Advances in Consumer Research Volume 5, 1978 Pages 360-369
A FUNCTIONAL APPROACH TO CONSUMER ATTITUDE RESEARCH
Richard J. Lutz, University of California, Los Angeles
[Associate Professor of Marketing, UCLA. The present paper is an excerpt from a more comprehensive version (Lutz, 1977) which may be requested from the author.]
The potentials and current limitations of the so-called functional approach to the study of attitudes are delineated. In order to more fully illuminate the functional approach, the evolution of its research tradition is traced. The core of the functional approach, it is argued, can be captured by an expectancy-times-value reconceptualization of the theory. Implications for research and practice under the new conceptualization are discussed.
It has been over fifteen years since Daniel Katz published his classic article, "The Functional Approach to the Study of Attitudes" (Katz, 1960). In this paper, which was essentially the last in a series of studies espousing the functional approach, Katz outlined the basic notion of the approach--i.e., that people hold attitudes toward objects, events, issues and behaviors for various reasons. That is, attitudes fulfill functions for the individual, such as maximizing rewards, expressing one's values, etc. He went on to specify four basic functions that an attitude may perform for an individual--utilitarian, value-expressive, ego-defensive and knowledge--and presented a summary of conditions under which the various functions might be aroused and another set of conditions conducive to the change of attitudes serving the four functions.
Katz's article was a major conceptual contribution to the attitude literature in that it made explicit the motivational underpinnings of attitudes and drew attention to the possible individual differences (in attitude function) which may be obscured by a simple affective representation of attitude. Consumer behavior scholars have been quite cognizant of Katz's contribution, as evidenced by the frequency with which the functional approach is discussed in consumer behavior texts. Virtually every major book in the field includes some treatment of the functional approach, and generally concludes that a functional approach is fundamental to understanding (1) why people hold the attitudes they do, and (2) how those attitudes might be changed.
Unfortunately, consumer researchers, and attitude researchers in general, have not shown as much interest in actually applying the functional approach. In fact, recently the functional approach has come under attack by some researchers who cite the lack of operational procedures to allow its application (Kiesler, Collins and Miller, 1969; Day, 1973; Eagly and Himmelfarb, 1974). The currently fashionable position on the functional approach, therefore, is that it is" .... more important as a classificatory schema and general analytical framework than as a theory generating hypotheses for research" (Eagly and Himmelfarb, 1974. p. 598).
The purpose of the present paper is to develop a rationale for the measurement of attitude functions. To accomplish this, a review of the historical development of the functional approach will be undertaken, followed by a proposed reconceptualization. Implications of the reconceptualized functional framework for consumer attitude research will also be discussed.
EVOLUTION OF THE FUNCTIONAL APPROACH
The functional approach to attitudes subsumes two independent schools of thought, one fostered by M. Brewster Smith and his colleagues at Harvard, and the other represented by Katz, Sarnoff and others at the University of Michigan. Though developed independently, the two approaches are remarkably similar. Each will be reviewed briefly here, in order to highlight their similarities and differences, and to provide a flavor of the functional approach.
The Harvard Study
At the close of World War II, the American people were confronted with the emergence of Russia as a major world power. Communism was a dreaded menace, and the concern among the public was great. Smith and his colleagues set out to study the nature of attitudes toward Russia and, in particular, the reasons underlying these attitudes.
Smith (1947), in the first explicit statement of functional theory, posited a relationship between attitudes and "personality." Based on the results of a survey of 250 New England males, Smith identified five functions which attitudes could serve for the personality:
1) the value function, wherein the attitude expresses a value of the individual;
2) the consistency function, which maintains the relationship between attitudes and characteristic modes of reaction (e.g., personality traits);
3) the gratification function, whereby the attitude indirectly satisfies some basic need;
4) the meaning function, which allows the person to give order and stability to his world; and
5) the conformity function, which is served by attitudes expressing identity with, and fostering acceptance by, important social groups.
The essential thrust of Smith's functional formulation, then, rested on the postulate that attitudinal affect is in some way based on the relationship of the attitude object to underlying psychological constructs variously labeled values, motives and needs. The partitioning into five distinct functions is perhaps best seen as an attempt to specify some qualitative differences among the underlying needs. For example, the meaning function was seen as relating to the needs for "adequacy" and "simplicity," while the conformity function centered on "need for approval'' and "need for acceptance." The value function dealt with "what sorts of things people think are important in life" (Smith, 1947, p. 518) such as liberty and economic security, while the consistency and gratification functions analyzed the relationship between attitude and traditional personality variables (e.g., extrapunitiveness).
While Smith (1947) did not go into any detail regarding the mechanisms through which the various functions operate, it seems clear that the indirect satisfaction of needs as specified by the gratification function would differ qualitatively from more direct mechanisms. Thus, there would appear to be not only differences in the content of needs across the five functions, but also qualitative differences in the psychological mechanisms implied by the functions.
In a subsequent paper, Smith amplified on the nature of the value function:
A person will tend to perceive and judge the focus of an attitude in terms of one of his personal values to the extent that (a) the value is important to him, occupying a central position in his value hierarchy; (b) the information available to him about the focus contains a basis for engaging the value; and (c) the scope of the value and of the person's interest is broad enough to extend to the focus of the attitude. (Smith, 1949, p. 486).
Thus, although the other functions were not similarly analyzed, Smith did engage in theorizing with respect to the conditions necessary (and presumably sufficient) for the value function to be an underlying mechanism in support of an attitude. This represents at least a partial rebuttal to those who contend that functional theory is not really "theory," but rather is little more than simply a taxonomy of attitude-related needs and values (Eagly and Himmelfarb, 1974; Baron, 1968).
In 1956 Smith, Bruner and White authored Opinions and Personality, which reported on the entire study of attitudes toward Russia, but focused primarily on the results of in-depth interviews with 10 men. From this study emerged a modification of Smith's earlier functional theorizing, together with an elaboration of theoretical mechanisms embodied by the functional approach.
Rather than the five functions postulated by Smith (1947), Smith et al. (1956), conceived of three basic functions: object appraisal, social adjustment and externalization. The latter function is closely related to Smith's earlier gratification function, while social adjustment corresponds to his conformity function. The object appraisal function has no direct analogue in Smith's earlier work, although it does appear to include his meaning function. Oddly, the value function, which was so central in Smith's earlier work, seemingly is ignored in the Smith, et al. (1956) statement.
The basis for the object appraisal function is "reality testing," in that an attitude serving the function is seen as classifying the attitude object for the purposes of allowing the individual to adequately cope with the object. Stated differently, "...object appraisal is the process whereby the person develops attitudes that are a creative solution to the problems posed by the existence of disparate internal demands and external or environmental demands" (Smith, et al., 1956, p. 41). Again emphasizing the basic need-relatedness of attitudes, Smith, et al. further specify that objects are appraised in terms of their relations to the person's "motives, goals, values and interests" (p. 261). But they are careful to point out that they are not advocating "...that admirable fiction, the 'rational' man," (p. 265), but rather are postulating object appraisal as only one of the functions attitudes may serve.
Smith, et al.'s social adjustment function deals with "...facilitating, disrupting or simply maintaining an individual's relations with other individuals (p. 41). Such processes as identification, conformity, which represents a broadening of Smith's (1947) conformity function.
Finally, the externalization function is operating when "...an individual, often responding unconsciously, senses an analogy between a perceived environmental event and some unresolved inner problem" (Smith, et al., 1956, p. 43). This externalization helps the individual to reduce anxiety. Like the other functions the degree and importance of externalization are expected to differ greatly across both individuals and situations.
Turning to an analysis of attitude change implications of the functional approach, Smith et al. state that while any one function can be a "precipitating factor" in attitude change, the process of change must entail a "...shift in the balance of all factors" (p. 276). This implicitly assumes a weighting of the various functions in their relationship to attitude, and this weighting has important implications for the form that attitude change strategies might take. For instance, a predominantly object appraisal attitude would most likely be modified by the presentation of information, while an attitude serving the social adjustment function ought to be more susceptible to "prestige suggestion" (Smith, et al., 1956, p. 277). Attitudes which are primarily based on externalization are expected to be more rigid and not subject to modification through either information or social manipulation (p. 277). As Smith, et al. conclude:
If this analysis of the conditions of attitude change is correct...it becomes important to devise rigorous and objective means of assessing the contribution of the different functional supports to people's opinions on given issues. Ways of doing so seem entirely within the range of feasible attainment." (p. 278)
But Smith, Bruner and White did not attempt to develop the measures they specify, leaving these important attitude change prepositions untested and largely untestable. This lack of operational procedures remains the single most glaring weakness of functional theory, not only for Smith, et al., but also for the work of Katz and his associates at Michigan.
The Michigan Group
Just as the Harvard Study had centered on an important social phenomenon--attitudes toward Russia--the University of Michigan group focused its work on the problem of understanding and changing prejudicial attitudes. The earlier work of Adorno, et al. (1950) on the authoritarian personality and its relationship to anti-Semitism had a significant impact on the Michigan group. However, rather than concentrate solely on the personality antecedents of prejudice, the Michigan group broadened its study to include a number of motivational bases supportive of attitude formation and change.
In the first paper in the series Sarnoff and Katz (1954) outlined three "motivational contexts" for attitudes corresponding roughly to the major psychological research traditions of Gestalt theory, learning theory, and psychoanalytic theory. These three contexts were labeled reality testing, reward and punishment, and ego-defense, respectively.
Reality testing was described in terms of the need to explore, the need to know, the quest for information, curiosity and the like. In what today might be described as an information processing view, Sarnoff and Katz (1954) stated that attitudes serving this function are "...a function of the range of information which has been accessible to the individual in regard to certain target objects" (p. 117). Further, they specified that such attitudes could be changed via a "rational" approach involving the presentation of new information within an appropriate frame of reference.
Attitudes based on reward and punishment were seen as an amalgam of two basic sorts of rewards: (1) those based on social group norms, and (2) those emanating from the individual's internal value system. Sarnoff and Katz were more interested in the former type (because of the presumed social nature of prejudice) and therefore emphasized the use of normative group pressure in changing such attitudes. In addition, however, they noted that restructuring the value system "...through making the specific connections between the target object and a different set of values" (p. 122) was also a viable modification strategy.
Finally, attitudes based on ego-defense were viewed as serving "consciously unacceptable" motives through such ego-defense mechanisms as projection. Ego-defensive attitudes were regarded by Sarnoff and Katz as "inappropriate,'' in that the cognitive object...perceived contrasts with characteristics of the object" (p. 120). These distortions could only be corrected through attitude change procedures involving some form of catharsis, which would allow the individual to perceive and respond to reality in a more "appropriate" manner.
Similar to Smith, et al. (1956), Sarnoff and Katz (1954) noted that any given attitude may serve one or a combination of functions. They concluded their paper with a discussion of attitude change "programs" designed to modify attitudes serving the various functions. While they reported no empirical work, a number of studies followed in the next several years which were aimed at investigating some of Sarnoff and Katz' ideas.
Cohen, Stotland and Wolfe (1955) investigated the need for cognition (n-cog) as ..."the individual's tendency to organize his experience meaningfully'' (p. 291). This corresponds directly to Sarnoff and Katz' (1954) reality testing function. Using two independent measures of n-cog, they were able to demonstrate that students' feelings of "frustration" were greater when reading "ambiguous" stories than when reading unambiguous stories, and further, that the increment in frustration was greater for students who were high in n-cog. While they failed to support two hypotheses regarding the degree to which the students would attempt to "impose structure" on the ambiguous stories, their study was important in documenting that n-cog did exist and that affective reactions to a situation varied with both intrapersonal (i.e., n-cog) and environmental (i.e., manipulated ambiguity) factors.
In a subsequent study, Cohen (1957) investigated the relationship of n-cog to the order of presentation in persuasion. Specifically, he used two orders of presentation, one in which need-arousal was followed by information, with another which reversed the two. His hypothesis was that those subjects high in n-cog would be less affected by the order of presentation, since they would work harder to structure the situation for themselves. Using a forced-choice scale to measure n-cog, he divided a class of 35 students into high and low groups. A communication supporting "grading on the curve" was presented either before or after a "fear-arousing" discussion of the vagaries of the grading system.
Cohen (1957) found greater attitude change on an immediate posttest for the group receiving the need-arousal treatment first. He concluded:
"Need satisfaction and presumably, therefore, opinion change in this situation demanded the perception of the information as instrumental to the aroused need" (p. 94).
However, on a delayed posttest, Cohen found that the high n-cog subjects retained a significant change in attitude regardless of which group they were assigned to, while the low n-cog subjects showed no positive change, regardless of which order of communication they had received. The results of this study served to re-affirm the presence of a reality testing function and its mediating influence on attitude change.
Reward/punishment. In 1955, Helen Peak published her classic article, "Attitude and Motivation," in which she outlined an expectancy-times-value (ExV) approach to representing the motivational structure underlying attitudes. Peak was a member of the Michigan group and a part of the Attitude Change Project under Katz' direction, and she assumed directorship of the project for one year while Katz was away. During this time, a number of studies were conducted which, while closely related to the functional approach, did not explicitly position themselves within the Sarnoff-Katz functional framework.
In the first of these studies, Rosenberg (1956) investigated the "attitude structure" underlying attitudes toward free speech for Communists and the removal of Negro housing segregation. By measuring the "value importance" of each of 35 goals and values, as well as the "perceived instrumentality" of the two issues mentioned above in blocking or attaining those goals, Rosenberg was able to construct an ExV "cognitive structure index" which was closely related to overall affect. Rosenberg's work was later cited by Katz, et al. (1956) as "...a combination of the rational and reward approaches..." (p. 28). The "rational" portion derived from the measure of instrumentality, while the "reward" function was represented by the value importance hierarchy.
In a second study, Carlson (1956) built on Rosenberg's earlier work on Negro housing segregation in attempting to modify instrumentalities relating to the values of American prestige, protection of property values, equal opportunity, and being broadminded and worldly-wise. By having students write essays supporting the removal of segregation and then following this with an in-class discussion, Carlson was able to modify the individual instrumentalities, the overall ExV index, and to a somewhat lesser extent, attitude. Subjects with initially moderate attitudes changed the most while initially favorable subjects had little room to change, and initially negative subjects exhibited a "boomerang" effect. This latter finding suggests that some of Carlson's subjects were holding ego-defensive, rather than reward/punishment attitudes, but since Carlson included no measure of ego defense, the finding is only suggestive. Katz, et al. (1956) also cite Carlson's (1956) study as utilizing a combination of the reward and rational approaches, although it should be noted that neither Rosenberg (1956) nor Carlson (1956) characterized their own work in this way.
In the subsequent studies reported by Peak and her associates (Peak and Morrison, 1958; Peak, 1960; Rosenberg, 1960), the emphasis continued to be placed on the ExV representation of cognitive structure, with no explicit links being made to the functional approach.
Ego-defense. Wagman (1955) tested Sarnoff and Katz' (1954) propositions regarding the ego-defense function by attempting to modify students' anti-Negro attitudes through both authoritarian suggestion and a "non-authoritarian'' cognitive restructuring technique. Splitting his sample into thirds based on their F-Scale scores, Wagman was able to show a fairly consistent pattern across four different attitude measures:
1) low F-Scale subjects changed more than high F-Scale subjects under conditions of cognitive restructuring;
2) high F-Scale subjects changed more than low F-Scale subjects under conditions of authoritarian suggestion, regardless of which direction of change was advocated (i.e., more anti-Negro prejudice or less anti-Negro prejudice).
These results are consistent with the Sarnoff and Katz' (1954) reasoning that ego-defensive (in this case, authoritarian) attitudes will be less susceptible to change through a "rational" approach. However, Wagman's findings indicate that ego-defensive attitudes are not necessarily rigid in the face of less "rational" attitude modification strategies.
In two related studies which were aimed at directly investigating the ego-defensive function of attitude, Katz, Sarnoff and McClintock (1956) and Katz, McClintock and Sarnoff (1957) attempted to change attitudes toward Negroes. In the first study, both an informational (rational) and "interpretive" (ego-defensive) appeal were used, in order to determine their relative effectiveness in attacking a prejudicial and presumably ego-defensive attitude. In order to measure ego-defense, a TAT card was designed and codes were developed for projection, denial and extrapunitiveness, three ego-defense mechanisms. These were combined into a single score which was used to split the sample into thirds. It was predicted that the efficacy of the informational appeal would decrease with increasing ego-defense, while the interpretive appeal would be maximally effective for moderately defensive subjects. Neither prediction was confirmed, which led Katz et al. to question the ego-defense measure they had employed.
In their second study, Katz, et al. (1957) used four separate measures of ego-defense: a pair of TAT cards, an 18-item multiple choice sentence completion task, 45 items from the paranoia subscale of the MMPI, and six subscales from the F-Scale. The former two measures were found to be insensitive, while the latter two were judged to perform adequately. An interpretive appeal was administered to 131 college women, with the expectation that those moderate in defensiveness would exhibit more change than either the high or low defensiveness groups. This prediction was confirmed for the two standardized measures (MMPI and F-Scale), while being disconfirmed for the two more ad hoc approaches.
McClintock (1958) investigated the relationship between attitude change and two personality variables, ego-defensiveness and other-directedness. One hundred ninety-eight female college students were exposed to one of three forms of persuasive appeal (information, insight, ethnocentric) designed to change anti-Negro attitudes. Consistent with Katz, et al. (1957) McClintock found that moderately defensive subjects (as measured by a portion of the F-Scale) changed the most under the insight treatment, while no differences were observed among the groups under the informational appeal. Under the ethnocentric appeal, subjects high in ego-defense changed more than those low in defensiveness. McClintock (1958) concluded:
...the findings of this research indicated that a given influence may have little effect in moving the total population in the direction of that influence...Thus, knowing the personality structure of the individual seems necessary for... the devising of appropriate and effective methods for changing attitudes. (pp. 492-3).
In yet another study of the self-insight approach to changing ego-defensive attitudes, Stotland, Katz and Patchen (1959) attempted to manipulate racial attitudes. Contrary to the pattern of findings emerging from earlier studies, a monotonic decreasing relationship was obtained between ego-defense (as measured by the F-Scale) and attitude change. This failure to replicate the earlier findings was not readily explainable, but Stotland, et al. speculated that it could be due to an insensitive measure of ego-defense (even though the same measure had been used satisfactorily in a number of other studies), increased awareness by low defensive subjects of their "inappropriate" attitudes, or a greater amount of projection among those low in ego-defense than had been previously anticipated. This study serves to underscore the rather weak operational procedures characterizing the research on the ego-defensive function, even though this function was clearly the major focus of empirical research by Katz and his associates.
In his summarization of the ego-defensive function of attitude Sarnoff (1960a) states:
It begins to appear that--insofar as consciously unacceptable motives do form the basis of an individual's disposition to agree with prevailing anti-minority stereotypes--projection plays a larger role than any of the other mechanisms of ego defense. (p. 273)
He concluded this on the basis of the results of a number of studies utilizing projection and other defense mechanisms (Katz, et al., 1956, 1957; Sarnoff, 1951, 1960b; Sarnoff and Corwin, 1959; Cohen, 1956). Thus it would appear that any further attempt to investigate the ego-defense function of attitude should rely heavily on projection as a basis for assessing attitudinal function.
In the first major theoretical statement of functional theory since the Sarnoff and Katz (1954) article, Katz and Stotland (1959) outlined a theory of attitude formation and change based upon the research described above. At the heart of their theory were the motivational functions served by attitudes. These functions were classified as proximal, object-instrumental, ego-instrumental, and ego-defensive.
Proximal attitudes are those where the attitude toward the object satisfies some need of the individual directly. This is in contrast to attitudes which are held toward an object because it is instrumental in helping the individual to satisfy a need. Attitudes toward foods and sex objects are cited as possible examples of proximal attitudes, but perhaps the most important example is the attitude based on the need for understanding. This is directly analogous to Cohen's (1955, 1957) investigations of n-cog and re-asserts the notion that to understand, to explore, to search for meaning is an end in itself, and not merely a means to some other end-state. Proximal attitudes are seen as being reinforced through repeated interaction with the attitude object and are "...based on the principle that individuals put high value on objects which satisfy their needs and low value on objects which frustrate them" (Katz and Stotland, 1959, p. 437).
Object-instrumental attitudes are those based on the means-ends principle first explicitly stated by Peak (1955) and her associates. According to Katz and Stotland (1959), many social attitudes fall into this category. Positive attitudes are developed toward those social objects which are seen as a means to ultimate satisfaction of wants and needs. These sorts of attitudes may be learned through trial-and-error, but as noted by Katz and Stotland: "The concept of instrumental learning has been converted into the notion of perceived instrumentality by workers dealing with cognition, perception and attitudes" (p. 436). Thus, the characteristics of the attitude object and their relationship to need satisfaction act as the motivational underpinnings of object-instrumental attitudes.
In contrast, an ego-instrumental attitude "...arises from sources further removed from the attitude itself, from ego satisfactions" (p. 440). The key feature of this type of attitude is that it allows the person to express to others what kind of person he is. Thus, such constructs as "value" and "self concept" become relevant attitude bases. When compared with object-instrumental attitudes, the cognitions underlying ego-instrumental attitudes are "...elaborated less upon the basis of the objective characteristics of the attitudinal object than upon the individual's need to maintain his own self-image" (Katz and Stotland, 1959, p. 440).
The final category of attitudes, ego-defensive, is essentially the same as was discussed by Sarnoff and Katz (1954). Ego-defensive attitudes conform to the classic pattern of repressing consciously unacceptable sexual and aggressive impulses and subsequently projecting these impulses onto others, usually outgroups of some kind. In this sense, ego-defensive attitudes are maladaptive and leave the basic conflict within the individual unresolved. They are quite resistant to change through common forms of influence, which follows from the deep-seated nature of the motives being repressed.
Katz and Stotland viewed the motivational-attitudinal-action system as dynamic, characterized by constant interaction, such that no true distinctions can be made among independent and dependent variables. However, they did feel that certain variables would be more dominant in a causal sense:
...we would in general regard motives and environmental forces as independent variables, attitudes as intervening variables, and expression in behavior as the dependent variable (p. 468).
This view of the causal structure of the system has direct implications for the derivation of attitude change strategies; hence, Katz and Stotland recommend modifications of (1) the individual's value system; or (2) cognitions with respect to the attitude object as the most viable means of changing attitudes.
In the final major paper pertaining to functional theory, Katz (1960) offered what is perhaps the most systematic description of the proposed functions as well as the conditions conducive to their arousal and change. The four functions he outlined were similar, but not identical, to the ones described by Katz and Stotland (1959): utilitarian (also referred to as instrumental or adjustive), ego-defensive, value-expressive, and knowledge.
The utilitarian function was described by Katz as an attempt by the individual to maximize rewards and minimize punishment from the environment. Thus, attitude formation was seen as "...dependent upon present or past perceptions of the utility of the attitudinal object for the individual" (p. 171). This relates closely to the notion of object-instrumental attitudes presented by Katz and Stotland (1959).
The ego-defensive function, as conceived by Katz (1960), is essentially identical to the one originally conceived by Sarnoff and Katz (1954) and reiterated by Katz and Stotland (1959). It is through this function that the individual protects himself from either external threats or internal conflicts. In its linkage to Freudian psychology, the ego-defensive function is concerned primarily with consciously unacceptable motives.
The value-expressive function involves attitudes which express the individual's own personally-held values or his self concept. This is analogous to Katz and Stotland's (1959) ego-instrumental function. It is virtually opposite to the ego-defensive function, which serves to obscure aspects of the person from himself and others. Value-expressive attitudes instead reveal the self concept and in some cases help the individual to move closer to his "ideal" self concept.
The knowledge function is served by attitudes which help the individual to find meaning, organize the environment, and in general provide clarity and consistency to the person's view of the world. This is in agreement with the proximal attitude based on n-cog as described by Katz and Stotland.
Katz' (1960) major contribution was a formal specification of the factors underlying attitude formation, and change. His was the only systematic analysis of the nature of the various functions, together with an assessment of the key elements which could be brought to bear on attitude change efforts. While some later authors have criticized the conditions he outlined as being too general to allow empirical test, nevertheless the formal statement of these conditions was at least a good first step toward building a set of testable propositions, which is a requirement for any theory.
More disappointing than the generality of the determinants of attitude arousal and change was the lack of any clues as to how the approach might be operationalized. Katz (1960) cited the need for public opinion polls measuring needs and attitudes, but beyond a few references to the F-Scale, MMPI and n-cog, he gave no indication as to how the functions could be assessed. It would appear that at least some of the difficulty that arises in the assessment of attitude functions derives from a lack of conceptual clarity with respect to the various functions. By moving toward appropriate measurement procedures, some of the somewhat vague notions regarding attitude arousal and change conditions may crystallize into more clearly testable hypotheses.
Table 1 summarizes the various functions of attitude discussed in five major theoretical papers. As is evident from the table, there are some discrepancies among the articles in terms of both the number of functions and the exact nature of these functions. The two functions which were identified with the most consistency across the various studies were (in Katz' (1960) terminology) the ego-defensive and knowledge functions, although the latter was characterized by five different labels and in one paper (Smith, et al., 1956) was combined with the equivalent to Katz' (1960) utilitarian function into a single "object appraisal" function. The consensus on the ego-defensive function is not surprising, given the involving and apparently threatening nature of the attitudes under study (i.e., attitudes toward Russia and racial and ethnic minorities).
With the exception of Smith's (1947) article, there was agreement on a utilitarian function stressing the reward/ punishment notion of hedonism. Somewhat less agreement was found, however, on the existence of a value-expressive function. It appeared in the later work of the Michigan group, but was not considered by either Sarnoff and Katz (1954) or by Smith, et al. (1956). In fact, the latter authors explicitly rejected the notion of a value-expressive function (although Smith (1947) had earlier postulated one), and instead argued that value expression was not a "need" of the individual but rather just a by-product of the expression of an attitude (Smith, et al., 1956, p. 38).
The clearest difference between the Harvard and Michigan groups is apparent in their treatment of the social adjustment function. Smith, et al. (1956) strongly emphasized the social function of attitude, while Katz and his associates (1954, 1959, 1960) did not distinguish it as a function per se. To some degree, the social function of attitude was subsumed in Katz' (1959, 1960) discussion of the value-expressive function, but the discussion focused more on socialization as a source of values than on social needs as direct underpinnings of attitude.
While there was some disagreement as to the exact functions served by attitude, the theorists were unanimous in their characterization of attitudes as having multiple determinants. Although it was conceded that uni-functional attitudes may exist in some cases, the consensus was that most attitudes serve more than one function for the individual, and that only through an understanding of these functional bases could attitudes be diagnosed and effectively changed. Thus, the functional theorists agreed on an approach to attitude which was strongly based on individual differences; i.e., the same observed attitude (in terms of strength and direction of affect) for two different individuals may be reflecting two entirely different motivational structures in terms of the functions being served.
A COMPARISON OF THE FUNCTIONS OUTLINED IN MAJOR THEORETICAL PAPERS
The lack of operations for measuring functions of attitude has been identified as a major drawback to functional theory (e.g., Insko, 1967; Himmelfarb and Eagly, 1974). Some researchers have gone even further with this criticism suggesting that the theory is really nothing more than a broad conceptual integration and should not be viewed as a potential operational framework (Kiesler, Collins and Miller, 1969). However, the functional theorists themselves clearly regarded the theory as potentially operationalizable. Smith, et al. (1956), while offering no clues as to how to operationalize the theory, nevertheless pointed out the need to do so and voiced optimism that the measurement of attitude functions "...seem entirely within the range of feasible attainment'' (p. 278). Similarly, Katz (1960) discussed the assessment of attitude functions, pointing to the need for systematic sampling of target populations and advocating increasing reliance on methods other than the depth interview.
Yet, despite the early attempts to measure ego-defensive (e.g., Katz, et al., 1956, 1957) and knowledge function (e.g., Cohen, et al., (1955) attitudes, little systematic empirical work was devoted to the task of functional assessment. Perhaps the best explanation for this problem lies in the following quotation from Katz and Stotland (1959):
...we cannot systematically develop the conceptual properties of all our constructs, point to validated operational measures for them, or describe an appropriate mathematical model for handling the data in this field. (p. 425, italics added)
Thus, although the Katz and Stotland (1959) contribution was considered only a "preliminary" statement of functional theory, the final conceptual article was published the very next year by Katz (1960). In neither of these articles was there sufficient conceptual clarity to allow rigorous specification of operations for representing the various functions and their relationship to the expressed attitude of an individual. Yet, the kernel of a fundamental approach to solving the conceptual, and, concomitantly, the operational problems with functional theory was present in these and earlier articles. The following section attempts to explicitly draw out from the functional literature the roots of the proposed re-conceptualization.
An Expectancy-Times-Value View
[In the recent discussion, "expectancy" will be used to refer to any perceived relationship between an attitude object and some related attribute, property or value. No distinction is made between the level of the perceived relationship (e.g., degree of sweetness of a soft drink) and the probability that the level exists (e.g., it is very likely that the soft drink is sweet). While this distinction may be important in subsequent measurement of attitude functions, it is not critical to the conceptual development of an ExV approach.]
Weiner (1972) describes the expectancy-times-value (ExV) approach to human motivation as the conception that "...direction and intensity of behavior is a function of the expectation [E] that certain actions will lead to the goal, and the incentive value [V] of the goal object" (p. 8). The ExV notion is central to Murray's (1938) theory of personality, Rotter's (1954) social learning theory, Atkinson's (1964) theory of achievement motivation, and Peak's (1955) theory of attitude. A careful examination of the literature on the functional theory of attitude suggests that it, too, is strongly grounded in ExV principles.
In their discussion of the relationship of attitude to personality, Smith, Bruner and White (1956) posited that personality is marked by an "...adaptive striving after goals [V]," and further that the individual becomes "...selectively aware of objects related [E] to certain goals..." (p. 32). In a similar vein, Katz (1960) states:
Knowledge of need state [V] indicates the type of goal toward which the individual is striving. But the means for reaching this goal [E] may vary considerably...(P. 203).
Thus, the general notion of goal-oriented behavior is apparent in both of the major approaches to functional representation of attitude.
The basic idea of an ExV reconceptualization of functional theory is that the cognitive "component" of attitude serving a particular function becomes elaborated around the basic needs or values salient to that attitude function. The "index of cognitive structure" (ExV) becomes an index of the particular attitude function under consideration. The exact nature of the cognitive (E) and affective (V) elements making up the index would vary with the content of the various functions. A consideration of the individual functions serves to further illuminate this ExV orientation of functional theory.
Probably the clearest expression of functional theory's implicit reliance on ExV principles is found in Katz' (1960) discussion of the utilitarian or adjustment function:
Attitudes acquired in the service of the adjustment function are either the means for reaching the desired goal or avoiding the undesirable one .... In general, then, the dynamics of attitude formation with respect to the adjustment function are dependent upon present or past perceptions [El of the utility [V] of the attitudinal object for the individual...The closer these objects are to actual need satisfaction IV] and the more they are clearly perceived [E] as relevant to need satisfaction, the greater are the probabilities of positive attitude formation (p. 171, italics added).
According to the ExV formulation, then, the utilitarian function would be measured by expectations regarding the need-related attributes of the attitude object, weighted by the value or utility of these attributes to the individual. The degree to which the utilitarian ExV index relates to attitudinal affect is an assessment of the degree to which the attitude is serving the utilitarian function for the individual.
Value-expressive. As noted by Katz and Stotland (1959), the utilitarian function rests primarily on the properties of the attitude object and their relationship to goal attainment. On the other hand, value-expressive attitudes are seen as reflections of properties of the individual:
The cognitive component of the attitude becomes elaborated less upon the basis of the objective characteristics of the attitudinal object than upon the individual's need to maintain his own self-image (Katz and Stotland, 1959, p. 440).
Katz (1960) postulates that the expression of the self concept is rewarding to the individual. Presumably, in order for an attitude to serve the value-expressive function, then, there must be some expectation (E) on the part of the individual that holding that particular attitude gives expression to some aspect of self (V). Or, as stated by Smith (1947):
...intensity of attitudinal affect is a function of the extent to which a personal value is engaged [E] and of the importance [V] of this engaged value in the hierarchy of the person's central values (p. 519).
Whether dealing with aspects of the self concept or central values of the individual, the mechanism underlying the value-expressive function relies on the perceived relationship (E) of the attitude object to some characteristic or value, and the strength or importance (V) of that value to the individual. The strength of the relationship between an ExV index based on individual values and overall attitude serves as an indicator of the value-expressiveness of the attitude.
Knowledge. In an investigation of the knowledge function, Cohen, et al. (1955) implicitly used an ExV conceptualization in their hypotheses that:
1. An ambiguous situation [E] will be more frustrating [i.e., negative affect-producing] than a more structured one.
2. Hypothesis 1 will be more applicable to people with high need for cognition [V] than for those with low need. (p. 291)
Need for cognition was postulated as a desired end-state (V) which could be satisfied by perceptions (E) of the environment as unambiguous and structured. The assumption of multiplicative combination of the E and V terms is clearly present in Hypothesis 2. In their later discussion of the knowledge function, Katz and Stotland (1959) state:
Thus is would be expected that those objects in the environment which aid in understanding the world would be evaluated highly. Furthermore, if the object itself is clearly understood, it will be evaluated more highly than if it is understood but vaguely (p. 438).
Therefore, it is apparent that the perception of the environment as unambiguous, or as aiding in the reduction of ambiguity, is the major factor in satisfying the need for knowledge function attitudes. An ExV index based on needs for knowledge, understanding, cognition, etc., and the related environmental perceptions would form a measure of the knowledge function. The relationship between this index and attitude would indicate the degree to which the knowledge function was being served by the attitude.
Ego-defensive. The final function to be considered here, ego-defensive, is on the surface rather far-removed from an ExV formulation in its emphasis on unconscious motives. In fact, Locke (1975) criticized ExV theory as being incompatible with subconscious motives. However, there is reason to believe that this criticism may be somewhat overstated, in view of several observations made by functional theorists. For instance, in discussing their externalization function, Smith, et al. (1956) postulated that:
Covert strivings influence selectivity in the perception [E] of objects [p. 271] ...Attitudes toward objects are influenced by attitudes toward covert strivings [VI (p. 273).
Similarly, Sarnoff and Katz (1954), in discussing the distortion that is inherent in ego-defensive attitudes, noted that four types of distortion could be present:
In terms of distortion of the target object, the person with ego-defensive attitudes may:
1. Attribute to others [E] qualities which they do not possess.
2. Ignore [E] qualities which they do possess.
In terms of affect loading, he may:
1. Experience an excessive emotional reaction (either positive or negative) [V] to real or attributed qualities of the object.
2. Fail to experience an adequate emotional reaction (either positive or negative) [V] to real or attributed qualities of the target. (p. 120)
Finally, Sarnoff (1966) notes:
...a precise determination of the functional relationship between an attitude and a consciously unacceptable motive involves:
1. A postulation of which combination of consciously unacceptable motive [V] and ego defense [E] might plausibly account for the particular overt response from which the attitude is inferred.
2. After conceptualizing the most plausible combination of ego defense and consciously unacceptable motive, we must proceed to demonstrate empirically the relationship between that combination [ExV] and the attitude it is presumed to support. (p. 283).
It has been previously noted that the ego defense mechanism of projection is the only one which has been shown to consistently relate to social attitudes (Sarnoff, 1960a). To-the extent that the projection of unacceptable traits and motives onto outgroup members constitutes an expectation (E), however inappropriate, then the resultant satisfaction of ego-defensive motives (V) should lead to the formation of ego-defensive attitudes. Further, an ExV index composed of all projected traits and their associated evaluations should provide an index of the ego-defensive function. To the extend that this index is associated with a measure of overall attitude, the attitude can be regarded as serving the ego-defensive function.
The preceding references to, and quotations from, functional theory literature should point out the considerable compatibility of functional theory with an ExV formulation, although this linkage was not explicitly recognized by the functional theorists, nor by later attitude researchers. While it may be argued that the ExV notion is obvious in the writings of functional theorists, it is not so obvious as to have been carried through in systematic fashion. This is particularly evident in Katz' (1960) hypothesized conditions mediating attitude arousal and change and in the lack of operations for assessing attitude functions. It is therefore asserted that viewing functional theory from an ExV perspective constitutes a first step in reconceptualizing the theory along lines which explicitly note its compatibility with ExV theory and move it further toward the status of an empirically testable theory.
There are numerous important implications stemming from the proposed reconceptualization of functional theory. These implications can be broken down into implications for theory and implications for policy.
Most importantly, the ExV reconceptualization of functional theory provides a means for empirical testing of functional theory propositions. Thus this conceptually rich approach to the study of attitude and attitude change can, for the first time, be thoroughly investigated to determine its usefulness as a theory.
Secondly, the ExV framework proposed here, together with appropriate operational procedures (Lutz, 1977), can perform an integrative function in at least two respects. At one level, the reconceptualization has helped to expose the basic similarity of process underlying the various functions. At another level, it can help to achieve the integration of multiattribute models of attitude (Wilkie and Pessemier, 1973), as well as self-concept research (e.g., Ross, 1971) and the Rokeach (1968) value paradigm, under the functional umbrella, an important unifying step in current research on attitudes. Self-concept research, in particular, has been a stepchild in consumer behavior research, with no broader frame of reference. Functional theory provides that frame of reference.
Functional theory is a potentially powerful policy tool; until now, however, its usefulness was hampered by the lack of measurement procedures. The current approach should lead to a solution of the measurement problem, thus making the theory available for market application. Its contributions should be felt in at least three areas.
Attitude measurement. Functional theory, as reconceptualized, would provide a unified framework for the measurement of attitudes toward brands, products, social issues, etc. Depending upon the level of involvement, different weights should appear on the various functions. It should be of interest to the manager, for example, to know the functional basis of attitudes toward a particular brand in comparison with the functional basis of attitudes toward the product class in general. This knowledge could lead to an industry-based campaign directed at one function, while brand advertising may focus on another function altogether.
Market segmentation. McClintock (1958) noted that "... knowing the personality structure of the individual seems necessary for the devising of appropriate and effective means for changing attitudes" (p. 493). Since modern segmentation theory espouses a "building up" of individuals into homogeneous market segments, functional theory is well-suited for market segmentation based on the particular attitude function(s) being served for various groups in the market. This approach to segmentation has been attempted, with some success, by Locander and Spivey (1975).
Communications strategy. Clearly, the functional approach should be quite useful in the diagnosis of consumer attitudes for the purposes of prescribing attitude change strategy. By measuring the various functional supports and determining their relative weights, the most efficient means of change can be derived. The functional framework also provides a basis for communications copy pretesting, to determine if the appropriate cognitive-motivational structure is being impacted. Finally, the proposed functional approach could be used for monitoring the effects of a persuasive campaign in the marketplace (Lutz, 1978).
In this paper, the origins and propositions of the functional theory of attitude have been reviewed, evaluated and reconceptualized within an expectancy-times-value framework. The proposed reconceptualization has been shown to be consistent with theoretical statements by the original functional theorists.
Functional theory is a potentially powerful managerial tool for use in advertising and segmentation decisions. Thus, adoption of the proposed ExV framework would not only enable testing of the theory, but also application of the theory to policy decisions. The potential richness of the theory for decision-making has long been noted but never attained due to lack of measurement procedures. The proposed reconceptualization does not completely ameliorate that deficiency in view of the fact that no data have been reported, but it does serve to make it clear that measurement is feasible within the reconceptualized functional framework.
T. W. Adorno, Else Frenkel-Brunswik, Daniel J. Levinson, and R. Nevitt Sanford, The Authoritarian Personality (New York: Harper and Row, 1950).
John W. Atkinson, An Introduction to Motivation (New York: Van Nostrand Reinhold Co,, 1964).
Reuben M. Baron, "Attitude Change Through Discrepant Action: A Functional Analysis," in A. G. Greenwald, T. C. Brock, and T. M. Ostrom, (Eds.), Psychological Foundations of Attitudes (New York: Academic Press, 1968), 297-326.
Earl R. Carlson, "Attitude Change Through Modification of Cognitive Structure," Journal of Abnormal and Social Psychology, 52 (1956), 256-61.
Arthur R. Cohen, "Experimental Effects of Ego-defense Preference on Interpersonal Relations," Journal of Abnormal and Social Psychology, 52 (1956), 19-27.
Arthur R. Cohen, "Need for Cognition and Order of Communication as Determinants of Opinion Change," in C.I. Hovland (Ed.), The Order of Presentation in Persuasion (New Haven: Yale University Press, 1957), 79-97.
Arthur R. Cohen, Ezra Stotland, and Donald M. Wolfe, "An Experimental Investigation of Need for Cognition," Journal of Abnormal and Social Psychology, 51 (1955), 291-4.
George S. Day, "Theories of Attitude Structure and Change," in S. Ward and T. S. Robertson (Eds.), Consumer Behavior: Theoretical Sources (Englewood Cliffs, NJ: Prentice-Hall, Inc. 1973), 303-53.
Alice H. Eagly and Samuel Himmelfarb, "Current Trends in Attitude Theory and Research," in S. Himmelfarb and A.H. Eagly (Eds.), Readings in Attitude Change (New York: John Wiley and Sons, Inc. 1974), 594-610.
Samuel Himmelfarb and Alice H. Eagly, "Orientations to the Study of Attitudes and Their Change," in S. Himmelfarb and A.H. Eagly (Eds.), Readings in Attitude Change (New York: John Wiley and Sons, Inc., 1974), 2-49.
Chester A. Insko, Theories of Attitude Change (New York: Appleton-Century-Crofts, 1967).
Daniel Katz, "The Functional Approach to the Study of Attitudes," Public Opinion Quarterly, 24 (1960), 163-204.
Daniel Katz, Charles McClintock and Irving Sarnoff, "The Measurement of Ego Defense as Related to Attitude Change," Journal of Personality, 25 (1957), 465-74.
Daniel Katz, Irving Sarnoff and Charles McClintock, "Ego-Defense and Attitude Change," Human Relations, 9 (1956), 27-45.
Daniel Katz and Ezra Stotland, "A Preliminary Statement of a Theory of Attitude Structure and Change," in S. Koch (Ed.), Psychology: A Study of a Science, Vol. III (New York: McGraw-Hill, 1959), 423-75.
Charles A. Kiesler, Barry E. Collins, and Norman Miller, Attitude Change: A Critical Analysis of Theoretical Approaches (New York: John Wiley and Sons, Inc., 1969).
William B. Locander, and W. Austin Spivey, "A Functional Approach to Attitudes and Consumers' Brand Loyalty," unpublished manuscript, 1975.
Edwin A. Locke, "Personnel Attitudes and Motivation," in M. R. Rosenzweig and L. W. Porter (Eds.), Annual Review of Psychology, Volume 26 (Pale Alto, Calif.: Annual Reviews, Inc., 1975), 457-80.
Richard J. Lutz, "The Functional Approach to Attitudes: A Reconceptualization with Operational Implications," Working Paper No. 53, Center for Marketing Studies, UCLA, July 1977.
Richard J. Lutz, "A Functional Theory Framework for Designing and Pretesting Advertising Themes," in J.C. Maloney and B. Silverman (Eds.), Attitude Research Plays for High Stakes (Chicago: American Marketing Association, 1978), forthcoming.
Charles G. McClintock, "Personality Syndromes and Attitude Change," Journal of Personality, 26 (1958), 479-93.
Henry A. Murray, Explorations in Personality (New York: Oxford Press, 1938).
Helen Peak, "Attitude and Motivation," in M. R. Jones (Ed.), Nebraska Symposium on Motivation (Lincoln: University of Nebraska Press, 1955), 149-89.
Helen Peak, "The Effect of Aroused Motivation on Attitudes,'' Journal of Abnormal and Social Psychology, 61 (1960), 463-8.
Helen Peak, and H. William Morrison, "The Acceptance of Information into Attitude Structure," Journal of Abnormal and Social Psychology, 57 (1958), 127-35.
Milton Rokeach, Beliefs, Attitudes and Values: A Theory of Organization and Change (San Francisco: Jossey-Bass, Inc., 1968).
Milton J. Rosenberg, "Cognitive Structure and Attitudinal Affect," Journal of Abnormal and Social Psychology, 53 (1956), 367-72.
Milton J. Rosenberg, "An Analysis of Affective-Cognitive Consistency," in C.I. Hovland and M. J. Rosenberg (Eds.), Attitude Organization and Change (New Haven: Yale University Press, 1960), 15-64.
Ivan Ross, "Self Concept and Brand Preference," Journal of Business, 44 (1971), 38-50.
Julian B. Rotter, Social Learning and Clinical Psychology (New York: Prentice-Hall, Inc. 1954).
Irving Sarnoff, "Identification with the Aggressor: Some Personality Correlates of Anti-Semitism Among Jews," Journal of Personality, 21 (1951), 199-218.
Irving, Sarnoff, "Psychoanalytic Theory and Social Attitudes,'' Public Opinion Quarterly, 24 (1960a), 251-79.
Irving Sarnoff, "Reaction Formation and Cynicism," Journal of Personality, 28 (1960b), 129-43.
Irving Sarnoff, "Social Attitudes and the Resolution of Motivational Conflict," in M. Jahoda and N. Warren (Eds.), Attitudes: Selected Readings (Baltimore: Penguin Books, 1966), 279-84.
Irving Sarnoff, and S. M. Corwin, "Castration Anxiety and the Fear of Death," Journal of Personality, 27 (1959), 374-85.
Irving Sarnoff and Daniel Katz, "The Motivational Bases of Attitude Change," Journal of Abnormal and Social Psychology, 49 (1954), 115-24.
M. Brewster Smith, "The Personal Setting of Public Opinions: A Study of Attitudes Toward Russia," Public Opinion Quarterly, 11 (1947), 507-23.
M. Brewster Smith, "Personal Values as Determinants of a Political Attitude," Journal of Psychology, 28 (1949), 477-86.
M. Brewster Smith, Jerome S. Bruner, and Robert W. White, Opinions and Personality (New York: John Wiley and Sons, Inc., 1956).
Ezra Stotland, Daniel Katz, and Martin Patchen, "The Reduction of Prejudice Through the Arousal of Self-Insight," Journal of Personality, 27 (1959), 507-31.
Morton Wagman, "Attitude Change and Authoritarian Personality,'' Journal of Psychology, 40 (1955), 3-24.
Bernard, Weiner, Theories of Motivation: From Mechanism to Cognition (Chicago: Markham, 1972).
William L. Wilkie, and Edgar A. Pessemier, "Issues in Marketing's Use of Multi-Attribute Models," Journal of Marketing Research, 10 (1973), 428-41.