Reflections: A Conceptual Analysis and Synthesis
Irving E. Sigel Educational Testing Service
Elaine Blechman Albert Einstein College of Medicine
This chapter attempts a synthesis of the papers in this volume. The papers represent different models of research directed at a common problem of interest--the study of depression and aggression in families. The fact that different models for research exist poses some critical problems for both research and practice. Because the goal of research in this area is to seek answers to questions about the pervasive impact of depression on family members, and hence develop treatment programs to alleviate the potential risks to members of families with depressive parents, there is good reason to create a coherent model of the illness and its consequences. The variety of contexts, constructs, mechanisms, and methods listed in Table 11. 1 show the diverse areas of interest.
Searching for commonalities among the diverse models is, we believe, a step toward more extensive and inclusive construction of models guiding research and practice dealing with the role of depression in the family context, particularly if we can systematize the relationships among the relevant variables.
As is well known in the physical sciences and often in medicine, researchers build on the shoulder of the predecessors. There is, overall, a continuous growth process where the accumulation of knowledge results in building a relatively integrated edifice. To be sure, this may be an idealized notion of how other scientific endeavors work, but nevertheless it seems clear that the physical and biological sciences develop common procedures and share a somewhat agreed upon notion of method that lead to replicable results of experiments. The behavioral sciences are not so blessed--at least not at this time. Investigators have greater license to develop their own theories, their own sets of procedures, and their own analyses resulting in considerable diversity. Further, there is not too much sharing of procedures and methods and there is relatively little replication