Chapter 15: Ancova And Quasi-Experimental Design: The Legacy of Campbell and Stanley

Two of the most frequently cited references on the utility of quasi-experimental design are Campbell and Stanley (1963) and Cook and Campbell (1979). The latter defined quasi-experimental design as

Campbell and Stanley suggested the use of quasi-experimental designs for the “many natural social settings” (p. 34) in which there is no ability to randomly expose participants to experimental stimuli. Some commonly encountered examples of quasi-experimental designs are depicted in Figure 15.1. Note the absence of symbols (typically “R”) representing random selection and random assignment.

The inability to randomly assign participants to a treatment group is quite common in social and behavioral science research. Clinicians, teachers, social workers, therapists, scholars, doctoral and master’s students, and funded researchers are frequently permitted access to the classroom, hospital, or clinic for research purposes, only to find that the treatment and control groups available must be limited to intact groups. Having participated on over seven dozen doctoral committees, I have heard many unsatisfactory attempts to support the possibility that the two intact groups are as random as can be expected in field research. In perusing the applied research literature, the reader is inundated with a variety of heuristic arguments from authors struggling to find base-line equality among intact groups.