Attitude Change

The Competing Views

Anton Pelinka, Peter Suedfeld

• Publisher: Routledge
• Year: 2017
• Print ISBN:
  9781138519015
• eBook ISBN:
  9781351315340
• Topic: Psychology
• Subtopic: History & Theory in Psychology
• Index: Psychology

What are attitudes and how are they modified? The many opposing theories to answer this question reflects not only the complexity but also the importance of the field. A central concern of social psychology, attitude change is also relevant to the study of human behavior in general, and a matter of major significance to the world outside the laboratory. Valid and useful theories of attitude change are thus of far-reaching consequence. At the same time, the richness and flexibility of attitude structures and the numerous methodological problems involved in studying them make the development of a definitive theory difficult, if not impossible. For these reasons many explanations have been offered but none have been greatly accepted.

The essays included here give voice to a broad sampling of these competing viewpoints. For years attention has been directed mostly to the individual's need to maintain harmony within him, and several of the authors focus on this concept. Cognitive dissonance theory is evaluated in particular detail. Ideas derived from other areas of psychology and attitude change theories based on learning, perception, and cognitive motivation are also well represented in this volume.

In his introduction, Suedfeld evaluates these major approaches as well as several less well-known alternatives. In weighing the strengths and weaknesses of each system he considers the limits of the applications of the various theories and the problems the theorists face. This book will be welcomed in a variety of undergraduate and graduate courses, and by scholars and educated laymen seeking information on the current state of knowledge in this field.

1 Models of Attitude Change:
Theories That Pass
in the Night

Peter Suedfeld

The inclusion of a book on attitude theory in a series devoted to controversies is highly justified by the state of the field. In any controversy there are at least two opposing views on the subject in question—and a recent survey counted thirty-four distinguishable models of attitude change (Ostrom, 1968)! When the controversy occurs in a field of science, the ultimate outcome depends on which side proves to be closer to empirical fact; the tremendous amount of effort that has gone into attitude research in the past two decades is therefore not surprising.

Theoretical controversies are beneficial to scientific progress to the extent that they lead to critical experiments—experiments that clearly show that one theory predicts events more accurately than another. Platt (1964) describes the best course of development as a series of alternative hypotheses (this is the controversy stage) followed by empirical tests designed so that each possible outcome will support only one hypothesis—and so on to finer and finer predictions.

There are few psychological theories that can be posed against each other so neatly as to produce this logical tree. The complex array of statistical techniques needed by psychologists to distinguish between relevant and irrelevant influences on their data testifies to theoretical imprecision (Platt, 1964) and inadequate experimental control (Skinner, 1956). Attitude theories and experiments share these characteristics, which in a relatively young science, dealing with highly variable phenomena, may be unavoidable and may in fact help to generate novel and increasingly fruitful hypotheses (Hafner and Presswood, 1965).

The problem with attitude theories is that they seldom stand in clear opposition to each other. Some models are so vague that while almost any outcome can be related to the theory in some way, prediction is highly speculative. In such a situation, a critical test is impossible to achieve. Others are so similar to each other that experiments could test only trivial differences, or so dissimilar in scope, language, and approach that one is hard put to identify areas in which their predictions or methods are comparable. Even when such areas are found, and empirical data do exist, one can usually find alternative interpretations that save each of the theories from having to admit inadequacy. That is the reason for the title of this essay: models that can so seldom be pitted against each other conclusively are indeed like ships that pass in the night.¹ This may be (as Charlie Brown once put it) not a fault, but a character trait. Quite probably there is no metatheory that can ever explain every aspect of attitude change. In that case, many theories may have parts of the truth, and each can be utilized in the specific situation for which it is most useful (like the wave and quantum theories, which are used to explain different aspects of the physical nature of light). Such a multiplicity of theories has been proposed for social psychology in general (Deutsch and Krauss, 1965). Even if this is the appropriate solution, however, we must identify the domain in which each theory is most powerful; and for that effort we need more specificity and greater overlap than are available to us now.

One may ask whether, in view of these difficulties, it is worth while to study the theories of attitude change. While there is some disagreement on the value of theorizing per se, the weight of scientific opinion is favorable. Science strives to explain, predict, and control events. Theories are necessary as an explanatory framework, since if there are no explicitly stated conceptual linkages among events, their actual relationship to each other may easily be overlooked. Good empirical research tells us the facts; a good theory tells us what the facts mean. In the same way, theoretical constructs are used to predict the future: once we have an idea of the ways in which antecedent conditions influence the outcome, their pattern can be fitted into a theoretical framework that points to the most likely consequence. If the theory is explicit enough, its strong and weak points can be identified by the accuracy of its predictions. Finally, a good theory defines its components so precisely and is so closely related to experimental operations that the relevant antecedents can be deliberately manipulated to obtain a predicted result. Not until it has reached this stage has the theory attained its full power and utility. The fact that no attitude theory has arrived at this point so far indicates that our understanding of the phenomena of attitude change is at present inadequate; the actual shortcomings should serve to indicate areas in which more research and thinking are needed.

Some Areas of Disagreement

The very definition of “attitude” has been embroiled in controversy. To many workers the concept has three components: cognition, affect, and behavior. Some writers restrict the term to the first two dimensions, and view behavior as an independent dimension that may be but does not have to be a function of attitude. Still others use “attitude” to mean only emotional reactions, and refer to cognitive responses as “beliefs.” These distinctions are by no means trivial. For example, many cigarette smokers admit that smoking is harmful to their health. While they accept this idea intellectually and emotionally, they continue to smoke. If we define behavior as part of attitude, then we see this situation as an attitude with components that are either inconsistent (in which case we are moved to investigate how this internal conflict developed) or differentially relevant to a variety of circumstances (in which case we turn our attention to the determinants of relevance). From the second viewpoint, the attitude is internally consistent; the reason it did not lead to action resides in other attitudinal, personality, and environmental factors.²

Another area of lively disagreement is that of attitude measurement. Problems of scoring and interpreting evaluations and beliefs have vexed numerous investigators. Just what, for example, is the meaning of a response of “zero” or “neutral” on a “plus-minus” or “like-dislike” attitude scale? A number of answers are possible:

1. The subject has no attitude on the issue.
2. The subject is a middle-of-the-roader.
3. The subject is ambivalent.
4. The subject is being reticent or uncooperative.
5. The subject is afraid to answer because of the possible interpretation that might be put on his real feelings if he revealed them.

Quite obviously, these reasons for a neutral response have very different antecedents and consequences; yet no attitude instrument can differentiate adequately among them.

Scaling is the most common method of measuring attitudes, and many of the techniques developed for psychophysics find social application here. Other measures, such as the semantic differential (Osgood, Suci, and Tannenbaum, 1957), were specifically developed for evaluating concepts and symbols. As usual, there are several unresolved issues. An obvious one is whether a given measurement technique is accurate; or rather, how accurate it is in comparison with others. There is the question of how widespread certain response tendencies are (giving “extreme” answers on rating scales, for example), and how personality may be related to such tendencies. Still another problem is the effect on attitude scaling of the difference between the rater’s own position and other positions sampled by the test. How well can a person who takes a strong stand on a subject discriminate among items representing divergent viewpoints?³

Some attitude researchers concentrate on persuasive technology. Whether one-sided or two-sided messages are better, whether the strongest arguments should be presented first or last, whether conclusions should be stated or only implied—these questions will not concern us here. They have been intensively investigated by a group of researchers at Yale University (Hov-land et al., 1957), and their findings as well as those of other investigators were summarized in a how-to book by H. I. Abelson (1959; revised, Karlins and Abelson, 1970). Neither shall we cover the factors of source prestige and credibility, of environments which facilitate the task of the persuader, or of the ways in which susceptibility to the message can be decreased.⁴ When we turn to the major theoretical systems on which this book focuses, it becomes clear that “major” is difficult to define and is to a large extent a matter of personal choice. There is one obvious dichotomy in the literature, between theories that propose that attitude change occurs because of a need for cognitive and emotional consistency, on the one hand, and theories based on other psychological concepts (mostly learning and perception) on the other. The first sort of explanation is the more widely used, and has led to the greater amount of research; four of the eight selections in this book concentrate on consistency theories (two of them on cognitive dissonance, the dominant construct in this group). Other chapters use the learning, perception, and utility models. The research relevant to the theories is covered only briefly.⁵

Consistency Theories

Explications of attitude change often invoke motivational constructs. Theorists assume that attitudes serve some function—to provide cognitive structure, to facilitate the processing of information, to fit the individual into his group, to maintain self-esteem, or whatever. Once an attitude succeeds in gaining the goal for which the individual intended it, it will not change without some reason. The reason must go beyond a mere encounter with new information; the new attitude must be more pleasing or useful in some way than the old. As in other motivational models, there are two alternative emphases: an internal need that is reduced by a response and an external reward that the response may make attainable. The parallel to drive and incentive theories in general is obvious. We would be well advised, however, to suppress the associations that terms such as “drive” are likely to call forth; the accumulated connotations of forty years of traditional experimental psychology should not be transferred blindly into the attitude area.

Motivational theories of attitude change have undoubtedly been the most important propositions offered over the past decade or so, and the experimental literature reflects their popularity. Among these approaches, consistency theories represent a dominant trend. The key assumption of these models is that human beings require their attitudinal systems to be internally consistent, and that a state of inconsistency is (by analogy) non-homeostatic. Attitude change then follows as a way to restore the stable state.

Theorists do not agree unanimously on the motivational basis of consistency. Some workers ignore the issue for various reasons (see Singer, 1966), while others propose a variety of answers to such questions as whether the “consistency motive” is one motive or several, absolute or relative, innate or learned, aimed at eliminating inconsistency or at maintaining some optimal level (McGuire, 1966). These questions need not delay us here, but the lack of agreement—or rather the lack of empirical resolution—is a serious flaw in the consistency literature.

In the paper reprinted as Chapter 2 of this book, Zajonc (1960) describes three important versions of consistency theory. He also discusses research based on these formulations, and therefore I shall not treat it in detail here.

Balance Theory

Balance theory, as it was originated by Heider (1946, 1958), analyzes systems consisting of two or three persons (or two persons and an object), with any two of these entities related to each other by either positive or negative sentiments, or not related at all. The model is most relevant when belongingness (the unit relation) does exist; then, the sentiment relations are balanced if they provide a good Gestalt—if a liked person agrees and a disliked person disagrees with one’s own sentiments toward the third entity. Newcomb (1953) enlarged the model to include communicative acts as linkages; Cartwright and Harary (1956) mathematicized it, clearing up some ambiguities and extending it to systems of more than three entities. Feather (1967) built further upon Cartwright and Harary’s improvements in a communication context.

While the basic idea was persuasive, the theory was vague and incomplete. For example, the three-entity relationship did not take into account the fact that liking and disliking usually flow in two directions (with only two people involved, mutual liking or disliking represents balance). Though in many situations there are both positive and negative feelings toward a person or object, and though affects vary greatly in intensity (both “like” and “adore” are positive, but they are hardly equivalent), there was no way to assess the relative or absolute strength of sentiments. Furthermore, as Insko (1967) points out, experimenters have inferred imbalance when a subject rates himself as feeling uneasy, tense, or unpleasant on a scale from pleasant to unpleasant; but a stable-unstable rating of the relationship would seem more relevant to the theory, and certainly to attitude change. Unfortunately, such ratings of stability may not be very reliable, and (a major flaw) there are no unequivocal behavioral measures of imbalance. But other consistency theories have not even tried to develop an independent measure of the hypothetical state; honor is due at least for the recognition of the problem.

Another problem is that the unit relationship introduces a high degree of vagueness. For one thing, “belonging together” can be inferred from many relationships: ownership, causation, proximity, responsibility, membership, and so on. If an experimental finding can be explained by a bond that unites two entities, the experimenter can surely find a definition of unity that is plausible in the particular situation. And there is another problem: while things may either belong together or not, there seems to be no equivalent of like-not like-dislike. The best candidate seems to be “belong apart,” and the effects of such a relationship on attitudes need to be investigated.

Insko (1967, pp. 162-63) cites some proverbs that contradict balance theory (“Opposites attract”; “Familiarity breeds contempt”) and gives Heider’s resolution of these dilemmas (for example, dissimilarity can result in liking if the characteristics of the two people involved complement each other). Other theories could offer less tortuous explanations; as an exercise, the reader might work out a balance interpretation of the Arab saying “The enemy of my enemy is my friend” as applied to the average American’s strong dislike of both Nazis and Communists, who are mutual enemies. This is a losing game: if it could not be done, the exercise would show a striking exception to a basic aspect of the theory (which supports the proverb). But it can (by emphasizing the Nazis’ and Communists’ joint hostility to the United States). The room for semantic maneuvering is obvious, and the consequent loss of precision greatly weakens the theory.

While Cartwright and Harary have resolved a few of these problems, most of them remain. In short, balance theory is relevant to a great number of areas—among them friendship development (Newcomb, 1961), conformity (Brown, 1965, pp. 673-77), and reactions to criticism (Pilisuk, 1962)—but it explains the past much better than it predicts the future.

Cognitive And Affective Consistency

A much more sophisticated version of balance theory was developed by M. J. Rosenberg and R. P. Abelson. An early version of the model emphasizes consistency between attitudes (affect) and beliefs (cognition) toward a given object (Rosenberg, 1956, 1960a). Objects are liked to the extent that they are believed to be helpful or harmful to the individual’s other significant attitude objects. If an object is positively valued and is seen as advancing other preferred objects (if an integrationist Democrat, for example, believes that his party is working for racial equality), the structure is consistent; but if the same liked object is believed to be hindering the attainment of other such objects (if our Democrat looks at some southern Democratic congressmen), there is inconsistency. And, according to the homeostatic model, when inconsistency exceeds the individual’s tolerance limits, change occurs to restore consistency. In a similar vein, McGuire (1960) has analyzed Socratic questioning to show how it makes the individual aware of logical inconsistency among his beliefs, with a consequent striving for consistency.

In another version (Abelson and Rosenberg, 1958; Rosenberg and Abelson, 1960), cognitive elements (mental representations) are affectively related to ea...

Table of contents

• Cover Page
• Halftitle Page
• Title Page
• Copyright Page
• Content Page
• 1 Models of Attitude Change: Theories That Pass in the Night
• 2 The Concepts of Balance, Congruity, and Dissonance
• 3 Symbolic Psycho-logic: A Model of Attitudinal Cognition
• 4 Cognitive Dissonance: Five Years Later
• 5 In Defense of Dissonance Theory: Reply to Chapanis and Chapanis
• 6 A Reinforcement Learning Model of Persuasive Communication
• 7 An Experimental Analysis of Self-persuasion
• 8 Processes of Opinion Change
• 9 An Uncertainty Model of Attitude Change
• Index