I have a response variable that lies between 0 and 1. It is a proportion. My dependent variables are categorical. Normally, I would run a two-way ANOVA but, in this case, would it be okay to run a beta regression? Or is there a two-ANOVA equivalent that I can run?