In designing studies, if randomization is not feasible, what can researchers do to ensure comparability between the experimental and control groups?

What will be an ideal response?

If randomization is not feasible, researchers have to use alternative methods to ensure comparability between the experimental and control groups. One approach is called the two new interventions, or two new programs, approach. Because new programs are often more attractive, researchers may add a second program as a control for the main intervention program. Subjects can be randomly assigned to one new program or the other. The second new program can be a version of the main intervention without the costlier components. It can also serve as a true competitor to the common intervention program that is used by others in the community.

Another approach is the borderline control-group strategy. This is appropriate when the people most in need must be served by a particular program. The measure of who is the neediest is not always accurate, however. The borderline group refers to those immediately above the need cutoff point, although many of them also need the program.

Other approaches include the taking turns, or delayed program, approach, in which the groups take turns receiving the intervention or program; matching, which involves making the control group as similar to the experimental group as possible by creating a quota matrix that includes all relevant characteristics to be matched; and stratified random sampling, in which researchers begin with a pool of subjects, create strata of subjects with similar characteristics to be matched, and randomly assign subjects from each strata to experimental or control groups.