We wrote our first book on research methods fresh out of graduate school. Having both trained at Northwestern University under the watchful eye and, at times, thumb, of Donald Campbell, we were heavily invested in the experimental method, and the particular mindset that Campbell and Stanley (1966) had championed in their classic monograph, Experimental and Quasi-Experimental Designs for Research. It seemed to us then that the most certain avenue to advance in the social sciences was via the experimental road. In some ways, we still adhere to this proposition. Despite a host of worthy competitors, we believe the experiment remains the single most certain method to uncover causal relationships. Further, the experimental model provides a useful standard against which to evaluate the quality and utility of research findings based on non-experimental techniques. As a reference point, the experiment is useful even in settings that do not admit to the experimental method.
At the same time that we learned and absorbed the critical importance of experimental techniques, we were learning about the developing quasi-experimental approaches, which Cook and Campbell (1979) elaborated so elegantly, and which in no small measure helped establish the sub-field of evaluation research. In so doing, development of the quasi-experimental approaches also contributed to the developing recognition in the field that applications of our methods in socially relevant field settings was not an unworthy activity. In this book, we discuss the research emphases that have developed on the basis of hard thinking about experiments and their limits, their potential for social good, their application, and their misapplication.
That the experiment is not the only method available to social scientists is abundantly clear, perhaps more so today than yesterday. Similarly, the strict conditions that govern the appropriate use of the experiment are perhaps more obvious and accepted than in earlier times in the field. And yes, we recognize more clearly now than before that features inherent in the method itself can cause serious problems in inference, if not controlled. These issues have become increasingly central features of methodological disquiet over the years, and this ferment has been beneficial. The apprehension regarding the proper use of the experiment, its weaknesses as well as its strengths, reflects the developing