FOUR CRITERIA FOR EVALUATING A THEORY
[References are to Slawski’s Social Psychological Theories text. This is an extract.]
With the preceding overview of the nature of social-scientific theory, theoretical problems, and the modes of theory construction, the student has the seeds of what to look for when approaching a new social theory or problem. Innumerable other considerations could be discussed which, for certain limited, say classificatory purposes, might shed some light on the philosophy, substance, or methodology of social science. The number of possibilities is so great that many theorists and philosophers have alternatively written books and been baffled by the problem of evaluating theories comparatively. Classification, however, is not enough. Nor is it sufficient to analyze one theory at a time. Because of the complexity and abstractness of theories and evaluative criteria, we need a point-by-point comparison of two or more theories at a time. This approach is not only theoretically sound, not to mention neglected, but also pedagogically crucial. It is quite clear in addition that to compare two complex theories as a whole is a task suitable only for genius. Thus, our approach will be to suggest comparison of the main points of two or three theories at a time, as applied to a single instance of personal or social change over time. This strategy is also contained in the instructions for student write-ups in the Appendix.
Our task has been to reduce the huge number of possible evaluative criteria to the most important ones, without leaving out anything really essential to the nature of theory. On these grounds, four criteria seem necessary (cf. Slawski: 1974). In a natural chronological order of use for evaluating theory, they are: (1) ease of application, (2) information value, (3) predictability, and (4) explanatory power.
Ease of Application
In common-sense applications of a theory, the relative ease of application depends on the understandability of the theory to the persons using it, and the degree to which the terms and hypotheses of the theory fit the facts of the case under analysis. Social scientists, however, are more likely to go beyond the intuitive and examine the question of testability of a theory. It involves the extent to which the variables can be separately and reliably measured by different researchers. If we could not measure the concepts, our theory would not be falsifiable, but rather a tautology or mere opinion. We could not then gather evidence for and against the validity of the hypotheses contained in the theory. A result of testability is the degree to which the theory is productive for those doing research. That is, how fruitful is it in generating evidence? Testability also involves the applicability of the theory to real-life events. Ideally, a tested theory should be applicable to policy decisions of administrators, or perhaps to recommendations and therapeutic methodologies of counselors and organizational change agents. Furthermore, a testable hypothesis is methodologically sound. The form of the theory as a whole must be clear, simply or economically stated, complete in application to the events to be explained. It should be elegantly stated as well (Mullins: 1971). Finally, the indicators of each concept of the theory should be observable in a way that will allow us to quantify them. As a final word, it should be clear that for a student who has not yet had a course in social-research methods, these points about testabilit) can be safely passed over lightly. He should instead focus simply on the ease of application of the perspectives presented to events and experiences in his own group life.
Once we discern that a theory is applicable to the case under consideration, it is
natural to look at the criterion of information value. Here we ask the questions: How well does the theory help u~ to describe what actually happens, the nature of the events and their sequence? Do the range and quality of the concepts and hypotheses of the theory focus our attention on the meaningful and significant, on facts and solid ideas rather than on mere opinion, the trivial or tautologous? An informative theory is also proximate to experience rather than purely conceptual.
The third crucial criterion for judging a theory is its predictive potential. Are the variables related in causal or functional statements? Do the statements tell us, for example, that if and when A happens, B will follow? Naturally, if we can predict, then we can more easily control our destiny, or at least set limits to it. Another aspect of prediction is postdiction, which is prediction of events occurring at time 2 (after the turning point) from the events or circumstances at time 1 (before the turning point), but making that prediction at a point in time after both time 1 and time 2 events have already transpired, or even where data have already been collected on both points or periods of time. We may be able to predict, however, on the basis of past experience alone, through correlations, even without being able to explain why the prediction held true. This brings us to the fourth, and most important, of the criteria for judging a theory, namely, explanatory power.
Explanatory power is the essence of a theory. It tells how well a theory shows why there was change or stability in a person or group over a given period of time. More broadly, it tells why what actually happened did in fact occur. Finally, it evaluates how well the hypotheses of the theory order the data, the basic concepts, relationships, and assumptions. Explanation is another word for theory. It goes beyond prediction. Prediction without reasons, without knowledge of conditions, causes, or motives is a useful happening, but a poor substitute for understanding. If we understand, we will also, of course, be more able to predict. Thus, with these four criteria, we have a very general but very powerful means of comparing and evaluating all perspectives treated in this text.