Almost more of conceptual question, but I am wondering if one should use the svy prefix with specified strata when analysing data from a stratfied randomized controlled trial using regression techniques?