Which of the following is generally true when analyzing data from a clinical trial?

	A.	Subjects who do not comply with the treatment protocol are usually excluded from analysis
	B.	The possibility of confounding is reduced when multiple subgroups are analyzed
	C.	Multiple comparisons are performed to minimize the possibility of making a type II error
	D.	Planned subgroup analyses specified before study initiation (a priori) are viewed with less skepticism than post-hoc subgroup analyses