I'm looking at the impact of 4 two level independent categorical variables on a continous dependent variable. How to I choose what effect to use for a posthoc analysis? If I choose forexample size*treatment*shelfcode how do I consider the effect color to explain dependent variable Firmness?
Do I do posthoc on all interactions? If yes, how do I interprete separate posthocs on the same variable firmness
Thank you for your responses.