Processes of Opinion Change [Foundations of Communication Theory]

Home | Audio Magazine | Stereo Review magazine | Good Sound | Troubleshooting






THE STUDY OF SOCIAL INFLUENCE

Social influence has been a central area of concern for experimental social psychology almost since its beginnings. Three general research traditions in this area can be distinguished:

(1) the study of social influences on judgments, stemming from the earlier work on prestige suggestion; (2) the study of social influences arising from small-group interaction; 2 and (3) the study of social influences arising from persuasive communications. 3 In recent years, there has been a considerable convergence between these three traditions, going hand in hand with an increased interest in developing general principles of social influence and socially induced behavior change.

From Herbert C. Kelman, "Processes of Opinion Change," Public Opinion Quarterly, 1961, 25, 57-78. Reproduced with permission of the author and publisher.

--------------------

1. See, for example, S. E. Asch, Social Psychology, New York, Prentice-Hall, 1952.

2. See, for example, D. Cartwright and A. Zander, editors, Group Dynamics, Evanston, Ill., Row, Peterson, 1953.

3. See, for example, C. I. Hovland, I. L. Janis, and H. H. Kelley, Communication and Persuasion, New Haven, Yale University Press, 1953.

------------------

One result of these developments has been that many investigators found it necessary to make qualitative distinctions between different types of influence. In some cases, these distinctions arose primarily out of the observation that social influence may have qualitatively different effects, that it may produce different kinds of change. For example, under some conditions it may result in mere public conformity-in superficial changes on a verbal or overt level without ac companying changes in belief; in other situations it may result in private acceptance--in a change that is more general, more durable, more integrated with the person's own values. Other investigators found it necessary to make distinctions because they observed that influence may occur for different reasons, that it may arise out of different motivations and orientations. For example, under some conditions influence may be primarily informational-the subject may conform to the influencing person or group because he views him as a source of valid in formation; in other situations influence may be primarily normative-the subject may conform in order to meet the positive expectations of the influencing person or group. My own work can be viewed in the general context that I have outlined here. I started out with the distinction between public conformity and private acceptance, and tried to establish some of the distinct determinants of each. I became dissatisfied with this dichotomy as I began to look at important examples of social influence that could not be encompassed by it. I was especially impressed with the accounts of ideological conversion of the "true believer" variety, and with the recent accounts of "brainwashing," particularly the Chinese Communist methods of "thought reform." It is apparent that these experiences do not simply involve public conformity, but that indeed they produce a change in underlying beliefs. But it is equally apparent that they do not produce what we would usually consider private acceptance--changes that are in some sense integrated with the person's own value system and that have become independent of the external source. Rather, they seem to produce new beliefs that are isolated

------------

4. See, for example, L Festinger, "An Analysis of Compliant Behavior," in M. Sherif and M. O. Wilson, editors, Group Relations at the Crossroads, New York, Harper, 1953, pp. 232-256; H. C. Kelman, "Attitude Change as a Function of Response Restriction," Human Relations, Vol. 6, 1953, pp. 185-214; J. R. P. French, Jr., and B. Raven, "The Bases of Social Power," in D. Cartwright, editor, Studies in Social Power, Ann Arbor, Mich., Institute for Social Research, 1959, pp. 150-167; and Marie Jahoda, "Conformity and Independence," Human Relations, Vol. 12, 1959, pp. 99-120.

5. See, for example, M. Deutsch and H. B. Gerard, "A Study of Normative and Informational Social Influence upon Individual Judgment," Journal of Abnormal and Social Psychology, Vol. 51, 1955, pp. 629-636; J. W. Thibaut and L Strickland, "Psychological Set and Social Conformity," Journal of Personality, Vol. 25, 1956, pp. 115-129; and J. M. Jackson and H. D. Saltzstein, "The Effect of Person-Group Relationships on Conformity Processes," Journal of Abnormal and Social Psychology, Vol. 57, 1958, pp. 17-24.

6. For instance, R. J. Lifton, "'Thought Reform' of Western Civilians in Chinese Communist Prisons," Psychiatry, Vol. 19, 1956, pp. 173-195.

------------------

from the rest of the person's values and that are highly dependent on external support.

These considerations eventually led me to distinguish three processes of social influence, each characterized by a distinct set of antecedent and a distinct set of consequent conditions. I have called these processes compliance, identification, and internalization.7

THREE PROCESSES OF SOCIAL INFLUENCE

Compliance can be said to occur when an individual accepts influence from an other person or from a group because he hopes to achieve a favorable reaction from the other. He may be interested in attaining certain specific rewards or in avoiding certain specific punishments that the influencing agent controls. For ex ample, an individual may make a special effort to express only "correct" opinions in order to gain admission into a particular group or social set, or in order to avoid being fired from his government job. Or, the individual may be concerned with gaining approval or avoiding disapproval from the influencing agent in a more general way. For example, some individuals may compulsively try to say the expected thing in all situations and please everyone with whom they come in contact, out of a disproportionate need for favorable responses from others of a direct and immediate kind. In any event, when the individual complies, he does what the agent wants him to do-or what he thinks the agent wants him to do-because he sees this as a way of achieving a desired response from him. He does not adopt the induced behavior-for example, a particular opinion response--because he believes in its content, but because it is instrumental in the production of a satisfying social effect. What the individual learns, essentially, is to say or do the expected thing in special situations, regardless of what his private beliefs may be. Opinions adopted through compliance should be expressed only when the person's behavior is observable by the influencing agent.

Identification can be said to occur when an individual adopts behavior de rived from another person or a group because this behavior is associated with a satisfying self-defining relationship to this person or group. By a self-defining relationship I mean a role relationship that forms a part of the person's self-image.

-------------------

7 A detailed description of these processes and the experimental work based on them will be contained in a forthcoming book, Social Influence and Personal Belief: A Theoretical and Experimental Approach to the Study of Behavior Change, to be published by John Wiley & Sons.

-------------------

Accepting influence through identification, then, is a way of establishing or maintaining the desired relationship to the other, and the self-definition that is anchored in this relationship.

The relationship that an individual tries to establish or maintain through identification may take different forms. It may take the form of classical identification, that is, of a relationship in which the individual takes over all or part of the role of the influencing agent. To the extent to which such a relationship exists, the individual defines his own role in terms of the role of the other. He at tempts to be like or actually to be the other person. By saying what the other says, doing what he does, believing what he believes, the individual maintains this relationship and the satisfying self-definition that it provides him. An influencing agent who is likely to be an attractive object for such a relationship is one who occupies a role desired by the individual-who possesses those characteristics that the individual himself lacks-such as control in a situation in which the individual is helpless, direction in a situation in which he is disoriented, or belongingness in a situation in which he is isolated.

The behavior of the brainwashed prisoner in Communist China provides one example of this type of identification. By adopting the attitudes and beliefs of the prison authorities-including their evaluation of him-he attempts to regain his identity, which has been subjected to severe threats. But this kind of identification does not occur only in such severe crisis situations. It can also be observed, for example, in the context of socialization of children, where the taking over of parental attitudes and actions is a normal, and probably essential, part of personality development. The more or less conscious efforts involved when an individual learns to play a desired occupational role and imitates an appropriate role model would also exemplify this process. Here, of course, the individual is much more selective in the attitudes and actions he takes over from the other person.

What is at stake is not his basic sense of identity or the stability of his self-concept, but rather his more limited "professional identity." The self-defining relationship that an individual tries to establish or maintain through identification may also take the form of a reciprocal role relationship--that is, of a relationship in which the roles of the two parties are defined with reference to one another. An individual may be involved in a reciprocal relationship with another specific individual, as in a friendship relationship be tween two people. Or he may enact a social role which is defined with reference to another (reciprocal) role, as in the relationship between patient and doctor. A reciprocal-role relationship can be maintained only if the participants have mutually shared expectations of one another's behavior. Thus, if an individual finds a particular relationship satisfying, he will tend to behave in such a way as to meet the expectations of the other. In other words, he will tend to behave in line with the requirements of this particular relationship. This should be true regard less of whether the other is watching or not: quite apart from the reactions of the other, it is important to the individual's own self-concept to meet the expectations of his friendship role, for example, or those of his occupational role.

Thus, the acceptance of influence through identification should take place when the person sees the induced behavior as relevant to and required by a reciprocal-role relationship in which he is a participant. Acceptance of influence based on a reciprocal-role relationship is similar to that involved in classical identification in that it is a way of establishing or maintaining a satisfying self defining relationship to another. The nature of the relationship differs, of course.

In one case it is a relationship of identity; in the other, one of reciprocity. In the case of reciprocal-role relationships, the individual is not identifying with the other in the sense of taking over his identity, but in the sense of empathically reacting in terms of the other person's expectations, feelings, or needs.

Identification may also serve to maintain an individual's relationship to a group in which his self-definition is anchored. Such a relationship may have elements of classical identification as well as of reciprocal roles: to maintain his self-definition as a group member an individual, typically, has to model his behavior along particular lines and has to meet the expectations of his fellow members. An example of identification with a group would be the member of the Communist Party who derives strength and a sense of identity from his self-definition as part of the vanguard of the proletarian revolution and as an agent of historical destiny. A similar process, but at a low degree of intensity, is probably involved in many of the conventions that people acquire as part of their socialization into a particular group.

Identification is similar to compliance in that the individual does not adopt the induced behavior because its content per se is intrinsically satisfying. Identification differs from compliance, however, in that the individual actually believes in the opinions and actions that he adopts. The behavior is accepted both publicly and privately, and its manifestation does not depend on observability by the influencing agent. It does depend, however, on the role that an individual takes at any given moment in time. Only when the appropriate role is activated-only when the individual is acting within the relationship upon which the identification is based-will the induced opinions be expressed. The individual is not primarily concerned with pleasing the other, with giving him what he wants (as in compliance), but he is concerned with meeting the other's expectations for his own role performance. Thus, opinions adopted through identification do remain tied to the external source and dependent on social support. They are not integrated with the individual's value system, but rather tend to be isolated from the rest of his values--to remain encapsulated.

Finally, internalization can be said to occur when an individual accepts in fluence because the induced behavior is congruent with his value system. It is the content of the induced behavior that is intrinsically rewarding here. The individual adopts it because he finds it useful for the solution of a problem, or because it is congenial to his own orientation, or because it is demanded by his own values-in short, because he perceives it as inherently conducive to the maximization of his values. The characteristics of the influencing agent do play an important role in internalization, but the crucial dimension here--as we shall see below--is the agent's credibility, that is, his relation to the content.

The most obvious examples of internalization are those that involve the evaluation and acceptance of induced behavior on rational grounds. A person may adopt the recommendations of an expert, for example, because he finds them relevant to his own problems and congruent with his own values. Typically, when internalization is involved, he will not accept these recommendations in loto but modify them to some degree so that they will fit his own unique situation. Or a visitor to a foreign country may be challenged by the different patterns of behavior to which he is exposed, and he may decide to adopt them (again, selectively and in modified form) because he finds them more in keeping with his own values than the patterns in his home country. I am not implying, of course, that internalization is always involved in the situations mentioned. One would speak of internalization only if acceptance of influence took the particular form that I described.

Internalization, however, does not necessarily involve the adoption of induced behavior on rational grounds. I would not want to equate internalization with rationality, even though the description of the process has decidedly rationalist overtones. For example, I would characterize as internalization the adoption of beliefs because of their congruence with a value system that is basically irrational. Thus, an authoritarian individual may adopt certain racist attitudes be cause they fit into his paranoid, irrational view of the world. Presumably, what is involved here is internalization, since it is the content of the induced behavior and its relation to the person's value system that is satisfying. Similarly, it should be noted that congruence with a person's value system does not necessarily imply logical consistency. Behavior would be congruent if, in some way or other, it fit into the person's value system, if it seemed to belong there and be demanded by it.

It follows from this conception that behavior adopted through internalization is in some way-rational or otherwise-integrated with the individual's existing values. It becomes part of a personal system, as distinguished from a sys tem of social-role expectations. Such behavior gradually becomes independent of the external source. Its manifestation depends neither on observability by the influencing agent nor on the activation of the relevant role, but on the extent to which the underlying values have been made relevant by the issues under consideration. This does not mean that the individual will invariably express internalized opinions, regardless of the social situation. In any specific situation, he has to choose among competing values in the face of a variety of situational requirements. It does mean, however, that these opinions will at least enter into competition with other alternatives whenever they are relevant in content.

It should be stressed that the three processes are not mutually exclusive. While they have been defined in terms of pure cases, they do not generally occur in pure form in real-life situations. The examples that have been given are, at best, situations in which a particular process predominates and determines the central features of the interaction.

---------------

SUMMARY OF THE DISTINCTIONS BETWEEN THE THREE PROCESSES

Compliance Identification Internalization

Antecedents:

1. Basis for the importance of the induction

2. Source of power of the influencing agent

3. Manner of achieving pre-potency of the induced response Consequents:

1. Conditions of performance of induced response

2. Conditions of change and extinction of induced response

3. Type of behavior system in which induced response is embedded

Concern with social effect of behavior

Means control Limitation of choice behavior Surveillance by influencing agent Changed perception of conditions for social rewards External demands of a specific setting Concern with social anchorage of behavior

Attractiveness Delineation of role requirements

Salience of relationship to agent

Changed perception of conditions for satisfying self-defining relationships

Expectations defining a specific role

Concern with value congruence of behavior

Credibility

Reorganization of means-ends framework

Relevance of values to issue

Changed perception of conditions for value maximization

Person's value system

---------------

ANTECEDENTS AND CONSEQUENTS OF THE THREE PROCESSES

For each of the three processes, a distinct set of antecedents and a distinct set of consequents have been proposed. These are summarized in the table below. First, with respect to the antecedents of the three processes, it should be noted that no systematic quantitative differences between them are hypothesized. The probability of each process is presented as a function of the same three determinants: the importance of the induction for the individual's goal achievement, the power of the influencing agent, and the prepotency of the induced response. For each pro cess, the magnitude of these determinants may vary over the entire range: each may be based on an induction with varying degrees of importance, on an influencing agent with varying degrees of power, and so on. The processes differ only in terms of the qualitative form that these determinants take. They differ, as can be seen in the table, in terms of the basis for the importance of the induction, the source of the influencing agent's power, and the manner of achieving prepotency of the induced response.

1. The processes can be distinguished in terms of the basis for the importance of the induction, that is, in terms of the nature of the motivational system that is activated in the influence situation. What is it about the influence situation that makes it important, that makes it relevant to the individual's goals? What are the primary concerns that the individual brings to the situation or that are aroused by it? The differences between the three processes in this respect are implicit in the descriptions of the processes given above: (a) To the extent that the individual is concerned-for whatever reason-with the social effect of his behavior, influence will tend to take the form of compliance. (b) To the extent that he is concerned with the social anchorage of his behavior, influence will tend to take the form of identification. (c) To the extent that he is concerned with the value congruence of his behavior (rational or otherwise), influence will tend to take the form of internalization.

2. A difference between the three processes in terms of the source of the influencing agent's power is hypothesized. (a) To the extent that the agent's power is based on his means control, influence will tend to take the form of compliance.

An agent possesses means control if he is in a position to supply or withhold means needed by the individual for the achievement of his goals. The perception of means control may depend on the agent's actual control over specific rewards and punishments, or on his potential control, which would be related to his position in the social structure (his status, authority, or general prestige). (b) To the extent that the agent's power is based on his attractiveness, influence will tend to take the form of identification. An agent is attractive if he occupies a role which the individual himself desires 8 or if he occupies a role reciprocal to one the individual wants to establish or maintain. The term "attractiveness," as used here, does not refer to the possession of qualities that make a person likable, but rather to the possession of qualities on the part of the agent that make a continued relationship to him particularly desirable. In other words, an agent is attractive when the individual is able to derive satisfaction from a self-definition with reference to him. (c) To the extent that the agent's power is based on his credibility, influence will tend to take the form of internalization. An agent possesses credibility if his statements are considered truthful and valid, and hence worthy of serious consideration. Hovland, Janis, and Kelley distinguish two bases for credibility: expertness and trustworthiness. In other words, an agent may be perceived as possessing credibility because he is likely to know the truth, or because he is likely to tell the truth. Trustworthiness, in turn, may be related to over-all respect, like-mindedness, and lack of vested interest.

---------------

8 This is similar to John Whiting's conception of "Status Envy" as a basis for identification. See J. W. M. Whiting, "Sorcery, Sin, and the Superego," in M. R. Jones, editor, Nebraska Symposium on Motivation, Lincoln, University of Nebraska Press, 1959, pp. 174 195.

---------------------

3. It is proposed that the three processes differ in terms of the way in which prepotency is achieved. (a) To the extent that the induced response becomes pre potent--that is, becomes a "distinguished path" relative to alternative response possibilities--because the individual's choice behavior is limited, influence will tend to take the form of compliance. This may happen if the individual is pressured into the induced response, or if alternative responses are blocked. The induced response thus becomes pre-potent because it is, essentially, the only response permitted: the individual sees himself as having no choice and as being restricted to this particular alternative. (b) To the extent that the induced response becomes pre-potent because the requirements of a particular role are delineated, influence will tend to take the form of identification. This may happen if the situation is defined in terms of a particular role relationship and the demands of that role are more or less clearly specified; for instance, if this role is made especially salient and the expectations deriving from it dominate the field. Or it may happen if alternative roles are made ineffective because the situation is ambiguous and consensual validation is lacking. The induced response thus becomes pre-potent because it is one of the few alternatives available to the individual: his choice behavior may be unrestricted, but his opportunity for selecting alternative responses is limited by the fact that he is operating exclusively from the point of view of a particular role system. (c) Finally, to the extent that the induced response becomes pre-potent because there has been a reorganization in the individual's conception of means-ends relationships, influence will tend to take the form of internalization. This may happen if the implications of the induced response for certain important values-implications of which the individual had been un aware heretofore-are brought out, or if the advantages of the induced response as a path to the individual's goals, compared to the various alternatives that are available, are made apparent. The induced response thus becomes pre-ponent be cause it has taken on a new meaning: as the relationships between various means and ends become restructured, it emerges as the preferred course of action in terms of the person's own values.

Depending, then, on the nature of these three antecedents, the influence pro cess will take the form of compliance, identification, or internalization. Each of these corresponds to a characteristic pattern of internal responses--thoughts and feelings--in which the individual engages as he accepts influence. The resulting changes will, in turn, be different for the three processes, as indicated in the second half of the table. Here, again, it is assumed that there are no systematic quantitative differences between the processes, but rather qualitative variations in the subsequent histories of behavior adopted through each process.

1. It is proposed that the processes differ in terms of the subsequent conditions under which the induced response will be performed or expressed. (a)

----------

9. op. cit., p. 21.

-------------

When an individual adopts an induced response through compliance, he tends to perform it only under conditions of surveillance by the influencing agent. These conditions are met if the agent is physically present, or if he is likely to find out about the individual's actions. (b) When an individual adopts an induced response through identification, he tends to perform it only under conditions of salience of his relationship to the agent. That is, the occurrence of the behavior will depend on the extent to which the person's relationship to the agent has been engaged in the situation. Somehow this relationship has to be brought into focus and the individual has to be acting within the particular role that is involved in the identification. This does not necessarily mean, however, that he is consciously aware of the relationship; the role can be activated without such awareness. (c) When an individual adopts an induced response through internalization, he tends to perform it under conditions of relevance of the values that were initially involved in the influence situation. The behavior will tend to occur whenever these values are activated by the issues under consideration in a given situation, quite regardless of surveillance or salience of the influencing agent. This does not mean, of course, that the behavior will occur every time it becomes relevant. It may be out-competed by other responses in certain situations. The probability of occurrence with a given degree of issue relevance will depend on the strength of the internalized behavior.

2. It is hypothesized that responses adopted through the three processes will differ in terms of the conditions under which they will subsequently be abandoned or changed.

(a) A response adopted through compliance will be abandoned if it is no longer perceived as the best path toward the attainment of social rewards. (b) A response adopted through identification will be abandoned if it is no longer perceived as the best path toward the maintenance or establishment of satisfying self-defining relationships. (c) A response adopted through internalization will be abandoned if it is no longer perceived as the best path toward the maximization of the individual's values.

3. Finally, it is hypothesized that responses adopted through the three processes will differ from each other along certain qualitative dimensions. These can best be summarized, perhaps, by referring to the type of behavior system in which the induced response is embedded.

(a) Behavior adopted through compliance is part of a system of external demands that characterize a specific setting.

In other words, it is part of the rules of conduct that an individual learns in order to get along in a particular situation or series of situations. The behavior tends to be related to the person's values only in an instrumental rather than an intrinsic way. As long as opinions, for example, remain at that level, the individual will tend to regard them as not really representative of his true beliefs.

(b) Behavior adopted through identification is part of a system of expectations defining a particular role-whether this is the role of the other which he is taking over, or a role reciprocal to the other's. This behavior will be regarded by the person as representing himself, and may in fact form an important aspect of himself. It will tend to be isolated, however, from the rest of the person's values and to have little interplay with them. In extreme cases, the system in which the induced response is embedded may be encapsulated and function almost like a foreign body within the person. The induced responses here will be relatively in flexible and stereotyped.

(c) Behavior adopted through internalization is part of an internal system. It is fitted into the person's basic framework of values and is congruent with it. This does not imply complete consistency: the degree of consistency can vary for different individuals and different areas of behavior. It does mean, however, that there is some interplay between the new beliefs and the rest of the person's values. The new behavior can serve to modify existing beliefs and can in turn be modified by them. As a result of this interaction, behavior adopted through internalization will tend to be relatively idiosyncratic, flexible, complex, and differentiated.


Also in Part 4:


Prev. | Next

Top of Page  | Index |  Home

Updated: Monday, 2021-08-30 13:00 PST