The SAGE Handbook of Persuasion
Page 13
Of course, attitudes that are more accessible from memory are more likely to be activated and to influence behaviors. Fazio (1986) proposed a process model of the attitude-behavior relationship that basically argues that accessible attitudes are more likely to be activated, and once activated, they will bias how information within the environment is perceived, which then influences a person’s behavior. If you have a highly accessible negative attitude toward cockroaches and you see a cockroach, that attitude will be activated. The activated negative attitude will lead to the judgment that this is an aversive situation and you should act accordingly by either leaving the situation or doing something to remove the cockroach, such as stepping on it.
Consistent with the MODE model, accessible attitudes most often affect behavior in situations that allow or require spontaneous decisions (e.g., deciding which gas station to turn into when your tank is on empty), or in ambiguous social situations (Fazio, 1986, 1990). With situations that facilitate or require spontaneous behaviors, highly accessible attitudes will likely determine which evaluations or objects come to mind when a behavioral decision must be made. For example, when research participants were asked to which charity they would donate monetary compensation for research participation, they chose the charities for which their attitudes were most accessible, even though they had evaluated other charities just as or more favorably in a pretest (Posavac et al., 1997). However, in situations that require accurate decisions or a consideration of norms, we might reflect on the underpinnings or cognitive components of our attitude toward the behavior, social norms regarding the behavior, and the unique characteristics of the situation and then act accordingly (Ajzen, 1991; Fazio, 1990). Hence, our behavior might not be strongly correlated with our attitudes in more deliberative situations that give us the opportunity to reflect on those attitudes. However, even in situations that involve careful, deliberative decision making, accessible attitudes can influence decisions (Roskos-Ewoldsen, 1997; Roskos-Ewoldsen, Yu, & Rhodes, 2004). Consistent with this, Roskos-Ewoldsen et al. (2004) found that women with more accessible attitudes toward breast self-exams were more likely to intend to perform breast self-exams in the future.
According to the transactive model of attitude accessibility, accessible attitudes tend to operate in a manner that maintains their accessibility through frequent activation and through elaboration. For example, because accessible attitudes orient our attention to objects in our environment, they are reactivated or rehearsed via the automatic judgment that occurs when one attends to a liked or disliked object (Roskos-Ewoldsen, 1997). Likewise, direct experience with an attitude object makes attitudes more accessible from memory, which can easily and consistently initiate the process from accessibility to attention to behavior that constitutes the remaining section of the transactive model (DeBono & Snyder, 1995; Fazio, 1986). For example, in a recent study, cigarette smokers showed more biased processing of antismoking messages than nonsmokers (Rhodes & Ewoldsen, 2009). In addition, as the smokers judged the antismoking ad to be more biased, they also indicated they were less likely to quit smoking. In other words, the antismoking ad created reactance in the smoker, which reinforced their desire to smoke. In this study, it is important to note that the accessibility of the smokers’ attitudes mediated this process. It was smokers with more accessible attitudes that judged the antismoking ads as more biased and were less likely to want to quit smoking. The accessible prosmoking attitude operated as a defense mechanism for these smokers to protect their smoking behavior from threats such as antismoking ads, and their accessible attitudes also strengthened the behavioral response to continue smoking (Arpan, Rhodes, & Roskos-Ewoldsen, 2007).
Implicitly Measured Attitudes
There has been a surge of interest in the social sciences in the study of implicitly measured attitudes (e.g., Bassili & Brown, 2005; Petty, Fazio, & Brinol, 2009; Wittenbrink & Schwarz, 2007). The designation of “implicit attitude” means that the attitude is measured without asking someone to explicitly indicate their attitude, in contrast to commonly used paper-and-pencil measures of attitudes (DeHouwer, 2006; Fazio & Olson, 2003). An example of an implicit measure of an attitude is the affective misattribution procedure (AMP; Payne, 2009), which involves presenting participants with a picture of a critical item (e.g., a bottle of beer), which is quickly followed by a picture of a Chinese ideograph. The participants’ task is to indicate whether they like or dislike the ideograph, participants are never asked to indicate their attitude toward the critical item (the bottle of beer). Payne, Cheng, Govorun, and Steward (2005) demonstrated that participants’ attitude toward the target (the beer) influenced how they responded to the ideograph. The attitude toward the target influenced responses to the target even when participants were explicitly told they should not allow that to occur. However, as DeHouwer (2006) has noted, none of the implicit measures can ensure that participants are not considering their attitudes toward the target item when they are responding. The critical point is that implicit measures either never make it explicit that people’s attitudes toward a certain object is being measured or (as is the case with some research using the AMP) instruct participants to ignore their attitude toward the target item when responding.
In theory, implicit measures of attitudes may have advantages over explicit measures for topics where social desirability concerns may result in participants not wanting to express their true attitudes. A consistent finding in the literature is that explicit and implicit measures of attitudes, although moderately correlated overall, diverge when the attitude target is one where there are strong social motivations for reporting a particular attitude (Greenwald & Nosek, 2009). For example, implicitly measured attitudes have been used quite extensively in the study of racial attitudes because explicitly measured racial attitudes have been shown to be strongly influenced by motivational factors (e.g., Amodio & Devine, 2009; Fazio, Jackson, Dunton, & Williams, 2005).
The extensive research on implicit measures of attitudes has resulted in some scholars arguing that there are two separate attitude systems—an explicit and an implicit system of attitudes (Greenwald & Nosek, 2009; Wilson, Lindsey, & Schooler, 2000). According to this dissociative view, attitudes that are measured implicitly are not strongly correlated with explicitly measured attitudes because the underlying attitudes measured by the different procedures represent distinct attitudes within two distinct cognitive systems. More specifically, implicitly measured attitudes are assumed to reflect a system of attitudes that operate outside of people’s awareness, are automatically activated from memory (see earlier discussion on attitude accessibility), and are beyond people’s control. In contrast, explicitly measured attitudes are hypothesized to reflect a more deliberative system that is available for people’s introspection and can be controlled. Evidence for these distinct attitudes comes from research suggesting that the two different types of attitudes develop via different mechanisms. For example, research suggests that implicitly measured attitudes are more influenced by socialization, as well as associations with the self and culture, than are explicitly measured attitudes (Devos, 2008).
A more parsimonious explanation is found in the MODE model (Fazio & Olson, 2003; Olson & Fazio, 2009). When attitudes are measured with different methods, it is important to consider how motivation and opportunity might influence those measurements (Olson & Fazio, 2009). With self-report or explicit measures of attitudes, individuals are more aware of the content of the questions and are rarely placed under time constraints. Neither motivation nor opportunity to respond are constrained, which should result in more deliberative judgments of the attitude object, and motivational factors can influence the attitude that is reported. In contrast, with implicit measures of attitudes, individuals are typically asked to make rapid judgments of seemingly unrelated stimuli. Both opportunity and motivation to respond are limited, which results in more spontaneous judgments of the attitude object (Olson & Fazio, 2009). For most topics, explicitly measured attitudes
are correlated with implicitly measured attitudes because respondents are not motivated to report an attitude that is different from the attitude that is activated from memory.
However, for socially sensitive topics such as prejudice, explicitly and implicitly measured attitudes often diverge. According to the MODE model, this divergence reflects individuals’ motivation and opportunity. When responding to implicit measures of attitudes, prejudiced respondents are not motivated to change their response because they are unaware of what attitude is being measured. Further, they have no opportunity to control their response because most implicit measures of attitudes place participants under time pressure to respond. However, on explicit measures where it is clear the goal of the scale is to measure racial attitudes, individuals with racist attitudes are motivated to mask their true feelings so as not to appear prejudiced (Eno & Ewoldsen, 2010; Fazio, Jackson, Dunton, & Williams, 1995). In addition, they have the opportunity to control their response because explicit measures of attitudes typically allow participants unlimited time to complete the scale. Consequently, respondents will make more deliberative judgments and respond with a non-prejudiced attitude.
This discussion of explicitly and implicitly measured attitudes probably seems far removed from the central concern of this chapter—the outcomes of persuasion. Research on persuasion-based attitude changes has almost exclusively focused on explicit measures of attitudes, so why address implicitly measured attitudes in this chapter? We discuss this distinction because we believe this distinction will become increasingly important for persuasion scholars. Recent research suggests that implicitly measured attitudes can be changed through communication (Brinol, Petty, & McCaslin, 2009; Eno & Ewoldsen, 2010; Maio, Haddock, Watt, & Hewstone, 2009). For example, Eno and Ewoldsen (2010) demonstrated that people’s implicitly measured attitudes toward African Americans became less prejudiced after watching a movie that attacked racist attitudes toward African Americans. Perhaps more important, given the focus of this chapter on deliberative versus spontaneous processes as outcomes of persuasion, research very clearly demonstrates that, consistent with the MODE model, implicitly measured attitudes are more predictive of spontaneous behaviors and explicitly measured attitudes tend to be more predictive of deliberative behaviors (Devos, 2008; Fazio et al., 2005).
Norms as Outcomes
* * *
There are a number of reasons to consider the role of norms in persuasion contexts. First, norms are important in models of behavior maintenance and change (e.g., theory of reasoned action/theory of planned behavior/integrative model, Ajzen, 1991; Fishbein & Yzer, 2003; process model, Fazio, 1986), thus considering their role both in the persuasion process and in the maintenance of behavior has theoretical value. More fully understanding these processes holds promise in a range of contexts. For example, a young teen who is offered a cigarette at a party may have to quickly determine the relative social costs and benefits of taking the cigarette or not. However, relatively little research has explored the psychological mechanisms underlying social influence (Cialdini, Kallgren, & Reno, 1991; Glynn, 1997). Indeed, recent efforts to change health risk behaviors through normative appeals has emphasized the importance of studying the processes through which norm appeals are (or are not) effective.
Deliberative Approaches: Social Norms Campaigns
An intriguing line of research that began in the context of trying to reduce episodes of binge drinking on college campuses focused on people’s normative beliefs about engaging in a behavior. Early work in this stream observed that there tends to be an exaggerated norm for many risky behaviors young people might consider performing, and especially for drinking on college campuses (Miller, Monin, & Prentice, 2000; Prentice & Miller, 1993; Schroeder & Prentice, 1998), with the general finding that college students individually tend to be less accepting of alcohol use than they perceive their peers to be. It is important to note that those who see the norm as most accepting of alcohol use tend to drink the most (Borsari & Carey, 2003; Park, Smith, Klein, & Martell, 2011).
The consistency with which these findings have been found led to the logical extension that one way to reduce college drinking behavior is to inform college students of the actual norm, specifically, that the norm for drinking alcohol is less permissive than they believe it to be. On this assumption, many campuses began social norms campaigns, in which it is emphasized that the majority of students on campus do not engage in binge drinking. In spite of some initially promising results (Haines & Spear, 1996), these norms campaigns have produced mixed findings (e.g., Campo & Cameron, 2006), and have been heavily criticized (Wechsler et al., 2002). There are problems with both the implementation of these programs and the evaluation of them, which renders making judgments about their effectiveness problematic, however. Specifically, there is a great deal of ambiguity surrounding the concept of norms. Various researchers have discussed descriptive norms, injunctive norms, subjective norms, moral norms, and customs. Each of these ways of looking at norms has its strengths, but overall, the use of these diverse operational definitions makes it difficult to draw any firm conclusions about how norms affect behavior (Rimal & Real, 2003; Schultz, Nolan, Cialdini, Goldstein, & Griskevidius, 2007).
Additionally, it is unclear which norms are being targeted in the various campaigns, and whether one would expect a change in such norms to affect individual behavior. Specifically, most of the campaigns have emphasized a campuswide descriptive norm (e.g., 80% of students at X University drink three or fewer drinks when they go out). It is unclear that students make decisions about drinking on the basis of a descriptive norm (how many other students drink) as opposed to an injunctive norm (what social costs and benefits accrue from drinking). It is likely that judgments about the norms for the campus subgroup that is central to one’s identity are more likely to play a role in one’s own behavior (Terry & Hogg, 2001). In particular, it is clear from much of the work on campus drinking patterns that certain subgroups of students are at significantly elevated risk (e.g., fraternities), and that those subgroups are characterized by both descriptive and injunctive norms that are strongly supportive of high levels of drinking (Phua, 2011). Certainly the publication of a campuswide norm showing less drinking would be unsuccessful for such groups. Further work to clarify how norms affect behavior will help to target norms’ messages more effectively.
Spontaneous Approaches: Norm Accessibility and Implicit Norms
Work by Cialdini and colleagues on the focus theory of norms (Kallgren, Reno, & Cialdini, 2000) found that social norms can be a strong determinant of behaviors such as littering when the norm is salient in the situation (Cialdini et al., 1991; Reno, Cialdini, & Kallgren, 1993). In this research, the manipulations of norm salience temporarily increase the accessibility of the norm in memory. In several studies, the injunctive norm against littering was primed through a manipulation of the content of leaflets that were left on participants’ windshields, which affected littering behavior. Cialdini and his colleagues (1990) interpret this finding by arguing that the priming manipulation increased the temporary accessibility of the injunctive norm against littering. As described in the section on accessible attitudes, a specific behavior is more likely to be driven by mental constructs that are activated, or made accessible, at the time the behavioral act is performed (Fazio, 1986, 1990; Kallgren et al., 2000). This idea has led to the examination of norm accessibility, or how quickly some can activate from memory his or her judgment of the social costs and benefits of engaging in a particular behavior (Rhodes & Ewoldsen, 2009).
The concept of norm accessibility also draws from social identity theory (and related literatures) that emphasize the extent to which membership in a specific group is central to the self-concept of the actor will determine how accessible the associated norm is for that behavior (Terry & Hogg, 2001). Research has shown that groups that are important to one’s identity are more influential on individual attitudes and behaviors (Terry, Hogg, & White, 2000). Norms associ
ated with important in-groups are likely to be chronically more accessible than those associated with less important groups, and thus are more likely to be activated and guide behavior. Thus, important in-groups may be associated with both chronically and temporarily accessible norms.
As in our discussion of attitude accessibility, norm accessibility appears to play an important role in spontaneous processes related to persuasion. The process model (Fazio, 1986) focuses on the role of attitude accessibility in predicting behavior. In the fully described model, accessible attitudes toward the object, along with social norms, influence how people perceive and define social situations (Fazio, 1986). How people define the situation influences how they act in that situation. Numerous studies have demonstrated that accessible attitudes influence how people perceive and interpret social information (e.g., Fazio et al., 1994; Roskos-Ewoldsen, Yu, & Rhodes, 2004). However, the influence of norms has only recently been explicitly tested within this model. This stream of research has found that accessible norms are predictive of behavior and intention to engage in behavior (Rhodes & Ewoldsen, 2009). In a sample of college students, Rhodes et al. (2008) showed participants four different antismoking public service announcements (PSAs). Using cognitive response data, we found that smokers with accessible prosmoking subjective norms avoided processing antismoking messages and instead focused on heuristic cues to discount the message, suggesting that health risk behaviors that are supported by accessible norms may be particularly resistant to change.
Extensive research has demonstrated the important role that social norms play in influencing people’s behavior. The lack of research specifying the cognitive and affective mechanisms underlying the influence of norms is unfortunate given the importance of norms in explaining social behavior. We propose that norm accessibility, that is, the ease of accessing normative information from memory, provides an explanatory mechanism for understanding the influences of norms on social behavior. Specifically, when people can easily bring to mind that there are important people in their lives who support health risk behaviors such as cigarette smoking, they are less likely to even think about a message opposing the behavior that is supported by that norm. This has important implications for health promotion campaigns: addressing health risk behaviors prior to the establishment of a supportive norm is the key to making that message heard by the audience. Additionally, as was recently found by Rhodes and colleagues, young teenagers who have accessible antirisk family norms engage in less risk behavior (Rhodes et al., 2011). This emphasizes the role of the family in teenagers’ lives and suggests that parents of young teens should talk frequently with their teens about their desire that the teen avoid engaging in risky behavior.