The Tradition of
Experimentalism in Psychology
By John Darley
It sometimes makes sense to examine a discipline in terms of what I am tempted to call its "tribal customs." By this I mean its habitual, frequently unexamined, ways of doing its everyday activities and communicating about the results of those activities both to itself and to adjacent tribes. Continuing the metaphor, it is useful to compare the customs of psychology with scientifically adjacent tribes such as sociology, cognitive science, and anthropology, and even more distant tribes such as literary criticism and philosophy.
One reason for this sort of examination is celebratory - to revel in the superiority of the customs of our discipline over those of the adjacent disciplines. Once that is completed to the satisfaction of all, and the moral and intellectual superiority of our ways demonstrated to all of open mind, then it is possible to grudgingly consider whether a thing or two might be learned or a practice or two adopted from the competition.
I propose to go forward exactly in this vein, coming to the conclusion that scientific psychology is in good order, and that its everyday practices and ideological commitments to ways of doing things are generally contributing to its strengths. But I will manage to discover some difficulties, artfully casting them as the reverse side of our strengths, and see whether some friendly amendments to our customs might not be worth considering. In this month's column, I will explore the strengths of our traditions, in the next months we will see how those traditions might be limiting us.
The central field of psychology, the one that set the paradigm to which we all aspired, has been "experimental psychology." The remarkable fact here is that it was seldom necessary to specify what processes the experimentation was about; we were just "experimentalists." The topics on which we were experimenting differed over time, but our commitment to the experiment as the preferred technique for unlocking the secrets of the universe has been constant.
We were able to privilege experimentation because of the derivations that followed from another tenet of our ideology, which was that our task was to discover the universal laws of learning, or of perception, or of reasoning. The ways in which experimentation could strip down a complicated, messy situation to its apparent essence fit well with an emphasis on the project of discovering the universals of process.
With those universal processes in hand, we could then understand the complexities of, for instance, college students learning, experienced pilots perceiving, or experts reasoning. The laws of learning could be discovered by experimentation on infra-human species, and of memory by having subjects - people this time - learn nonsense syllables, and so on.
Other fields of psychology looked at problems that were less tractable to an experimental approach - or seemed so. But triumphs could be achieved by bringing the experimental method to those fields; many readers will remember the stir caused by Leon Festinger's first set of dissonance experiments, exactly because they demonstrated that it was possible to do experiments in a field in which they had not previously seemed possible. (Lewin carried out several studies that are broadly referred to as "experimental," but looked back on don't support that claim very well.)
And it is useful to remember that Festinger, who had a remarkably broad competitive streak, chose a reinforcement theory as the one he would challenge in those experiments. Reinforcement theory at that time was the paradigmatic center of the search for universals in psychology and so in challenging it, one announced the arrival of a new kid on the block, ready to challenge the system. But and again, the challenge was carried out by experimentation. In personality psychology, a number of scales of single constructs were designed, and validated by a procedure that resembled experimentation; groups of individuals who scored high or low on a particular personality construct were put through, for instance, an experimental procedure that measured conformity-proneness.
Ideological elements intertwine and support each other. Our preference for the experimental method fit well with our emphasis on the analysis of causality. (As introductory textbooks often proclaim, experimentation is the one method that can unequivocally establish causality.) And of course we were aware of the difficulties attendant on other modes of inference. We were aware, for instance, of the demonstrations that introspection gave a rather poor account of perceptual phenomena.
For me the power of experimentation to identify cause with immediate lucidity was learned in my undergraduate laboratory courses with Hans Wallach, who gently led us to hypotheses we could test about Gestalt perceptual phenomena. And when we had the hypothesis right, and the experiment properly designed, we could instantly see which stimulus alterations controlled the perceptual phenomena. I have talked to many other psychological researchers and it is amazing the frequency with which some similar story of discovering the joys of doing clean experimental designs in the laboratory of a beloved mentor is the life-altering event that led to our career choice.
So our program, to sum it up, involved discovering through experimentation, the basic laws of human functioning, laws that would identify the true causal events in a confused world of multiple possibilities. The laws would be universals, would transcend differences among persons brought about by culture or upbringing, would transcend the multiple differences brought about by the details of the contexts in which people existed. And we have the intellectual characteristics that are the internal representations of that program. We insist on clear causal thinking, we are skeptical about unproven (that is, experimentally untested) claims, and we scorn those whose theories seem to be nothing but long lists of complex and underspecified assertions. Metaphorically, psychologists pride themselves on being from Missouri.
This ideological program, or paradigm to stray into Kuhnian terms, has stood us in good stead. That is, it has led us to some remarkable and important discoveries. Anyone concerned with learning who ignores the far reaching discoveries of reinforcement theory is making a serious mistake. And we have outlined the workings of the perceptual system, and are now making important discoveries about the ways in which sensory and motor information are merged in the cortex. From judgment and decision making research we have prospect theory, and a bias- and heuristic-driven account of human decision making. Using priming and response latency measurement techniques, and the power of modern computers, we have made some fascinating discoveries about cognitive associationistic structures and their workings, and on the ways in which the mind represents information, events, and perceptions of people.
The paradigm has also protected us from what many would now consider fads and excesses. To say that the post-modernist perspective, that all "readings of a situation," all interpretations of an event, have equal validity (or perhaps equally no validity) had little influence on psychology is putting it mildly. Even though we hold that many aspects of an individual's view of the world are social constructions, we hold that those constructions are determined by causal circumstances, and constrained by the external realities of both the physical and the social worlds. And our commitment to experimentation as revealing causal truth has made us a far too rocky soil for notions of science being merely another cultural ritual, similar to augurs consulting chicken entrails or astrological signs to determine "truth," to take root. Compared to some other sciences, psychologists have spent little time in these debates and that seems to me to be on balance to be a good thing.
Each reader will have many more examples to add to a list of successes our scientific ideology has brought about, and the excesses it has allowed us to avoid. So with the security of "a job well done," it might now make some sense to consider what limits or difficulties the adherence to what we might call the causal/experimental paradigm has visited on us. Next column, I will make some suggestions about that.