On the impact of decision rule assumptions in experimental designs on preference recovery