Repeated measures ANOVA + multiple comparisons corrections

[rohanp16]

Hello everybody,

 

I had a question in regards to ANOVAs and multiple comparisons corrections (Bonferroni mostly).

 

I am running a 3 way repeated measures anova (2 within-subjects factors - one factor with two levels, another factor with most likely more than 3 levels + 1 between subject factor). Lets name this ANOVA #1

 

I am running this with EEG spectral power as the dependent variable. Now if I divided the power into the traditional power bands (delta, theta, alpha, beta and gamma), then I will have 5 ANOVA's (one for each band).

 

Now where I stand confused, theoretically, is that do I do multiple comparisons corrections for

 

a) the 5 different ANOVAs I am conducting for each frequency band

 

AND/OR

 

b) for the different number of pairwise comparisons that will arise due to the within subject factor having multiple levels in EACH anova ?

 

I realize SPSS does have an option for selecting the multiple comparisons corrections (LSD, Sidak and Bonferroni). So the output that comes is that already adjusted and has taken care of "alpha value (0.05)/number of tests" for the Bonferroni correction (if selected) or do I have to do that ?

 

Thanks a lot for reading, any guidance or pointing to any relevant research papers/books/websites would be highly appreciated !!

 

Rohan