VI. Attitude Measurement as a Method in Social Psychology
Daniel Katz
Princeton University
THE main uses of attitude measurement to date have been of a practical and applied variety. From the use of straw ballots to detect the varying currents of public opinion to controlled studies of our movie-made children, the trend of attitude research is unmistakable. This practical emphasis is only natural in social psychology which is so intimately bound up with the problems of everyday existence. Moreover, it is highly desirable for psychologists not to try to stand
( 480) above the battle of life. Nevertheless, there is no reason why attitude measurement should not be applied more extensively to problems in social psychology proper. If the absorption in immediate and practical issues continues to the neglect of fundamental problems, then attitude measurement will go the way of mental testing and become the exclusive property of educational, vocational, and other applied psychologies. It is the thesis of this paper that attitude measurement has a place in social psychology as well as in the applied fields.
Of the many uses which may be made of attitude measurement I shall mention only three. The first is the accurate and reliable recording of the ideologies or attitudes of people. This, rather than the study of personality, gives us the so-called content of culture. It has many advantages over the anthropological approach. Anthropologists bring back a description of a culture taken from a few informants and tell us very little of the range and frequency of the attitudes among the people studied. Moreover little is said about the extent of the deviations from particular norms of belief. Progress has been made in our own culture with this use of attitude measurement. As a result of these studies we have acquired considerable knowledge concerning the incidence and distribution of religious attitudes, moral judgments, political stereotypes, and other social and economic beliefs.
In an accurate recording of the ideologies of a people, the techniques to be employed merit careful consideration. The most refined measuring devices for attitudes are the psycho-physical scales contributed by Professor Thurstone. These scales give a finely graded series of opinions ranging from one extreme of an attitude variable to the opposed extreme. Their most appropriate use is to test the intensity of affect for or against a social symbol or proposal, since feeling can be conceived of in terms of quantitative gradation. Be-fore we measure the intensitive or affective dimension of an attitude, however, we generally want to know the ideational content of the attitude in relation to possible alternative attitudes. Now attitudes as the content of social situations do not consist of evenly graded positions along a scale. They represent abrupt jumps and even discrete qualities. Instead of the psycho-physical scales, therefore, they can be studied more adequately by means of the a priori type of scale used by Allport and Hartman. Once the important attitudes are brought out by this method the affective intensity of each attitude can be measured by the psycho-physical scale.
The second outstanding function of attitude measurement is the search for significant relationships between ideologies and psychological, physiological, and environmental factors. In spite of a number of informative studies, this search for significant relationships has not yielded rich returns. The results generally hold only for a particular place and a particular time and permit of little generalization. One reason is that the studies have been too piecemeal and too restricted in scope. Investigators have failed to appreciate a fundamental problem of methodology here. They tend to confuse attitudes and personality as Professor G. W. Allport has pointed out. The implicit assumption is made that the attitude measured is a reflection of some deep psychological trait in the individual's personality. Attitude measurement, however, tells us little about the psychological motivation back of men's beliefs. It gives us across-sectional comparison of many individuals. The knowledge of the psychological significance of an idea, or of a bit of behavior, in the individual's make-up comes from a
( 481) long-sectional study of that individual's personality. Now attitude measurement can go a certain way toward this same goal, if it is applied extensively enough to a group of individuals. If our cross-sectional study of a group covers a great many of their attitudes in relation to a great variety of other factors, we are really in a position to make significant generalizations, even though we have not studied the unique integration of every personality in the group. In other words if we find a high correlation between a single attitude and a single environmental factor, its value for predictive purposes is limited. Other factors which have not been considered may totally blot out the relationship discovered. The Gestalt psychologists are correct, I believe, in their emphasis upon the dynamic interplay of all the forces affecting the mind of the individual. Our arbitrary atomistic approach in the study of attitudes needs to be replaced by more ambitious studies in which we take ac-count of the complexity of factors related to the ideology of men. It might be bet-ter, therefore, to study a small group intensively than to study a large group superficially. Professor Schanck's study of a small community of two hundred people has yielded more social psychological data than some attitude studies involving thou-sands of cases.
A third use of attitude measurement is the delimitation or definition of the very groups which we wish to study. The stock way of selecting groups for comparison is to take natural divisions based upon sex and age or upon the conventional sociological groupings in terms of membership in organizations, institutions, and classes. Then we seek to discover differences in attitudes between the groups so defined. A far more profitable procedure is to delimit our groups psychologically from the start. This can be done by measuring the individual's attitude of identification with the ideology of his group. For example, it is better to take the identification of the individual with an economic class, (that is to measure his class attitude) than to put him into an economic class on the basis of his actual income. In the United States most people who derive 90 per cent of their income from wages and 10 per cent from investments identify themselves with economic groupings not on the basis of the 90 per cent of their income but on the basis of 10 per cent of their income. It will be remembered that in Germany the dispossessed middle class still clung to a middle class identification rather than to the ideology of their actual income level.
Professor F. H. Allport's emphasis upon group fictions is of particular relevance here. The attitudes to look for in defining groups are the beliefs in the reality of the concept of the group. Affiliation with political parties is less important than the individual's acceptance of the institutional fiction of the party. Professor Robinson found that on most issues the supporters of Hoover and the supporters of Roosevelt in the 1932 election did not differ greatly. It would be interesting to know whether more clear cut differences might not be found if we compared institutionally-minded Republicans with institutionally-minded Democrats.
In conclusion, the real significance of attitude measurement in social psychology is that it provides a method of getting at the affective and ideological-or if you prefer the subjective-side of man. A social psychology which omits the study of attitudes omits a good share of its rightful problems. Social interaction is highly verbal and symbolic in nature. Men react not only to an objective stimulating situation but in terms of their subjective additions to the situation. So much is this
( 482) true that the long search by reformers for an ideology to end ideologies still marches from defeat to defeat. Even if we do not follow John B. Watson in defining social psychology as the study of attitudes, we must still recognize the importance of this field. Adequate techniques for its study are one of the most pressing problems in social psychology. Attitude scales have been criticized by Professor H. M. Johnson and others from the rigid logic of physical measurement. Nonetheless they constitute the best techniques thus far available. Without them we are left to study the ideologies of men by intuitive methods which cannot be checked for reliability and accuracy, and we become social philosophers rather than social psychologists.
