[FieldTrip] Post hoc test following significant interaction

Paul Dhami pdhami06 at gmail.com
Thu Nov 1 21:21:38 CET 2018


Dear FieldTrip community,

I have a 2 x 2 mixed study design (within factor being pre post, and
between factor being group 1 and group 2) in which I am hoping to use an
ANOVA in the cluster based permutation framework to analyze.

Since I have a 2 x 2 factorial design, I have been following the tutorial:

http://www.fieldtriptoolbox.org/faq/how_can_i_test_an_interaction_effect_using_cluster-based_permutation_tests

as well as various previous responses on the mailing list.

In the context that I have an interaction in certain electrodes at certain
time points, my questions are:

1. How would I go about testing the simple effects to disentangle the
interaction. Would it simply be a case of skipping the main effects and
running the appropriate indep/dep t-test between each 4 of the possible
contrasts for simple effects? (i.e. group 1 vs group 2 at pre, group 1 vs
group 2 at post, group 1 at pre vs group 1 at post, group 2 at pre vs group
2 at post).

2). If an interaction is initially found, should the post-hoc (e.g. simple
main effects) be limited to the electrodes and time in which the
interactions are present, or would the simple effect contrasts be run
across all electrodes and time points as were run in the initial
interaction test.

Any insight would be greatly appreciated.

Thank you,
Paul
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20181101/f094c81b/attachment-0002.html>


More information about the fieldtrip mailing list