Checking for baseline differences

Question

I have a 2x2x2 design where the variables are treatment (CBT, medication), information given (high, low), and time (baseline, post-treatment). I am going to do a mixed ANOVA. However, I would also like to compare the baseline groups to see if there are any differences between the four groups (CBT-high, CBT-low, medication-high, medication-low).

How do I test these differences? Using an one-way ANOVA before doing the mixed ANOVA? Or using post-hoc tests after I have done the mixed ANOVA? Thanks!

Answer 1

It depends on the kind of variables you want to compare on baseline. Are they categorical or continuous (and then normally distributed or not normally distributed)?

Categorical data 2 or more groups might simply be tested with the Chi-squared test
Continuous data with normal distribution might be compared using a one-way ANOVA, as you said.
Continuous data without normal distribution might be compared using the Kruskal-Wallis H test.

Does that answer your question?