The ideal control group is one that is identical to the experimental group in every way except that they do not receive the experimental manipulation or "treatment". In experimental research, random assignment helps to insure that the groups are comparable. We also must pay careful attention to our control to make sure that it is isolating out the most important factor.

Clinical experiments often use a "wait-list" control group as a standard against which to measure treatment. Some researchers object to such a control because our results essentially compare treatment to "doing nothing". A more appropriate comparison is one that includes some activity, just not the one that is considered important for effective behavior change.

For example, if we were evaluating a new reading program for elementary school children, it would seem absurd to compare this treatment to not teaching children to read at all. The most appropriate comparison would be to the standard instructional method. Likewise for psychological interventions, we would want to compare a new treatment to the standard approach. If one did not exist, we would want to have a comparison that involved some type of learning activity and some contact between the experimenter and the participant.