Scientific Method in the Social Sciences (with commentary by Ray Bergner).
The 19th and 20th centuries have been good ones for the physical sciences (e.g. physics, chemistry, molecular biology). Our ability as a race to understand, predict, and control our physical environment has advanced by leaps and bounds. At the same time, the social sciences (e.g. psychology, sociology, political science) have made more modest gains. Attempts to carry the successful scientific method of the physical sciences over to the social sciences have met with mixed success at best. Admittedly, persons are more difficult to understand, predict and control than molecules. Are the differences in the results of the two kinds of science due to some essential difference between the objects of study? Or have we just not found the right formula for making the social sciences more productive?
In this paper I am going to argue in favor of the former position—that there are essential differences in the two kinds of sciences that are rooted in their respective subject matters, and which impact what you can reasonably consider to be scientific method. However, once you have accounted carefully for the essential differences, you end up with set of requirements for a more general scientific method for the social sciences of which scientific method for the physical sciences is a special case. The analysis undertaken here will also show how to join the two kinds of scientific method for subject matter like human emotions or psychopathology for which both have something to offer.
The principle technical resource that I will use for this investigation is a discipline called Descriptive Psychology (hereafter DP). DP has been developed over the past 35 years by Dr. Peter Ossorio of the University of Colorado, and two generations of his students. Ossorio’s approach was to develop a large set of agreements, formats, and principles for talking about human behavior. He has thereby provided practical semantics for a wide range of concepts needed to do science, including object, process, event, and state of affairs (to support explanation in the physical sciences, Ossorio, 1978), and person, intentional action, significance, status, and community (to support explanations of behavior, Ossorio, 1998). It is important to this paper that there exist robust semantics for these terms and several others of similar ilk, while the details of the semantics are less important here and can be left to the references.
I will not attempt to survey DP here, others have done that (Shideler, 1988). A small but active group of researchers and practitioners has embraced DP and has, over the past three decades, produced a significant body of literature (Advance in Descriptive Psychology, Vols. 1-6, 1981-1991). I will, however, freely use intellectual resources from DP, citing a source for each. These resources will be particularly useful in establishing the outer limits on the kinds of situations that scientific method for the social sciences, especially psychology, is going to have to deal with. So lets begin.
In the American Heritage Dictionary (Second College Edition) "science" is defined as "The observation, identification, description, experimental investigation, and theoretical explanation of natural phenomena." The words and their corresponding concepts in this definition are close to everyday experience and non-controversial with two exceptions: what constitutes a legitimate experiment and what constitutes an explanation. Consider the concept of explanation in two kinds of contexts. On the one hand there is the physical science context, the context that physicists and chemists use when speaking professionally. Their language features objects, processes, events, and states of affairs, but not persons, intentions, communities, or significance. Explanation in such contexts typically has a reductionistic flavor: if you ask a physicist "why" you will get an elaboration of the laws of physics applied to the case at hand to imply that the observed behavior was the only one possible. Contexts of this sort are Newtonian Physics, Relativity, Quantum Theory, and Evolution.
On the other hand, there are contexts in which human behavior takes center stage, as in the worlds of the politician, the businessman, the playwright, and the clergyman. In these worlds, person, intention, community, and significance all play leading roles. If you ask a playwright "why" you will get an explanation involving intention, community, and values that elucidates the significance of a certain action in the life of a character. These behavioral worlds include the possibility of creation. Only persons are eligible to create, and persons include homo sapiens (if you are not a reductionist) and God (if you are a theist). On the face of it, there is little connection between the two kinds of worlds, and there are even linguistic and conceptual dangers in confusing the two different approaches to answering "why".
Having different standards in different communities for what constitutes an explanation is neither new nor particularly harmful. In this case, however, there is a problem. The problem lies in coming up with comprehensive explanations—that do justice to both perspectives—for phenomena for which both types of explanations have light to shed on the phenomenon. Such phenomena include emotions, psychopathology, and the behaviors of ancient cultures, to name just a few. For example, a psychological researcher interested in emotions may need to take into account the hormonal activities of her subjects, while simultaneously considering the appraisals made by those subjects of the emotion-producing stimuli and the significance, in the worlds of those subjects, of the situations appraised.
Let us consider some of the guiding principles of scientific method in the physical sciences, and how they might be generalized to admit the additional methods needed for explanations of behavior. In this we will grow the domain of application of the scientific method of the physical sciences, successively approximating from below the scope needed for the social sciences.
(See Bergner comment 1) Success of a theory is measured by its ability to predict phenomena accurately (with the promise of being able to control those phenomena). The behavior of persons and groups can seldom be predicted accurately, much less controlled. On the other hand, after-the-fact explanations of behaviors can be constructed that are helpful for guiding future action. Two kinds of anticipation of future behaviors are possible: (a) designating certain behaviors as likely rather than certain, and (b) designating a certain range of behaviors as the only ones possible, with the individual or group under study choosing freely within this range. Note that both these kinds of predictions are subject to empirical test, both quantitatively through statistical studies, and qualitatively, through the judgment of an expert that "Yes, that agrees with my clinical experience."
(See Bergner comment 2) Only that which has been rigorously empirically verified is considered to be true. If persons, in the course of everyday life, chose only to act on those facts that had been rigorously empirically verified, they could never get anything done. On the contrary, "A person takes it that things are as they seem unless he has reason to think otherwise" (Ossorio, 1998, Maxim A8). In this context the distinction between "real" and "true" (ibid. pp. 126-127) is useful. A state of affairs is real—for me—if I am prepared to act on it. In any community there exists a shared world consisting of everything that is real for that community (Putman, 1981). Individual community members may harbor minor disagreements with this shared world but major disagreement would call into question the membership of the individual in the community. A state of affairs is true—for a community—if it is a fact in the shared world of that community. The standards for truth of the community of physical scientists are well-known and rigorous. (For another kind of community for which truth is important, and for which very different standards apply, consider the legal community.)
Part of my intention here is to point out that the standards for truth in the community of social scientists need to admit truths that do not meet the current standards of empirical verification in the community of physical sciences, for example, consider the diagnosis of a client as having Obsessive-Compulsive Disorder. In cases like this the test of truth is agreement by a panel of competent community members in good standing, given the same raw empirical data. Although this may seem a bit soft compared to, say, measuring the specific heat of aluminum, notice that even in the case of the physical sciences, a panel of competent community members in good standing may be needed to judge the validity of an empirical result. So for both communities a combination of raw data and judgment by competent community members is required. The main difference is that in the physical sciences, there are fewer disagreements among the subcommunities, and therefore more emphasis on the raw data than in the social sciences, where there is more opportunity for different subcommunities to judge the same raw data differently due to differences in their respective real worlds.
Note in passing that, in the community of physical sciences, everything that is true is also real. The converse is not the case, even in this very truth-oriented community. For example the question of who is eligible to judge whether another’s claim of empirical validation is itself justified could not be answered empirically. Attempting to do so would lead to an infinite regress.
Thus reasonable methods are available within one community for attaining objectivity, for both truth and reality, but what of the relativism of these methods across communities? The bad news is that there is no avoiding it: the real world of the Australian Aborigine really is different from the real world of the Swiss Burgher. The good news is that the impact of these differences is softened in two ways: 1. Some communities cut across others: If the Australian Aborigine and the Swiss Burgher both happen to be Computer Scientists, or Catholics, then in each case there is a community with its corresponding real world that they share. 2. There is a process—negotiation(Shideler, 1988, pp. 80-85)—by means of which agreement among real worlds may be increased: If the Australian Aborigine and the Swiss Burgher are called upon to cooperate on some project, they are likely to negotiate just enough shared world to move forward on the project.
In the light of all the above considerations, a psychologist in the consulting room is in an essentially different situation from the biologist in the laboratory. The psychologist is acting on a human relationship, taking it that things are as they seem unless he has reason to think otherwise, and calling on his own real world, the real world of his professional community, and the truths that have met the standards of that community. He gets frequent opportunities to create therapeutic moves out of whole cloth, based on his own experience and the tricks of the trade. A major part of his task is to assist the client in the reconstruction of the client’s real world into one in which living is more satisfying. As a human being conducting her life, many of the same things are true of the biologist, even in the laboratory, but in her life in the laboratory, the emphasis is all different. The truths that have met the standards of her scientific community take center stage, creation is done less frequently and in bigger chunks. The real world of the laboratory is so well agreed upon that it seldom comes up for discussion. She is trying to add a brick to the edifice of biological truth within that single world.
Results only qualify as empirical if they can be reproduced in other than their original circumstances. This principle must hold in the social sciences as well, else what would be the use of the results? Yes, but care is called for in understanding "results" properly. They are the results of testing mutually understood hypotheses by means of agreed-upon methods. Both this mutual understanding and this agreement are subject to failure due to the differing real worlds of different communities. Negotiation on these points is necessary before results can even be compared. Further care is called for because in the social sciences many of the most interesting phenomena, e.g. dreams and headaches, are essentially irreproducible. The social scientist must be careful not to throw up her hands at the outset when dealing with irreproducible phenomena, but instead frame hypotheses wisely so that they are amenable to independent test in spite of the irreproducibility of the underlying phenomena. For example, a dream researcher might be able to invent a scheme for classifying dream content in such a way that the statistics of dream content for an individual predict personality characteristics that are measurable in other ways.
For both the physical and the social sciences, the underlying logic of the scientific method is the same: truth is established empirically, and empirical method is established in the practices of the community. That which is true is also real, but reality is broader than truth, by virtue of Maxim A8: "A person takes it that things are as they seem unless he has reason to think otherwise" (Ossorio, 1998).The differences between the two communities, which call for a broader view of what is empirical and what constitutes an explanation in the social sciences, come from two main sources: 1. The presence of subject matter—persons—that feature both free will and private experience, and 2. A greater diversity of subcommunities, calling for more negotiation to bring differing real worlds into agreement.
Most communities have taken the logical position of, for the most part, embracing as real whatever they deem true: truths are to be acted upon, and the negations of untruths are also to be acted upon. (States of affairs classified as superstitions that are nevertheless widely acted upon represent possible exceptions to this logical position.) Many communities have imported, lock, stock, and barrel, the truths of the community of physical scientists, since that community has such strong and widely respected standards for truth. An example of a community that has not done this is the community of Christian Creationists. Its position, although within its rights as a community to have its own real world, carries with it two challenges, one logical within the community and one political outside it. The logical challenge is that some of the truths of physical science—e.g. those supporting the practices of medicine and food production—are important parts of the Creationists’ world. Where to draw the line? Physical science is pretty much of a piece; it’s a pretty overwhelming logical job to take it apart in order to pick and choose. (Incidentally, reductionists have the same challenge going the other way: they need behavior description in order to talk about science as a human endeavor, yet they try to avoid the rest of the implications of being human.) The political challenge has to do with the society in which Creationists are living and its aspirations in the area of universal education. Universal education calls for a baseline of truth and reality that can be taught to all comers, regardless of their community. Agreement on this baseline is an important objective for negotiation among the various cultures within a society.
Shideler, M. M. (1988). Persons, Behavior, and the World. Lanham, MD: University Press of America, Inc, 1988. ISBN 0-8191-6787-8.
Ossorio, P. G. (1971/1975/1978). "What Actually Happens". Columbia, South Carolina: University of South Carolina Press, 1975, 1978. (Original work published 1971 as LRI Report No. 10a. Whittier, CA & Boulder, CO: Linguistics Research Institute. Later listed as LRI Report No. 20.)
Advances in descriptive psychology, Vols. 1-4. Greenwich, CN: JAI Press 1981-1984, and Vols. 5-6, Boulder, CO and Ann Arbor MI: Descriptive Psychology Press. 1990,1991.
Putman, A. O. (1981) Communities. In K.E. Davis (Ed.) Advances in descriptive psychology, Vol. 1. Greenwich, CN: JAI Press.
Ossorio, P. G. (1982/1998). Place. Volume III, The Collected Works of Peter G. Ossorio. Ann Arbor, MI: Descriptive Psychology Press. Originally published as Place (LRI Report No. 30a). Boulder, CO: Linguistic Research Institute.