Bolender.com

Concepts are essential to theory formation, but they are not theories in themselves. Adult, male, wage earner, unemployment are concepts; each may bear a relation to the others in certain contexts. But their mere array side by side does not constitute a theory. However, we could use them to formulate one. In fact, we could not theorize without concepts. And just as we could not theorize without concepts, so we could not accumulate knowledge in sociology without theories.

A sociological theory--or, more generally, a scientific theory--consists of a set of concepts and generalizations arranged in such a way that it explains some phenomenon or a relationship between phenomena (Dressler, 1973, p. 23). We observe a great many cases in which African Americans vote Democratic in presidential elections. With the help of voting statistics, survey data (interviews with voters), and a sense of the political and social history of African Americans in the United States, we put together a theory to explain these findings (or relationships between events). The theory should "save the date," that is, provide plausible explanation for all the findings. It should not ignore certain facts and emphasize others, for if it does, it probably will not prove to be a valid theory of African American politicization. In this instance, we might conclude that Africa Americans vote for the political party that seems to them to be most sympathetic to the plight of African Americans, keeping in mind the historical roles of both parties and the impact of Roosevelt's New Deal coalition during the 1930s (which put together minorities, farmers, intellectuals, and unionized workers). Using this theory, we can then predict more or less accurately the political responses of categories of African Americans. We further refine our theory as we go along, making it more efficient in explaining African American voting behavior. If our theory does not "work," that is, does not provide significant predictive power, we may abandon it in favor of some alternative theory. If our theory is comprehensive, so will the range of its application; if it is narrow, so will be the amount of date we can "save" by the theory.

Sociological theory not only leads to understanding of immediate present events, conditions, and other phenomena; it additionally improves our ability to predict human interaction generally, with respect to the given subject. Thus, we may theorize from immediate observation and analyze that in a group of people who are on the verge of starvation because there is a shortage of available food, individuals will to a greater or lesser extent abandon traditional ideas about approved behavior and fight each other for what food is available. This theory, if proven sound, will enhance our ability to predict what will occur whenever a group of people is on the verge of starvation because of a shortage of available food.

This example suggests that theories vary according to how efficient and comprehensive they are as aids in predicting the relationship between events. We theorized above that behavior of individual members of the group would change "to a greater or lesser extent." This indicates that not every member is expected to depart in the same degree from traditional modes of behavior. Similarly, not all African Americans vote Democratic, and those who do may not do so for the reasons expressed in our theory. It is desirable, therefore, to classify theories on the basis of the predictability factor.

Scientific theory is educated and well-reasoned conjecture based upon available facts. However, the unavailability of all of the pertinent facts at a given time places some theories in the category of hypothesis. Hypotheses are tentative explanations of the relationship between events. They are in the nature of "hunches" still to be fully proven: for instance, a hypothesis that states hungry mice will eventually learn to run a maze to the point where food awaits them.

Some theories are referred to as laws. Laws are invariable explanations of the relationship between events. An example is the so-called law of gravity. Gravitational pull seems to be present only and always when certain factors are present. (Actually, we cannot ever be sure that even this is a law. Any so-called law may require revision or total repudiation on the basis of later evidence.) Sociologists have discovered no laws of human interaction. Can you think of one? What law of courtship makes it certain that marriage will follow? When a person from an advanced society introduces her/his advanced technology into a society less developed, what law states precisely what the effect will invariable be upon family life in that society? 

The hypothetical hunch and the supposedly invariable law are connected by a continuum of predictability. Although no sociological theories are laws in the strict sense that we use the term, they do contribute to our understanding of human interaction, and for this reason it is necessary for us to inquire into how the theories are constructed.

1. A proposition is advanced. A proposition in this context, is an assumed relation between one factor and one or more others. Here is an example from sociological literature: "...the occurrence of aggressive behavior always presupposes the existence of frustration and, contrariwise...the existence of frustration always lead to some form of aggression" (Dollard, Doob, Miller, Mowrer, & Sears, 1939, p. 1).

 2. Data on the subject are accumulated. The sociologists may collect his own or rely wholly or partly on what already exists bearing on her/his problem. Either way, in our example a logical procedure would be to observe persons who are frustrated in their attempts to achieve a desired goal.

3. The data so secured are classified and analyzed.

4. The original proposition is tested against the data and is confirmed, rejected, or modified.

5. A generalization, or theory, is formulated concerning the problem under investigation.

To illustrate with the frustration-aggression proposition referred to above, every element in it would have to be so clearly and precisely defined that the investigator knows exactly what she/he is dealing with (this is the process of concept formation in science). Moreover, another investigator, wishing to repeat the test, must be enabled to do so by the fact that the definitions are so clearly stated they have the same meaning to her/him as to the original investigator (recall that science must be replicable). In our frustration-aggression investigation it would be essential to define such terms as occurrence, frustration, and aggression.

The original investigator seeking to test her/his proposition would next seek data in research reports that concern her/his problem. We know, as a matter of fact, that she/he would learn from such reports that many, but not all, persons react to frustration by some form of aggression. Further, we know that aggression, when present, takes several characteristic forms. 

From these data we would conclude that the proposition is not to be entirely rejected. Neither has it been completely confirmed. Rather, it seems generally valid but subject to modification.

We emerge finally with a generalization, or theory, which may be stated as follows: 

When an individual seeking to achieve a certain goal finds an external barrier standing between him and that goal, he normally attempts to (a) remove the barrier, (b) bypass it, or (c) otherwise master it. Failing in this, he is frustrated. In that condition, he may take some action that is not to be classified as aggression: He may flee from, rather than attack, the barrier (Berelson & Steiner, 1964, pp. 266-271).

But instead of taking flight, the frustrated individual may indulge in some other characteristic form of aggressive behavior. He may assault the invulnerable barrier. He may displace his aggression by directing it, not at the barrier, but rather at a more vulnerable, although innocent, object. He may revert to an earlier mode of behavior that is ill adapted to removal of the present barrier.

By way of review, any theory would ideally include the following characteristics:

1. The concepts employed are defined with painstaking exactitude. (In our illustration, as we have already noted, occurrence, frustration, and aggression must be defined. So must goal, barrier, displaced aggression, and others.)

2. The propositions are consistent with each other. ("The occurrence of aggressive behavior always presupposes the existence of frustration." If we hope to prove this, we must expect the proposition to be consistent with its corollary, which is that "the existence of frustration always leads to some form of aggression.")

3. The final generalization or theory, is an outcome of careful observation, facilitated by efficient measuring devices. (In our example, observations, facilitated by measuring devices, corroborated the proposition.)

4. The theory can be verified by other investigators. (This is true of the studies contributing to the frustration-aggression generalization.)

5. The theory is tentative. No theory is ever final. It is always subject to modification and even repudiation, if and when further evidence is discovered.