Xem mẫu

P A R T I V INTERPERSONAL SYSTEM IANttTitEuRdPe ECRhSaOngNeAL SYSTEM C H A P T E R 2 4 Attitude Change GREGORY R. MAIO GEOFFREY HADDOCK Attitudes are important because they involve large parts of our mental life. When we really like something or someone, our relevant feelings, beliefs, and behaviors are all swayed in a positive direction, and it is difficult for people to be “objective” and ignore these tendencies. This potent effect of attitudes is the reason why people often think of attitudes as the principal cause and the principal consequence of important social problems. For instance,ifpeoplewishtoexplainbitterfightingbetween two ethnic groups (e.g., Shia and Sunni Muslims or Serbs and Croats), the first explanation usually considered is “ethnic hatred,” which is simply a negative attitude to-ward the other ethnic group. Consequently, if people seek a means to reduce the conflict, they will suggest waystotryandamelioratethebitterintergroupattitudes. The success of such interventions to change attitudes may vitally depend on developing a thorough under-standing of the persuasion process. Attempts at persuad-ing people to like another group may work well on some occasions but many terribly backfire on other occasions. To facilitate successful interventions, persuasion re-searchers have long searched for the factors that influ-ence attitude change and for models that can explain how these factors are influential. This search has led to a large corpus of data spanning a variety of domains of ap-plication, including social groups (e.g., Blacks), policies (e.g., tuition fees), social values (e.g., equality), health-relevant behaviors (e.g., smoking), religious behavior (e.g., attending church or synagogue), environmental be-haviors (e.g., recycling), food items (e.g., jams), house-hold utensils (e.g., razors), simple tasks (e.g., choices be- tween objects), and even humorous stimuli (e.g., car-toons). This chapter attempts to summarize some basic principles of persuasion that operate across these do-mains, drawing heavily on well-researched models of per-suasion. To begin, the chapter describes the content, structure, and function of attitudes, which helps to lay the founda-tion for understanding principles of attitude change. We then begin to foreshadow specific principles by review-ing models of attitude change and noting their points of agreement. Next, the chapter considers four basic princi-ples in detail and briefly highlights other potential princi-ples.Finally,wenoterelevantissuesforfutureresearch. ATTITUDE CONTENT, STRUCTURE, AND FUNCTION The first step to understanding attitude change involves comprehending the content, structure, and function of attitudes. These terms have had varied uses in the past. For example, the phrase “attitude structure” has been used to reflect both the capacity of attitudes to reflect dif-ferent beliefs and emotions and the capacity to reflect different combinations of positivity and negativity (Eagly & Chaiken, 1993). Similarly, the term “attitude function” is used to denote specifically the needs served by express-ing or possessing attitudes, which are often related to the function served by the object of the attitude, but not al-ways (Kruglanski & Stroebe, 2005; Shavitt, 1990). In this chapter, the phrase “attitude content” refers to the ways 565 566 INTERPERSONAL SYSTEM in which attitudes may express different psychological constructs, such as beliefs and emotions. The phrase “at-titude structure” refers to the ways in which attitudes summarize positivity and negativity toward the attitude object. The phrase “attitude function” refers to the ways in which the attitudes themselves may serve different psy-chological motivations, such as needs to be correct or to be liked by others. By understanding the content, struc-ture, and function of attitudes, it becomes clearer what attitude change attempts must achieve in order to claim success at eliciting attitude change. So, what contents, structure, and functions must be addressed? Attitude Content Research on the content of attitudes has been dominated by two perspectives: the three-component model and the expectancy-value model. According to the three-component model, attitudes express people’s feelings, beliefs, and past behaviors regarding the attitude object (Zanna & Rempel, 1988). For example, both authors of this chapter disliked the first two Oscar-winning films in the “Lord of the Rings” series because they thought the films were boring and abhorrent (affective component) and believed that the films were not particularly innova-tive and overly violent for children (cognitive compo-nent). Nonetheless, one of the authors sat through the third film in this series (behavioral component), despite his negative affective and cognitive reactions. The three-component model predicts that his behavior is also an important part of his attitude toward these films because, to some extent, it is difficult to believe that a person com-pletely dislikes something that he or she has endured three times. Through the process of self-perception (Bem,1972;Olson,1990,1992)andcognitivedissonance (Festinger, 1957), people tend to decide that they like something when they can recall doing it often. This example points out an interesting feature of atti-tudes. Specifically, it is not necessary that all the attitude components express the same sentiment toward the ob-ject of an attitude. People’s feelings, beliefs, and behav-iors may imply conflicting evaluations of the attitude ob-ject, and this situation is often labeled “intercomponent ambivalence” (MacDonald & Zanna, 1998; Maio, Esses, & Bell, 2000). There may also be intracomponent ambiva-lence such that people possess both positive and negative feelings,beliefs,andbehaviors.Theamountofevaluative conflict within and between components can vary as a function of the attitude object (Bell, Esses, & Maio, 1996), the attitude holder (Thompson & Zanna, 1995), and the situation (Bell & Esses, 2002). Moreover, the net attitude itself may match or mismatch the evaluations in the components on which it is based (Chaiken, Pomerantz, & Giner-Sorolla, 1995). Nevertheless, there is also a press toward “synergy” among the attitude components (Eagly & Chaiken, 1993; Zanna & Rempel, 1988). That is, despite the frequent ex-istence of ambivalence, people also tend to have positive attitudes toward an object when their feelings, beliefs, and behaviors all express favorability toward an object. In contrast, people tend to have negative attitudes to- ward an object when their feelings, beliefs, and behaviors express unfavorability toward the object. This consis-tency may occur because, in general, internal psy-chological conflict is aversive to us (Festinger, 1957). Consequently, we seek information that confirms our preexisting beliefs, feelings, and behaviors (Frey, 1986; Frey & Stahlberg, 1987), while also processing new infor-mation in a manner that might help to reduce ambiva-lence (Jonas, Diehl, & Bromer, 1997; Maio, Bell, & Esses, 1996; Maio, Greenland, Bernard, & Esses, 2001). If a result of these information-processing biases is consistency between the components, it could be argued that we should adopt a model of attitude content that fo-cuses simply on one of the components. Some research-ers also argue that a model focusing on one component alone has the advantage of simplicity, because there is a danger in making attitudes too much of an all-encompassing entity. These factors buttress the argu-mentformodelsofattitudesthatfocusonthebeliefcom-ponent. According to models with this focus, attitudes are based on beliefs that are evaluative in nature, and the most interesting issue is how people combine the beliefs to form an attitude (McGuire, 1960b; Wyer, 1970). This issue is central to the expectancy-value models of atti-tudes, such as the theory of reasoned action (Ajzen & Fishbein, 1980). This perspective proposes that an atti-tude is the sum of all the evaluative beliefs regarding the attitude object. For instance, if people believe that a new snack product is healthy and good tasting, they should hold a positive attitude toward it. Nevertheless, beliefs are rarely held with perfect certainty. For example, a per-son may be only 70% certain that the snack is healthy but 100% certain that it is good tasting. According to the expectancy-value model, the effect of each belief is a functionofthecertaintywithwhichitisheld.Thisnotion is expressed using a well-known formula: A = Σ be, where A is the total attitude toward the attitude object, b is the subjective belief that the object possesses attribute i (e.g., the probability that the snack is healthy) and e is the eval-uation of attribute i (e.g., the positive value attached to health promotion). Research has examined the utility of the three-component and expectancy-value models in different ways but has obtained support for both approaches. With regard to the expectancy-value approach, evidence indicates that people’s reports of their own attitudes are at least moderately correlated with the summed products of the attitude-relevant expectancies and values (Bagozzi, 1981; Budd, 1986; van der Pligt & de Vries, 1998), al-though there have been statistical and methodological criticisms of these findings (Sparks, Hedderley, & Shep-herd, 1991). In addition, it is clear that persuasive mes-sages influence evaluative beliefs and that these beliefs influence attitudes when people are motivated and able to process persuasive messages in a systematic manner (Chaiken, Liberman, & Eagly, 1989; Petty & Cacioppo, 1986). With regard to the three-component model, evi-dence indicates that people’s beliefs, feelings, and behav-iors toward an attitude object are correlated but distinct (Breckler, 1984; Crites, Fabrigar, & Petty, 1994; Haddock & Zanna, 1998). There is also evidence that attitude- Attitude Change 567 relevant feelings and beliefs are clustered separately in memory (Trafimow & Sheeran, 1998) and that the affec-tive and cognitive components each account for unique variance in attitudes (Breckler & Wiggins, 1989; Brinol, Petty, & Tormala, 2004; Crites et al., 1994; Trafimow & Sheeran, 1998). The findings in support of the three-component model could be rebutted by the argument that affective reactions and past behaviors are simply different types of beliefs about the attitude objects. For example, people may form a positive attitude toward a snack that tastes good (an affective response) because, as the expectancy-value model proposes, people believe that the bar makes them happy and value their own happiness. Neverthe-less, this “affective belief” is made salient only by con-sidering the three-component model. Thus, the three-component model at least spurs the expectancy-value approach to consider different types of beliefs, which must be altered during the persuasion process. Furthermore, as we shall see, there is evidence that ap-peals that are based on affective or cognitive information may be differentially effective at changing attitudes that are based primarily on affective information or cognitive information. Specifically, appeals can elicit more attitude change when they contain content that matches the con-tent of the message recipient’s attitude than when they contain content that mismatches the content of the atti-tude. For example, affective interventions influence affectively based attitudes more than cognitively based attitudes (Edwards & von Hippel, 1995; Fabrigar & Petty, 1999; Huskinson & Haddock, 2004). These effects form the basis for the third principle of persuasion that we de-scribe below, while also providing further support for two components of the three-component model. Attitude Structure An important issue is how positive and negative evalua-tions are organized within and between components. The intuitive and most frequently assumed possibility is that the existence of positive feelings, beliefs, and behav-iors inhibits the occurrence of negative feelings, beliefs, and behaviors. In other words, the positive and negative elements are stored in memory at opposite ends of a sin-gle dimension, and people tend to experience either end of the dimension or a location in between. This one-dimensional view is opposed by a two-dimensional view, which suggests that positive and nega-tive elements are stored along two separate dimensions (Cacioppo, Gardner, & Berntson, 1997; Thompson, Zanna, & Griffin, 1995). One dimension reflects whether people have few or many positive elements, and the other dimension reflects whether people have few or many negative elements. This view proposes that people can possess any combination of positivity or negativity in their attitudes. Consistent with the one-dimensional view, attitudes may subsume little positivity and high negativity, little negativity and high positivity, or no positivity or negativity (i.e., a neutral position). Inconsistent with the one-dimensional view, attitudes might occasionally subsume high amounts of positivity and negativity, leading to attitudinal ambivalence. The bidimensional perspective explicitly allows for this ambivalence to occur, whereas the one-dimensional per-spective does not. As a result, the bidimensional perspec-tive challenges the meaning of the neutral point in one-dimensional scales for assessing attitudes (Kaplan, 1972). In these scales, a neutral response (e.g., neither like nor dislike”) is halfway between the most extreme positive re-sponse option (e.g., “extremely likable”) and the most ex-treme negative response option (e.g., “extremely dislikable”). People could choose this option because it is a compromise between many positive and negative ele-ments of their attitude or because they have no positive or negative elements whatsoever. This failure to distinguish between these two reasons for the neutral selection is consequential, because mea-sures that directly assess ambivalence predict a variety of outcomes. The best known outcome is response polariza-tion (Bell & Esses, 2002; Katz & Hass, 1988; MacDonald & Zanna, 1998): People who are highly ambivalent to-ward an object are more strongly influenced by features of their environment that make salient its positive or neg-ative attributes, causing them to behave more favorably toward it when the positive elements are salient than when the negative elements are salient; nonambivalent people are less strongly influenced by the acute salience of the positive or negative attributes. More relevant to this chapter, ambivalence predicts attitude change, be-cause ambivalent attitudes tend to elicit more message scrutiny (Jonas et al., 1997; Maio et al., 1996). Hence, am-bivalent people exhibit more favorable attitudes in re-sponse to strongly argued messages than in response to weakly argued messages, whereas nonambivalent partici-pants show little sensitivity to argument strength. Thus, ambivalence may enhance the motivation and ability to carefully consider the merits of persuasive information, causing more relevant information to be relied on more than less relevant information. Ambivalence is therefore one of many variables that is relevant to another basic principle of persuasion that we describe next. This principle describes a tendency for the motivation to be correct to elicit greater use of relevant information in the persuasion context. Indeed, there is evidence that people who tend to experience ambiva-lence are more likely to fear invalidity in their judgments (Thompson & Zanna, 1995). Moreover, it is possible that ambivalence is relevant to the matching principle men-tioned previously. That is, ambivalent individuals may be more persuaded by messages that recognize their ambiv-alence by presenting both sides of the arguments than by one-sided messages, whereas nonambivalent individuals may be more persuaded by one-sided messages than by two-sided messages. At present, however, we know of no evidence directly examining this hypothesis. Attitude Function Theorists have proposed that attitudes fulfill a variety of psychological needs or functions (Katz, 1960; Smith, Bruner, & White, 1956). Smith and colleagues’ (1956) object-appraisalfunctionisoftenconsideredthemostba- ... - tailieumienphi.vn
nguon tai.lieu . vn