Heterogeneous Political Judgment, p. 1
On the Motivational Bases of Heterogeneity in Political Judgment Strategies
Abstract - What factors prompt citizens to switch from a partisan cue-taking judgment strategy, one in which they reflexively side with the in-group in policy and electoral contests, to a more thoughtful one, in which they pause to consider more valuable information? Previous work suggests that heterogeneity in political reasoning is triggered by the experience of anxiety. In this research, we examine a broader consideration: whether partisan cues provide adequate confidence in the quality of one’s judgments. Using ANES panel studies, we examine how the emotions of anxiety and enthusiasm influence the manner in which voters appraise presidential candidates, update opinions on policy issues, and form perceptions of the economy. The results consistently indicate that heterogeneity does not depend exclusively on anxiety, but on whether a given emotion harmonizes or conflicts with one’s partisan identity. Findings are discussed in terms of the paths to adaptive partisanship and responsible citizenship.
Word Count – 8,490
Heterogeneous Political Judgment, p. 2
Over the past half century, two broad questions have dominated the study of mass political behavior. First, how – by what reasoning processes – do ordinary citizens form their political judgments? And second, how “good” are they from a normative perspective? To address these questions, political scientists have focused on variation in two key dimensions of mass political thought: its depth and objectivity (Lau and Redlawsk 2006; Sniderman, Brody and
Tetlock 1991). Consistent with the portrayal of the American public as awash in political ignorance and lacking in political interest (Bennett 1989; Delli Carpini and Keeter 1996), empirical studies indicate that judgments about policies and candidates tend to be quick, reflexive, and shallow, often bearing little relation to the most diagnostic information (Druckman
2001; Kam 2005; Rahn 1993). Moreover, research on motivated reasoning demonstrates that citizens are often more interested in gratifying partisan expectations than in forming evenhanded judgments that respect the available “evidence” (Bartels 2002; Jacobson 2010; Lodge and Taber
2012; Redlawsk 2002; Taber and Lodge 2006). In summarizing their work on political information processing, Lau and Redlawsk (2006: 13) conclude that “most decisions are better understood as semiautomatic responses to frequently encountered situations than as carefully weighed probabilistic calculations of the consequences associated with the different alternatives.”
At times, however, citizens are more thoughtful, critical, and open-minded in their approach to politics (e.g., Basinger and Lavine 2005; Hillygus and Shields 2008; Kam 2006;
Marcus et al. 2000). Rather than digging in their partisan heels and ignoring readily available information, individuals are occasionally willing to devote substantial cognitive resources in making their judgments, and to show a stronger desire to be accurate than to shore up their partisan identities. For example, Hillygus and Shields (2008) find that voters are responsive to
Heterogeneous Political Judgment, p. 3 campaign information when they disagree with their party on a personally important policy issue.
Under these circumstances, citizens turn out not to be hard-headed motivated reasoners, but rational (i.e., prospectively-oriented) policy voters. In related work, Kam (2006) finds that intense campaigns for the House and Senate increase the extent to which voters form balanced reactions toward the candidates. She concludes that “intense campaigns not only give citizens food for thought, but they also encourage them to digest it more slowly – to ruminate and engage in open-minded thinking” (p. 939). This variation in political thinking has led to the identification of two idealized strategies by which people learn about, evaluate and engage with the political world: resolute partisanship and reflective deliberation (MacKuen, Wolak, Keele and Marcus 2010).
The two forms of reasoning serve different goals and lead to judgments of varying normative quality. Reflective deliberation – in which citizens “consider alternative views, switch perspectives, [and] consider new and possibly contradictory evidence” (Kam 2006: 932) – is a validity-seeking strategy in which political beliefs reflect an attempt to grapple with the substantive merit of competing claims. By contrast, resolute partisanship serves directional goals
(i.e., coming to pre-fabricated conclusions designed to uphold standing commitments), facilitating political engagement and commitment to the partisan in-group. It can also promote a tribal “us” versus “them” mindset, enhancing psychological equanimity at the cost of accurately perceiving political reality or understanding the implications of one’s preferences. Steadfast partisan cue-taking also provides an opportunity for elites to engage in manipulation, i.e., to craft particular phrases and presentations to “change public opinion and create the appearance of responsiveness as they pursue their [own] desired policy goals” (Jacobs and Shapiro 2000: iv).
Heterogeneous Political Judgment, p. 4
The major question we wish to address in this article is when— under what conditions— will citizens be motivated to think in a deliberative manner about their political choices, and more importantly, why ? Specifically, what prompts citizens to switch from a partisan cue-taking strategy, one in which they reflexively side with the in-group in policy and electoral contests, to a more thoughtful one, in which they pause to consider more valuable information? For example, what determines whether citizens rely on their material circumstances in forming preferences on salient economic issues? Or when will they reason: “I favor a progressive tax system because I value social equality,” rather than “I favor a progressive tax system because I am a Democrat and that’s what Democrats favor.” While it might be argued that the two forms of reasoning produce the same outcomes – especially as the electorate has become increasingly “sorted” along partisan-ideological lines (Abranowitz 2010; Levendusky 2009) – the connections among party identification, income, policy preferences, and core values are not that strong.
1 Therefore, perceptions and preferences reached on the basis of partisan cues will perforce be quite different from those rooted in one’s values and interests.
Political scientists have often acknowledged that in constructing their judgments, individuals take account of different considerations and attach different weights to them (e.g.,
Basinger and Lavine 2005; Gomez and Wilson 2001; Lau and Redlawsk 2006; Sniderman et al.
1991; Zaller 1992). To date, research has focused most frequently on dispositional differences in cognitive ability. Whether referred to as “level of conceptualization,” “sophistication,”
“awareness,” or “knowledge,” or measured simply as years of education, the traditional view is that a small number of “able” citizens are more likely than the ill-informed masses to form their
1 Data reported by Levendusky (2009: 46-49) and Abramowitz (2010: 45) indicate that the percentage of the electorate that is sorted on policy issues has increased only weakly between 1972 and 2004. As examples, consider that the zero-order correlations between party identification and tax preferences (using data from the 2008 ANES
Panel Study) is .35; the correlation between party ID and income (using the same 2008 data) is .16; and the correlation between party ID and support for limited government in 2004 is .35.
Heterogeneous Political Judgment, p. 5 political judgments using complex decision rules that focus on the most diagnostic information
(Delli Carpini and Keeter 1996; Kam 2005; Sniderman et al. 1991).
However, sophistication turns out to be a double-edged sword. While it facilitates political understanding, it also makes it easier for citizens to engage in biased reasoning to defend their political attitudes (Bartels 2008; Gaines et al. 2007; Jacobson 2010; Lodge and
Taber 2012; Kuklinski et al. 2008; Taber and Lodge 2006). Traditional ability-centered approaches also ignore citizens’ motivation to acquire and use political information. If wellinformed citizens can manage to attain sufficient confidence in their political judgments through reliance on cost-saving cues alone (e.g., party identification), they should rationally choose to ignore other relevant information, even if that information is highly diagnostic to the decision, and even if they are capable of acquiring it (Downs 1957; Lupia and McCubbins 1998). Finally, as MacKuen et al. (2010) have noted, ability-related approaches assume that individuals apply the same decision strategies across contexts. By contrast, a large body of research in psychology
(and a growing one in political science) indicates that people reach their political judgments through a variety of strategies, depending on the situation (for reviews, see Chaiken and Trope
1999; Druckman and Lupia 2000; Lavine et al. 2012; Payne, Bettman and Johnson 1993).
In the next section, we review research that focuses on voters’ incentives to behave more as resolute partisans or reflective deliberators.
2
In particular, we contrast two broad motivational perspectives on the circumstances under which voters are expected to rely principally on partisanship in making their policy and electoral judgments, and when they are expected to think more deeply and evenhandedly about their choices: affective intelligence theory (Marcus et al.
2000, 2011) and partisan ambivalence theory (Basinger and Lavine 2005; Hillygus and Shields
2 We are not suggesting that these two strategies constitute discrete categories; rather, they are theoretical endpoints of a continuum along which individuals rely increasingly on costly information (see Chaiken et al. 1989; Petty and
Cacioppo 1986).
Heterogeneous Political Judgment, p. 6
2008; Lavine et al. 2012). We then present several empirical tests to distinguish between these two perspectives, and conclude that responsible citizenship need not be precipitated by the experience of anxiety, but may come about through other pathways.
Motivational Theories of Heterogeneous Political Judgment
The hypothesis that political judgments depend on citizens’ informational incentives can be traced to Downs (1957). However, it was Marcus and colleagues who provided both a powerful empirical demonstration of the primacy of motivation as well as a rigorous intellectual framework to explain it (MacKuen et al. 2010; Marcus 1988; Marcus and MacKuen 1993;
Marcus et al. 2000). Relying on insights from neuroscience (e.g., Damasio 1994; Grey 1981;
LeDoux 1996), they argue that the nature of political reasoning lies in the experience of emotion
Specifically, Marcus and his colleagues have developed a theory of affective intelligence, in which the key emotions of enthusiasm and anxiety – viewed as the products of functionally distinct bio-behavioral regulatory systems – determine how citizens construct their political judgments. The “disposition system,” which mediates the emotion of enthusiasm, responds to positive incentives by directly initiating habit-based behavior. A second system, referred to by
Marcus et al. (2000) as the “surveillance system,” detects threat and danger in the environment.
Activation of the surveillance system results in the emotion of anxiety, which interrupts ongoing habitual action and promotes increased thoughtfulness and greater motivation for learning.
Applied to the realm of mass political judgment, enthusiasm is expected to motivate political participation and to increase voters’ reliance on their political habits (e.g., party identification). Anxiety, by contrast, is expected to lead voters to turn away from habitual responses and devote more attention to diagnostic information in the environment. In line with affective intelligence theory (hereafter AIT), Marcus et al. (2000) have shown that anxious
Heterogeneous Political Judgment, p. 7 voters (a) pay greater attention to the political environment and acquire more information about candidates’ policy stands; (b) rely less on partisanship and more on policy preferences and assessments of candidate character in forming candidate evaluations; and (c) defect at higher rates in presidential elections (see also Marcus and MacKuen 1993). Brader (2006) provided more powerful evidence of the role of anxiety by manipulating it experimentally in the context of political advertising. He found that compared to an emotionally pallid ad, a fear ad – which was intended to produce anxiety – led to greater attentiveness and more persuasion. Finally, anxiety – though not anger or enthusiasm – has been found to increase the quantity and quality of information processing, including the desire for balance and compromise (MacKuen et al. 2010;
Valentino, Hutchings, Banks and Davis 2008). Taken together, this work suggests that anxiety leads to political decisions that are more informed by contemporary information and less by partisan loyalty. In the language of dual-process theories of cognition, anxiety reduces heuristic processing and stimulates systematic processing.
The affective intelligence model has gained wide acceptance in political science.
However, reflective deliberation may arise through pathways other than anxiety. If opportunities for citizens to take a sober second look derive from a broader set of dynamics, we should expect to observe good citizenship (in the form of high quality judgments) more of the time. A large body of research on persuasion and social judgment in psychology indicates that people are motivated to attain a certain degree of confidence in the “correctness” of their decisions, and that they will seek out (and deliberate about) information until actual confidence matches or exceeds the desired level (Chaiken and Trope 1999). Most relevant here, when party behavior and performance undermine the reliability of partisan brand names (e.g., profligate Republican
Heterogeneous Political Judgment, p. 8 spending under George W. Bush), partisan cues become less trustworthy, and thereby lose their power as judgment guides (Lupia and McCubbins 1998).
Partisan ambivalence theory (Basinger and Lavine 2005; Lavine et al. 2012) proposes that the degree of confidence citizens may derive from partisan cues depends on the extent to which contemporary evaluations of party behavior and performance dovetail with expectations derived from long-term partisan identifications.
In line with the general theory of cognitive consistency (Festinger 1957), citizens should have a strong preference for harmonizing their affective attachments and their contemporary perceptions of performance. However, when the information flow about one’s own party becomes persistently negative (or when the other party performs consistently well), efforts to maintain equilibrium inevitably break down. Under normal circumstances, partisans may be quick to blame the out-party and credit the in-party, but they are not entirely immune to how economic swings, scandal, and the quality of domestic and foreign policy management reflect on party competence and leadership.
Expressed in psychological terms, the process of motivated reasoning is not unbounded . A disequilibrium is therefore likely to arise when perceptions of the parties’ behavior are out of step with normal expectations, for example, when the in-party (i.e., one’s own party) is plagued by scandal, when it embraces issue positions that are inconsistent with core ideological principles, or when it fields poor candidates, presides over economic downturn, or mismanages an international conflict or a domestic emergency.
The central idea is that when identity-based expectations are disconfirmed by perceptions of party behavior, the habit of partisanship loses much of its heuristic value. Specifically, this disjuncture undermines the belief that partisan cues can effectively substitute for more detailed
(but costlier) information. To pick up the judgmental slack, ambivalent partisans should engage
Heterogeneous Political Judgment, p. 9 in deeper, more effortful thought about their political options. For example, they may devote more attention and thought to candidates’ policy stands in deciding whether to support them, and may rely more on their personal economic circumstances in forming public policy preferences.
Moreover, ambivalent partisans should hold judgments that are less demonstrably plagued by partisan bias, as they are less likely to view political reality through a partisan lens.
Experimental research in psychology corroborates that ambivalence reduces judgment confidence, motivates individuals to devote greater cognitive resources to judgment tasks, and increases accuracy. In one study, Jonas, Diehl and Bromer (1997) manipulated ambivalence by associating a new consumer product with either evaluatively consistent or inconsistent features.
They found that those assigned to the latter (ambivalent) condition held less confidence in their initial attitudes toward the product, and as a result engaged in greater elaboration of its attributes in deciding whether to purchase it. In another persuasion experiment, Maio, Bell and Esses
(1996) linked ambivalence with deliberative thinking by presenting individuals with either strong or weak arguments in favor of increased immigration of a minority group. They found that among those with ambivalent attitudes toward the group, persuasion depended on the cogency of the substantive arguments, such that more opinion change was observed in the strong (vs. weak) arguments condition (indicating deliberative reasoning). By contrast, non-ambivalent respondents ignored the substantive merit of the arguments and simply relied on their prior group attitudes in expressing their policy opinions, indicating heuristic (habit-based) thinking.
Studies by political scientists on electoral perception and judgment provide converging evidence on the impact of ambivalence. For example, Meffert, Guge and Lodge (2004) found that ambivalence toward presidential candidates was associated with less confidence but greater accuracy in placing them on policy issues. Basinger and Lavine (2005) and Hillygus and Shields
Heterogeneous Political Judgment, p. 10
(2008) found that when voters held beliefs that conflicted with their partisan identities, they relied more on policy and less on partisanship in their voting decisions. And Lavine et al. (2012) found that ambivalent partisans relied more on real economic signals (e.g., changes in GDP and unemployment) and less on partisanship in forming retrospective judgments of economic performance. In sum, by motivating citizens to think more deeply and more objectively about their political options, partisan ambivalence may (1) be a broad determinant of whether citizens act more like resolute partisans or reflective deliberators, and (2) account for the effects of anxiety in past research (e.g., Marcus et al. 2000).
When Do Citizens Rely on Costly Information?
Affective intelligence holds that the motivation to engage in reflection and reevaluation hinges specifically on activation of the surveillance system, that is, on the perception of threat and the experience of anxiety. However, negative evaluations of one’s own party represents only one route by which a disjuncture between partisan identity and party evaluations may occur. It may also result from holding positive evaluations of the other party. For example, in signing
NAFTA and welfare reform and in acknowledging that “the era of big government is over,”
President Clinton improved the Democratic Party brand among moderate Republicans (Lebo and
Cassino 2007).
3
As approval of the out-party does not logically entail the activation of the surveillance system or the experience of anxiety (just the opposite we would think), its influence in stimulating deliberative thought is difficult to square with AIT. If positive evaluations of the out-party enhance deliberative reasoning just as do negative evaluations of the in-party, then: (1)
AIT is in need of theoretical revision; and/or (2) adaptive political behavior may arise from other circumstances.
3 As, of course, did the economic super-boom that Clinton presided over during his second term.
Heterogeneous Political Judgment, p. 11
The predictions of the two frameworks in the context of candidate evaluation are depicted in Figure 1.
4 According to AIT (see top panel), only feelings of anxiety modulate the degree to which party identification and issue preferences drive candidate evaluations. Enthusiasm toward the in-party candidate, by contrast, is hypothesized to exert a direct (positive) effect on voters’ electoral preferences. According to AIT, then, enthusiasm and anxiety exert functionally distinct effects on political judgment. According to partisan ambivalence theory (hereafter PAT; see bottom panel), the political effects of enthusiasm and anxiety depend entirely on whether they constitute identity-conflicting or identity consistent reactions. Anxiety experienced in relation to the candidate of one’s own party or enthusiasm for the candidate of the opposing party – hereafter “in-candidate anxiety” and “out-candidate enthusiasm” – should undermine judgment confidence on the basis of partisan cues, as both emotions (so directed) are in conflict with one’s partisan identity. In-candidate enthusiasm and out-candidate anxiety, however, reinforce one’s partisan identity, and should thus heighten citizens’ confidence in standing (partisan) decisions.
4 The illustrations are inspired by Ladd and Lenz (2008).
Heterogeneous Political Judgment, p. 12
Figure 1. Graphical Depictions of AIT and PAT
A. Affective Intelligence Theory
+
Enthusiasm
Candidate
Evaluation
+
Anxiety
-
Party ID
Issue
Preferences
Out-
Anxiety and In-
Enthusiasm
In-Anxiety and Out-
Enthusiasm
+
-
+
-
B. Partisan Ambivalence Theory
Party ID
Candidate
Evaluation
Issue
Preferences
-
Heterogeneous Political Judgment, p. 13
Hypotheses
We conducted several empirical tests to determine the factors that prompt citizens to move from a reflexive partisan cue-taking strategy to one in which they pause to consider more valuable (but costly) information. In each case, AIT predicts that only anxiety (especially toward the in-party candidate) instigates systematic thinking, i.e., less reliance on party and more reliance on costlier information. By contrast, PAT predicts that the experience of identityconflicting emotions (whether in the form of anxiety or enthusiasm) leads to more systematic
(and less habit-based) thinking, and that the experience of identity-consistent emotions has the opposite effects (i.e., increasing reliance on party and decreasing reliance on costlier information). Our first tests examine the pattern of heterogeneity in the context of electoral choice. We then examine the bases of preference updating on policy issues. Here we determine the extent to which updating reflects partisan polarization or increased alignment with material interests, and whether AIT or PAT better captures the pattern of observed heterogeneity. Last, we examine changes in economic perceptions from January to November of 2008. Although the economy was in steep decline throughout the year, Republicans – as in-partisans – were less likely than Democrats to hold veridical perceptions. We examine whether variation in judgmental accuracy among Republicans is better captured by AIT or PAT.
Data
Data and Variables
We rely on three datasets to distinguish between AIT and PAT: the American National
Elections Studies (ANES) cumulative file; the 2008 ANES Panel Study; and the 1992-1996
ANES Panel Study. The cumulative file merges all election studies from 1948-2008, and
Heterogeneous Political Judgment, p. 14 contains most items asked in at least three election years. The variables required for our analyses appear in all presidential election years from 1980-2008.
Variables
Candidate Emotions . Beginning in 1980, the ANES has asked respondents about their emotional reactions toward the presidential candidates. The question asks: “Has [candidate] – because of the kind of person he is, or because of something he has done – made you feel
[emotion term]?” In the February wave of the 2008 panel study, the question read: “When you think about (Barack Obama/John McCain), how [emotion term] does he make you feel?”
Responses were dichotomous (yes/no) in the 1980-2008 cumulative file surveys and in the 1992-
1996 panel study, and ranged from one (“not at all”) to five (“extremely”) in the 2008 panel study. In all surveys, respondents were asked about four discrete emotions: “afraid,” “angry,”
“hopeful,” and “proud.” Consistent with Marcus et al. (2000), we combined the proud and hopeful items separately for each candidate as a measure of candidate-specific enthusiasm, while responses to the afraid item were used as a measure of candidate-specific anxiety.
5
Partisanship and Partisan Strength. In all surveys, partisan identification was a dichotomized measure coded “1” for Republicans and “0” for Democrats (leaners included). In the 2008 panel study, we utilized the measure from the January wave of the survey. Partisan strength was measured by folding the 7-point ANES partisanship item at its midpoint, so that
0=leaners, 1=weak partisans and 2=strong partisans (independents excluded).
Comparative Issue Distance. Following Marcus et al. (2000), issue-based candidate evaluation was measured via comparative candidate issue distance. Specifically, we calculated the absolute distance between the respondent’s position on a given issue and his or her
5 While Marcus et al. (1993, 2000) originally measured candidate anxiety with both the “afraid” and “angry” items, they have recently reconceptualized the latter as “aversion” (Marcus, MacKuen, Wolak and Keele 2006; MacKuen et al. 2010; see also Valentino et al. 2008). We thus exclude the anger item from the measure.
Heterogeneous Political Judgment, p. 15 perception of each candidate’s position. For each issue, we then subtracted the absolute distance from the Republican candidate from the absolute distance from the Democratic candidate, and averaged across issues. In the cumulative file analysis, the issues were defense spending, foreign policy orientation toward the USSR, women’s role in society, government aid to black
Americans, government versus private health insurance, government spending and services, and government guaranteed jobs and income. We use the average for all available issues for each respondent in this calculation. In the 2008 panel study, the issues (assessed in the January wave for respondents’ positions, and in the June wave for candidate perceptions) included a constitutional amendment banning gay marriage, raising taxes on those making more than
$200,000 per year, raising taxes on those making less than $200,000 per year, governmentprovided prescription drugs for seniors on low income, government-provided healthcare for all
Americans, imprisonment of suspected terrorists without charging them with a crime, warrantless wiretapping, temporary work visas for illegal immigrants, and a path to citizenship for illegal immigrants.
Economic Retrospections . Economic perceptions were measured in both the January and November waves of the 2008 panel study. In each wave, the question read: “Now thinking about the economy in the country as a whole, would you say that as compared to one year ago, the nation’s economy is now better, about the same, or worse?” Responses on the final scale ranged from one (“much worse”) to five (“much better”).
Candidate Choice . Following Marcus et al.’s (2000, 2011) recommendation, candidate choice was assessed by voting intentions in the pre-election wave of the cumulative file surveys.
This measure was not available in the 2008 panel study (nor was vote choice). Therefore, in the
February and October waves of the study, we used a measure of comparative candidate
Heterogeneous Political Judgment, p. 16 evaluation, taking the difference between respondents’ attitudes toward McCain and Obama. The question read: “Do you like [Barack Obama/John McCain], dislike him, or neither like nor dislike him?” Responses ranged from one (“dislike a great deal”) to seven (“like a great deal”), with values of four corresponding to “neither like nor dislike.”
Controls.
In all analyses, we control for age, gender, education, income, and an indicator for black respondents. We also controlled for political knowledge and partisan strength, both as first-order terms and in interaction terms with party identification, comparative issue proximity
(in the candidate choice analyses), and material interest (in the preference updating analysis).
6
This provides some assurance that any contingent effects of emotion are not a function of dispositional differences in these traditional engagement variables. All variables were recoded on a zero to one scale prior to analysis.
Results
Heterogeneity in Candidate Appraisal Processes
We begin by modeling vote intentions in the ANES cumulative file, pooling the data from 1980 (when the emotion terms debuted) through 2008. Specifically, using ordered probit, we modeled voting intentions (Democrat=-1; Undecided=0; Republican=1) on the basis of the four emotion items (in- and out-candidate anxiety and enthusiasm), partisanship, comparative
6 In 2008, political knowledge was gauged using six civics questions: How many times can one be elected U.S. president? For how many years is a U.S. senator elected? How many U.S. senators are there from each state? For how many years are U.S. representative elected? What is the presidential succession after the Vice-President? What is the veto override percentage needed in the U.S. House and Senate? (α= .57). To maintain consistency across time, we utilize two items to measure knowledge in the ANES cumulative file: the interviewer’s subjective rating of the respondent’s political knowledge, and the respondent’s knowledge of which party controlled the House of
Representatives. In 1992, we rely on the interviewer’s subjective rating of the respondent’s knowledge, as the civics items asked in that survey were coded in such a way that missing data were conflated with incorrect and “don't know” responses.
Heterogeneous Political Judgment, p. 17 issue proximity and the controls. To gauge whether and how the emotion terms conditioned the bases of intentions, we also included eight interactions: party ID x each emotion term, and issue proximity x each emotion term. We also included separate interactions of political knowledge and partisan strength with both party ID and issue proximity (as well as the first-order terms). As
Figure 2 indicates, seven of the eight interaction terms involving emotion were statistically significant (all but for the interaction of issue proximity and in-candidate anxiety), with the overall pattern strongly supporting PAT (see column labeled “vote”). As the coefficients indicate, both types of identity-conflicting reactions – out-candidate enthusiasm and in-candidate anxiety – depressed voters’ electoral reliance on party; moreover, the former had the corresponding effect of heightening reliance on issues. By contrast, the two identity-consistent reactions – in-candidate enthusiasm and out-candidate anxiety – had the opposite effects: both factors heightened voters’ reliance on party and diminished their reliance on issues.
In Figure 3, we translate the coefficients into changes in the predicted probability of voting Republican as a function of partisanship (top panel) and issue proximity (bottom panel).
As the top panel shows, out-candidate enthusiasm reduced the marginal effect of partisanship by nearly 80 percentage points, whereas in-candidate enthusiasm had the opposite effect – increasing the role of partisanship by about 75 points. The role of anxiety si milarly depended on
Heterogeneous Political Judgment, p. 18
Figure 2. Regression Coefficients, All Models
Vote Evals08 Insure92 Taxes08 EconRetros08
Party ID
Issues/Econ
PID Strength
Sophistication
IA
OA
IE
OE
PID x Strength
PID x Soph
PID x IA
PID x OA
PID x IE
PID x OE
Iss/Econ x Strength
Iss/Econ X Soph
Iss/Econ x IA
Iss/Econ X OA
Iss/Econ x IE
Iss/Econ x OE
Lagged Prefs
Age
Gender
Black
Education
Income
Cut 1
Cut 2
Cut 3
Cut 4
Cut 5
Cut 6
-5 0 5 -1 -.5 0 .5 1 -2 0 2 4 -2 0 2 4 -2 0 2 4
Notes: IA=In-Party Anxiety, OA=Out-Party Anxiety, IE=In-Party Enthusiasm, OE=Out-Party
Enthusiasm.
Heterogeneous Political Judgment, p. 19
Figure 3. Conditional Effects of Partisanship and Issue Proximity on Presidential Voting,
1980-2008
A. Partisanship
In-Anxiety Out-Anxiety
Low
In-Enthusiasm
High
Out-Enthusiasm
B. Issues
In-Anxiety Out-Anxiety
Low
In-Enthusiasm
High
Out-Enthusiasm
Heterogeneous Political Judgment, p. 20 whether it constituted an identity conflicting or consistent reaction. Anxiety toward the in-party candidate reduced the effect of partisanship by 28 percentage points, whereas anxiety toward the out-party candidate magnified the role of party by 17 percentage points ( p s < .01). The bottom panel of Figure 3 shows the corresponding dynamics for issues. Per PAT, both identityconflicting emotions substantially heightened respondents’ reliance on issues: out-candidate enthusiasm by 73 percentage points and in-candidate anxiety by about 30 points. By contrast, both identity-consistent emotions substantially decreased respondents’ reliance on issues: incandidate enthusiasm by over 60 percentage points, and out-candidate anxiety by nearly 40 points. As AIT holds that only anxiety (i.e., not enthusiasm) conditions party and issue voting, it appears that PAT may provide a better explanation for the dynamic effects of emotion on political judgment.
However, because the emotion and vote intention variables were measured at the same time, it is possible that the former is endogenous to the latter ( i.e., disliking a candidate leads to anxiety; liking a candidate leads to enthusiasm).
Ladd and Lenz (2008) proposed such an endogenous affect mechanism to account for the indirect effect of anxiety in a pooled crosssectional analysis using the same ANES data presented in the leftmost numerical column of
Figure 2 (minus the elections of 2000, 2004, and 2008). Specifically, Ladd and Lenz argued that the positive interaction between anxiety and comparative policy distance could occur in the absence of anxiety playing a causal role. To test their endogenous affect explanation, Ladd and
Lenz modeled third wave candidate evaluation in the 1980 ANES Major Panel Study as a function of second wave explanatory variables (including anxiety and a lagged measure of candidate evaluation). If endogenous affect is responsible for the interactions (between anxiety
Heterogeneous Political Judgment, p. 21 and partisanship and anxiety and issue proximity), they should disappear when anxiety is measured in a prior wave. This is exactly what Ladd and Lenz found.
Perhaps, then, the pooled cross-sectional results we reported in Figures 2 and 3 can be explained by endogenous affect. To determine whether this is the case, we turn to the 2008
ANES Panel Study. In this analysis, we use lagged measures of party ID and issue proximity, and a lagged measure of candidate evaluation (measured in June). By controlling for prior candidate evaluations , we can examine how voters’ electoral preferences (in October) changed over the course of the campaign, and determine the extent to which such changes are driven by partisans versus policy considerations (measured in January). More importantly, by using panel data, we rule out endogenous affect by design (see Ladd and Lenz 2008). Specifically, we modeled comparative candidate evaluations in October as a function of evaluations in June, the four emotions items, partisanship, issue proximity, the relevant interaction terms, and all controls. The model was estimated via ordinary least squares with robust standard errors.
The statistical results, shown in Figure 2 (under the column “Evals”), indicate that the shift from party- to issue-based candidate evaluation is driven by enthusiasm (not anxiety), and in ways that (again) conform to the expectations of PAT. Predicted candidate evaluation scores, calculated across levels of the constituent terms of the interactions (holding all other variables at their means), are displayed in Figure 4. As the figure indicates, voters who are unenthusiastic about the other party’s candidate rendered their evaluations on the basis of partisan loyalty, with policy considerations making no statistically appreciable contribution: the difference in candidate evaluation between Republicans and Democrats is .29 (i.e., reflecting an effect of 29 percentage points), whereas the impact of issues (i.e., the difference between those at the 5 th and 95 th percentile values of the comparative issue proximity scale) is .05 (effectively
Heterogeneous Political Judgment, p. 22
Figure 4. Conditional Effects of Partisanship and Issue Proximity on Candidate
Evaluations in the 2008 Presidential Election
Partisanship Comparative Issue Distance
In-Enth Out-Enth
Low
In-Anx
High
Out-Enth
Heterogeneous Political Judgment, p. 23 zero). By contrast, those who do express enthusiasm for the other party’s candidate are issue voters. Among these respondents, the impact of party and policy are reversed: the difference in candidate evaluation between those with policy stands closer to Obama and those with policy stands closer to McCain is .40; the corresponding impact of party is only .11. Thus, all else equal, experiencing positive emotion toward the other party’s candidate substantially reduces voters’ electoral reliance on partisanship, and substantially increases their reliance on policy agreement. Moreover, voters who are enthusiastic about their own party’s candidate rely on party
ID at more than twice the rate of those lacking such enthusiasm, all else equal ( b s = .32 and .15; p < .01). Finally, and perhaps most tellingly, AIT’s signature effect – that in-candidate anxiety shifts voters’ electoral reliance from party to issues – failed to obtain.
7
In sum, none of these findings follow from AIT; however, they are consistent with PAT.
As voters experience more identity-conflicting feelings (in this case, enthusiasm), they derive less confidence in their electoral judgments on the basis of partisan loyalty. To pick up the slack, they turn to more diagnostic information, namely their issue preferences. By contrast, the more voters experience identity-consistent feelings, the more confidence they derive from partisan loyalty, and thus the more they use it as a judgmental yardstick (and the less they rely on the more difficult calculus of issue distance). Moreover, none of these findings can be accounted for by endogenous affect.
Preference Updating on Economic Issues: Partisan Polarization or Material Interest?
In this section, we examine systematic variation in how citizens update their policy preferences on economic issues in the wake of intense national debate. We focus first on how attitudes toward healthcare reform evolved over the first two years of Clinton’s presidency. We
7 Indeed, surprisingly, we find a marginally significant decline in the use of issues as a function of in-candidate anxiety. Given the inefficiency of the estimate, however, this effect should be treated with caution.
Heterogeneous Political Judgment, p. 24 then examine preference updating on taxes during the 2008 presidential campaign. Healthcare reform was the dominant policy emphasis during 1993-1994: the president devoted several major policy speeches to the topic; the issue was extensively covered in the media; and the insurance lobby funded an expensive television advertising campaign to defeat the eventual bill. The issue was thus highly salient to the public (Dancey and Goren 2010; Jacobs and Shapiro 2000).
We contrast two basic processes by which citizens may have responded to this flood of information. First, they may have chosen to receive and/or accept policy messages only from fellow partisans, leading to preference adjustments resulting in greater alignment with copartisan elites (Zaller 1992). Alternatively, they may have been less attuned to the partisan source of the message than to its actual content . If so, citizens may have given greater consideration to how the alternatives resonated with their own material circumstances.
According to AIT, anxiety should depress the influence of party-based preference and heighten the influence of contemporary information – in this case, how the parties’ positions fit with citizens’ current economic predicaments. PAT, by contrast, predicts that the switch from partisanship to material interest as the basis of preference updating will depend, as in the case of electoral judgment, on the broader pattern of identity conflicting and consistent emotions.
Specifically, the experience of identity-conflicting emotions should shift the locus of attitude change from partisan cues to personal economic predicaments, whereas the experience of identity-consistent emotions should depress the role of personal economics and facilitate partisan polarization.
Preference Updating, 1992-1994. The ANES included a battery of items in the 1992 wave of the 1992-1996 Panel Study to assess personal economic insecurity. These items ask whether in the past year, one (and one’s family) (a) was better or worse off than a year ago; (b)
Heterogeneous Political Judgment, p. 25 put off making planned purchases (including medical and dental treatments); (c) borrowed money from relatives, friends or a financial institution to make ends meet; (d) dipped into savings; (e) looked for a second job or worked more hours at one’s present job; (f) saved money
(reverse coded); and (g) fell behind on a rent or house payment. Together, the items provide a broad and internally consistent portrait of an individual’s current economic predicament (α =
.76). We averaged them to form a composite measure of personal economic insecurity
(0=secure; 1=insecure).
We used this measure to predict attitudes toward government-provided national healthcare in 1994. The item in both the 1992 and 1994 waves of the panel study read:
There is much concern about the rapid rise in medical and hospital costs. Some people feel there should be a government insurance plan which would cover all medical and hospital expenses for everyone.
Others feel that all medical expenses should be paid by individuals, and through private insurance plans like Blue Cross or other company paid plans. Where would you place yourself on this scale, or haven't you thought much about this? (1=government insurance plan; 7=private insurance plans)
The analysis includes the same controls as before (all measured in 1992), as well as a lagged
(1992) measure of health care preference.
Unsurprisingly, citizens use partisanship to update their preferences over time: controlling for baseline differences in 1992, Republicans were 17 percentage points more conservative in their preferences than Democrats by 1994, holding all other variables at their central tendencies. By contrast, the corresponding effect for personal economic insecurity is -.02 (effectively zero).
The key question we wish to address is how the emotion terms altered the influence of these two focal explanatory variables on preference updating. Most importantly, do the effects of partisanship and material circumstances (which are only weakly correlated, r
1992
= .26, with
Republicans being more secure) depend on anxiety (as AIT claims) or on whether the emotions constitute identity conflicting or consistent reactions (as PAT claims)? Figure 2 (column labeled
Heterogeneous Political Judgment, p. 26
Figure 5. Conditional Effects of Partisanship and Personal Economic Condition on Health
Insurance Preferences
Partisanship Personal Economic Condition
Out-Anxiety
Low High
Out-Enthusiasm
Heterogeneous Political Judgment, p. 27
“Insure92”) reveals two significant conditional effects of emotion: anxiety toward the other party’s candidate heightened the impact of partisanship on preference change, and enthusiasm toward the other party’s candidate heightened the impact of personal economic insecurity. Both findings follow from PAT (but not AIT). As Figure 5 shows, a positive change in identityconsistent emotion (qua out-candidate anxiety) more than tripled the impact of partisanship on preference updating (from .10 to .30, p <.05), with Democrats moving to the left and Republicans moving to the right. Strikingly, policy updating among these individuals was entirely unrelated to their contemporaneous economic predicaments. Thus, for example, economically insecure
Republicans who experienced anxiety toward Bill Clinton in 1992 ignored the policy implications of their pocketbooks and toed the party line in opposing health care reform.
Figure 5 also shows that identityconflicting emotion (in this case, out-candidate enthusiasm) substantially heightened the effect of personal economic insecurity. All else equal, preference updating among these individuals was driven primarily by material considerations, the effect of which increased from -.01 to .42 (indicating that the economically secure are 42 percentage points more opposed to reform than the insecure, controlling for baseline preferences in 1992). This is a substantively meaningful and statistically significant shift.
Preference Updating, January-October 2008. We now consider the extent to which partisanship shaped preferences over taxes during the 2008 presidential campaign. Both candidates clearly signaled their preferences on the issue, with Obama vowing to let the Bush tax cuts expire on the top two percent of income earners, and McCain pledging to maintain them.
We model October tax preferences on the wealthy (>$2000,000) as a function of lagged preferences in January, partisanship, the emotion items, their interactions with partisanship, and all controls. As the 2008 ANES panel study does not contain items sufficient to
Heterogeneous Political Judgment, p. 28
Figure 6. Conditional Effects of Partisanship on Tax Preferences in the 2008 Presidential
Election
In-Enth Out-Anx
Low High
Out-Enth
Heterogeneous Political Judgment, p. 29 operationalize economic insecurity, we examine the conditional influence of partisanship only.
Specifically, we gauge the extent to which Republicans and Democrats polarized with respect to their preferences over the course of the campaign. The key interaction estimates, shown in
Figure 2 (under the column “Taxes08”) support PAT. Both identity-consistent emotions (outcandidate anxiety and in-candidate enthusiasm) facilitate partisan polarization, while the identity-conflicting emotion of out-candidate enthusiasm decreases polarization. The corresponding predicted probabilities are plotted in Figure 6. For the two identity-consistent emotions, moving from low to high increases polarization by 15 and 26 percentage points, respectively. Conversely, for the identity-conflicting emotion (out-candidate enthusiasm), a change from low to high decreases polarization by 16 points.
Perceptions in Flux: The Economic Collapse of 2008
The 2008 ANES panel study provides an opportunity to examine the bases of individuallevel change in economic perceptions over the course of a presidential campaign in which the outlook changed dramatically. Of particular interest here is perceptual heterogeneity among
Republicans (as partisan supporters of the president during a precipitous downtown). We model changes in perceptions between January and November, a period that straddled several highprofile economic events, including a plunge in the Dow Jones of more than 5,000 points (a nearly 40 percent drop), the collapse of two major Wall St. financial institutions, and dire warnings in late September by the Fed Chairman and Treasury Secretary (and ultimately the president) of a complete meltdown in the U.S. economy. Our primary question is whether variability in the pattern of belief updating among Republicans is better predicted by AIT or
PAT. To address this question, we model citizens’ economic retrospections in November as a function of their retrospections in January, partisanship, the four
Heterogeneous Political Judgment, p. 30
Figure 7. Conditional Partisan Bias in Economic Retrospections, November 2008
Democrats Republicans
MW SW SS MW SW SS
Low Out-Enthusiasm
MW SW SS MW SW SS
High Out-Enthusiasm
Heterogeneous Political Judgment, p. 31 emotions items, the relevant interactions and all controls. By controlling for lagged economic perceptions and by measuring the explanatory variables at lagged time points relative to perceptions in November, the potential for endogenous affect to distort the estimates is removed by design.
We estimated the model via maximum likelihood with an ordered probit link. The key interactions, displayed in Figure 2 (under the column labeled “Retros08”), indicate a substantial decline in the effect of partisanship as out-candidate enthusiasm increases, a result consistent with PAT. Indeed, neither of the anxiety items conditions partisan bias in economic evaluations.
In Figure 7, we plot the predicted probabilities of membership in each of the three populated retrospection categories (“much worse,” “worse,” “stayed the same,” denoted in the figure as
MW, W, and SS; very few respondents perceived the economy to have improved during 2008).
The left side of the figure shows the predicted probabilities for Democrats, the right side for
Republicans. As can be seen, out-candidate enthusiasm has no impact on the retrospections of
Democrats (i.e., the black and grey bars are virtually identical). Nor should it. Whether due to partisan bias or because they are paying attention, Democrats are expected to hold negative perceptions of economic change in November of 2008.
The picture is quite different for Republicans. At low levels of out-candidate enthusiasm, we observe substantial amounts of partisan bias in retrospections. Indeed, at the lowest level of this moderator (i.e., those with no enthusiasm for candidate Obama, which represents more than one-third of Republicans), only 66 percent are expected to report that the economy had gotten
“much worse” over the past year. Even more surprisingly, more than a quarter of the remaining
40 percent (about 8% overall) were expected to report that the economy “stayed the same” during this period. In line with PAT, Republicans who were enthusiastic about Obama in January
Heterogeneous Political Judgment, p. 32 were substantially more attentive to the economic environment, with more than 90 percent perceiving the economy to have gotten “much worse.”
Discussion and Conclusions
An important thrust of contemporary work on political judgment is that perceptions and preferences are reached through a diverse and flexible set of cognitive strategies. The strategy adopted is likely to depend on the political context and on the individual’s capabilities and goals.
Political reasoning is thus contingent: sometimes voters rely primarily on their partisan habits, ignoring more diagnostic information (even if they are capable of acquiring it), while at other times they taken in more (or better) information and engage in more effortful reflection and reevaluation. The purpose of our investigation was to identify and test a relatively general mechanism by which this contingency in political thinking works, and to use our findings to clarify when, why, and how frequently we might expect to observe adaptive partisanship and responsible democratic citizenship.
Past work on the contingent nature of political judgment has focused on the idea that voters will abandon party cues and think more carefully about issues and candidates when they experience anxiety (Marcus and MacKuen 1993; Marcus et al. 2000). We contrasted this expectation with a broader framework based on two longstanding insights about the nature of mass political thought: (1) party identification is the most important predisposition in the political systems of ordinary Americans; and (2) people make strategic use of their cognitive resources: when possible, they will minimize decision effort by relying on simple rules, but will step up their thinking when cognitive shortcuts do not provide for sufficient judgment confidence
(Chaiken and Trope 1999; Lupia and McCubbins 1998).
Heterogeneous Political Judgment, p. 33
We acknowledge that voters can and do use a wide range of informational shortcuts in deciding what policy options to support and what electoral choices provide the most attractive outcomes. In many and probably most political contexts, however, partisan loyalty serves as a general anchor from which adjustments, typically insufficient ones, are made in judging the political landscape. Whether because partisanship is a product of early affective socialization
(Campbell et al. 1960); because the party system constrains the public’s menu of choices and attaches well-known, easily identifiable brand names to the competing options (Jackman and
Sniderman 2002); or because citizens simultaneously attempt to make it easy, get it right, and maintain cognitive consistency when forming judgments – or for all of these reasons – partisanship as a heuristic cue stands out in terms of its breadth, inferential power, and cognitive efficiency.
However, responsible democratic citizenship (e.g., holding office-holders accountable, maintaining policy control) requires that individuals be willing – at least at key times – to judge the political landscape independently of their partisan commitments. Given the fact that partisan identities and material circumstances are nearly orthogonal (the correlation in the 2008 ANES panel is .16; it averages .20 over the entire length of the ANES series), delegating one’s preferences to partisan elites while ignoring the policy implications of one’s material interests will often lead to sub-optimal judgments (at least from a simple economic perspective). More perniciously, reflexive cue-taking diminishes the responsiveness of elites to the public’s policy wishes. In
Politicians Don’t Pander,
Jacobs and Shapiro (2000) argue that politicians conduct extensive public opinion polling and focus groups to identify the most powerful “language, arguments and symbols” to persuade the public to support its extremist policy goals (see also
Luntz 2007). They (2000: 5) write that “competing efforts to tip public evaluations of policy
Heterogeneous Political Judgment, p. 34 proposals transformed ‘public opinion’ from an autonomous, external influence into a product of an endogenous process.” Good citizenship therefore requires a mechanism for determining when simple partisan cue-taking is a reasonable strategy, and when it needs to be supplemented by some independent thought about the details.
Thus, our empirical effort focused on what moves citizens to eschew toeing the party line and to consult more substantive considerations (such as material interests). We distinguished between two theories in this regard: the anxiety-focused perspective of affective intelligence theory, and partisan ambivalence theory, in which a disjuncture between contemporary party evaluations and affective partisan attachments provides a signal about the reliability of partisan cues. We contrasted the two perspectives by examining how two key emotions that have been the mainstay of research on AIT– anxiety and enthusiasm – influence how voters engage in three basic tasks of democratic citizenship: evaluating presidential candidates, updating opinions on policy issues in the wake of a national debate, and perceiving changes in the macro economy. All but one of our results was based on panel data, in which the explanatory variables were measured prior to the dependent variable, and in which a lagged measure of the latter was included as a predictor. This provides a good measure of protection against the scourge of endogeneity, as well as a conservative test of our hypotheses.
The results, based on three sets of independent samples, provided converging evidence that heterogeneity does not depend exclusively on anxiety, but on the broader consideration of whether a given emotion harmonizes or conflicts with one’s partisan identity. In each analysis, the experience of (one or more) identity-conflicting emotions decreased the role of partisanship and increased the role of more costly information, whereas the experience of (one or more) identity-consistent emotions had the opposite effect. Moreover, in direct contradiction to AIT,
Heterogeneous Political Judgment, p. 35 the effects were stronger and more reliable for enthusiasm than anxiety. As the former does not logically entail activation of the surveillance system or the experience of anxiety, these results do not fit neatly into the AIT framework. As we carefully controlled for both political knowledge and partisan strength (both as first-order terms and in product terms with focal predictors), the observed effects cannot be attributed to these standard engagement factors. Moreover, as Figure
2 attests, once identity conflicting and consistent partisan evaluations are in the model, political sophistication and partisan strength failed to stratify respondents’ reliance on costly information.
What seems to matter, then, is not whether voters experience one particular discrete emotion (e.g., anxiety), or whether they are highly engaged in politics, but whether they can derive adequate judgment confidence on the basis of low-effort cues alone . Contemporary partisan evaluations that conflict with the judgmental implications of party cues undermine confidence, leading citizens to think more carefully. By contrast, partisan evaluations that dovetail with party cues enhance confidence, fostering shallower (and potentially biased) judgments. Where does this leave AIT? Despite the fact that our critical tests seem to cast doubt on its validity, we believe this is an unwarranted conclusion. Rather, we believe that both theories are tenable but under different circumstances. AIT may provide a better explanation of heterogeneity in the moments when anxiety is experienced as a full-throttle, physiological event
(as opposed to indicating in a survey whether a candidate has “ever made you anxious”). When in the active throes of anxiety, it is eminently adaptive to set aside habit and focus on the most important information (e.g., Kahneman 2011). Over time, however, the urgency of an emotional reaction decays, leaving a residual evaluative response. If this evaluative response is out of step with one’s partisan attachment, it should produce little in the way of judgment confidence, and thereby instigate a change in the process by which politics is perceived and evaluated. Under
Heterogeneous Political Judgment, p. 36 these circumstances, and all of those in which anxiety is irrelevant, PAT would appear to provide the better explanation.
8
The question of how much and what kind of knowledge citizens need to fulfill their role in a democratic society has been hotly debated by political scientists (Bennet 1988; Delli Carpini and Keeter 1996; Fishkin 1991; Habermas 1989; Lupia and McCubbins 1998). There would appear to be a consensus that the Progressive Era ideal of the “informed citizen” – one who regularly keeps abreast of important political developments and who evaluates politics independent of party elites – is unattainable. Given the far more immediate and personal concerns that people face in their day-to-day lives (e.g., raising children, working, engaging in social relationships), politics rarely rises to the level of a compelling spectacle. At its very core, however, the concept of democracy implies the constraint of representatives by citizens. In a political world where partisan cues are utilized unthinkingly, citizens effectively become the agents of elite interests, as the latter learn that they may take actions (e.g. make policy statements) with impunity. The dynamic between citizens and elites then becomes one of, in the words of Shapiro and Jacobs (2000; 2010), “simulated representation,” whereby elites, rather than seeking to bend their positions to the median, simultaneously shape opinion and claim responsiveness to a majoritarian will. Our research suggests that citizens – or least the 90 percent who identify with (or lean toward) one party or the other – are equipped with a broadly functional tool for protecting themselves against this form of tyranny.
8 However, the relevance of PAT is bounded by circumstances in which partisanship itself is relevant. Several recent tests of AIT were conducted in either explicitly non-partisan contexts (e.g., primary elections; Brader 2005) or in ones in which partisanship was not directly invoked (MacKuen et al. 2010; Valentino et al. 2008).
Heterogeneous Political Judgment, p. 37
References
Abramowitz, Alan. 2010. The Disappearing Center. New Haven: Yale University Press.
Bartels, Larry. 2002. “Beyond the Running Tally: Partisan Bias in Political Perceptions.”
Political Behavior 24(2): 117-150.
Bartels, Larry. 2008. Unequal Democracy: The Political Economy of the New Gilded Age .
Princeton, NJ: Princeton University Press.
Basinger, Scott, and Howard Lavine. 2005. “Ambivalence, Information and Electoral Choice.”
American Political Science Review 99(2): 169-184.
Bennett, Stephen. 1989. “Trends in Americans’ Political Information, 1967-1987.”
American
Politics Research 17(4): 422-435.
Brader, Ted. 2006. Campaigning for Hearts and Minds: How Emotional Appeals in Political Ads
Work. Chicago: The University of Chicago Press.
Campbell, Angus, Philip E. Converse, Warren E. Miller and Donald E. Stokes. 1960. The
American Voter , New York: Wiley & Sons.
Chaiken, Shelly, Akiva Liberman, and Alice Eagly. 1989. Heuristic and Systematic Processing
Within and Beyond the Persuasion Context. In Unintended thought: Limits of awareness, intention, and control , eds. Jim Uleman and John Bargh. New York: Guilford, 212-252.
Chaiken, Shelly, and Yaacov Trope. 1999. Dual-process theories in social psychology . New
York: Guilford Press.
Cutout - Cohen, Geoffrey. 2003. “Party Over Policy: The Dominating Impact of Group Influence on Political Beliefs.” Journal of Personality and Social Psychology 85(5): 808-822.
Damasio, Antonio. 1994. Descartes Error: Emotion, Reason and the Human Brain. New York:
G.P. Putnams Sons.
Dancey, Logan, and Goren, Paul. 2010. “Party Identification, Issue Attitudes, and the Dynamics of Political Debate.”
American Journal of Political Science 54(3): 686-699.
Delli Carpini, Michael, and Scott Keeter. 1996. What Americans Know about Politics and Why it is Important. Yale University Press: New Haven.
Downs, Anthony. 1957. An Economic Theory of Democracy . New York: Harper and Row.
Druckman, James. 2001. “Evaluating Framing Effects.”
Journal of Economic Psychology 22(1):
91-101.
Heterogeneous Political Judgment, p. 38
Druckman, James, and Arthur Lupia. 2000. “Preference Formation.”
Annual Review of Political
Science 3: 1-24.
Duch, Raymond, Harvey Palmer, and Christopher Anderson. 2000. “Heterogeneity in
Perceptions of National Economic Conditions.”
American Journal of Political Science
44(4): 635-652.
Fazio, Russell H. 1990. “Multiple Processes By Which Attitudes Guide Behavior: The MODE
Model as an Integrative Framework.” In
Advances in Experimental Social Psychology 23, ed. Mark Zanna. San Diego, CA: Academic Press, 75-110.
Fishkin, James. 1991. Democracy and Deliberation: New Directions for Democratic Reform .
New Haven: Yale University Press.
Gaines, Brian, James Kuklinski, Paul Quirk, Buddy Peyton, and Jay Verkuilen. 2007. “Same
Facts, Different Interpretations: Partisan Motivation and Opinion on Iraq.” Journal of
Politics 69(4): 957-974.
Gomez, Brad, and J. Matthew Wilson. 2001. “Political Sophistication and Economic Voting in the American Electorate: A Theory of Heterogeneous Attribution.” American Journal of
Political Science 45(4): 899-914.
Gray, Jeffrey A. 1981."The Psycho-physiology of Anxiety." In Dimensions of Personality, ed.
Richard Lynn. New York: Pergamon, 233-252.
Habermas, Jurgen. 1989. The Structural Transformation of the Public Sphere . Cambridge: MIT
Press.
Hillygus, D. Sunshine, and Todd G. Shields. 2008. The Persuadable Voter: Wedge Issues in
Presidential Campaigns . Princeton: Princeton University Press.
Jacobs, Lawrence, and Robert Shapiro. 2000.
Politicians Don’t Pander: Political Manipulation and the Loss of Democratic Responsiveness . Chicago: University of Chicago Press.
Jacobson, Gary. 2010. “Perception, Memory, and Partisan Polarization on the Iraq War.”
Political Science Quarterly 125(1): 31-56.
Jonas, Klaus, Michael Diehl, and Philip Bromer.1997. “Effect of Attitudinal Ambivalence on
Information Processing and Attitude-Intention Consistency.” Journal of Experimental
Social Psychology 33(2): 190-210.
Kahneman, Daniel. 2011. Thinking Fast and Slow. New York: Farrar, Straus and Giroux.
Kam, Cindy 2005. “Who Toes the Party Line? Cues, Values, and Individual Differences.”
Political Behavior 27(2): 163-182.
Heterogeneous Political Judgment, p. 39
Kam, Cindy. 2006. “Political Campaigns and Open-Minded Thinking.”
Journal of Politics 68(4):
931-945.
Kruglanski, Arie and Donna Webster. 1996. “Motivated Closing of the Mind: Seizing and
Freezing.”
Psychological Review 103(2): 263-283.
Kuklinski, James, Paul Quirk, and Buddy Peyton. 2008. "Issues, Information Flows, and
Cognitive Capacities: Democratic Citizenship in a Global Era." In Democrary in the 21st
Century: International Perspectives ed. Peter Nardulli. Champaign: University of Illinois
Press.
Ladd, Jonathan, and Gabriel Lenz. 2008. “Reassessing the Role of Anxiety in Vote Choice.”
Political Psychology 29(2): 275-296.
Lau, Richard, and David Redlawsk. 2006. How Voters Decide: Information Processing During
Election Campaigns . Cambridge: Cambridge University Press.
Lavine, Howard, Christopher Johnston, and Marco Steenbergen. 2012. The Ambivalent Partisan:
How Critical Loyalty Promotes Democracy . New York: Oxford University Press.
Lazarsfeld Paul.F., Bernard R. Berelson, and Hazel Gaudet, H. 1948.
The People’s Choice
: How the Voter Makes Up his Mind in a Presidential Campaign . New York: Columbia
University Press.
Lebo, Matthew, and Daniel Cassino. 2007. “The Aggregated Consequences of Motivated
Reasoning and the Dynamics of Partisan Presidential Approval.”
Political Psychology
28(6): 719-746.
Ledoux, Joseph. 1996. The Emotional Brain: The Mysterious Underpinnings of Emotional Life.
New York: Simon and Schuster.
Levendusky, Matthew. 2009. The Partisan Sort: How Liberals Became Democrats and
Conservatives Became Republicans . Chicago: University of Chicago Press.
Lodge, Milton and Charles Taber. 2012. The Rationalizing Voter . New York: Cambridge
University Press.
Luntz, Frank. 2007. Words that Work: Its Not What You Say, Its What People Hear . New York:
Hyperion.
Lupia, Arthur, and Anthony McCubbins. 1998. The Democratic Dilemma: Can Citizens Learn what they Need to Know?.
Cambridge: Cambridge University Press.
MacKuen, Michael, Jennifer Wolak, Luke Keele, and George Marcus. 2010. “Civic
Engagements: Resolute Partisanship or Reflective Deliberation.” American Journal of
Political Science 54(2): 440-458.
Heterogeneous Political Judgment, p. 40
Maio, Gregory, David Bell, Victoria Esses. 1996. “Ambivalence and Persuasion: The Processing of Messages about Immigrant Groups.” Journal of Experimental Social Psychology
32(6): 512-516.
Marcus, George. 1988. “The Structure of Emotional Response: 1984 Presidential Candidates.”
American Political Science Review 82(3): 737-761.
Marcus, George. 2002. The Sentimental Citizen: Emotion in Democratic Politics . University
Park: Pennsylvania State University Press.
Marcus, George and Michael MacKuen. 1993. “Anxiety, Enthusiasm and the Vote: The
Emotional Underpinnings of Learning and Involvement During Presidential Campaigns.”
American Political Science Review 87(3): 672-685.
Marcus, George, Michael MacKuen, and W. Russell Neuman. 2011. “Parsimony and
Complexity: Developing and Testing Theories of Affective Intelligence.”
Political
Psychology 32(2): 323-336.
Marcus, George, W. Russell Neuman, and Michael Mackuen. 2000. Affective Intelligence and
Political Judgment.
Chicago: The University of Chicago Press.
Marcus, George, Michael MacKuen, Jennifer Wolak and Luke Keele. 2006. “The Measure and
Mismeasure of Emotion.” In Feeling Politics: Emotion in Political Information
Processing , ed. David Redlawsk. New York: Palgrave Macmillan, 31-45.
Marcus, George, Michael MacKuen, and W. Russell Neuman. 2009. “Measuring Subjective
Emotional Response: New Evidence on an Alternative Method.” Presented at the
International Society of Political Psychology Annual Scientific Meeting, Dublin.
Meffert, Michael, Michael Guge, and Milton Lodge. 2004. “Good, Bad, and the Ambivalent: The
Consequences of Multidimensional Political Attitudes.” In
Studies in Public Opinion:
Attitudes, Nonattitudes, Measurement Error, and Change , eds. Saris, William and Paul
Sniderman. Princeton: Princeton University Press, 69-100.
Payne, John, James Bettman, and Eric Johnson. 1993. The Adaptive Decision Maker. New York:
Cambridge University Press.
Petty, Richard. and Cacioppo, John. 1986. Communication and Persuasion: Central and
Peripheral Routes to Attitude Change . New York: Springer-Verlag.
Rahn, Wendy. 1993. “The Role of Partisan Stereotypes in Information Processing about Political
Candidates.” American Journal of Political Science 37(2): 472-496.
Redlawsk, David. 2002. “Hot Cognition or Cool Consideration? Testing the Effects of Motivated
Reasoning on Political Decision Making.”
Journal of Politics 64(4): 1021-1044.
Heterogeneous Political Judgment, p. 41
Sniderman, Paul, Richard Brody, and Philip Tetlock. 1991. Reasoning and Choice: Explorations in Political Psychology . New York: Cambridge University Press.
Taber, Charles and Milton Lodge. 2006. “Motivated Skepticism in the Evaluation of Political
Beliefs.” American Journal of Political Science 50(3): 755-769.
Valentino, Nicholas, Vincent Hutchings, Antoine Banks, and Anne Davis. 2008. “Is a Worried
Citizen a Good Citizen? Emotions, Political Information Seeking, and Learning Via the
Internet.”
Political Psychology 29(2): 243-273
Watson, David, and Auke Tellegen. 1985. “Toward a Consensual Structure of Mood.”
Psychological Bulletin 98(2): 219-235.
Zaller, John. 1992. The Nature and Origins of Public Opinion . New York: Cambridge University
Press.
Heterogeneous Political Judgment, p. 42
Supporting Information
The supporting information includes the regression tables for each of the models included in figure 2. Table 1 corresponds to column 1 of figure 2, table 2 to column 2, etc. We include one additional model in the supporting information that was excluded from the main body of the paper for space considerations.
This additional model is included in column 2 of table 4. It is a model of preference formation over the course of the 2008 presidential campaign on the issue of governmentprovided health insurance. The setup of the model is the same as with the tax preference model.
We model October preferences as a function of lagged preferences in January, partisanship, the emotion items, their interactions with partisanship, and all controls The only significant moderator of partisanship is in-candidate enthusiasm. As predicted by PAT, this emotion increases the marginal effect of partisanship from .14 to .50.
Heterogeneous Political Judgment, p. 43
Table 1. Vote Intention, ANES Cumulative File
_______________________________________________
Variable B SE p
_______________________________________________
Age .00 .01 .85
Gender .04 .05 .38
Black -.53 .08 .00
Hispanic -.18 .09 .03
Education .18 .09 .04
Income .11 .09 .22
Interest -.04 .08 .62
Republican .75 .13 .00
Relative Issue Distance 4.31 .71 .00
Strength -.35 .27 .19
Sophistication -.74 .39 .06
In-Anxiety .02 .30 .94
Out-Anxiety .46 .25 .07
In-Enthusiasm -.36 .28 .20
Out-Enthusiasm
-.38 .28 .18
Rep X Strength
Rep X Sophistication
.38
.32
.12
.16
.00
.04
Rep X In-Anxiety -1.13 .12 .00
Rep X Out-Anxiety
Rep X In-Enthusiasm
.97
3.02
.12
.12
.00
.00
Rep X Out-Enthusiasm
Issues X Strength
Issues X Sophistication
-2.95
.35
.67
.13
.56
.78
.00
.53
.39
Issues X In-Anxiety
Issues X Out-Anxiety
.83
-1.55
.58
.52
.15
.00
Issues X In-Enthusiasm -2.64 .58 .00
Issues X Out-Enthusiasm
Intercept
3.85
-1.98
.57
.37
.00
.00
N 7395
_______________________________________________
Notes: Entries are ML ordered probit coefficients and standard errors..
The link function is ordered probit.
Heterogeneous Political Judgment, p. 44
Table 2. Comparative Candidate Evaluations, ANES 2008 Panel
_____________________________________________________
Variable B SE p
_____________________________________________________
Lagged Evaluations
Age
Male
.50
-.02
-.01
.04
.02
.01
.00
.49
.25
Black
Education
Income
Economic Retrospections
Republican
.00
-.03
-.04
.08
-.02
.02
.02
.03
.03
.06
.91
.16
.18
.00
.74
Relative Issue Distance
Partisan Strength
Sophistication
In-Anxiety
Out-Anxiety
In-Enthusiasm
.19
-.03
-.05
.15
-.09
-.10
.18
.03
.07
.09
.04
.05
.27
.37
.45
.10
.06
.06
Out-Enthusiasm
Rep X Strength
Rep X Soph
Rep X In-Anx
Rep X Out-Anx
Rep X In-Enth
Rep X Out-Enth
-.06
.14
.09
.03
.06
.17
-.25
.08
.04
.06
.07
.06
.05
.07
Issues X Strength
Issues X Soph
Issues X In-Anx
Issues X Out-Anx
Issues X In-Enth
Issues X Out-Enth
Constant
-.06
-.06
-.35
.17
-.12
.46
.24
.09
.17
.20
.13
.14
.19
.08
.54
.73
.09
.18
.39
.01
.00
R^2 .81
N 725
_____________________________________________________
Notes: Entries are OLS coefficients and robust standard errors. The dependent variable is measured in Oct. Partisanship and economic retrospections are measured in January. Emotions and lagged evaluations are measured in
February. Comparative issue distance is calculated as a function of respondents’ issue positions in February and their perceptions of candidate positions in June.
.44
.00
.12
.68
.32
.00
.00
Heterogeneous Political Judgment, p. 45
Table 3. Preference Formation on Health Insurance, ANES 1992-1994 Panel
____________________________________________________________________
Variable B SE p
_________________________________________________________
Lagged Preferences 1.13 .19 .00
Age
Male
Black
-.14
.09
-.14
.25
.12
.19
.60
.44
.47
Education
Income
Unemployed
Republican
Personal Economic Conditions
Partisan Strength
-.15
.37
-.14
.40
-.05
.06
.24
.24
.23
.43
.76
.34
.54
.13
.55
.35
.95
.87
Sophistication
In-Anxiety
Out-Anxiety
In-Enthusiasm
Out-Enthusiasm
Rep X Strength
Rep X Soph
Rep X In-Anx
Rep X Out-Anx
Rep X In-Enth
Rep X Out-Enth
Economic X Strength
Economic X Soph
Economic X In-Anx
Economic X Out-Anx
Economic X In-Enth
.27
-.63
-.34
-.19
-.96
.34
-.16
.23
.53
-.21
-.35
-.18
-.82
.79
.31
.64
.53
.38
.29
.37
.39
.29
.49
.31
.25
.34
.35
.59
.94
.62
.49
.61
Economic X Out-Enth
Cut 1
Cut 2
Cut 3
Cut 4
1.37
-.75
-.38
.08
.73
.62
.44
.44
.44
.44
.03
Cut 5
Cut 6
1.16
1.71
.44
.44
Pseudo R^2 .10
N 407
____________________________________________________________________
Notes: Entries are ML ordered probit coefficients and standard errors. The dependent variable is measured in 1994, and all independent variables are measured in 1992.
.03
.54
.32
.76
.38
.20
.52
.29
.62
.10
.24
.61
.01
.24
.75
.45
Heterogeneous Political Judgment, p. 46
Table 4. Preference Formation on Taxes and Health Insurance, ANES 2008 Panel
_______________________________________________________________________
Taxes on >$200K Govt. Health Insurance
Variable B SE p B SE p
_______________________________________________________________________
Lagged Preferences 2.54 .18 .00 2.10 .16 .00
Age
Male
Black
-.10
.09
.21
.27
.11
.27
.71
.41
.44
-.32
.05
.17
.26
.11
.27
.23
.62
.54
Education
Income
Republican
Partisan Strength
Sophistication
In-Anxiety
Out-Anxiety
In-Enthusiasm
.32
.59
-.90
-.27
-.96
.44
-.65
-.36
.22
.28
.43
.19
.35
.34
.30
.30
.14
.04
.04
.16
.01
.19
.03
.23
.05
.75
.38
-.03
.30
.30
.44
-.76
.22
.28
.42
.18
.34
.33
.27
.28
.82
.01
.36
.87
.37
.36
.10
.01
Out-Enthusiasm
Rep X Strength
Rep X Soph
Rep X In-Anx
Rep X Out-Anx
Rep X In-Enth
Rep X Out-Enth
.23
.39
1.51
-.45
.92
1.06
-.84
.35
.27
.47
.54
.38
.41
.48
.50
.15
.00
.40
.02
.01
.08
.62
.28
-.09
-.72
.03
1.07
-.71
.34
.27
.46
.52
.36
.41
.48
Cut 1
Cut 2
Cut 3
Cut 4
Cut 5
Cut 6
-.03
.85
1.12
1.95
2.12
2.70
.35
.35
.35
.36
.36
.37
.60
1.17
1.31
2.22
2.34
3.03
.33
.33
.33
.34
.34
.35
Pseudo R^2 .27 .24
N 513 512
_______________________________________________________________________
Notes: Entries are ML ordered probit coefficients and standard errors. The dependent variable is measured in October, lagged issue preferences and partisanship are measured in January, and emotions and sophistication are measured in February.
.07
.29
.84
.17
.93
.01
.14
Heterogeneous Political Judgment, p. 47
Table 5. Economic Retrospections, ANES 2008 Panel
________________________________________________
Variable B SE p
________________________________________________
Lagged Retrospections
Age
Male
2.01
-.48
-.02
.23
.23
.10
.00
.04
.86
Black
Education
Income
Republican
Partisan Strength
.27
-.33
.04
.66
.06
.23
.19
.24
.40
.20
.24
.09
.85
.09
.75
Sophistication
In-Anxiety
Out-Anxiety
In-Enthusiasm
Out-Enthusiasm
-.37
.09
.05
-.15
-.12
.34
.37
.29
.30
.36
.27
.82
.85
.61
.73
Rep X Strength
Rep X Soph
Rep X In-Anx
Rep X Out-Anx
Rep X In-Enth
Rep X Out-Enth
Cut 1
-.25
.39
-.23
-.42
.36
-1.13
.93
.26
.42
.49
.35
.38
.47
.34
Cut 2
Cut 3
Cut 4
1.90
3.18
3.73
.34
.39
.49
Pseudo R^2 .15
N 982
________________________________________________
Notes: Entries are ML ordered probit coefficients and standard errors. The dependent variable is measured in
November, lagged retrospections and partisanship are measured in January, and emotions and sophistication are measured in February.
.32
.36
.63
.23
.35
.02