say. The statistical 
functions in EEGLAB can support group of equal size (usually paired) or 
group of unequal size (unpaired). The command line output indicates you 
what type of statistic is being used by default.

Best regards,

Arno

> Thanks for any help you can give me with this.  Rob 
>  
> ** 
>  
> Rob Anderson, Ph.D.
> Research Assistant Professor
> ------------------------------------------------------------------------
>
> _______________________________________________
> eeglablist mailing list eeglablist 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

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20070511/326a661f/attachment.html>