Exactly, it's not about always doing X and never doing Y, but designing a procedure that best answers your research question (within the constraints of practicality) and reporting results in a way that they accurately reflect what you really did.