-pair subjects who are identical on relevant variables and assign one to the control group and one to the experimental group -create groups that are equivalent in terms of their averages on some relevant variables -randomly assign subjects to the experimental and control groups -match subjects on relevant variables -(all of these choices)