<div dir="ltr">Hi All<div><br></div><div>Sorry for the very simplistic question here but I am finding it difficult to construct an answer - if indeed there is one - from previous posts.</div><div><br></div><div>I have a study which is a 2x2x3 repeated measures design, event-related using 64 channel EEG. The design is a culmination of a number of previous studies, so in that sense is justified, and eyeballing the data suggests that the effects are pulling in the directions we have hypothesised. I would really like to attempt to pull interaction effects, main effects and then specified (a priori and justified) contrasts from my EEG data. So, is it even possible using cluster permutation methods to report 3-way, 2-way and main effect stats from such a design, and if so, what is the best/most justifiable method? Secondly, considering the amount of *possible*n simple contrasts available in a 2x2x3 design - even if I don;t run them all because it is not meaningful to do so - what is the best method to correct for multiple comparisons in the simpel contrasts I do end up running?</div>
<div><br></div><div>Thans for any help in this confusing matter!</div><div><br></div><div>best,<br clear="all"><div> Jim <br></div><div><br></div>
</div></div>