29
ON THE RELATION BETWEEN THE EXPERIMENTAL ANALYSIS OF HUMAN BEHAVIOR AND APPLIED BEHAVIOR ANALYSIS SAMUEL M. DEITZ
Georgia State University
In the beginning of the science of behavior, there was only the experimental analysis of behavior. Not many years later, however, this experimental analysis of behavior begat the field of applied behavior analysis. B. F. Skinner (e.g., 1953; 1978), who has consistently argued that a complete science of behavior would tell us not only about the variables that control behavior but also about ways we could work for the survival of the culture, could only be pleased. Both the experimental analysis of behavior and applied behavior analysis grew and prospered. But all was not happy in this behavioral Eden. Beginning in the second decade ), articles appeared which of The Journal of Applied Behavior Analysis (1968 were critical of this child of the experimental analysis of behavior. He or she had strayed from the path of righteousness. No one argued that this child was evil, just that he or she was wayward and needed correction (not necessarily overcorrection, but correction, nonetheless). It is, in part, from the issues raised by these critics that the experimental analysis of human behavior grew (see Hake, 1982); and it is on these issues that the relation between the experimental analysis of human behavior and applied behavior analysis rests. To understand this relation, then, we must carefully examine the several charges these voices raised. They are the rather serious charges of inappropriate purposes, insufficient data, and inadequate consideration of theory. Inappropriate Purposes The experimental analysis of behavior was designed to increase our understanding of the "causes"1 of behavior (e.g., Sidman, 1960). Since its beginning, studies conducted within this area of science were conducted to achieve this purpose. Initially, applied behavior analysis was also supposed to strive toward this same objective. Applied behavior analyses were to teach us about the causes of behavior, but not just of any behavior; we were to learn of the causes of socially significant human behavior (see Baer, Wolf, & Risley, 1968). This was a significant advance since the experimental analysis of behavior had clearly helped achieve only one part of Skinner's goal for a science of behavior. The applied analysis of behavior was a needed step to teach us about important 'The term "cause" is used throughout this paper with some caution. As Skinner (1953) explained, "There is no particular danger in using "cause" and "effect" in an informal discussion if we are always ready to substitute their more exact counterparts" (p. 23). Since my paper is written in a somewhat informal way and since I am sure the reader is quite capable of substituting more exact meanings, I will trust the reader to make any necessary translation.
30
DEITZ
human behavior. From these studies we could learn how to improve the world and possibly even how to assure the survival of our culture. But early successes in fixing real human problems led to a change in applied behavior analysis. Many studies in the field began to be conducted only to improve the human condition. These studies did not attempt to teach us about the causes of important human behaviors. They used what was already known to improve behavior; new facts about causes were not discovered. Much of the field began to apply the existing knowledge of behavior analysis rather than to conduct applied behavior analyses. The analysis, in other words, was disappearing from applied behavior analysis (see Hayes, Rincover, & Solnick, 1980). It was to this point that the voices (e.g., Deitz, 1978; 1982; Pierce & Epling, 1980) began their critique. They claimed that the purpose of applied behavior analysis had changed. They said that this sort of applied behavior analysis would not teach us very much and there was more we needed to know. The causes of behavior needed to be studied and applied behavior analysis was no longer contributing to that purpose. Of course many of the new applied behavior analysts disagreed (Azrin, 1977; Bear, 1981). Their argument was that there are many, many problems in the world which need to be solved right now. They argued that most of those problems could be solved with what we already know. They admitted that we needed to know more about causes but their jobs could not wait for those who wished to find out about them. A brief visit to a school or a mental retardation institution would surely show that they were right. Many problems exist and some of them can be solved with our current knowledge. So what was the problem? This leads us into the second charge against these new applied behavior analysts. Insufficient Data To use existing knowledge to solve social problems implies that there is enough knowledge existing to solve those problems. This is an arguable point. How much do we really know about the various ways to fix social problems? Some behavior analysts would argue that we really know very little about solutions to complex problems (e.g., Ribes, 1977). Sure, we can get children to be quiet in a classroom or retarded clients to pee in the potty. Existing knowledge is quite sufficient to meet these types of objectives. The problem arises when we ask more difficult questions. Do we currently know enough to design effective prison reforms so that prisoners, after they have served their terms, do not eventually end up back in prison? The data on that question are fairly clear-the answer is no (M. Milan, personal communication, August, 1985). We do not know enough about fixing criminals or for that matter about the causes of crime and its maintaining variables to be able to design such a system. Do we currently know enough to design an effective school curriculum on levels above elementary skills? In other words, can we effectively teach such complicated subjects as chemistry or history or music? The current data are again clear and the answer is again no. In this case, we are lacking the necessary knowledge of instructional stimulus control techniques to establish or transfer control over a complicated set of behaviors, none of which we understand very well either. On this topic, too, many applied behavior analysts disagreed. There are those who said we now know enough about education, for example, to completely fix what is wrong with our schools (Greer, 1982). These people, as well as those who claim we have enough knowledge to fix any major social flaw, are, in my opinion, quite wrong. Others (Baer, 1981), however, made a different argument. They said,
EXPERIMENTAL ANALYSIS OF HUMAN BEHAVIOR SYMPOSIUM
31
sure there is more to be learned and we cannot do everything. But th9re is much we can do and we must proceed to do it. We don't have time to wait and what we can already do is very important and very useful. These people are, in myopinion, quite right. We should do what we can do but we need to remember always how much we do not yet know. So once again we must ask, what's the problem? This leads us into the third and final charge against the new applied behavior analysts. Inadequate Theory When applied behavior analysts studied the causes of socially significant human behavior, many interesting facts were discovered which could be incorporated into a growing theory of behavior. As applied behavior analysis became a field predominately dedicated to solving p'roblems, however, the kind of facts necessary to build such a theory were no longer generated. Various therapies were used to improve some behavior but the practitioners involved in those studies found no need to relate their findings to the existing and inductively growing body of behavioral theory (see Deitz, 1978). The experimental analysis of behavior continued to provide these kinds of facts but data about humal) issues were missing. One could argue, perhaps, that facts obtained specifically from human studies are unnecessary to building a complete theory of behavior. I do not think such an argument would get very far, however. Much of what we need to know concerns issues in which other animals are rarely involved. Analog studies provide some information but there are always dangers in drawing ana:logies. Behavior analysts cannot learn enough from non human animals just as cognitive psychologists cannot learn enough from computers. This would be less of a problem if all we were interested in learning about was a set of simple responses under a set of fairly simple contingencies. We hope we are interested in much more than that. Humans do things which are very complicated-they remember, solve problems, create, and talk. These sorts of activities are probably under the control of very complicated contingencies. For the sake of this theory, we need to be able to understand these types of behaviors and the contingencies which establish and maintain them; we cannot be satisfied just to teach them or to fix them (see Marr, 1983). Applied behavior analysts often examine these types of complex behaviors but rarely in order to be able to draw theoretical implications from their work. But as with other criticisms, applied behavior analysts replied that this is true but irrelevant (Baer, 1981). It is true that they do not do their work to add to the body of behavioral theory. They do their work to solve pressing problems. It is not important if what they find is theoretically important as long as it solves the problem. Again, since there are so many important social problems in our world, it is hard to argue that they should not be involved in such activities. So, for a final time I must ask, what's the problem? This time the question leads to my conclusions and the topic of the experimental analysis of human behavior. Conclusions If applied behavior analysis has become a field with improvement of human behavior as its purpose, with little concern for answers to questions about the variables of which complex human behavior is a function, and with little interest in theory, and if this is such a good thing, is there really a problem? Well, for those
32
DEITZ
who think those issues are not very important, the answer is "no." But early behavioral work with humans (e.g., Weiner, 1983) showed clearly that a set of very different and difficult problems existed to be understood. These were not problems that needed to be solved so that someone could be made better, they were problems of a conceptual nature that related to the causes of human behavior. If this is so, the problem is, "Who will study those causes of human behavior? Who will provide the data about human behavior? Who will build the theory?" Only a very few traditional experimental analysts of behavior have shown any interest in the causes of human behavior-they continue to study the causes of behavior but with rats and pigeons and apes. That too is a good thing but what about humans? Two ends of a continuum defining the field of behavior analysis is obviously insufficient; we are in need of a middle. The field of behavior analysis is fortunate because a set of investigators is beginning to fill that middle. This middle ground of behavior analysis has been formed to investigate human behavior. Some have come to this middle from the end defined by applied behavior analysis; others have come from the end defined by the experimental analysis of behavior; some have come from both; still others have been working in this area for many years. This middle ground has defined a field which has come to be called the experimental analysis of human behavior. The purpose of this new field is to investigate causes, not just of any behavior but of human behavior, and not just of SOCially important human behavior but of any kind of human behavior. The field does not exist to replicate work in the nonhuman animal laboratory (see Hake, 1982); rather studies within this field can tell us much that we do not know about the variables of which both simple and complex human behavior is a function. The data from these studies may even lead us to the design of better ways to improve socially significiant behavior. Finally, the data should be useful in building a more complete theory of behavior. So what is the relation between the experimental analysis of human behavior and applied behavior analysis? At the present time, there's not a very large one. The experimental analysis of human behavior has filled some gaps left by the necessary and important changes in applied behavior analysis. This new field will provide some answers about the causes of human behavior and about theoretical concerns that applied behavior analysts no longer seek. Those in the experimental analysis of human behavior will relate to those in the applied analysis of behavior much as they do to those in the experimental analysis of behavior. There will be three overlapping areas defining a complete behavior analysis; each will most likely work somewhat independently but will watch the others for good ideas and good data. With an experimental analysis of human behavior added to the experimental analysis of behavior and applied behavior analysis, almost all of Skinner's objectives for a science of behavior will be possible to reach. This complete field of behavior analysis can proceed on many significant fronts making many different kinds of contributions. The gaps are filled, the criticisms of any part of behavior analysis can end, and the answer to my repetitive question, "What's the problem?", is that there is no longer a problem. We can all look upon the various deeds and accomplishments of these three parts of behavior analysis with pride.
EXPERIMENTAL ANALYSIS OF HUMAN BEHAVIOR SYMPOSIUM
33
References AZRIN, N. H. (1977). A strategy for applied research: Learning based but outcome oriented. American Psychologist, 32, 140-149. BAER, D. M. (1981). A flight of behavior analysis. The Behavior Analyst, 4, 85-91. BAER, D. M., WOLF, M. M., & RISLEY, T. R. (1968). Some current dimensions of applied behavior analysis. Journal of Applied Behavior Analysis, 1,91-97. DEITZ, S. M. (1978). Current status of applied behavior analysis: Science versus technology. American Psychologist, 33, 805-814. DEITZ, S. M. (1982). Defining applied behavior analysis: An historical analogy. The Behavior Analyst, 5, 53-64. GREER, R. D. (1982). Countercontrols for the American Educational Research Association. The Behavior Analyst, 5, 65-76. HAKE, D. F. (1982). The basic-applied continuum and the possible evolution of human operant social and verbal research. The Behavior Analyst, 5, 21-28. HAYES, S. C., RINCOVER, A., & SOLNICK, J. V. (1980). The technical drift of applied behavior analysis. Journal of Applied Behavior Analysis, 13, 275-285. The Journal of Applied Behavior Analysis. (1968 ). Lawrence, KS: Society for the Experimental Analysis of Behavior, Inc. MARR, M. J. (1983). Memory: Models and metaphors. The Psychological Record, 33, 12-19. PIERCE, W. E., & EPLlNG, W. F. (1980). What happened to the analysis in applied behavior analysis? The Behavior Analyst, 3, 1-9. RIBES, E. (1977). Relationship among behavior theory, experimental research, and behavior modification techniques. The Psychological Record, 27,417-424. SIDMAN, M. (1960). Tactics of scientific research. New York: Basic Books. SKINNER, B. F. (1953). Science and human behavior. New York: Macmillan. SKINNER, B. F. (1978). Reflections on behaviorism and society. Englewood Cliffs, NJ: Prentice-Hall. WEINER, H. (1983). Some thoughts on discrepant human-animal performances under schedules of reinforcement. The Psychological Record, 33, 521-532.