ABSTRACT
Observational methodology contrasts with experimental and quasi-experimental methodologies inasmuch as it does not use the design and statisti cal controls that they employ to warrant the validity of research findings. Lack of controls is both a blessing and a handicap. On one hand, it makes observation the quintessential naturalistic methodology. On the other hand, it creates a dilemma: How can observational studies be performed rigorously without controls over the research setting? In this chapter, I ad dress this question by drawing a distinction between two conceptions of human observers: the observer as recorder, who is required to emulate me chanical data collection devices by refraining as much as possible from bringing in his or her own interpretation into the observation process, and the observer as interpreter, who is encouraged to convey his or her inter pretations provided he or she makes the process of moving from data to conceptualization transparent to others. The two conceptualizations entail different notions of "good data," validation of findings, and rigorous obser vation. Following a critique of the first conception in terms of arguments in spired by the second conception, I discuss how the latter can be practiced by decision researchers compatibly with scientific rigor. I begin with a vi gnette from a review by Thorngate (1978) of an edited volume of group deci sion making that has stuck in my mind ever since I came across it quite a few years ago:
What made this episode into a sort of epiphany for me was on one hand Rowell's conviction that "there is so much more to learn by watching," to which I, as an admirer of Darwin, could easily relate, and the ready accep tance by Thorngate (who is, I assume, an experimenter by training like myself) of this conviction, nicely put as the dictum "Search before [you] research, [and] trust [y]our senses as much as [y]our designs." Still, ow ing to my training (or possibly to some character flaw), some doubts persistently clouded my delight in Thorngate's (1978) call for naturalists to step forward in psychological research. Does "trusting] our senses" mean that everything goes? Put differently, how can we make warranted asser tions based on observational studies? What I wish to do in this presenta tion is share with you some tentative conclusions to which my preoccupa tion with this question has led me to date. These conclusions assume that rigorous research, irrespective of methodology, satisfies two principles-"collect good data" and "draw sound conclusions"—and basically unpack the implications of these principles for observational methodology. To be gin with, I consider some general observations on the problem of rigor in this methodology.
