The SAGE Handbook of Persuasion

by James Price Dillard

  Ego-Defensive

  According to Katz (1960), attitudes that serve an ego-defensive function (externalization for Smith et al., 1956) allow people to avoid acknowledging unpleasant aspects of self or of the external world. Ego-defensive attitudes can promote protecting self-esteem. Threats to self-esteem can include personal failings or poor behavior. Katz argued that such attitudes often involved rationalizing and other defense mechanisms. One scenario in which an attitude toward BMWs serves an ego-defensive function would be one in which someone had a positive attitude toward BMWs because of frustration resulting from an inability to earn occupational advancement. The car could instill a feeling of importance despite being passed over for promotion. Katz’s research in this area focused on those who held negative attitudes toward minority groups as a means of increasing their own self-esteem. He and his colleagues argued that these attitudes might be changed by helping the audience gain self-insight into the defense mechanisms supported by these attitudes (Stotland, Katz, & Patchen, 1959).

  Knowledge

  Attitudes that serve a knowledge function help people gain greater understanding of the structure and operation of their world (Katz, 1960). Katz noted that the world that people inhabit is extremely complex, and that attitudes might facilitate making sense of that world without serving needs other than understanding. Herek (1987) argued that Smith and colleagues’ (1956) object appraisal function represents a combination of both Katz’s utilitarian and the knowledge functions because Smith et al. argued that people organize the world for the purpose of reaching utilitarian goals. Other scholars propose that the knowledge function drives attitudes that serve for no purpose other than learning about the world as an end in itself, predicated on a basic need to know (Locander & Spivey, 1978). Someone might have an attitude about candidates in a presidential election in Bolivia not because it affects their interests or expresses their values but because they try to make sense of South American politics. Similarly, an attitude toward BMWs may emerge because of an interest in the auto industry. Katz argued that these attitudes can be changed by explaining how an audience’s understanding of the world is inadequate and how a different attitude provides a more accurate world view.

  Summary

  It is important to emphasize that although this list contains the most commonly described attitude functions, it may not be comprehensive. Both Katz (1960) and Smith et al. (1956) hint at the existence of functions that they did not identify. In this spirit, Clary, Snyder, Ridge, Copeland, Stukas, Haugen, and Miene (1998) suggest that the ego-defensive function may be divided into attitudes that defend self-esteem from attacks and attitudes that promote self-esteem. Furthermore, Locander and Spivey (1978) attempted to measure Katz’s functions and found that there were people whose attitude toward some objects did not serve any of the functions mentioned previously. Although both Katz (1960) and Smith et al. (1956) suggested the possibility of an attitude serving multiple functions, this possibility was not investigated systematically until techniques of measuring and inducing the various functions were developed in order to successfully target functions with persuasive messages. The second wave of theory and research that developed methods of identifying functions for the purpose of matching messages to functions to increase message persuasiveness is discussed next.

  Functions as Variables

  * * *

  Some scholars argue that the study of attitude functions was neglected for two decades because the seminal initial work (Katz, 1960; Smith et al., 1956) failed to supply useful methods of studying the phenomenon (Snyder & Debono, 1985). Specifically, Smith et al. (1956) conducted lengthy interviews, and such a methodological approach does not lend itself to empirical investigations with samples sufficient to provide precise estimates of effects. Moreover, it is viewed commonly as being more subjective than required for a rigorous study of the phenomenon. Katz and colleagues attempted to explore the ego-defensive function using authoritarianism as an indicator of the ego-defensive function (Katz et al., 1956; Sarnoff & Katz, 1954). These studies yielded inconsistent results, and perhaps for that reason others did not pursue this line of inquiry. One influential attempt at remedying this state of affairs was Herek’s (1986) neofunctional approach, and it involved reconceptualizing attitude functions in a way that would encourage empirical investigation.

  Herek’s Neofunctional Approach

  Herek (1986) proposed that the concept of attitude functions should be expanded beyond the personality centered-approach of Smith et al. (1956) and the categorical approach discussed by Katz (1960). In his reconceptualization of attitude functions each function is construed as a quantitative dimension such that any attitude serves all functions to varying degrees. Thus, for Herek, attitudes serve multiple functions and the extent to which each attitude serves multiple functions can be measured. Consequently, the relationships among attitude functions can be investigated empirically and quantitatively.

  Herek (1986) also proposed a new typology of attitude functions. Specifically, he posited that conceptual clarity may be gained by distinguishing two categories of attitude functions. The first category includes the evaluative functions, which are associated with gaining rewards and avoiding punishments. These attitude functions provide people with a means of organizing objects and events by self-interest. Herek distinguished three types: experiential and specific, experiential and schematic, and anticipatory-evaluative. Experiential and specific attitudes are formed by and directed at a single object of a category of attitude objects based on one or more encounters with that specific object. Varda might encounter a poorly maintained BMW driven by her friend Erin, and from that experience Varda might develop an experiential and specific negative attitude toward Erin’s BMW as a means of assisting her in avoiding accepting rides in dangerous and unreliable automobiles. On the other hand, Varda’s attitude toward Erin’s BMW would not necessarily generalize to all BMWs. If Varda formed a positive view of BMWs based on having driven and ridden in many others that she found to be safe and dependable, she would have developed an experiential-schematic attitude. Such attitudes represent the inductive development of an attitude toward a category of attitude objects based on personally encountering and evaluating many specific cases of the attitude object. In contrast, the anticipatory-evaluative function serves to inform people concerning the rewards or punishments that might accompany attitude objects, but they do not arise from any actual experience with these objects. Instead, one estimates the rewards and punishments, and the extent of them, that would result from an encounter with an attitude object. If Varda had never encountered a BMW, but had she heard positive reviews of its safety and reliability from trusted sources, she might develop a positive attitude that functions in an anticipatory-evaluative manner. Notably, these three functions that are grouped under the evaluative category are all similar to Katz’s (1960) utilitarian function.

  The other category of attitude functions identified by Herek (1986) is termed “expressive.” They include the social-adjustive, the value-expressive, and defensive functions identified by other theorists. Specifically, the social-adjustive and the value-expressive functions are similar to those to which Katz (1960) attaches the same labels, and the defensive function is similar to Katz’s ego-defensive function. The expression to others of these attitudes, rather than the utilitarian character of attitudes that serve evaluative functions, provides benefit for those who hold them. Herek’s use of this categorization scheme to understand more clearly the attitudes persons have toward those with AIDS is a particularly useful application of this set of distinctions (see Herek, 2000 for a review).

  In addition to conceiving of the functions as quantitative variables and providing an original classification scheme, Herek (1986) rejected the idea that functions are stable within individuals but across contexts. He argued that functions may be related to personality variables, and thus, that certain personality characteristics would be associated with a higher likelihood of an attitude se
rving a particular function. For example, Snyder and Debono (1985) proposed that people who are likely to engage in extensive self-monitoring are also likely to have attitudes that serve a social-adjustive purpose. Moreover, Herek extended the personality approach by proposing that characteristics of the attitude topic might influence the likelihood that a particular topic would be associated with a particular function. For example, Shavitt (1990) found that attitudes toward air-conditioners were most likely to serve utilitarian functions. Finally, Herek (1986) argued that situations could also make particular attitude functions salient. For example, he argued that social-adjustive functions were likely to predominate with attitudes toward one’s friends. In one application of this approach, Shavitt, Swan, Lowrey, and Wanke (1994) found that particular functions can be primed. These various approaches that conceive of attitude functions as variables have been used to craft tailored messages. The tailoring research using these methods will be examined next.

  The Functional Matching Hypothesis

  Katz (1960) argued that an important application of research on attitude functions is tailoring messages to functions. He believed that knowing what function an attitude served would enable the persuader to focus suasory effort on arguments relevant to the needs of the audience. His proposal has come to be known as the functional matching hypothesis. Since the 1980s, a great deal of research has been conducted examining the effectiveness of matching messages to functions using strategies based on three methods of targeting attitude functions identified by Herek (1986).

  Personality

  The most common method of targeting attitude functions began with Smith and colleagues’ (1956) assumption that attitude functions are associated with personality. The earliest research in this area was done by Katz and his colleagues (Katz, McClintock, & Sarnoff, 1957; Stotland et al., 1959), and their work attempted to measure authoritarianism using a variety of measures (California F-Scale, MMPI conformity items, thematic apperception tests, and sentence completion measures). They believed that those who were more authoritarian would be more likely to have ego-defensive attitudes toward racial minority groups. The variety of tests did not offer convergent validity of their construct; some were consistent with the effectiveness of targeting ego-defensive attitudes and some were not.

  More consistent results have been found using self-monitoring as an indicator of social-adjustive attitudes (see Debono, 2000, for a review). Initially, it is worth noting that the self-monitoring scale (Snyder, 1974; Snyder & Gangestead, 1986), the primary means of assessing the trait, lacks construct and content validity (Dillard & Hunter, 1989). Despite these measurement problems, a number of studies have used the scale to predict attitude functions. Snyder and Debono (1985) conducted the initial experiment, arguing that because self-monitoring is a personality construct that is associated with adapting oneself to gain the social approval of others, those high in this trait would be more likely to have highly social-adjustive attitudes. They hypothesized that those low in self-monitoring would hold more value-expressive attitudes because low self-monitors tend to look inward for their attitudes rather than to the expectations of others. They used scores on the self-monitoring scale to create two groups, one high in self-monitoring and one low in self-monitoring. They found that for low self-monitors, advertising messages that focused on product quality were substantially more persuasive than those that focused on social status; but that for high self-monitors, advertising that focused on social status was more persuasive than advertising that focused on product quality.

  Following their seminal work, a number of studies have used self-monitoring to explore the types of messages that match social-adjustive and value-expressive attitudes (e.g., Debono, 1987; Debono & Rubin, 1995; Debono & Snyder, 1989). Debono and Rubin (1995) found that high self-monitors preferred cheese from France to cheese from Kansas regardless of taste, but that low self-monitors preferred the cheese with superior flavor, being relatively uninfluenced by the country of origin. They argued that high self-monitors responded to the country of origin because imported cheese was perceived to be of higher status than domestic cheese, whereas low self-monitors were influenced by their reaction to the cheese’s taste. Debono and Snyder (1989) found that high self-monitors preferred an attractive automobile, whereas low self-monitors preferred an ugly automobile. They reasoned that the high self-monitors had social-adjustive attitudes and therefore chose their automobiles based on what would impress others. They claimed that the low self-monitors chose the less attractive car because they associated less attractive cars with more reliable cars.

  Although it makes intuitive sense that high self-monitors would be more likely to have social-adjustive attitudes, it is less clear what to expect from low self-monitors. Most scholars investigating self-monitoring and attitude functions argue that low self-monitors will be more likely to have value-expressive functions (Debono, 1987; Debono & Harnish, 1988; Debono & Telesca, 1990; Petty & Wegener, 1998; Snyder & Debono, 1985). Yet, these studies target value-expressive attitudes with utilitarian appeals that focus on the quality of products rather than on an abstract value. Given the results that most investigators have reported using utilitarian appeals with low self-monitors, some (Dutta-Bergman, 2003, Shavitt, Lowrey, & Han, 1992) have argued that low self-monitors’ attitudes are more likely to be utilitarian than value-expressive. On the other hand, Lavine and Snyder (1996) were able to show that for low self-monitors a value-based appeal was more successful than an image-based appeal. It appears that although high self-monitors are consistently able to be persu aded using messages targeting the social-adjustive function, the functions of attitudes served by low self-monitors are both conceptually and empirically less clear. Furthermore, Hullet and Boster (2001) found that measures of value-expressive functions and social-adjustive functions were uncorrelated. This finding suggests that there are people who are able to possess attitudes that are both social-adjustive and value-expressive, one but not the other, or neither. Additionally, as noted previously, Herek’s (1986) neofunctional approach emphasized the possibility of multifunctional attitudes.

  Attitude Objects

  Although the personality approach to studying attitude functions remains the most popular, some have examined Herek’s (1986) suggestion that attitudes toward some topics or objects are inherently likely to serve single functions. For example, Shavitt and colleagues have pursued this path (Shavitt, 1990; Shavitt, Lowrey, & Han, 1992). Shavitt noted that in Snyder and Debono’s (1985) self-monitoring studies, there were several consumer products that did not produce the results predicted by the functional matching hypothesis. She proposed that some objects (e.g., aspirin) are likely to serve, and only serve, a utilitarian function; whereas, other objects (e.g., automobiles) might serve a variety of functions, at which point one’s personality might be more likely to predict attitude functions. She reasoned that for unifunctional objects, persuasive messages that matched those functions would be more persuasive than those that did not. After using a thought-listing task coupled with questionnaire data to identify the predominant function served by various objects, she gave subjects the task of reading advertisements for some of those objects. The advertisements either matched or mismatched the objects’ function. Results indicated that the typical functional matching effect occurred; the advertisements were more persuasive when they matched the functions associated with their products. In subsequent experiments, Shavitt et al. (1992) found that self-monitoring exerted a strong effect on the type of functions associated with particular consumer products only when the product was multifunctional. It appears that personality differences are only useful for predicting attitudes for products that are likely to serve multiple functions. No matter how high one might score on the self-monitoring scale, one wants aspirin to alleviate a headache, not to impress friends.

  Situation

  Some investigators have attempted to determine if situations can be structured to increase the salience of attitude functions. Shav
itt, Swan, Lowry, and Wänke (1994) administered a survey that either highlighted sensory experiences or social concerns in order to induce either utilitarian functions or social-adjustive functions. Utilitarian functions were primed by asking the subjects to rate how good or bad a variety of experiences made them feel on scales ranging from “makes me feel very bad” to “makes me feel very good” (p. 143). Social-adjustive functions were primed by asking the subjects to rate a number of events based on how good of an impression the event would make on hypothetical observers of the event. They found evidence consistent with a functional matching effect for the functions they primed.

  Katz (1960) noted, “The most general statement that can be made concerning attitude arousal is that it is dependent upon the excitation of some need in the individual, or some relevant cue in the environment” (p. 176). Following Katz, Julka, and Marsh (2005) argue that the best way to induce any specific attitude function is to heed Katz’s (1960) suggestion that attitudes change when the need that the operative attitude function serves is frustrated. Rather than attempting to measure pre-existing functions, they suggest frustrating the operative attitude function, and then providing a persuasive message (organ donation in this case) that would allow the audience to satisfy the need that had been frustrated. To induce a need to express attitudes consistent with their values, some subjects were given a rigged values survey that indicated that they were not living up to their values. The investigators believed that this tactic would increase their motivation to couple their attitudes and their values. To induce the knowledge function, some subjects were given confusing instructions for a card game. They found evidence consistent with the functional matching effect such that targeting messages at the value function was more persuasive for those who had value-expressive motivation induced than those who had knowledge motivation induced. Alternatively, the knowledge targeted message was more persuasive for those who had the knowledge motivation induced than those who had the value-expressive motivation induced. Furthermore, they found this functional matching effect was stronger for the conditions in which these functions were frustrated than other conditions in which these functions were merely made salient by using a simple values survey with no feedback (value-expressive functions were salient but not aroused) or giving them easy to understand instructions for a card game (knowledge function were salient but not aroused).