Seeing versus Feeling Threats: Group Cues, Emotions, and Activating ...

5 downloads 6951 Views 190KB Size Report
Group Cues, Emotions, and Activating Opposition to Immigration ..... York Times report about a governors' conference on immigration.10 We also manipulated ..... outgroup cues, because what we call anxiety is just another way of measuring ...
Seeing versus Feeling Threats: Group Cues, Emotions, and Activating Opposition to Immigration Ted Brader Nicholas A. Valentino Elizabeth Suhay University of Michigan

Abstract Prior research suggests that threat is a potent source of attitude change. Few studies, however, examine the psychological mechanism underlying these effects or explain what makes some threats more effective than others. We posit that not all threats are equal. Information may alter awareness of a threat, yet fail to elicit feelings of anxiety. We argue that the capacity of threatening information to influence behavior depends greatly on eliciting negative emotions and that group cues are common attributes of political communication with the potential to trigger such emotions. Focusing on immigration, we test these hypotheses by embedding an experiment in a national survey. Results confirm that cueing a stigmatized outgroup (Latino immigrants) generates greater anxiety in response to negative information about immigration and this anxiety mediates changes in preferences and action. A second experiment traces the anxiety-arousing power of these cues to stereotypes of immigrants as ethnically Latino and low-skilled.

Acknowledgements Data collection for this paper was made possible by Time-sharing Experiments for the Social Sciences (TESS) and grants from the Howard R. Marsh Center for the Study of Journalistic Performance and the Horace H. Rackham School of Graduate Studies. Earlier versions of this paper were presented at the 2004 annual meetings of the Midwest Political Science Association and American Political Science Association. assistance.

Antoine Banks, Andrea Benjamin, and Eric Groenendyk provided excellent research

Political scientists have uncovered considerable evidence that news, campaigns, and other forms of political communication can influence citizens. We still know little, however, about what makes some messages more persuasive than others. Many studies emphasize the volume and distribution of massmediated information, on the idea that attitude formation is a function of the accessibility of relevant considerations (Iyengar and Kinder 1987; Zaller 1992). Others suggest that certain types of information are especially effective. Research in a variety of domains regards “threat” (i.e., the implication of an undesirable outcome) as a particularly potent source of attitude change and action (Campbell 2003; Feldman and Stenner 1997; Gordon and Arian 2001). In addition, a growing body of work suggests that the capacity of politically-relevant stimuli to persuade citizens depends heavily on evoking the feeling of being threatened, namely anxiety or fear (Brader 2005; Marcus, Neuman, and MacKuen 2000; Nadeau, Niemi, and Amato 1995). The fact that research on “threat” is scattered across disparate fields of politics and largely disconnected from recent work on emotions begs the question: Are studies of threat and studies of fear investigating different phenomena, the same phenomena with different labels, or different aspects of the same phenomena? We argue that threatening stimuli, perceptions of threat, and fearful reactions are distinct elements of the same causal process and that these conceptual distinctions are important for understanding the impact of many forms of political communication. Previous work has largely assumed equivalence between the perception of a threat and the emotional experience of feeling threatened. Researchers typically label an implication of present or future harm as a threat and predicate belief in its efficacy on the notion that certain effects follow so long as the “threat” is perceived by the individual. Few, if any, published studies demonstrate that this assumption holds. We wonder, therefore, whether perceptions and emotions are part of the same mechanism, equivalent alternative pathways, or distinct mediators of the impact of threat on attitudes and action. Does perceiving an event or trend as potentially harmful inexorably trigger an emotional state such as anxiety? Are changes in attitudes and behavior caused by the perception of a threat or the emotional experience? We hope to shed light on when and how threatening messages influence citizens by distinguishing between perceptions of threat and emotional responses to threat as potential mediating variables. We posit

1

that all threats are not equal in their capacity to motivate a political response. Threats are most likely to cause attitudinal and behavioral change when they elicit negative emotions such as anxiety and anger. However, this also begs a further question: Which threatening messages are more likely to elicit such feelings? Emotional arousal may be tied directly to information regarding the severity and proximity of the undesirable outcome. But even when information about potential consequences is constant, certain attributes of a message may lend it a greater capacity to arouse emotions. We suspect that group cues are a prime example. Research on mass politics has found few factors as salient and powerful in their ability to shape opinion and behavior as group identities and attitudes (Citrin and Green 1990; Hutchings and Valentino 2004). When cues link threats to stigmatized groups, they should trigger anxiety and thereby render attitudes and behavior more responsive to communication about potential threats. We investigate these propositions in the context of public reactions to news about increasing immigration in the United States. Immigration is an issue with the potential for generating material and symbolic threats, as well as benefits, to citizens. Mass-mediated discourse about immigration also is rich in group cues. We begin by clarifying our view about the psychology underlying responses to threat. This leads to several hypotheses about the effects of threatening messages and the psychological process that mediates them. We test the hypotheses with experiments conducted on both local convenience and representative national samples.

PERCEPTION AND EMOTION IN RESPONSE TO THREAT Research has left little doubt about the power of threats in social and political life. Threats can trigger intolerance (Marcus et al. 1995), authoritarianism (Feldman and Stenner 1997), pessimism (Huddy et al. 2002), vigilance (Green, Williams, and Davidson 2003; Marcus, Neuman, and MacKuen 2000), attitude change (Brader 2005; Hovland, Janis, and Kelley 1953; Way and Masters 1996), and political action (Campbell 2003; Miller and Krosnick 2004).1 Some of these conclusions are based on

1

Scholarship on the effects of threat is undergirded by evidence that individuals weigh negative information more heavily than positive information (Cacioppo, Gardner, and Bernston 1997) and are more sensitive to losses than to gains (Kahneman and Tversky 1979).

2

demonstrations of the impact of threat-related stimuli, while others are based on the impact of perceptions or feelings of threat. Few studies document the full path from (a) a threatening stimulus to (b) the perception or feeling of being threatened to (c) a change in attitudes or behavior. Survey studies that focus entirely on self-reports to test the second causal link (i.e., b → c) often cannot say how much people are responding to the world around them—and, if so, precisely which element of it—versus how much their responses emerge exclusively from their “inner demons.” Similarly, studies that focus entirely on behavioral changes caused by threatening stimuli (i.e., a → c) often fail to clarify the psychological mechanism underlying the effect. In most cases, there is an assumption that external stimuli are causing a threat to be perceived or felt and that this is why behavior changes. We seek to decouple elements of political communication that are potentially threatening from the individual perceptions or feelings of threat they may (or may not) generate. In other words, we hope to demonstrate, rather than assume, whether external forces influence behavior by means of triggering a change in perceptions, an emotional reaction, or both. One reason to “connect all the dots” is intellectual fastidiousness. A second reason is that studies have turned up quite distinct, even contradictory, effects (Brader 2002). Part of the confusion may stem from the fact that investigators have operationalized threat in a wide variety of ways and made different assumptions about the psychological mechanism. Some speculate that fear or other emotions mediate the impact of threats on behavior, while others stress the mediating role of perceptions and do not posit any role for emotions.2 Moreover, we know that threats have the potential to elicit other emotions such as anger, which sometimes, though not always, have different consequences (Feldman, Huddy, and Cassese 2004; Lerner et al. 2003; Marcus et al. 2003). We believe the mechanism matters, but which mechanism—perceptions or feelings—is at work? The type and extent of reaction provoked by a threatening message may depend critically on whether it produces a change in perceptions or an emotional response. Although both are plausible consequences of threat, there is evidence to suggest that threats do not always trigger both (Eagly and Chaiken 1993; Witte

2

Few studies explicitly deny that perceptions or emotions are relevant. Most simply focus on one to the exclusion of the other or rely on measures that blur the boundaries between perceived threat and fear/anxiety.

3

and Allen 2000). In addition, “cognitions” and emotions are strong, yet independent, predictors of political choice (Abelson et al. 1982).3 Attitudes and behavior may logically flow from the ways people see the world, that is, from their perceptions (Hochschild 2000). At the same time, however, the function of emotions is to enable individuals to adapt their behavior to changes in their environment (Damasio 2000; Lazarus 1991; Scherer 1994). In fact, because they are both a signal of relevance and a source of motivation, emotions may be closely tied to behavior (Kinder 1994). Affective intelligence theory posits that anxiety, as a response to threat, stimulates a search for information, attitude change, and political action (Marcus, Neuman, and MacKuen 2000). We argue that the capacity of threatening information to change relevant attitudes and behavior occurs primarily through the emotional channel. For now, we do not rule out the possibility that a person may respond to a change in perceptions alone, but we believe a change in emotions is the stronger, more reliable path to attitudinal and behavioral change. Therefore, our central proposition is that threat in the political environment will be particularly effective at bringing about a shift in behavior and attitudes to the extent it elicits emotions such as fear or anger rather than merely changing perceptions about the nature or severity of the threat itself.4 Obviously, at some level, “perceptions” of threat and “feeling threatened” must be related. But how? Offering a complete answer to that question lies beyond the scope of this paper, because a full test of any predictions would require access to more intrusive types of data. However, we do not mean to imply that perceptions and emotions are entirely independent processes. At the subconscious level, we know that a situation can trigger anxiety only when the brain “perceives” a threat; yet emotions themselves often occur without conscious awareness and therefore without triggering conscious perceptions of threat (Damasio 2000). We also know that emotions and perceptions are likely to exert reciprocal influence, especially by the time they manifest themselves consciously. As a result, it is impossible to test the 3

Scholars often use the term “cognition” to distinguish perceptions, beliefs, and other elements of thought from affective drives and reactions (Zajonc 1998). However, evidence from neuroscience suggests cognition, understood as information processing, is involved in the production of emotions, perceptions, and conscious reasoning. 4 We suspect perceptions may be more closely tied to other “cognitions,” such as causal or attributional beliefs, but our focus in this paper is on the impact of threat on attitudes and behavior.

4

direction and extent of causal influence between emotions and perceptions when relying on data from self-reports, as we do here.5 In fact, psychologists using a wide variety of methods continue to debate the issue of “which comes first” (Lazarus 1991; Zajonc 1998). So while we expect perceptions and feelings of threat to be highly correlated, we must remain agnostic here about the precise manner and extent to which they influence one another. For now, we focus on the potential mediating roles of conscious perceptions and self-reported feelings. Our goal is to determine whether these two mechanisms, perceptions and feelings, operate similarly or differently in mediating the relationship between exposure to threats and attitudinal and behavioral change.

GROUP CUES AND EMOTIONAL RESPONSES Threats come in many forms. Scholars have examined the impact of threats as far ranging as deadly diseases, terrorist attacks, job losses, undesirable policy changes, violations of norms, group conflict, and even scary pictures. Are some stimuli more likely to affect a person’s perception of threat while others are more likely to elicit an emotional response? This is a broad empirical question, and categorizing such stimuli is beyond the scope of the present effort. However, a couple of ready distinctions seem promising. Credible information about the potential for adverse consequences seems likely to change how a person perceives a threat, regardless of whether emotions are aroused. Stimuli of a more symbolic nature—say, scary music in a political ad or images of a stigmatized outgroup—may or may not change the perceived level of threat, but seem especially likely to provoke an emotional response such as fear or anger. Separate from information describing negative consequences of some political event or trend, group cues may have a powerful influence on the way a message is interpreted by the audience. Cues regarding which groups in society are vying for rights and resources are common in political communication, and may be particularly likely to trigger strong emotions. Decades of research underscore the centrality of group dispositions to public opinion formation (Citrin and Green 1990; Converse 1964; Nelson and 5

In addition, self-reported perceptions and feelings of threat both reflect not only a response to contemporary information, but also a synthesis of past experience and predispositions (Brader and Valentino 2004). However, we can still make good use of self-reports in this study because, by manipulating exposure to threat in an experiment, we can observe changes in self-reported perceptions and emotions along with changes in attitudes and behavior.

5

Kinder 1996; Sears and Funk 1991). Theories of prejudice and group conflict often suggest those phenomena are rooted in a sense of threat from disliked or stigmatized outgroups. Although there is evidence that such a sense of threat can emerge from realistic group competition (Bobo 2000), the “symbolic politics” perspective contends that feelings toward social groups become socialized into predispositions at an early age, thereby encoding group cues with lasting emotional significance (Sears 1993). A growing body of research on “group priming” also suggests that subtle, stereotypical racial cues can boost the impact of group attitudes on opinions (Gilens 1999; Gilliam and Iyengar 2000; Mendelberg 2001; Valentino 1999; Valentino, Hutchings, and White 2002). The strongly affective nature of racial and ethnic group categories leads us to expect that the presence of stigmatized group cues may heighten the emotional impact of media messages concerning potential threats. The literatures on priming and emotion, however, seem to make divergent predictions for the ease with which people process information. Research on group priming has presumed, and found some evidence for, an accessibility-based process underlying group priming effects: Subtle racial cues facilitate racial schemas in memory which are then applied to evaluations of candidates and or issues (Valentino, Hutchings, and White 2002). As a result of priming, responsiveness to related concepts is quicker and more automatic. But this seems to be, at least on its face, at odds with psychological theories of emotion that suggest negative emotions are likely to trigger more deliberate reflection about the issue at hand (Bless 2001; Marcus, Neuman, and MacKuen 2000). As a result of arousing these emotions, individuals are more attentive to related matters and engage in more effortful processing. We wonder whether stigmatized group cues function in at least two distinct, albeit related ways. Some cues, implicit and stereotype reinforcing, might operate via an automatic cognitive process, boosting the accessibility of racial schemas in memory. Others, perhaps more explicit, might trigger negative emotions which then lead to attitude and behavioral change.6 To study this possibility, we will examine the accessibility of

6

One recent body of work on emotion produces findings more similar to those of group priming. Proponents of the “motivated reasoning” approach find that affectively-charged cues heighten the accessibility of similarly-valenced concepts in memory, a process known as affective priming (Lodge and Taber 2000). However, the cues used in

6

constructs in memory to determine whether anxiety-arousing group cues speed up or slow down responsiveness to immigration-relevant schemas (i.e., suggesting the facilitation of a more automatic or a more thoughtful process, respectively). We suspect this more conscious, effortful process may be quite common in political communication, given the situations in which citizens typically encounter and process information about public, issue-based threats. In sum, we see a potential connection between the literature on emotion and that involving group priming: We propose that group cues can infuse media portrayals of threat with the power to elicit negative emotions. In the absence of such cues, threatening information may be recognized as such, but will be less likely to cause anxiety.7 Furthermore, we believe these negative emotional reactions are critical mediators in attitudinal and behavioral reactions to media portrayals of threat.

HYPOTHESES: PUBLIC REACTIONS TO THE THREAT FROM IMMIGRATION We turn to the politics of immigration in the U.S. to examine the preceding propositions. Public opinion about immigration tends to run heavily toward opposition (Simon and Lynch 1999) and appears to have both realistic and symbolic origins (Citrin et al. 1997; Kinder 2003a; Sniderman et al. 2000). Although there has been work on the correlates of immigration opinion, we know little about how political discourse affects public opinion in this domain. As with most issues, we also know significantly less about what motivates citizens to act on their attitudes. In response to the mounting pressures of globalization, a number of politicians and activists have voiced concerns about the economic and cultural threat immigration poses to Americans (Cohen 2001; Scheve and Slaughter 2001). In truth, of course, rising levels of immigration carry both costs and benefits for citizens, but the emphasis of public discourse tends to be on the threat side of the equation (Friedman 2000). Anti-immigration rhetoric at times also takes group overtones and, explicitly or implicitly, makes a distinction between currently stigmatized ethnic or racial groups, such as Latinos, and “good” immigrants such as those from Canada, affective priming studies are also subtle, even subliminal, and thus may be part of an automatic process that is distinct from that set in motion by conscious consideration of threatening stimuli. 7 Although we focus on group cues here, we believe there are potentially many such elements that may enhance the emotional power of media representations of threat.

7

Ireland, or Poland (Huntington 2004). Therefore, the politics of immigration is an apt arena in which to explore propositions about threat and group cues. We are interested in the ways common group cues alter the emotional reactions to the threats inherent in immigration. We speculate that negative emotions aroused by particular groups, not just the perception of economic and cultural threats, determines the impact of political discourse on public reactions to immigration. In other words, we believe that not all realistic threats cause the same emotional reactions, and therefore carry the same political consequences. Citizens may recognize realistic threats from immigration when information about potential costs is in the media environment regardless of the group cues present. However, threats linked to stigmatized minorities, such as Latinos, might be more effective at changing opinions and behavior because they may stimulate powerful emotions such as anxiety and anger. Using experimental evidence, we test the following hypotheses:

H1: Public discourse (e.g., news stories) emphasizing the realistic costs to Americans posed by immigration will boost the perception that immigration is costly, compared to stories that emphasize the benefits of immigration. H2: Public discourse on immigration will be more likely to influence negative emotions, such as anxiety and anger, when paired with cues identifying Latinos as the immigrant group rather than white-skinned Europeans. H3: Public discourse on immigration, when paired with Latino cues, will be more likely to produce significant, message-consistent shifts in political attitudes and behavior. H4. These attitudinal and behavioral effects will be mediated by the changes in negative emotions, not by changes in either positive emotions or perceptions of threat.

8

RQ1. When public discourse on immigration and Latinos cues trigger anxiety, does cognition about immigration speed up or slow down?

EXPERIMENT 1 For our first tests of these hypotheses, we embedded an experiment in a nationally representative survey of 354 white, non-Latino adults. The survey was conducted via WebTV by Knowledge Networks, a web survey company that maintains a large, randomly-contacted respondent pool by offering free Internet access in exchange for occasional participation in surveys. The company randomly selects participants from its pool for each survey.8 Subjects from 46 states took part in our survey, which was conducted between October 21 and November 5, 2003. The median interview was 16 minutes. The median subject was 46 years old with some college and a $45,000 household income. Fifty-two percent of the subjects were women. Thirty-one percent identified as Republicans, and 28% as Democrats.9 Prior to the experiment, subjects were told they would take part in “a survey regarding a number of current social issues.” During the first half of the survey, subjects answered questions regarding their political predispositions and feelings about the economy. Immediately prior to the stimulus, subjects were told that we wanted their reactions to a story that “had been in the news lately.” After reading the story, subjects answered a number of questions about immigrants and immigration. This web-based survey platform is essential for our purposes because it allows us to deliver political stimuli that match those in the respondents’ real information environment (in this case, news coverage). The study employed a simple 2 × 2 design with a control group. We manipulated the ethnic cue by altering the picture and name of an immigrant (white European versus Latino) featured in a mock New York Times report about a governors’ conference on immigration.10 We also manipulated the tone of the story, focusing either on the positive consequences of immigration for the nation (e.g., strengthening the 8

Among people originally contacted by Knowledge Networks to join the pool, 56% agreed to participate. The completion rate for our survey was 77%. 9 Data collection for this experiment was made possible by Time-sharing Experiments for the Social Sciences (NSF Grant 0094964, Diana C. Mutz and Arthur Lupia, principal investigators). 10 At the end of the surrvey, subjects were debriefed as to the fictional nature of the news articles. The complete text of the articles used as stimuli for this study are available upon request from the authors.

9

economy, increasing tax revenues, enriching American culture) or the negative consequences (e.g., driving down wages, using up public resources, undermining American values). Regardless of how consequences were framed, every article included the following information: the fact that immigration to the U.S. is increasing and will continue to increase; possible reasons for the increase; portrayal of governors as either glad or worried about immigration; quotes from fictional citizens who had had either positive or negative interactions with immigrants.11 As part of the ethnic cue manipulation, a picture of an alleged recent immigrant to the U.S. who appears to be either Latino or European accompanied the article. A caption below the photo read: “[Jose Sanchez/Nikolai Vandinsky] is one of thousands of new immigrants who arrived in the U.S. during the first half of this year.” The text of the article included a quote (identical across all conditions) from the men on their feelings about coming to the U.S. and identified their national origin as Mexico or Russia. To maximize control over the ethnic cue, a graphic artist altered photographs so that only the heads differed between the two pictures (i.e., their bodies, dress, background were identical). The selection of these faces from over a dozen possibilities was based on ratings by an independent panel of eight naïve judges. Photos were rated for how much each male appeared “stereotypically European” or “stereotypically Latino,” as well as in terms of wealth, attractiveness, friendliness, and law-abidingness. The faces chosen for our manipulation were maximally distinct on the dimension of ethnicity, but statistically indistinct on the other dimensions.

Perceptions of Threat and Emotional Reactions Our first hypothesis concerned the impact of news about immigration on perceptions of the potential costs and benefits of the issue to Americans. Shortly after reading the news story, subjects offered their perception of the likelihood that immigration will have a “negative financial impact on many Americans” and/or a “negative impact on the way of life in many American communities” (see the appendix for full question wording). We added their responses together to form a perception of threat scale (α = .82). The

11

Subjects in the control group read an article on the relationship between cell phone use and car crashes.

10

top graph in Figure 1 demonstrates the simple effect of positive versus negative news about immigration on perceptions of threat. Regardless of the ethnic cue present, respondents were significantly more likely to perceive immigration as an economic and cultural threat when the news was negative than when it was positive (F = 12.83, p