TOWARD A CONSTRUCTIONAL APPROACH TO SOCIAL PROBLEMS: ETHICAL AND CONSTITUTIONAL ISSUES RAISED BY APPLIED BEHAVIOR ANALYSIS

Accordingly, I shall now consider some research implications of the model.

RESEARCH IMPLICATIONS OF CONSTRUCTIONAL MODELS

The constructional models we have been discussing are based on the commonalities between laboratories of the experimental analysis of behavior and programed instruction, its derivative. These make change procedures and outcomes explicit. By so doing, they make it possible to identify which procedures effectively contribute to contracted progressions. The practices described have led not only to the development and production of successful programs of practical importance, but also to a research methodology on how to program change. Further, the requirement of explicitness makes it possible to communicate readily the research and programing methodology (cf. Skinner, 1968).

It is not generally realized that such practices may simultaneously be used for all the ends that govern basic research. The notion that development of technology and of technical competence defines the endeavor is far from the truth.

The reasoning is not complex. The program, no matter how successful, is not identical to the variables which produced the outcome, since it also contained elements extraneous to the outcome. These, so to speak, were incorporated in the successful package. Refinement consists of successful substitution of procedures considered more relevant. The practical effect is that the program becomes less costly in time and effort. The scientific effect is that the program approaches identity to the variables of which the outcomes (terminal and way-stations) are functions. Carefully controlled changes within a program for an individual make it possible to describe functional relations between dependent and independent variables, using the classic strategy of single organism research (Sidman, I960).81 Other cases may similarly serve as new experiments addressed to replication or extension, or to unresolved questions.82 Changes made for such investigative purposes are, of course classic scientific procedures and may contribute to general scientific knowledge.83

Such changes may also contribute to the personal knowledge of the client (person or system) about his own (personal or social) contingencies. This holds to the extent that he becomes an explicit coinvestigator in this scientific task. In conventional psychotherapy, as most often practiced today, the therapist typically informs the patient that "we both" will work toward understanding of what is going on. However, the deduced relations are stated in terms and are produced under conditions which make them difficult to validate. When necessary controls are introduced, the relation of the results to change procedures is usually tenuous.84 The constructional models suggested utilize a strategy which makes possible the derivation of validated relations functional for knowledge and treatment, for the investigator. The strategy may serve the same purposes for the client. The strategy is in accord with the classic psychotherapeutic aim noted. Such a schema poses no practical difficulties for reconciling practice and research. The schema also poses no ethical problems regarding the use of a patient as an experimental subject in a research project. He may be a coinvestigator, not only in the classic psychotherapeutic sense, but also in joint discussions on what should be manipulated to find out what was going on, in the explicit collection of data for this purpose, in the use of categories which make quantification possible (our patients often draw their own graphs), in analysis of functional relations, and in suggestions for future directions.