brader

"Affective Intelligence and Beyond: Next steps in research on emotion in politics"

After decades of slow but steady development, the systematic study of emotion in politics is growing rapidly. Researchers have shown that emotions powerfully predict public opinion and behavior, and that political messages can elicit emotions that alter the very process by which decisions are made. In this essay, I suggest three promising focal points for the next phase(s) of this evolving agenda and highlight some recent moves in the right direction.

Earlier research has laid a firm foundation on which we may now build a more refined theoretical and empirical understanding. In exploring the role of emotion in politics, scholars have applied many leading psychological approaches, including theories of emotional expression (32), affect and information processing research (14, 19, 27), cognitive appraisal theories (15, 16, 30), and affective neuroscience (2, 24, 26).1 These are not necessarily competing views as much as perspectives focusing on different aspects or levels of emotional response.

Affective intelligence theory (AIT), formulated by Marcus, Neuman, and MacKuen, represents the broadest and best-known effort to construct an over-arching model of emotion and politics (26). It is based primarily on affective neuroscience, but it incorporates insights from and is compatible with many aspects of other approaches (23). AIT posits two basic emotional systems in the brain that continually monitor the environment to allocate thinking and behavior efficiently in accordance with situational needs. The disposition system translates feedback on the success of current pursuits into enthusiasm or depression; the surveillance system translates feedback on threats/novelty into anxiety or calm. The former manages reliance on learned routines, while the latter interrupts that reliance to redirect attention and thought processes. AIT has received considerable empirical support to date (1, 2, 7, 17, 22, 26, 28, 33, 35).

We are now poised to move beyond the basic claims of AIT to explore its more dimly lit corners and shed light on how it connects to other psychological models of emotion. For example, there is plenty of room to enlarge and refine AIT to encompass the fuller range of human emotions, to clearly identify when and how politics triggers specific emotions, and to refine our expectations about when we should observe particular consequences of particular emotions. This is not to say additional replications or extensions of AIT in other contexts are unimportant. Such work is likely to spur further refinements.2 Indeed some recent studies yield inconsistent or partially consistent results that illustrate the need for building on AIT. I now discuss three focal points for doing so as part of the next generation of research on emotion and politics.

Consistent with AIT, there is a good deal of evidence that anxiety can provoke attention and information seeking (2, 7, 8, 26, 28, 33), motivate action (2, 7, 26, 34), and precipitate a change of mind based more on contemporary information and less on predispositions (2, 7, 17, 26, 35). In addition, largely consistent with AIT, a number of studies have found that enthusiasm promotes decision-making consistent with political predispositions (2, 12, 13).3 But other findings appear to be at odds with AIT or one another. I look briefly at one such emerging “debate” on learning.

Marcus and colleagues stress the role of anxiety in motivating learning (26), and many take that to be a central tenet of AIT. Even in the original AIT study, however, Marcus et al. acknowledge that enthusiasm may indirectly contribute to learning by fueling interest. They initially find that anxiety and enthusiasm predict interest, caring about the outcome, attention, and opinionation, but only anxiety predicts accurate awareness of candidate positions. In the analysis of panel data, they find sharper distinctions: enthusiasm predicts increases in interest and caring, while anxiety predicts increases in attention and accuracy. Thus, enthusiasm seems to promote general engagement, while anxiety produces actual learning.

Results from other studies sometimes, but not always, support these conclusions. I cannot summarize all of the research in detail here, and the way forward lies to some extent in the details. Nonetheless, Table 1 offers a cursory overview, and I will try to draw attention to a few key details underlying these patterns.4 On balance, AIT holds up pretty well, but there are unexpected null results and, more rarely, findings that run contrary to predictions. First, note the presence of a third emotion, anger, which has drawn increasing attention from scholars in this area. Second, all three emotions have at times stimulated the desire to pay attention and think about political events or issues, though anxiety has done so most consistently and strongly.5 This runs parallel to an emerging pattern in research on the effects of emotion on political participation (not shown). Whereas AIT emphasizes anxiety as an impetus to participation (26), others have found that enthusiasm (1, 2, 4, 34) or anger (4, 34) spur political action as well. Thus, all three of these basic emotions can give rise to a broad family of outcomes including psychological engagement, motivation to attend, and actual participation. The critical question ahead for researchers in the field is when each emotion will do so and why. Answering that question will move us past existence claims and toward a stronger theory (20).

Third, the evidence on actual information seeking and learning is more mixed. It is also more subject to qualifications. There two facts are likely related. In expanding AIT to incorporate anger, MacKuen and colleagues find that, while all three emotions increase the intention to pay attention, each emotion has a different impact on search behavior (22): enthusiasm does not affect actual searching, anxiety expands the depth and breadth of the search, and anger contracts search depth and breadth. Other studies have also found that anxiety broadens search behavior while anger shrinks it (28, 33), but these findings are conditional. The expected effects of anxiety were more likely to appear when subjects considered a preferred candidate (28), when the overall level of threat in the environment was high (28), and when subjects knew they would need to defend their views (33). These studies also confirm the positive impact of anxiety on learning under similar conditions, but one study does not. Drawing on other work in psychology, Feldman and Huddy argue that anxiety can impair cognitive performance, making it harder for even attentive citizens to learn (8). They find anxious citizens display less accurate knowledge of the Iraq war, and speculate that the discrepancy with AIT may be due to the source of anxiety in their study posing a stronger personal threat than the candidate and issue sources in other studies.

Table 1. Recent Evidence on the Impact of Emotions on Engagement and Learning

Note: Entries in the table indicate evidence that an emotion has a positive impact (+), negative impact (-), or no impact (0) on the outcome variable at the top of the column. Blank cells indicate the relationship was not tested or reported. Entries separated by a slash indicate mixed results across separate studies (e.g., multiple experiments) and/or separate measures of the outcome variable.

In sum, when it comes to the effects of emotions on politics, current research has found AIT decidedly more right than wrong. A growing number of inconsistencies, however, present an opportunity to improve our understanding. Moreover, recent studies provide clues about the need to formulate more precise expectations for when we will observe learning, action, or other behaviors as a result of each emotion. One way to do so is to build insights from appraisal theories on top of the neuropsychological foundation of AIT (mimicking somewhat the layering of cognitive and emotional architecture in the brain itself). Although appraisal theories primarily focus on the antecedents that give rise to emotions, they also call our attention to subtle differences in the relationship between an individual and her situation—salience, proximity, blame, certainty, control, available options. We need not believe that emotions work differently across situations; we may simply discover we have been too broad in defining outcomes. Enthusiasm, anxiety, and anger may promote different types of learning and participation that are directed at distinct objectives. Nonetheless, it is also possible for emotions to be present but not always lead to further action. Human emotions are impulses, not reflexes, and individuals can decide whether to act on those impulses or not in most cases of interest to students of politics. One of our tasks for future theory building is to predict when the impulses will win out.6

The other focal points for future research primarily involve areas of neglect, not inconsistency. The first is the need to strengthen theoretical predictions and research on sources of emotions. AIT describes the basic inputs of the disposition and surveillance systems: feedback on progress toward our goals and feedback on the safety and familiarity of the environment, respectively (26). With these general guidelines in mind, we can begin to identify recurring elements of politics that should evoke particular emotions. However, not all threats, novelty, and success (subjectively or objectively defined) elicit the expected emotions to the same degree. At present, AIT does not offer much guidance on how to predict the intensity of arousal, a lack of expected arousal, or individual differences. For example, we know intensity of anxiety should vary with the level of threat or novelty, but not how it will vary for a given person across a range of “threats” such as lingering health problems, the possibility of a terrorist attack, financial stress, the passage of a constitutional amendment on abortion, or the intrusion of a disliked group. In a new study, colleagues and I find that news stories causing equal increases in perceived threat only triggered anxiety when the story included images of a negatively stereotyped group (7).

Part of the path to theory building in this area may lie, as it often does, with an initial growth in empirical research. Fortunately, such research is underway. Some insights may come from the differential effectiveness of stimuli in experiments, especially as researchers examine a broader range of stimuli. Recent experiments have manipulated candidate issue distances (13, 28), favorability of policy outcomes (7, 22), news frames (10), group cues (7), and emotive imagery or music (2, 12). Insights may also come from direct analysis of emotional cues in the political environment, such as in news broadcasts (9) or campaign ads (2, 3).

Our main goal is not to compile an inventory of the emotional significance of all political stimuli, but rather to identify key qualities of political stimuli and situations that shape their emotional power. Thus, in making sense of the empirical patterns, we will be well served by drawing again from psychological theories. Affective neuroscience sheds light on how objects, ideas, events, and people become emotional triggers, and has already informed research on motivated political reasoning (19) and my own investigations of political advertising (2). Appraisal theories are instructive on how personality and situation intersect to produce a diverse array of emotions. Indeed, this is likely the domain in which appraisal theories will be most useful to students of politics. Whereas other approaches say more about the impact of emotion on cognition and behavior, the comparative advantage of appraisal theories is their focus on the conscious and preconscious interpretations (i.e., appraisals) that evoke specific emotions.

There is another, complementary avenue for improving our understanding of when emotions are and are not triggered in politics. It concerns the strategic use of emotions. Many politically relevant emotions arise incidentally or unexpectedly from events and communications, but others are intentionally or at least knowingly evoked. By combining empirical knowledge of how emotions work with knowledge of actor incentives and relationships, game theoretic models can help us predict not only when attempts to arouse emotions will be effective, but also when political actors will make such attempts in the first place. This approach can also shed light on common normative concerns about the political manipulation of emotion. Lupia, for example, applies game theory to ask when politicians can and will use fear to convince the public to support undesirable policies (20, 21). Empirical studies should inform and test such models by examining when and where political actors appeal to certain emotions and how they respond to one another’s appeals (3).

The third focal point for the next phase of research is a conscientious expansion of the emotions under study. By conscientious, I mean guided by theory about meaningful distinctions and mindful of the distinctions and labels used by other scholars. This involves both semantics and substance. For example, if I wish to label the output of the surveillance system “fear,” I should clarify I mean the same emotion as when Marcus and colleagues say “anxiety.”7 More critically, we must resist letting language substitute for theory by assuming every emotion word represents a distinct emotional state. We might “discover” distinct effects of three similar emotions—say, joy, enthusiasm, and satisfaction—by throwing all three into a series of regressions as separate variables. This data-driven approach, however, risks carelessly confusing noisy measures and other stochastic processes with real emotional differences.8 Are anger and disgust synonyms or distinct emotions? Some psychologists treat them as distinct, but they do so on the basis of well-defined conceptions of how the antecedents and effects differ. This is the model to follow.

Which emotions should be investigated? Researchers should make that decision based on their own theoretical and substantive interests. Studies to date have heaped the most attention on enthusiasm and anxiety, in part due to their central place in AIT. As noted earlier, scholars have recently shown increased interest in anger (4, 11, 28, 30, 34), and MacKuen and colleagues have begun to elaborate its place in AIT under the label “aversion” (22). Within the original AIT framework, the field has so far neglected the low arousal or “negative feedback” states of the two systems, namely depression/sadness and calm. At best, political research has treated these emotions as baselines against which to assess the effects of their high arousal counterparts. But shifts in the opposite direction (e.g., from enthusiasm to sadness) may be just as or more politically significant at times. For example, one might wish to study how the disappointment of losing an election or a war affects citizens, or study government efforts to instill calm in the face of terror threats or disease outbreaks.

Other emotions are relevant to political life but lie beyond the edges of the existing AIT dual system model. For example, Valentino and I are beginning to study the political consequences of sympathy and amusement by drawing on related research in psychology (5, 6). The ability and desire of graduate students to break free of the confining ideas set by their advisors’ research is often a fruitful source of innovation. One of my students, Elizabeth Suhay, is investigating how shame and pride serve as emotional mechanisms for the adoption and application of political values (31). Another student, Yanna Krupnikov, has begun to study the role of regret in the formation and dissolution of political habits (18). Taking this growing list of emotions seriously is not necessarily incompatible with AIT. It simply remains unclear whether scholars should eventually try to refine AIT to incorporate these emotions (e.g., by conceptualizing additional affective “systems” or, alternatively, elaborating on how secondary cognitive appraisals transform the basic inputs of the two systems into other emotions), or should abandon AIT in favor of a more comprehensive alternative account.

Advances in theory and methods have propelled an emerging field of scientific inquiry into emotion and politics. There is now widespread recognition that discrete emotions play an important role in political communication and decision-making. Affective intelligence theory (AIT) has helped to organize both research and understanding of the topic. A growing number of studies, including by scholars not adopting the AIT perspective, lend solid empirical support to many AIT propositions. At the same time, a growing number of studies, including by scholars adopting the AIT perspective, have found unexpected or inconsistent results. We should not see such variance as a source of confusion as much as a puzzle that stimulates us to refine our theoretical understanding.

In addition, while AIT provides a decent foundation, it also remains vague or even silent on several important questions: When and how intensely will politics arouse specific emotions? When will politicians try to elicit emotions, and will they succeed? What roles do other politically relevant emotions such as sympathy and shame play? In the next stages of development, therefore, the field should invest effort in (1) identifying the qualities that cause some emotional triggers to be effective and others to misfire, (2) studying the strategic decisions of political actors to manipulate emotions (or not) for their own advantage, and (3) expanding the emotions under consideration. Scholars have taken initial steps in each of these directions. I believe these provide useful focal points for future research, though there are other exciting areas for further research as well. Regardless of which directions we take, I believe the field will continue to flourish so long as we endeavor to tie new ideas and research explicitly to the existing edifice of theory and findings, whether by way of addition, amendment, or refutation.

Ted Brader is Research Associate Professor at the Center for Political Studies and Associate Professor of Political Science at the University of Michigan.

1. Brader, Ted. 2005. “Striking a Responsive Chord: How Political Ads Motivate and Persuade Voters by Appealing to Emotions.” American Journal of Political Science 49(2): 388-405.

2. Brader, Ted. 2006. Campaigning for Hearts and Minds: How Emotional Appeals in Political Ads Work. University of Chicago Press.

3. Brader, Ted, and Bryce Corrigan. 2005. “Emotional Cues and Campaign Dynamics in Political Advertising.” Paper presented at the annual meeting of the American Political Science Association in Washington, DC.

4. Brader, Ted, and Bryce Corrigan. 2006. “How the Emotional Tenor of Ad Campaigns Affects Learning, Choice, and Participation.” Paper to be presented at the annual meeting of the American Political Science Association in Philadelphia, PA.

5. Brader, Ted, and Nicholas A. Valentino. N.d. “A Political Tempest: Emotions, Accountability, and Action in the Aftermath of Katrina. Research in progress, University of Michigan.

6. Brader, Ted, and Nicholas A. Valentino. N.d. “A Laughing Matter? How Satire Affects Political Behavior by Eliciting Amusement.” Research in progress, University of Michigan.

7. Brader, Ted, Nicholas A. Valentino, and Elizabeth Suhay. 2004. “Seeing versus Feeling Threats: Group Cues, Emotions, and Activating Opposition to Immigration.” Paper presented at the annual meeting of the American Political Science Association in Chicago, IL.

8. Feldman, Stanley, and Leonie Huddy. N.d. “The Paradoxical Effects of Anxiety on Political Learning.” Unpublished manuscript, Stony Brook University.

9. Graber, Doris. Forthcoming. “The Road to Public Surveillance: Breeching Attention Thresholds.” In The Affect Effect: Dynamics of Emotion in Thinking and Behavior, edited by W. R. Neuman, G. E. Marcus, A. Crigler, and M. B. MacKuen. University of Chicago Press.

10. Gross, Kimberly, and Lisa D’Ambrosio. 2004. “Framing Emotional Response.” Political Psychology 25(1): 1-29.

11. Huddy, Leonie, Stanley Feldman, and Erin Cassese. Forthcoming. “On the Distinct Political Effects of Anxiety and Anger.” In The Affect Effect: Dynamics of Emotion in Thinking and Behavior, edited by W. R. Neuman, G. E. Marcus, A. Crigler, and M. B. MacKuen. University of Chicago Press.

12. Huddy, Leonie, and Anna H. Gunnthorsdottir. 2000. “The Persuasive Effects of Emotive Visual Imagery: Superficial Manipulation or the Product of Passionate Reason?” Political Psychology 21(4): 745-78.

13. Hutchings, Vincent L., Nicholas A. Valentino, Tasha S. Philpot, and Ismail K. White. 2006. “Racial Cues in Campaign News: The Effects of Candidate Issue Distance on Emotional Responses, Political Attentiveness, and Vote Choice.” In Feeling Politics, edited by David Redlawsk. NY: Palgrave Macmillan.

14. Isbell, Linda M., and Victor C. Ottati. 2002. “The Emotional Voter: Effects of Episodic Affective Reactions on Candidate Evaluations. In The Social Psychology of Politics, ed. Victor C. Ottati et al. NY: Kluwer, 55-74.

15. Just, Marion R., Ann N. Crigler, and Todd Belt. Forthcoming. “Don’t Give Up Hope: Emotions, Candidate Appraisals, and Votes.” In The Affect Effect: Dynamics of Emotion in Thinking and Behavior, edited by W. R. Neuman, G. E. Marcus, A. Crigler, and M. B. MacKuen. University of Chicago Press.

16. Kinder, Donald R. 1994. “Reason and Emotion in American Political Life.” In Beliefs, Reasoning, and Decision Making, edited by Roger Schank and Ellen Langer. Mahwah, NJ: Lawrence Erlbaum Associates, 277-314.

17. Kinder, Donald R., and Lisa D’Ambrosio. 2000. “War, Emotion, and Public Opinion.” University of Michigan. Unpublished manuscript, University of Michigan.

18. Krupnikov, Yanna. N.d. “Can a Voter Regret a Vote? Why Post-Decision Processes Explain Voting Habits.” Research in progress, University of Michigan.

19. Lodge, Milton, and Charles Taber. 2000. “Three Steps toward a Theory of Motivated Political Reasoning.” In Elements of Reason, edited by Arthur Lupia, Mathew McCubbins, and Samuel Popkin. New York: Cambridge University Press, 183-213.

20. Lupia, Arthur, and Jesse Menning. Forthcoming. “Politics and the Equilibrium of Fear: Can Strategies and Emotions Interact?” In The Affect Effect: Dynamics of Emotion in Thinking and Behavior, edited by W. R. Neuman, G. E. Marcus, A. Crigler, and M. B. MacKuen. University of Chicago Press.

21. Lupia, Arthur, and Jesse Menning. N.d. “When Can Politicians Scare Citizens Into Supporting Bad Policies?” Unpublished manuscript, University of Michigan.

22. MacKuen, Michael, George E. Marcus, Luke Keele, Jennifer Wolak, and W. Russell Neuman. 2005. “Crossing the Private/Public Boundary: Emotion as Perception and Motivation.” Paper presented at the annual meeting of the International Society of Political Psychology in Toronto, Canada.

23. Marcus, George E. 2003. “The Psychology of Emotion and Politics.” In Oxford Handbook of Political Psychology, edited by David Sears, Leonie Huddy, and Robert Jervis. New York: Oxford University Press, 182-221.

24. Marcus, George E., and Michael B. MacKuen. 1993. “Anxiety, Enthusiasm, and the Vote: The Emotional Underpinnings of Learning and Involvement during Presidential Campaigns.” American Political Science Review 87:672-85.

25. Marcus, George E., Michael MacKuen, Jenny Wolak, and Luke Keele. 2006. “The Measure and Mismeasure of Emotion.” In Feeling Politics, edited by David P. Redlawsk. NY: Palgrave-Macmillan.

26. Marcus, George E., W. Russell Neuman, and Michael MacKuen. 2000. Affective Intelligence and Political Judgment. Chicago: The University of Chicago Press.

27. Rahn, Wendy M., and Brian Kroeger, and Cynthia M. Kite. 1996. “A Framework for the Study of Public Mood.” Political Psychology 17(1): 29-58.

28. Redlawsk, David P., Andrew J.W. Civettini, and Richard R. Lau. Forthcoming. “Affective Intelligence and Voting: Information Processing and Learning in a Campaign.” In The Affect Effect: Dynamics of Emotion in Thinking and Behavior, edited by W. R. Neuman, G. E. Marcus, A. Crigler, and M. B. MacKuen. University of Chicago Press.

29. Rudolph, Thomas J., Amy Gangl, and Dan Stevens. 2000. “The Effects of Efficacy and Emotions on Campaign Involvement.” Journal of Politics 62:1189-97

30. Steenbergen, Marco R., and Christopher Ellis. 2006. “Fear and Anger in Candidate Evaluation: Context, Traits, and Negative Candidate Affect.” In Feeling Politics, edited by David P. Redlawsk. NY: Palgrave-Macmillan.

31. Suhay, Elizabeth. 2006. “Self-Conscious Emotions and American Political Values: How Pride and Shame Motivate the Adoption of Peer-Group Ideals.” Paper to be presented at the annual meeting of the American Political Science Association in Philadelphia, PA.

32. Sullivan, Denis G., and Roger D. Masters. 1988. “Happy Warriors: Leaders’ Facial Displays, Viewers’ Emotions and Political Support.” American Journal of Political Science 32:345-68.

33. Valentino, Nicholas A., Vincent L. Hutchings, Antoine J. Banks, and Anne K. Davis. 2006. “Is a Good Citizen a Worried Citizen? Emotions, Political Information Seeking, and Learning via the Internet.” Paper presented at the annual meeting of the International Society of Political Psychology in Barcelona, Spain.

34. Valentino, Nicholas A., Vincent L. Hutchings, Krysha Gregorowicz, and Eric W. Groenendyk. 2006. “Election Night’s All Right for Fighting: The Participatory Impact of Negative Emotions.” Paper presented at the annual meeting of the Midwest Political Science Association in Chicago, IL.

35. Way, Baldwin, and Roger Masters. 1996. “Political Attitudes: Interactions of Cognition and Affect.” Motivation and Emotion 20:205-36.34.

1 The relevant literatures in psychology are immense. Some prominent figures include Paul Ekman and John Lanzetta for expression; Herbert Bless, Joseph Forgas, and Norbert Schwarz for affect and processing; Craig Smith, Richard Lazarus, and Klaus Scherer for cognitive appraisal; and Antonio Damasio, Jeffrey Gray, and Joseph LeDoux for affective neuroscience.

2 For example, most systematic research on emotion and politics has been carried out in the United States. Studies in other countries offer opportunities not only for replication, but also for new insights. We should expect basic emotions and emotional processes to be similar across countries, as they are foundational elements of human psychology, but the manner of their expression and activation may well differ in ways significant for politics.

3 I say “largely consistent” because, in promulgating AIT, Marcus and colleagues actually contrast judgments of anxious voters to those of “complacent” voters who report low overall emotional arousal (26). My interpretation of AIT and psychological research on affect and information processing is that enthusiasm increases reliance on habits, just as anxiety decreases reliance (2). I see this as an example of a “friendly amendment” to AIT.

4 Please note that I meant “emerging” debate quite literally, as much of this research has yet to appear in final published form. Most or all of the work has been presented publicly, but readers should keep in mind that results mentioned here were only in preliminary form and may change. If interested, contact the respective authors for more information.

5 I have grouped what I see as related classes of outcome variables for these purposes. Attention in Table 1 refers to the inclination to obtain more information or desire to pay attention (none of these studies measures actual attention in the form of neural or retinal activity, though the information search findings are based on actual search behavior). Opinionation refers to the expression of thoughts or attitudes (2, 26, 33), self-reported time spent thinking and talking about an issue (8, 11), and time spent processing information (28).

6 Although I have not addressed measurement or research design issues here, these features do vary across the studies in Table 1 and could play a role in explaining apparent discrepancies. Nonetheless, we would still need to explain why two measures or designs produced different results and this leads us back to theory in most cases.

7 I pick this example deliberately. Scholars disagree on whether the “fear family” of emotional terms encompasses more than one emotional state or simply refers to differing intensities of a single emotional state (with or without possible non-linearities in effects). Vernacular usage tends to make distinctions based on intensity only, a point worth recognizing in light of our heavy reliance on self-reports.

8 The same holds for the somewhat opposite situation when factor analysis or similar tools suggest that what we believe are distinct emotions actually cluster together on a single scale. Theory of course should govern our use and interpretation of such methods. Nonetheless, we can also derive substantive lessons from these cases, which may reveal weaknesses in measurement and/or challenges posed by the capacity of humans to report their emotions (e.g., an emotion checklist may pick up discrete emotional states in some cases but only diffuse positive or negative moods in other cases). One issue currently challenging scholars in the field to think harder about both theory and measurement is the fact that anger terms often load onto anxiety scales in political surveys, but sometimes emerge as a distinct anger scale (25).