Political Psychology and Choice
Abstract and Keywords
This article discusses political psychology and choice, starting with an overview of the recent emphasis on the importance of emotion in understanding political choices. This is followed by a discussion of the research that deals with the ability of citizens to process information without any bias. It then highlights the contributions of methodological innovations to an understanding of political psychology. The article concludes with several reflections on the political psychologists' emphasis on the importance of information, cognition, and rationality in research for the past few decades.
Political psychology is, at heart, concerned with the characteristics of individuals and of situations that are most conducive to a successful political system. For most political psychologists whose work is reviewed in this chapter, the ideal political system is a western-style democracy, with individual rights and responsibilities for self-governance, combined with varying degrees of protection of minority interests. For these reasons, the kinds of citizen choices that are most valued and most widely studied are ones that reflect these emphases. They include, but are not limited to, high levels of political information, active political participation, fair-minded evaluation of political alternatives, and so forth.
Given the sheer volume of work in this burgeoning area, I cannot hope to do a thorough review of the many contributions of political psychology in recent years. Moreover, another recent volume in this same series, the Oxford Handbook of Political Psychology, does an admirable job in summarizing the many developments in this field (see Sears, Huddy, and Jervis 2003). Thus, I have chosen to highlight three of the more recent trends and most promising new areas of investigation in political psychology that have emerged over the last few decades. I explore these particular themes not only because they are recent, but also because they hold some promise of changing, in some fundamental way, how we think about political psychology.
This chapter begins with an overview of the recent emphasis on the importance of emotion in understanding political choices. Next, I turn to research dealing with the ability of citizens to process information in an unbiased fashion. This (p. 346) category includes studies of motivated reasoning and selectivity, as well as research on the effects of partisanship and ideology on the processing of information. Third, I highlight the contributions of methodological innovations to our understanding of political psychology. While no one method is a cure-all, recent advances in the field of neuroscience are opening up new approaches with the potential to help us better understand the black box psychological processing of political stimuli.
Finally, I conclude by reflecting upon political psychologists’ emphasis on the importance of information, cognition, and rationality in research over past decades, examining rationality’s use as a standard (both empirical and normative) for judging the quality of decision-making processes. It is ironic that political psychology so often defines itself in opposition to rational choice approaches, and yet its standard for normative judgments is virtually the same.
1 The Role of Emotion in Political Choice
Over the last few decades, political psychologists have enriched our understanding of choice by incorporating emotion into models that were formerly almost exclusively cognitive in describing political decision-making processes. In order to describe the progress (and lack thereof) in this domain, it is useful to first discuss several terms that are used more or less interchangeably within contemporary political psychology, including mood, affect, feeling, and emotion. As Kuklinski (2001) has noted, the study of these concepts within political psychology is still in its infancy, and “[we] do not always adopt the same conception of identically labeled psychological phenomena.” As a result, it is less clear than one might think what is and is not known about the role of emotion in political behavior. I begin by sorting through some of the most frequently used terms and operationalizations, and then turn to the difficulty of differentiating emotions from other phenomena.
Within political psychology, the term affect often is used to describe whether an individual likes or dislikes some political object, or whether it is positively or negatively valenced, or “affectively charged,” to use a popular terminology. Common measurement techniques such as feeling thermometers or Likert scales are used to ascertain an individual’s positive or negative evaluation of some political person, policy, or object.
Unfortunately, this operationalization of affect is often difficult or impossible to distinguish from political judgments and opinions more generally. Few doubt that affect influences political attitudes and the processing of political information, but as it is usually measured by political scientists, such positive or negative judgments need not necessarily result from emotional reactions. After all, one may feel (p. 347) positively or negatively toward a political object for reasons that are wholly cognitive in nature.
It has long been acknowledged, for example, that the strongest predictor of candidate choice in the American National Election Studies (ANES) comes from the feeling thermometer ratings of presidential candidates (see e.g. Bartels 1988). Such measures are often referred to as indicators of affect toward the candidates, and yet this evidence is a weak basis on which to claim that emotion plays an important role in political choice. Thermometer ratings may instead represent running tallies of respondents’ likes and dislikes about the candidate over time, which is a far cry from the kind of visceral reaction to a political event that the study of emotion promises to help us understand.
Just as like or dislike for political objects and measures drawn from feeling thermometers should not be considered synonymous with emotion, another seemingly related concept—mood—is also frequently conflated with emotion. Whereas emotions tend to be fleetingly experienced in response to a specific stimulus, and then dissipate, mood refers to a much longer-lasting phenomenon. Moods are also less focused in their target than are emotional reactions (see Bless 2001).
Because of the inconsistent use of terms in the study of emotion and politics, and because of highly variable operationalizations of those same terms, it is difficult to draw a clear line between research on political attitudes and studies of political emotion. Researchers have proposed a variety of theories of emotion over the last century, but almost all define emotion in terms of physiological arousal, which is often (though not necessarily) combined with a cognitive label of some kind. To be consistent with most psychologists’ definitions, political emotion should involve some kind of negative and/or positive reaction to a political object, along with a concurrent experience of arousal. This visceral reaction may occur below the level of conscious recognition, and is relatively automatic, that is, it need not be mediated by cognition.
Conceptually, emotions also are different from attitudes in that emotional reactions are relatively short-lived and highly focused. Perhaps because emotion involves well-known physiological symptoms, it is often assumed that people must know it when they feel it. But emotions need not be particularly pronounced or obvious to the person experiencing them. Although the natural tendency in studies of emotion and politics is to treat the political object that evokes the emotion as if it were the sole cause, the kind of cognitive label that people give to emotion is determined at least in part by cues present in the environment at the time. Likewise, when arousal is artificially induced unbeknownst to experimental subjects, they will nonetheless report experiencing an emotion and attribute it to something even though it was not the actual cause of their arousal.
A great deal of research within political science has focused on particular types of emotions, such as anxiety, anger, fear, or enthusiasm. This focus most likely results from the steady supply of self-report measures of these emotions in the ANES and other election surveys. Others have focused more on the extent to which emotional arousal occurs, without respect to the subspecies of emotion being experienced.
(p. 348) Both approaches are relevant so far as they lead to an understanding of how emotions are involved in political attitudes and behavior. Unfortunately, the traditional survey method has made it difficult to disentangle the experience of emotional arousal from the cognitive assessment of the object and the labeling of the specific emotion.
To date, the most prominent theory tying emotion to political psychology is Marcus, Neuman, and MacKuen’s (2000) theory of affective intelligence, which posits that affect ultimately serves to make citizens more sophisticated. When anxious about how things are going in the political world, this generalized anxiety drives a search for more information, and for better use of existing information resources. Thus greater political “intelligence” is induced by emotion, at least this specific variety. Drawing on ANES data, Marcus and colleagues argue that generalized anxiety about politics causes people to engage in more effortful information gathering and processing. As a result, they are less likely to rely on default heuristics such as party identification in informing their vote preferences, and more likely to seek out and rely on substantive information. According to their formulation, emotion plays an indirect role in promoting more effortful processing by motivating citizens to seek out and use more information. In other words, emotion is the driving force behind a process that ultimately improves the quality of political decision making. More specifically, Marcus and colleagues argue that a specific positive emotion—enthusiasm—elicits greater participation, whereas the negative emotion labeled anxiety elicits an information search.
The theory of affective intelligence has undoubtedly played an important role in renewing consideration of emotion in a field that has been heavily cognitive throughout its brief history. Perhaps even more importantly, this work has brought about reconsideration of the normative perspective on emotion that is common to most political psychology. Much of political theory has disdained the role of emotion in political decision making and, until recently, political psychologists have largely followed suit. Psychologists have recognized the important role emotion plays in intelligent functioning, and how cognition alone leads to serious dysfunction. Political psychologists have been slower to take up the defense of emotion as a potentially positive force in political decision making.
The theory of affective intelligence is not without its critics. Although few argue with the general logic of the theoretical framework, nor that emotions may serve useful (as well as potentially harmful) purposes in the political world, the empirical evidence supporting affective intelligence has been criticized as limited and inconclusive. For one, evidence is limited to retrospective self-reports of emotional reactions. Evidence of affective intelligence hinges on the validity of survey questions asking respondents to tell the interviewer whether a given political figure has ever made them feel angry, afraid, anxious, enthusiastic, and so forth. While such measures have face validity, studies outside the political realm raise doubt that they provide accurate recall of previously experienced emotions. Without the presence of the emotion-inducing event or object, such reports tend to be heavily mediated by cognitions (Breckler 1984). Likewise, induced emotion is quite different from semantically (p. 349) activated reports of emotion. As Niedenthal and colleagues (2003, 327) suggest, “affect infusion … requires that the perceptual aspects of an emotion are experienced, not merely the semantic aspects.”
In a related critique, Ladd and Lenz (2004) point out that while the theory of affective intelligence suggests that a generalized anxiety among members of the electorate drives greater engagement and the search for more information, empirical evidence is based on whether anxiety is reported to have been produced by specific candidates. Thus it is not a general emotional state that is operationally tapped in examinations of affective intelligence, but rather how one feels about a candidate or candidates. Using ANES data, Ladd and Lenz show, not surprisingly, that candidate preference and vote choice are related to comparative emotions toward the two candidates. The extent that one candidate produces more anxiety than another is strongly related to candidate preference. They argue that those reporting anxiety may, indeed, be more engaged, but only spuriously so, either because intensely held preferences drive both anxiety and engagement, or because political engagement leads to still stronger reactions to the campaign. As Ladd and Lenz note, the results seen thus far are consistent with evidence of affective intelligence, but they do not rule out other possible interpretations.
Clearly, some doubt exists regarding the specifics of affective intelligence, but few doubt that politics can be emotion provoking, nor that emotion matters to the political choices that people make. Although affective intelligence focuses our attention on the benefits of emotion for political behavior, emotion is also widely acknowledged to be potentially manipulative. As Brader and Corrigan (2005, 1) point out in their study of the emotional content of political advertisements, “The full significance of emotions for politics comes not because emotions influence the political behavior of citizens, but rather because political actors know that they do and try to capitalize on the power of emotions to achieve their goals.” Most consultants believe in the importance of emotional appeals, though these lay theories have not been validated by empirical evidence (e.g. Kaid and Johnston 2001).
Methodologically, political scientists find it difficult to study emotion as distinct from cognition. Survey data alone cannot make a strong case for emotions as a cause of most politically relevant outcomes (e.g. Glaser and Salovey 1998; Isbell and Ottati 2002). But even in experimental settings, efforts to manipulate emotion without changing the informational content of messages prove quite difficult. For example, in two experiments on the role of emotion in political advertising, Brader (2005) compares the reactions of subjects exposed to ads that include emotional cues for enthusiasm and fear to those that do not. Operationally, he does this by comparing a relatively negative script to a similar one that includes evocatively fearful images and music, and a relatively positive ad to one that includes enthusiastic music and images. He suggests that imagery and music are critical to emotional appeals, whereas verbal content is processed in highly cognitive ways. While there is some evidence that pictures are particularly good at inducing emotional responses relative to words, like most scholars, Brader relied on the post hoc report of emotion.
(p. 350) It would be fairly simple to interpret the results of Brader’s study if one could validate that information is entirely contained within the verbal content of communications, whereas changing the visual content and music alters only emotions. As psychological studies suggest, some words carry far more emotional content than others do, just as some pictures do (see Lang, Bradley, and Cuthbert 1997). But just as a picture is often said to be worth a thousand words, there is no clear way to change images and music within a presentation without also changing the information that viewers are given, and the context in which they are interpreting it. Within psychology, many researchers use standardized sets of words and pictures that allow them to roughly equate stimuli as strongly or weakly positive, negative or neutral in the emotions they elicit. But standardized stimuli like these have yet to be developed for political psychology. Moreover, to do so would be quite difficult. Whereas smiling babies and cute bunnies are consensually regarded as producers of positive affect in the psychology lab, George Bush could be one person’s positive stimulus and another’s strong negative one.
How else might researchers manipulate emotion without inadvertently changing other variables in their designs? In one study, subliminal cues were used to induce emotional reactions without viewer awareness and thus also without changing the visual or verbal information of which subjects were cognitively aware (see Weber, Lodge, and Taber 2005). This approach has the advantage of holding information constant, but it probably also mutes the potential effects that emotion might have relative to real world examples of emotion-inducing messages.
Furthermore, even if one does not seek to manipulate emotion, but instead measures it as an outcome, our usual methodological toolbox is limited in what it has to offer. The heavy reliance on emotion as reported by subjects after the fact casts serious doubts on the appropriate interpretation of many studies. If, as many psychologists suggest, affect is most often experienced extremely quickly and often in the absence of conscious cognitive awareness (see Zajonc 1980; Bargh and Chartrand 1999), then the usual approaches to measurement will not do. People only become aware of their emotions if they are very strong emotions, and most directed at the political world probably do not reach that level. As Alford and colleagues (2005, 20) summarize, “Emotion produces choices and behavior without much in the way of controlled cognitive deliberation that is introspectively transparent.” Even if one trusts self-reports, there is the additional hurdle of getting subjects to accurately recall felt emotions. Civettini and Redlawsk (2005) find that when affect is reported immediately after a stimulus, and then recalled later in the same experiment, there are nonetheless high levels of error in their self-reports.
All of this is problematic for what we political scientists ask of our survey respondents and experimental subjects. There is no easy solution, but it seems doubtful that post hoc self-reports of emotion will continue to be defensible as the standard measure of emotional response. If political psychologists are convinced—as we seem to be—that automatic, preconscious emotional reactions precede and shape the kind of subsequent cognitive processing that transpires, then there is little choice but to pursue alternative approaches. If we are to further an understanding of emotion (p. 351) and politics that is more than simply a repackaging of studies of political cognition, then we need to sort out our terminological inconsistencies and improve methods of measurement. Despite progress, we know far too little about the extent to which emotions are involved in political judgment. At best we can say that we have studied the effects of some emotions that citizens are aware of and can label, and can respond to in some purposive way. But that points to a huge limitation on current knowledge.
2 The Psychology of Biased Processing
Because of the ever-increasing range of choice offered to citizens and consumers, one of the most active areas of political psychology research is the study of whether people are biased versus fair-minded processors of political information. Do people assimilate information in a rational way, or do they raise the bar for convincing evidence when new information contradicts their existing views? Are they simply rational updaters who take new information and add it to their existing mix in order to formulate a new opinion? Or are they selective in what they expose themselves to and to what extent they revise their views accordingly?
This research is triggered in part by renewed interest in parties and partisanship in American elections. The early research suggesting that partisanship was declining in the 1970s gave way to a consensus of “renewal” in the 1980s and 1990s (Fiorina 2002). The strength of the statistical relationship between party identification and vote choice rose continuously from 1972 to 1996, but this new consensus diffused relatively slowly throughout the discipline (see Bartels 2002). In addition, even widespread acceptance of the increased strength of this relationship has not necessarily meant that everyone agrees that party identification is now a stronger predictor of vote choice. As Fiorina (2002) points out, if party identification now works in concert with other determinants of vote choice that once predicted in opposite directions or not at all, then there may be good reason to call this new consensus into question.
More recently, Levendusky (2005) showed that party identification and ideology are much more tightly aligned now than in the 1970s. Whereas party ID and ideology were once largely orthogonal, liberals are now predominantly Democrats and conservatives are predominantly Republicans. This sorting process, he argues, has occurred as a result of elite polarization. When elites are ideologically polarized and send homogeneous signals about what it means to be a Democrat/Liberal and a Republican/Conservative, then the electorate “sorts” themselves into more consistent categories, largely by changing ideology to align with party identification.
Interestingly, what it means precisely to “identify” with a political party remains an unanswered question. Party identification is easily the most widely used concept in all of political psychology if not political science, but it has been reified to such an extent that its meaning is seldom questioned, except in comparative contexts.
(p. 352) Moreover, the extent to which people in various countries will self-identify with a party hinges precariously on how the question is asked. In a study comparing a variety of approaches to asking about party attachments in Canada, the US, and Britain, Blais and colleagues (2001) found that the extent of these publics willing to adopt these labels went from 76 percent to 48 percent, based on a minor change in the wording of the question.
Despite some skepticism about the newfound power of partisanship in the United States electorate, the strengthening of this statistical relationship has spawned a resurgence of interest in the extent to which partisanship biases the processing of political information. Whereas twenty-five years ago one was more likely to read about partisanship in the academic journals as a source of high levels of political knowledge, mobilization, and attitude consistency, many contemporary political psychologists study partisanship as a source of bias in the processing of political information. Political parties have been at the root of the debate over biased assimilation from the very beginning of election research. As Angus Campbell and colleagues (1960, 133) argued, “Identification with a party raises a perceptual screen through which an individual tends to see what is favorable to his partisan orientation.” The theme of partisan resistance to new information persists in contemporary models of the vote, and it is argued to cause people to selectively consume information and/or selectively interpret the implications and importance of new information, so that it does not threaten their existing views.
Interest in selective perception and selective exposure has been with us since the earliest election studies (e.g. Lazarsfeld, Berelson, and Gaudet 1944), but only recently have these basic ideas taken root in more complex models of information processing. Selectivity and biased processing represent one of the most active areas of research in recent political psychology. As the number of avenues for obtaining political information has increased, political psychologists want to know whether citizens select sources that are more likely to reinforce their existing views. Further, to what extent is new information interpreted and processed so as to reinforce existing beliefs, and to what extent are citizens responsive to new information?
One prominent example of the emphasis on motivated reasoning is Lodge, Taber, and colleagues’ work suggesting that all political concepts are affectively charged as positive or negative, and that this information is stored in long-term memory (see e.g. Taber, Lodge, and Glathar 2001; Lodge and Taber 2005). New information is not necessarily retained, but it is used to update the affective tags that are attached to these concepts in memory. When asked for an evaluation of a political concept, citizens are said to recall the affective tally attached to the concept. Feelings serve as a summary of information that is no longer accessible in memory. This model represents a relatively rational approach to choice, though not necessarily in the Bayesian sense of rational updating.
However, as Lodge, Taber, and colleagues (e.g. Taber, Lodge, and Glathar 2001) have pointed out, an accurate model of political reasoning must take into account that it is often motivated by goals other than accuracy. In their motivated reasoning model, the online tally is not simply an unbiased account of previously (p. 353) encountered information. Instead, directional goals continually alter the processing and integration of new information into the tally. To the extent that the goal is to maintain one’s prior beliefs (as opposed to pursuing accuracy), people may ignore or devalue contrary information. They may also seek evidence selectively, biasing the considerations they draw from memory, exercising different levels of scrutiny for disconfirming evidence, and/or altering the weights attached to different criteria in a way that is less threatening to the prior belief. According to this model, directional goals “emerge spontaneously as the affective tags associated with elements of the problem represented in long-term memory are brought into working memory (hot cognition).”
According to this model, the direction and strength of affect toward a political person or idea will cause most citizens to be “biased reasoners” who fail to treat new evidence fairly: “Most citizens most of the time will be decidedly ‘partisan’ in what and how they think about and reason about political leaders, groups, events, and issues” (185). Interestingly, advocates of this model suggest that it is neither wholly a vice nor a virtue. On the one hand, an online tally provides a better summary of one’s past evaluations than preferences based on the recollection of specific pros and cons that happen to come to mind at any given point in time. The online model thus implies that choices are based on more information than is evident in assessments of knowledge made at the time of the decision. On the other hand, that same affective tally biases the processing of subsequent information, and is, in that sense, normatively undesirable.
Lau and Redlawsk (2006) have constructed a closely related model of motivated reasoning based on behavioral decision theory. In their model of the vote choice, they focus on the process of decision making and how individual motivations influence the extent to which voters choose correctly. They begin by accepting the notion that preexisting preferences bias subsequent assimilation of information, but they attempt to determine where such motivations enter into this process. Using an interactive information board/computer screen that allows people to seek out information in order to make decisions, they suggest that bias enters into information gathering and processing at many points along the way to decision-making. Surprisingly, voters who use a classically rational decision-making process, that is, one involving a deep and balanced information search, “were in many circumstances less likely to make a correct decision compared to voters using an intuitive or fast and frugal strategy” (Lau and Redlawsk 2005, 23). Barker and Hansen (2005) likewise question whether more information and deeper cognitive processing is the answer to what ails citizens. They found that subjects who engaged in systematic cognitive processing had weaker and less consistent attitudes than subjects in a control group.
Two recent studies stake out the ground on both sides of this important debate over whether citizens ultimately make good use of information. Gerber and Green (1999) use aggregate opinion data to argue that selectivity and perceptual bias are actually not the norm when citizens take in new information. Using over-time aggregate data, they argue that Republicans, Democrats, and Independents all basically change their views in the same direction and to the same extent as a result of new information. (p. 354) Based on an analysis of presidential approval among Republicans, Democrats, and Independents, Gerber and Green (1999, 205) conclude that all three groups tend to go up and down together over time: “Only the faintest traces of selective perception are evidence from partisan tends in presidential approval. All three partisan groups move together—sometimes markedly—as party fortunes change.” They applaud this pattern as rational in both the colloquial and Bayesian sense of the term. In other words, citizens appear to demonstrate Bayesian learning, with all groups making equally good use of new information as it comes along. If people were truly biased processors, they argue, their views would not move in parallel in response to ongoing political events.
If Gerber and Green’s claim is correct, it has far-reaching consequences for some of the most widely believed tenets of mass political behavior. Partisanship, in this view, is simply a running tally of information and judgments that have occurred over time. It summarizes information efficiently but has no influence on choice independent of the information and value judgments that it encapsulates. This conceptualization stands in sharp contrast to the traditional idea of partisanship as a driving force in how people perceive, interpret, and respond to the political world. According to Gerber and Green, information is key to understanding the political fortunes of candidates and policies, and the public responds roughly as if it were updating its views accordingly.
For most political psychologists, Gerber and Green’s conclusion is shocking if not implausible. How could so many studies, laboratory and otherwise, demonstrate findings of resistance to counter-attitudinal information, particularly in the context of political views that have been relatively stable throughout a person’s lifetime? If prior views do, in fact, bias the processing of new information, one would expect this pattern to be observable in the realm of political decision making if it happens at all.
Interestingly, using the same standard model of Bayesian updating as the basis for his conclusion, Bartels (2000) suggests that biased processing is alive and well in the American public, with partisanship as its driving force. Bartels suggests that when oppositional partisan groups adjust their views in the same direction and to roughly the same extent over time, it is anything but evidence of Bayesian learning.
To help explain the basis for this difference of opinion, Figure 18.1 illustrates the same kind of over-time evidence that convinced Gerber and Green that political psychologists’ assumptions about biased processing were greatly exaggerated. As new information becomes available to all three groups—say, for example, news that the economy has improved—all three partisan groups move toward higher levels of presidential approval. The trendlines in Figure 18.1 exemplify this parallel movement in presidential approval, though obviously from groups that began with very different attitudes toward a Republican president in this hypothetical example. Downturns due to bad news such as economic decline would cause all three groups’ approval levels to plummet, as they do in this illustration between 1985 and 1988.
In contrast, Figure 18.2 provides an illustration of what Bartels thinks Bayesian learning should look like in over-time public opinion data. As new information (p. 355) becomes available—perhaps news that the economy has worsened—the three groups of partisans update their presidential approval ratings in light of their initial views. In contrast to the Gerber/Green expectation, downward movement of approval due to negative information is not even across all groups, but is more pronounced in groups that begin with higher levels of approval. This occurs in a Bayesian model because (p. 356) the new information is more significant to the extent that it contradicts initial expectations. So, for example, in Figure 18.2 the decline in approval between 1981 and 1982 produces a shallower slope for Democrats, whose expectations for the Republican president were quite low to begin with. For Republicans, negative information of this kind is more of a surprise given their generally positive expectations, thus the extent of impact is greater for this group as shown by the steeper downward slope between 1981 and 1982. Most importantly, the net effect of Bayesian updating is some convergence of opinion. Whether the news is positive or negative, the three lines ultimately move closer and closer together over time. And even when the new information that citizens must incorporate is outside the range of expectation—better than even what the most supportive expect, or worse than what the most oppositional political group expects—the differential change in light of expectations should still bring the groups closer together if they are processing via Bayesian learning.
Thus Bartels suggests that we should not find real data that mimics Figure 18.1 particularly reassuring in its implications. It substantiates, rather than refutes, the hypothesis of biased processing. Moreover, Bartels’s conclusion comports with the bulk of evidence in political psychology—that is, that partisans are indeed biased assimilators and that patterns of Bayesian convergence such as what is illustrated in Figure 18.2 are uncommon.
Neither model, however, takes us through the full range of possibilities for how citizens respond to new political information. Thus far we have discussed these models in terms of events and information with clear positive or negative implications that all citizens would share. News that pollution levels have increased, or that unemployment is down, for example, would be received as negative and positive news, respectively, by all citizens. But new information about position issues as opposed to valence issues could easily create polarization within a Bayesian framework. For instance, if the “new information” about the president is that he vetoed a gun control bill, then Republicans should move in the more positive direction, if at all, and Democrats in a more negative direction. In this scenario, Bayesian learners should, quite rationally, polarize.
Whatever their differences, biased processing models are typical of contemporary political psychology in that they share an underlying skepticism that information is the cure for all that ails the quality of political decisions. If people are not passive recipients of information, but rather active choosers, interpreters, and rationalizers, then the limitations of information become apparent.
We are, in one sense, at an early stage in research that models biased processing, still sorting out what qualifies as evidence and what does not. To understand this process more fully in the future, researchers must unpack the process of biased assimilation in order to understand how bias occurs in the selection of information sources, the credibility granted to those sources, the discounting of information, and the relative weights given to new information in updating preferences. These are all separate mechanisms by which new information could differentially affect partisan groups based on their initial predispositions. (p. 357)
3 Beyond Self-report: New Sources of Theory and Evidence
Methodologically, political psychology has been criticized for relying too heavily on cross-sectional survey data (e.g. Krosnick 2002). Although this criticism seems valid with regard to much of the past work in this subfield, a greater level of methodological pluralism is difficult to find in any other subfield within political science. Burgeoning pluralism is evident in the kinds of methods political psychologists use as well as in the types of measures they now employ to operationalize key concepts.
In comparing early research in political psychology with today’s studies, there is a striking difference in the extent to which political psychologists trust self-reports as a means of getting at the black box processes involved in formulating political choices. For example, when authors of classics such as The People’s Choice wanted to know why people voted the way they did, they simply asked them. In contrast, the consensus view today is that the reasons people offer for their decisions “are better understood as justifications of a decision that has already been made” (12). (See also Lau 1982; McGraw 2000; Rahn, Krosnick, and Breuning 1994.)
for better or worse, humans appear to have little ability to introspect about the actual causes of their attitudes and actions. Nonetheless, they are disturbingly facile at rationalizing the choices and actions that they make. I say “disturbing” because as social scientists, we may be led on many a wild goose chase by people’s abilities to rationalize their emotions and choices. In addition, it is disturbing to lose the comfort of believing that there is an accessible, transparent logic to individuals’ political choices.
A dramatic example of the need to be skeptical of self-report and introspective accounts of behavior is illustrated by Wegner (2002) in a study in which electrical stimulation was used unbeknownst to experimental subjects to force them to react involuntarily by standing up. Despite the fact that their decision to stand was completely outside of their control, a large percentage reported a logical reason why they did so. Our brains are apparently compelled to offer deliberate, conscious reasons for our actions, but these rationalizations may have little to do with what actually happens. If we cannot understand the origins of our decision to sit or stand, how can we possibly understand the origins of a far more complex decision such as a vote choice?
What options do intrepid explorers of the black box psychological processes underlying political choices have to turn to? The good news is that the methodological repertoire for political psychology has undoubtedly expanded over the past fifty years. In addition to the survey data that served as the initial springboard for interest in the psychology of political choices, scholars now make regular use of laboratory experiments as well.
But the expansion in methodologies has not been exclusively toward imitating the internal validity of psychologists’ laboratory studies. In addition, experimental (p. 358) designs embedded within surveys provide researchers with new insights into understanding the basis of sensitive and socially undesirable political opinions and behaviors such as non-voting (see Holbrook and Krosnick 2005) and negative attitudes toward racial minorities (Sniderman et al. 1991). What is more, field experiments have been brought back into the methodological mix as well, primarily by Green and his associates (see e.g. Green and Gerber 2002). Still others study the psychology of political decision making in the context of real world political choices, as Glaser (2002) did in his study of the effects of ballot structure on the outcome of school bond initiatives.
Recognizing that so much of what political psychologists want to know may transcend the realm of self-report or even self-awareness, scholars also increasingly pursue measures that do not require research participants’ conscious awareness or introspection. Response times in answering questions, for example, are used to better understand respondents’ associations between positive or negative attributes and racial groups. In the most sophisticated applications of these techniques, researchers use complex designs to understand the associative links that facilitate attitudes.
The two most widely used paradigms for evaluating implicit (as contrasted with explicit) attitudes, are the “implicit association test” (IAT; see Greenwald, McGhee, and Schwartz 1998), and the “bona fide pipeline” (BFP; see Fazio et al. 1995). The IAT measures the strength of an association between two target categories (e.g. black and white) and two attributes (e.g. good and bad) by having people categorize examples of the target and attribute categories at the same time. So, for example, respondents would be presented with a test stimulus (e.g. a picture of a flower), and asked to sort what they observe into one category if it is either Black or good, or a second category if the object is either White or bad. The speed with which they perform this task across a number of stimuli is then compared to the speed with which they perform the same task with the two groups switched so that they sort objects into either a Black/bad category or a White/good category. In this particular example, negative attitudes toward Blacks would be assessed by comparing response latencies on the Black-bad and White-good trials to the Black-good and White-bad trials. Interestingly, even when one knows how the test works and is aware of what is being measured, it is still next to impossible for respondents to falsify results by trying to respond more quickly to some pairings than others.
The BFP also measures implicit attitudes, but in this case a prime such as a Black or White face is presented before an adjective is shown. In this case, negative associations with Blacks would be demonstrated by faster latencies when Black faces and negative adjectives are shown, and slower latencies for Black faces followed by positive adjectives relative to the same latencies after the presentation of White faces.
Both techniques avoid the perils of self-report and solve social desirability biases. In studies of racial attitudes, they also predict race-related behaviors (Fazio and Olson 2003). Although these are controversial measures of racial prejudice and of negative attitudes toward groups (see e.g. Arkes and Tetlock 2004), they are uncontroversial as indicators of the associations that people maintain, whether they act on them or not. One might well ask whether they are really necessary to political psychology outside of (p. 359) a few particularly sensitive topics such as race. The answer to this question remains to be seen, but as political psychologists increasingly seek understandings of phenomena outside the realm of conscious awareness, techniques of this kind will undoubtedly become increasingly valuable.
Finally, another set of methods involving psycho-physiological approaches to political attitudes and behaviors has opened up new possibilities as political psychologists begin to see how social neuroscience and psycho-physiological measurement techniques may be useful for understanding political attitudes and behavior. Technological advances in our ability to observe physiological evidence of the processes underlying political choice have drawn a small group of scholars to incorporate the tools of neuroscience into their work. Although a thorough review of studies that employ psycho-physiological and social neuroscience approaches is beyond the scope of this chapter, a special issue of Political Psychology published in 2003 (Volume 24: 4) provides useful examples of how social neuroscience is increasingly incorporated into political psychology. Given the field’s focus on understanding real world political events, these techniques are not likely to replace traditional methods within political psychology, but they are a very promising means of augmenting our limited access to people’s internal states.
Recently political psychologists also have begun to draw on evolutionary psychology as a basis for understanding reactions to the political world. For example, Alford, Funk, and Hibbing (2005) use the results of twin studies to distinguish the environmental determinants of political attitudes from their inherited traits. They conclude that attitudes toward a wide variety of political issues, as well as affect toward the major parties, is significantly influenced by genetic predisposition. Likewise, Sidanius and colleagues’ theory about the role of gender in social dominance orientation is rooted in evolutionary psychology. Mutz and Reeves (2005) also draw on evolutionary psychology to understand viewers’ reactions to incivility in televised political discourse.
To be sure, the potential applications of these approaches to political choice are in their infancy, but they appear relevant to some of the very same questions political psychologists have been trying to answer for years. For example, brain imaging studies demonstrate that activity in one area of the brain can bias what goes on elsewhere in the brain, thus bolstering conclusions about biased processing. Moreover, there appears to be no centralized location in the brain for integrating information and making choices (see Alford, Hibbing, and Smith 2005). Thus there is unlikely to be any one calculus for political decision making.
To date, very little of this evidence is directed toward answering the kinds of questions that plague political psychology, but the implications are clear. For example, McClure et al. (2004) show that judgments made about immediate versus delayed gratification activate different areas of the brain. As Alford and colleagues explain, “the time element stimulated different parts of the brain that are associated with different functions. Specifically, the possibility of immediate gratification seems to activate the emotional part of the brain, but when immediate gratification is not an option, the more reflective and cognitive part of the brain is activated.” As (p. 360) political scientists ponder how promises of tax cuts influence choice relative to long-term promises to protect the environment, such findings may well become applicable.
4 Information as the Gold Standard
It is a profoundly erroneous truism … that we should cultivate the habit of thinking about what we are doing. The precise opposite is the case. Civilization advances by extending the number of operations which we can perform without thinking about them. Operations of thought are like cavalry charges in a battle—they are strictly limited in number, they require fresh horses, and must only be made at decisive moments. (Alfred North Whitehead, 1911)
If an extraterrestrial took a cursory glance at the books published in political psychology over the past fifteen years, she would come away with the impression that what we earth people value in our citizens is information, reason, and rationality. Consider, for example, Ferejohn and Kuklinski’s Information and Democratic Processes (1990), Popkin’s The Reasoning Voter (1991), Sniderman, Brody, and Tetlock’s Reasoning and Choice (1991), Lupia, McCubbins, and Popkin’s Elements of Reason: Cognition, Choice, and the Bounds of Rationality (2000), Page and Shapiro’s (1992) The Rational Public, and so forth. These books do not concur on all matters, but the desirability of rational, well-informed political choices resonates throughout all of these volumes.
A closer look would reveal that the bulk of studies concur that people do not have loads of information about politics—indeed, far from it. But this closer examination would nonetheless suggest that most political psychologists wish citizens had perfect information, and think the political process would be far better off if citizens could at least better approximate this goal. As Kuklinski (2002) has suggested, rational choice assumes citizens are even-handed processors of information, while political psychology tends to assume (and to find) that they are not, though it nonetheless argues that they should be.
In this respect, Whitehead’s statement above may seem an anathema from the perspective of political psychology. What could be more sacred than the idea that good citizens should put a great deal of thought into the political choices they make? Are we, indeed, depleting citizens’ resources by asking them to make too many political decisions? Or are we reaching the wrong conclusions by assuming that the best decisions are ones made based on the most information? It is worth remembering that the well-educated citizen was not always the gold standard in politics.
Contemporary political psychology is beginning to question whether a classic rational decision-making process is truly what political psychology should pursue as its gold standard. In all three of the areas discussed in this chapter, political (p. 361) psychologists are reconsidering the emphasis on information and cognition as the root of ideal political choice. Studies of emotion and politics suggest that emotion is equally, if not more, important to political choice than cognition, and they question whether that is necessarily a bad thing. Studies of information processing suggest that information is severely limited in its capacity to improve political choice given the extent of biased processing; moreover, rational decision making does not necessarily mean better choices. As new approaches to measurement are applied to political choice, they further suggest that much of human decision making—political or otherwise—may be driven by processes of which citizens are not aware.
Taken together, these trends suggest that one of the most long-lasting premises of political decision making—that information gathering, thinking, and reasoning make for superior political decisions relative to visceral, subconscious reactions—is being called into question. Whereas political psychologists in the past have thought of citizens as information processors, they are rapidly becoming seen as less purposeful and as having less conscious control over their preferences. Whether such a representation of citizen choice is more accurate than the citizen as rational processor and/or more normatively desirable remains to be seen. In an era when voters are being asked to make more individual political choices than ever before, the horses may indeed need rest.
Alford, J., Funk, C., and Hibbing, J. 2005. Are political orientations Genetically Transmitted? American Political Science Review, 99: 153–67.Find this resource:
——Hibbing, J., and Smith, K. 2005. The challenge evolutionary biology poses for rational choice. Paper presented at the Annual Meeting of the American Political Science Association, Washington, DC.Find this resource:
Arkes, H. and Tetlock, P. 2004. Attributions of implicit prejudice, or “would Jesse Jackson ‘fail’ the implicit association test?” Psychological Inquiry, 15(4): 257–78.Find this resource:
Bargh, J. and Chartrand, T. 1999. The unbearable automaticity of being. American Psychologist, 54: 462–79.Find this resource:
Barker, D. and Hansen, S. 2005. All things considered: systematic cognitive processing and electoral decision making. Journal of Politics, 67 (2).Find this resource:
Bartels, L. 1988. Presidential Primaries and the Dynamics of Public Choice. Princeton: Princeton University Press.Find this resource:
——2000. Partisanship and voting behavior, 1952–1996. American Journal of Political Science, 44: 35–50.Find this resource:
——2002. Beyond the running tally: partisan bias in political perceptions. Political Behavior 24: 117–50.Find this resource:
Blais, A., Gidengil, E., Nadeau, R., and Nevitte, N. 2001. Measuring party identification: Britain, Canada and the United States. Political Behavior, 23 (1): 5–22.Find this resource:
Bless, H. 2001. The consequences of mood on the processing of social information. Pp. 391– 412 in Blackwell Handbook of Sociology: Individual Processes, ed. A. Tesser and N. Schwartz Malden, Mass.: Blackwell.Find this resource:
(p. 362) Brader, T. 2005. Striking a responsive chord: how political ads motivate and persuade voters by appealing to emotions. American Journal of Political Science, 49: 388–405.Find this resource:
——and Corrigan, B. 2005. Emotional cues and campaign dynamics in political advertising. Paper presented at the Annual Meeting of the American Political Science Association, Washington, DC.Find this resource:
Breckler, S. 1984. Empirical validation of affect, behavior and cognition as distinct components of attitude. Journal of Personality and Social Psychology, 47: 1191–205.Find this resource:
Cacioppo, J. and Visser, P. 2003. Political psychology and social neuroscience: strange bedfellows or comrades in arms? Political Psychology, 24: 647–56.Find this resource:
Campbell, A., Converse, P., Miller, W., and Stokes, D. 1960. The American Voter. New York: Wiley.Find this resource:
Civettini, A. and Redlawsk, D. 2005. A feeling person’s game: affect and voter information processing and learning in a campaign. Presented at the Annual Meeting of the American Political Science Association, Washington, DC.Find this resource:
Fazio, R. H., Jackson, J. R., Dunton, B. C., and Williams, C. J. 1995. Variability in automatic activation as an obtrusive measure of racial attitudes: a bona fide pipeline? Journal of Personality and Social Psychology, 69: 1013–27.Find this resource:
——and Olson, M. A. 2003. Implicit measures in social cognition research: their meaning and use. Annual Review of Psychology, 54: 297–327.Find this resource:
Ferejohn, J. and Kuklinski, J. (eds.) 1990. Information and Democratic Processes. Urbana: University of Illinois Press.Find this resource:
Fiorina, M. 2002. Parties and partisanship: a 40-year retrospective. Political Behavior, 24: 93–115.Find this resource:
Gerber, A. and Green, D. 1999. Misperceptions about perceptual bias. Annual Review of Political Science, 2: 189–210.Find this resource:
Glaser, J. 2002. White voters, black schools: structuring racial choices with a checklist ballot. American Journal of Political Science, 46: 35–46.Find this resource:
——and Salovey, P. 1998. Affect in electoral politics. Personality and Social Psychology Review, 2: 156–72.Find this resource:
Green, D. and Gerber, A. 2002. Reclaiming the experimental tradition in political science. Pp. 805–32 in Political Science: The State of the Discipline, ed. H. V. Milner and I. Katznelson, 3rd edn. New York: W. W. Norton.Find this resource:
Greenwald, A. G., McGhee, D., and Schwartz, J. L. K. 1998. Measuring individual differences in cognition: the implicit association task. Journal of Personality and Social Psychology, 74: 1469–80.Find this resource:
Holbrook, A. and Krosnick, J. 2005. Vote over-reporting: testing the social desirability hypothesis in telephone and internet surveys. Paper presented at the 2005 Conference of the American Association for Public Opinion Research. Miami.Find this resource:
Isbell, L. and Ottati, V. 2002. The emotional voter. Pp. 55–74 in The Social Psychology of Politics, ed. V. C. Ottati, S. Tindale, et al. New York: Kluwer.Find this resource:
Kaid, L. and Johnston, A. 2001. Videostyle in Presidential Campaigns. Westport, Conn.: Praeger.Find this resource:
Krosnick, J. 2002. Is political psychology sufficiently psychological? Distinguishing political psychology from psychological political science. Pp. 187–216 in Kuklinski 2002b.Find this resource:
Kuklinski, J. (ed.) 2001. Citizens and Politics: Perspectives from Political Psychology. New York: Cambridge University Press.Find this resource:
——(ed.) 2002. Thinking about Political Psychology. New York: Cambridge University Press.Find this resource:
(p. 363) Ladd, J. and Lenz, G. 2004. Emotions and voting behavior: a critique. Paper presented at the annual meeting of the Midwest Political Science Association, Chicago.Find this resource:
Lang, P. S., Bradley, L. M., and Cuthbert, B. N. 1997. International Affective Picture System (IAPS): technical manual and affective ratings. NIMH Center for the Study of Emotion and Attention, retrieved on January 30, 2006 from www.unifesp.br/dpsicobio/adap/instructions.pdf
Lau, R. R. 1982. Negativity in political perception. Political Behavior, 4: 353–78.Find this resource:
——and Redlawsk, D. 2005. Toward a procedurally plausible model of the vote choice: decision strategies, information processing, and correct voting. Paper presented at the annual meeting of the American Political Science Association, Washington, DC.Find this resource:
——2006. How Voters Decide: Information Processing during Election Campaigns. New York: Cambridge University Press.Find this resource:
Lazarsfeld, P., Berelson, B., and Gaudet, H. 1944. The People’s Choice. New York: Duell, Sloan and Pearce.Find this resource:
Levendusky, M. 2005. Sorting, not polarization: understanding the dynamics of mass partisan change. Midwest Political Science Association Meetings, Chicago.Find this resource:
Lodge, M. and Taber, C. 2005. Implicit affect for candidates, parties and issues: an experimental test of the hot cognition hypothesis. Political Psychology.Find this resource:
Lupia, A., McCubbins, M., and Popkin, S. (eds.) 2000. Elements of Reason: Cognition, Choice, and the Bounds of Rationality. New York: Cambridge University Press.Find this resource:
McClure, S., Laibson, D., Loewenstein, G. and Cohen, J. 2004. Separate neural systems value immediate and delayed monetary rewards. Science, 306: 503–7.Find this resource:
McGraw, K. M. 2000. Contributions of the cognitive approach to political psychology. Political Psychology, 21: 805–32.Find this resource:
Marcus, G., Neuman, W., and MacKuen, M. 2000. Affective Intelligence and Political Judgment. Chicago: University of Chicago Press.Find this resource:
Mutz, D. C. and Reeves, B. 2005. The new videomalaise: effects of televised incivility on political trust. American Political Science Review, 99 (1): 1–15.Find this resource:
Niedenthal, P. M., Rohman, A., and Dalle, N. 2003. What is primed by emotion concepts and emotion words? Pp. 307–33 in The Psychology of Evaluation: Affective Processes in Cognition and Emotion, ed. J. Musch and K. C. Klauer. Mahwah, NJ: Erlbaum.Find this resource:
Olson, M. and Fazio, R. H. 2003. Relations between implicit measures of prejudice: what are we measuring? Psychological Science, 14 (6): 636–9.Find this resource:
Page, B. and Shapiro, R. 1992. The Rational Public: Fifty Years of Trends in Americans’ Policy Preferences. Chicago: University of Chicago Press.Find this resource:
Popkin, S. 1991. The Reasoning Voter: Communication and Persuasion in Presidential Campaigns. Chicago: University of Chicago Press.Find this resource:
Rahn, W., Krosnick, J., and Breuning, M. 1994. Rationalization and derivation process in survey studies of political candidate evaluation. American Journal of Political Science, 38: 582–600.Find this resource:
Sears, D., Huddy, L., and Jervis, R. 2003. The psychologies underlying political psychology. Pp. 3–16 in Oxford Handbook of Political Psychology, ed. D. Sears, L. Huddy, and R. Jervis. Oxford: Oxford University Press.Find this resource:
Sidanius, J. and Pratto, F. 2001. Social Dominance: An Intergroup Theory of Social Hierarchy and Oppression. Cambridge: Cambridge University Press.Find this resource:
Sniderman, P., Brody, R., and Tetlock, P. 1991. Reasoning and Choice: Explorations in Political Psychology. New York: Cambridge.Find this resource:
——Piazza, T., Tetlock, P., and Kendrick, A. 1991. The new racism. American Journal of Political Science, 35: 423–47.Find this resource:
(p. 364) Taber, C., Lodge, M., and Glathar, J. 2001. The motivated construction of political judgments. In Citizens and Politics: Perspectives from Political Psychology, ed. J. H. Kuklinski. New York: Cambridge University Press.Find this resource:
Weber, C., Lodge, M., and Taber, C. 2005. Subliminal priming and political campaigns: the impact of subliminally presented affective primes on campaign ad evaluations. Paper presented to the annual meetings of the American Political Science Association, Washington, DC, September.Find this resource:
Wegner, D. 2002. The Illusion of Conscious Will. Cambridge, Mass.: MIT Press.Find this resource:
Whitehead, A. 1911. An Introduction to Mathematics. New York: Holt.Find this resource:
Zajonc, R. 1980. Feeling and thinking: preferences need no inferences. American Psychologist, 35: 151–75.Find this resource: