[FieldTrip] Applying Bonferroni Correction to Alpha Parameter Instead of ANOVA

Paul Dhami pdhami06 at gmail.com
Sun Dec 30 18:42:06 CET 2018


Dear Fieldtrip community,

I have an ERP experiment with a 2 x 2 design (lets say with one factor
having levels A and B and the other having factors 1 and 2), both being
within groups measures. As stated in the interaction page, I can't use that
to test for an ANOVA design when both factors are within group measures.

Instead of running an ANOVA, I was thinking of running separate contrasts
of interest, which would be A1 vs A2 and B1 vs B2.

My question is:
1) to correct for now the multiple contrasts, would simply setting the
alpha parameter (which is used for the inference of accepting or rejecting
the null hypothesis) to 0.5/2 be acceptable? What if I wanted to test all
possible contrasts, would I simply then divide my alpha parameter by the
number of contrasts I have?

Any input would be appreciated.
Best,
Paul
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20181230/584df671/attachment-0001.html>


More information about the fieldtrip mailing list