Contempt of Congress : Do Liberals and Conservatives Harbor Equivalent Negative Emotional Biases Towards Ideologically Congruent vs . Incongruent Politicians at the Level of Individual Emotions ?

Prior research suggests that conservatives are more fear-motivated, disgust-sensitive, and happy than liberals. Yet when it comes to political targets (e.g., politicians), both liberals and conservatives can get very emotional. We examined whether the ideological differences in emotion seen in past research apply to emotions towards specific ideologically similar vs. dissimilar targets, or whether these emotions are instead equivalent between liberals and conservatives. Across two studies, liberals and conservatives rated their anger, contempt, disgust, fear, and happiness towards Democratic and Republican congresspersons. We compared participants’ levels of each emotion towards their respective ideologically dissimilar and ideologically similar congresspersons. Liberals and conservatives both experienced stronger negative emotions towards ideologically dissimilar congresspersons than they did towards ideologically similar ones. Neither liberals nor conservatives differed in negative emotions towards politicians overall (i.e., on average). However, there were ideological differences in emotional bias. In Study 1, liberals exhibited a greater contempt bias (i.e., a larger gap in contempt ratings between ideologically similar and ideologically dissimilar politicians) than conservatives did. In Study 2, liberals exhibited greater contempt, anger, disgust, and happiness biases than conservatives did. The need to consider context in the study of ideological differences in emotion is discussed.


Background
The combination of politics and emotion can bring out the worst in people.Arguments and clashes between political opposites can quickly become heated, leading to the experience and expression of intense negative emotions towards ideological opponents -especially towards politicians representing the opposing side.It is tempting for both liberals and conservatives to believe that those on the opposite side of the ideological spectrum have greater hostile emotional biases, and are nastier, more hateful, and more spiteful towards opposing-ideology politicians.But does one side objectively harbor greater negative emotional bias towards opposing-ideology politicians than the other?

Why was this study done?
A large and robust body of research in political psychology suggests conservatives are more fear-motivated and disgust-prone than liberals.This body of work might suggest that conservatives could indeed harbor more emotional negativity towards opposing-ideology politicians than liberals do -particularly in regard to fear and disgust.Yet politicians elicit intense negative and positive emotions in both liberals and conservatives.We wanted to investigate whether one side of the ideological spectrum harbors greater hostile emotional biases towards their opposing-ideology politicians than the other; or whether both sides of the ideological spectrum are "just as bad as the other" when it comes to negative In contrast, it is common for both liberals and conservatives to feel and express positive emotions toward ideologically-aligned politicians, often turning a blind eye to their misdeeds, scandals, and hypocrisies (Bhatti, Hansen, & Olsen, 2013).Although political talk shows and social media are replete with tit-for-tat accusations of hatred and emotional hostility from across the political aisle, the question of whether one side of the ideological spectrum is actually more emotionally biased towards politicians than the other has not been fully examined.Given that bias is often used as a rationale for discounting the political positions of ideological rivals, it is important to understand the nature and scope of these biases (Kim, 2015;Vallone, Ross, & Lepper, 1985).
The literature provides competing perspectives that could reasonably account for liberals and conservatives holding either similar or different emotional biases towards politicians.On the one hand, a robust body of literature describes conservatives, compared to liberals, as having greater predispositions towards experiencing certain emotions like fear and disgust (e.g., Jost et al., 2003), which could lead conservatives to harbor greater fear and disgust towards politically-opposing groups.On the other hand, a growing body of research has found that negative attitudes typically associated with conservatism such as intolerance (Brandt, Reyna, Chambers, Crawford, & Wetherell, 2014) and prejudice (Chambers, Schlenker, & Collisson, 2013) may become equivalent between liberals and conservatives in contexts of ideological conflict-when they are directed towards groups perceived as ideologically threatening.Initial research on emotion within this context has found that liberals and conservatives may indeed have similar levels of generalized negative affect towards ideological rivals and similar levels of generalized positive affect towards ideological allies (Roscoe & Christiansen, 2010;van Prooijen, Krouwel, Boiten, & Eendebak, 2015).However, whether or not this pattern is varied or consistent across specific (discrete, qualitatively distinct) emotions such as anger, contempt, disgust, fear, and happiness, has not been fully explored.
The present research will reconcile these competing perspectives by providing a balanced test of liberals' and conservatives' anger, contempt, disgust, fear, and happiness biases against ideologically-incongruent politicians (i.e.liberals towards Republican congresspersons, conservatives towards Democratic congresspersons) and towards ideologically-congruent politicians (liberals towards Democratic congresspersons, conservatives towards Republican congresspersons).By examining emotional biases at the level of specific discrete emotions, this approach provides a more definitive answer to the question of whether or not liberals and conservatives vary in emotional bias vis-à-vis the particular emotions previously associated with ideology (e.g., disgust and fear), or whether the ideological similarity versus dissimilarity of the target overrides these ideological differences in emotion.Furthermore, we extend existing literature by using politicians (Democratic vs. Republican congresspersons) as the targets of affective evaluation, and by including the politically consequential emotion of contempt, which has yet to be studied in this context.

Ideological Differences in Emotion
There have been robust findings that conservatives, compared to liberals, have more fear-based motivations and greater sensitivity to disgust (see Jost et al., 2003 for review).In turn, fear-based motives have been linked with intolerance of opposing political views (Jost et al., 2003) and disgust-sensitivity has been associated with negative attitudes towards politically-relevant groups seen as ideologically threatening (e.g., prejudice towards gays and immigrants; Hodson & Costello, 2007;Inbar, Pizarro, & Bloom, 2009;Terrizzi, Shook, & Ventis, 2010).This body of research could suggest that conservatives, compared to liberals, might have greater fear and disgust towards politicians; perhaps overall, but certainly towards those who are seen as ideologically threatening (e.g., Democratic congresspersons).

Contempt of Congress 102
Another body of research has found that conservatives, compared to liberals, tend to be happier and more satisfied with the political status quo (e.g., Napier & Jost, 2008;Schlenker, Chambers, & Le, 2012).These findings suggest that conservatives, compared to liberals, might report a greater degree of positive emotional biases towards politicians who they see as ideologically-congruent (e.g., Republican congresspersons).

The Ideological Conflict Hypothesis
Similar to the literature on ideological differences in emotion, a robust body of work on ideological differences in attitudes has linked conservatism with a greater propensity towards prejudice (e.g., see Sibley & Duckitt, 2008 for a meta-analysis) and political intolerance (e.g., see Jost et al., 2003 for review).However, a recent series of studies have found that liberals and conservatives may actually express similar degrees of intolerance and prejudice towards groups that are perceived as ideologically-incongruent (i.e., ideologically threatening and/or different from their own; Brandt, Reyna, et al., 2014;Brandt, Wetherell, & Reyna, 2014;Chambers et al., 2013;Crawford & Pilanski, 2014;Roscoe & Christiansen, 2010;Wetherell, Brandt, & Reyna, 2013).Some of these studies also found that liberals and conservatives have similarly positive attitudinal biases towards ideologically-congruent groups.
The ideological conflict hypothesis (Brandt, Reyna, et al., 2014) proposes that people, regardless of ideology, are more tolerant toward ideologically-congruent groups, and more intolerant toward ideologically-incongruent groups due to perceptions that ideological outgroups threaten important values.Given that politicians can be perceived as threats to-or defenders of-important values, the ideological conflict hypothesis framework could apply not only to attitudinal biases towards partisan groups, but to emotional biases as well.This framework suggests that despite the robust associations between conservatism and fear, disgust, and happiness (e.g., Jost et al., 2003), liberals and conservatives might have similar levels of negative emotions towards ideologically-incongruent politicians and similar levels of positive emotion towards their ideologically-congruent politicians.

Emotional Biases in the Ideological Conflict Hypothesis Framework
To the extent of our knowledge, it appears that no published studies have fully examined the distinctions between contempt, anger, disgust, fear, and happiness within the framework of the ideological conflict hypothesis.However, several studies have examined ideological similarities and differences in generalized positive and negative affect held towards ideologically-incongruent versus ideologically-congruent political groups.
Roscoe and Christiansen (2010) examined generalized positive versus negative affect that liberal and conservative partisans felt towards the Democratic and Republican parties by creating positive and negative emotion scales derived additively from categorical (yes/no) items assessing: anger, hope, pride, embarrassment, fear, safety, happiness, sadness, nervousness, and calm-but they did not analyze specific emotions.Similarly, van Prooijen and colleagues (2015) examined Dutch liberals' and conservatives' generalized negative affect towards ideologically-incongruent political parties using a composite scale using "anger," "afraid," "fearful," and "disgust" (p.489) averaged together-but did not analyze specific emotions.Both studies' findings were consistent with the ideological conflict hypothesis: liberals and conservatives felt similar levels of generalized negative affect towards ideologically-incongruent political parties and (for Roscoe and Christiansen) similar levels of generalized positive affect towards ideologically-congruent parties.
Iyengar, Sood, and Lelkes (2012) examined partisan Democrats' and Republicans' feeling thermometer ratings (a 1-100 measure of unfavorable/cold feelings vs. favorable/warm feelings) over time towards both the Democratic and Republican parties and their members.They found that feeling thermometer ratings towards ideologicallyincongruent groups varied over time but were predominately driven by contextual factors, such as which party was currently in power and whether or not data were collected during a presidential election season.This raises the possibility that results could be influenced by differences in political power at the time of data collection.At the time of data collection, Democrats controlled the presidency, which could influence the degree of hostile emotional biases conservatives hold towards ideologically-incongruent politicians.In contrast, Republicans controlled the legislature at the time of data collection, and the subsequent obstruction of the Democratic agenda could influence the degree of hostile emotional biases liberals hold towards ideologically-incongruent politicians.While these types of results are not entirely consistent with the ideological conflict hypothesis, they would indicate that generalized affect was predominately driven by intergroup conflict, as opposed to any stable ideological differences in negative emotionality.
The abovementioned studies were important in establishing that generalized negative and positive affect towards ideologically-incongruent and congruent groups may operate similarly to attitudes such as prejudice and intolerance within the ideological conflict hypothesis framework.However, since previous work scaled multiple emotions into generalized negative affect scales, potentially important differences in the pattern of results at the level of specific discrete emotions may have been averaged out, or in the case of using feeling thermometers were unavailable to begin with.

Contempt
Importantly, no previous studies have assessed the emotion of contempt in the context of liberals' and conservatives' emotions towards ideologically-incongruent versus ideologically-congruent groups.The experience of contempt is perhaps best defined as cold feelings of dislike, combined with psychological distancing (a loss of warmth, empathy, and respect) and viewing the target as unchangeably inferior (e.g., Steiger & Reyna, 2017).Given these features, the emotion of contempt could have very powerful consequences in a political context.Contempt is highly relevant to ideological conflict research, and research in political psychology in general, as it correlates with a variety of attitudinal and behavioral outcomes in intergroup contexts in ways that differ from other emotions (in turn, these emotions also differ from one another in intergroup contexts; see below).Yet large-scale political surveys such as the American National Election Studies (ANES, 2012) do not include measures of contempt.
Therefore, we include it here to fill this gap in the literature.

Specific Emotions Matter in Intergroup Contexts
Research on individual-level and intergroup emotions has shown that there are distinct attitudinal, behavioral, and relational outcomes for different discrete emotions that have interpersonal, intergroup, and political consequences.Therefore, separating out contempt, anger, disgust, fear, and happiness from each other might yield different patterns of results for different emotions.For example, contempt and anger have been found to differ in the behaviors they prompt and their long-term effects on relationships, such that "anger is characterized more by short-term attack responses but long-term reconciliation, whereas contempt is characterized by rejection and social exclusion… both in the short-term and in the long-term" (Fischer & Roseman, 2007, p. 103).Contempt and disgust at the group level have been found to predict moral exclusion and even dehumanization of certain outgroups (Esses, Veenvliet, Hodson, & Milic, 2008;Mackie & Smith, 2015).Anger tends to prompt mobilization and normative political action at the group level (e.g., Mackie & Smith, 2015;Tausch et al., 2011), whereas contempt tends to lead to political disengagement and the endorsement of non-normative (i.e., extreme) political action, potentially even including the endorsement of political violence (Tausch et al., 2011).
In conjunction with prior work linking conservatism with increased propensities towards disgust, fear, and happiness (e.g., Jost et al., 2003), the abovementioned literature on differences between discrete emotions in intergroup contexts suggests that examining these emotions in an ideological conflict context might yield differing patterns of results for different discrete emotions.However, work on intergroup emotions also indicates that emotions in an ideological conflict context may be distinct from more generalized ideological differences in emotion, given that priming different social category memberships (such as political membership) can sometimes influence the types of discrete emotions people feel towards outgroups (e.g., Mackie & Smith, 2015).

The Present Research: Ideological Differences in Emotion Versus Ideological Conflict
Our primary research questions are as follows: (1) When examining specific discrete emotions towards congresspersons, does one side of the ideological spectrum hold greater negative emotional biases against ideologically-incongruent congresspersons and/or greater positive emotional biases towards ideologically-congruent congresspersons?Or, in keeping with the ideological conflict hypothesis, are these emotional biases equivalent -such that liberals and conservatives have similarly high levels of anger, contempt, disgust, fear, and (un)happiness towards ideologically incongruent congresspersons and similarly low levels of each emotion towards ideologically congruent congresspersons?(2) Is this pattern of emotional bias consistent or varied across different discrete emotions?If the patterns of emotional bias vary across the different discrete emotions, are they in keeping with the ideological differences in emotion literature -such that conservatives (compared to liberals) have greater disgust and fear towards ideologically incongruent congresspersons and greater happiness towards ideologically congruent congresspersons?Or might there be novel findings of ideological differences in emotion bias within this context?For example, perhaps liberals might show greater emotional bias than conservatives do on one or more of the discrete emotions we measure.Across two studies, we provide a balanced test of these two competing perspectives from the literature.

Ideological Distance
For any given sample, it could be the case that there is an imbalance in how incongruently liberals vs. conservatives view the values of their ideologically dissimilar congresspersons from their own values.For instance, research has shown that Republican congresspersons have become more ideologically extreme than Democratic congresspersons over time (Lewis, Poole, & Rosenthal, 2017).Additionally, M-Turk samples (which we use) tend to lean left and thus may have a sampling bias (Huff & Tingley, 2015), such that the proportion of liberals who are extreme may be greater than the proportion of conservatives who are extreme.As such, it is possible that liberals-compared to conservatives-might view their respective ideologically dissimilar congresspersons as more ideologically "distant" from themselves.Alternatively, it could be the case that one side of the ideological spectrum has a greater tendency to view their ideologically similar congresspersons as either being too extreme or not being extreme enough relative to their own ideological values.In order to account for these possible influences, we decided to control for ideological distance in the present studies -the distance between participants' own level of liberalism-conservatism and their perceptions of how liberal/conservative both Democratic and Republican congresspersons are.

Why Congresspersons?
There are many different types of politicians.We chose to have participants rate their emotions towards Democratic and Republican congresspersons because they tend to consistently elicit strong, often negative emotions.
During the time of data collection, congressional approval ratings hovered around 15% (Jacobsen, 2014).One poll even found that, comparatively, congress was viewed less favorably than traffic jams, lice, colonoscopies, and even cockroaches (Jensen, 2013)!Yet despite this, the incumbent re-election rate for both the Senate and the House has consistently been 90% or greater (Jacobsen, 2014), indicating that positive biases towards ideologically-congruent congresspersons are influential.
Congress was ideal because it is a highly recognizable political group, but it has a large enough number of members that people are more likely to view them as a group ("congress") than as a collection of specific individuals.Given that the ideological conflict hypothesis is an intergroup theory pertaining to what drives attitudes towards ideologically-threatening groups, assessing emotions towards congress as a group rather than individual congresspersons (such as Nancy Pelosi vs. John Boehner) was the most appropriate way to test how emotions fit within this theory.This approach also helped minimize potential variance from individual differences in political knowledge about, or idiosyncratic attitudes towards, specific individual congresspersons.

House Versus Senate
At the time Study 1's survey was administered (January-February, 2014), Republicans held the majority in the House of Representatives while Democrats held the majority in the Senate.Additionally, a wave of especially conservative politicians had recently been elected the House of Representatives (i.e., the "Tea Party" affiliated House Freedom Caucus).In order to account for the possibility that one body of congress elicited negative emotions differently than the other did, in Study 1 we experimentally manipulated whether participants rated their emotions towards House members versus Senators.

Study 1 Methods
Study 1 used a mixed design.Participant's ideological leanings were analyzed continuously; the target's congressional political office (house vs. senate) was used as between-subjects variables, and participants' emotions towards ideologically congruent and incongruent congresspersons were treated as a within subjects variable.

Participants
We recruited 120 participants from Amazon's M-Turk.Seventy-three were women, 45 were men, and two did not report.The average participant age was 38.57, and ages ranged from 18 to 72.Participants were 84% White, 8% African American, 5.0% Latino, and 2% Asian, and 1% Native American or Alaskan Native.

Procedure
The emotion items analyzed in Study 1 were a subset of measures from a larger 2014 online political survey.
Participants were randomly assigned to one of two political office conditions (House vs. Senate).The survey's measures were divided into five different sections; the only section analyzed in the present study assessed emotions felt towards Democratic/Republican congresspersons.
i Each section was presented in a randomized order, and the order of items within each section was also randomized.Participants completed all five survey sections for congresspersons from one party, and then repeated the process for the second party.Finally, participants filled out demographic information that included political identification items, then ended the survey.
Demographics -Among the items in the demographic section, participants reported their age, gender, ethnicity, religion, employment status, income, and education.Participants also reported their general ideology; "How would you generally describe your political views?" (1 = very liberal, 4 = centrist/moderate, 7 = very conservative).The average rating on this scale was 3.52, SD = 2.07.anger, contempt, disgust, fear, happiness.Each emotion measure consisted of a photo of that emotion's unique facial expression, along with a label of the emotion's name underneath it.Participants rated the degree they felt each emotion towards the target using a 6-point scale (1 = not at all, 6 = very much).We utilized facial expression photos along with the emotion words as part of the measure because prior research has indicated that it helps participants more easily differentiate the emotions of contempt, anger, and disgust from one another (e.g., Rosenberg & Ekman, 1995;Rozin et al., 1999).All emotional expression photos were of the same female Caucasian model from the Japanese and Caucasian Facial Expressions of Emotion photoset-a validated set of emotional expression photos (Matsumoto & Ekman, 1988).

Emotion section -
We recoded emotion ratings into ideologically-congruent (i.e.similar) versus ideologically-incongruent (i.e.dissimilar) emotion ratings for anger, contempt, disgust, fear, and happiness.Ideologically-congruent emotion rating items consisted of liberal participants' emotion ratings towards Democratic congresspersons and conservative participants' emotion ratings towards Republican congresspersons.Ideologically-incongruent emotion rating items consisted of liberal participants' emotion ratings towards Republican congresspersons and conservative participants' emotion ratings towards Democratic congresspersons.
ii Congressional ideology and ideological distance discrepancy -At the beginning of the survey, participants responded to two questions asking: "How liberal/conservative do you think [Democratic/Republican] congresspersons are?"The items used 7-point scales (1 = very liberal, 7 = very conservative).We computed a variable representing the ideological distance discrepancy between each participants' own ideological position and their ratings of how liberal or conservative they rated their congruent and incongruent congresspersons.First, we created an ideologically-incongruent politician ideological distance variable.This variable was the absolute value of the difference between participants' own continuous liberal-conservative ideology rating and their incongruent congressional liberal-conservative rating (Republican congresspersons for liberals, Democratic congresspersons for conservatives).
We next created a congruent politician ideological distance variable.This variable was the absolute value of the difference between participants' own liberal-conservative rating and their respective congruent congressional liberal-conservative rating (Democratic congresspersons for liberals, Republican congresspersons for conservatives).
Finally, we computed an ideological distance discrepancy variable by subtracting each participant's congruent politician ideological distance score from their incongruent politician ideological distance score.In other words, this variable represents how wide a gap there is between participants' perceived ideological similarity with ideologically-congruent congresspersons compared to ideologically incongruent congresspersons, which could vary as a function of how extreme or moderate the participant is ideologically.This measure ranged from 1 to 6.The average rating on the distance discrepancy variable was 3.39, SD = 1.25.

Ideological Distance Discrepancy and Ideology
We first examined whether participants' ideology rating was related to the ideological distance discrepancy variable.
There was no relationship between our continuous ideology measure and the ideological distance discrepancy variable r(118) = .03,p = .76,indicating that neither liberals nor conservatives were exhibiting greater perceived ideological distance between congruent and incongruent congresspersons.

Emotions Towards Ideologically-Congruent Versus Ideologically-Incongruent Politicians
To examine whether or not the difference between emotion ratings towards congresspersons with a similar or dissimilar stance as participants differed based on participant ideology, we ran a series of mixed model ANOVAs with the difference between emotion ratings for congruent and incongruent politicians ("Target Politician") as a repeated measures factor and participant ideology as a continuous predictor.In these analyses a main effect of target politician without an interaction with participant ideology would mirror the Ideological Conflict pattern of symmetry across liberals and conservatives in their degree of emotional bias towards ideologically congruent vs. incongruent politicians.A main effect of Participant Ideology without an interaction with target politician would suggest that participants of a particular ideology felt more of a particular emotion overall regardless of the ideological similarity of the target.A Target Politician x Participant Ideology interaction suggests a greater emotional bias among one of our ideological groups such that the discrepancy in emotion felt towards similar and dissimilar politicians varies by participant ideology.
To examine whether or not the overall main effects and interactions were present, we midpoint-centered the continuous ideology variable, and used it to predict the difference between each separate emotion for same and opposing party congress people.This analytic strategy allows one to examine the main effects and interactions in the model when ideology is at its midpoint (Judd, McClelland, & Ryan, 2011).The results of the overall models are depicted in Table 1.
To examine the difference between emotion ratings between same and opposing party congress people among liberals and conservatives respectively, we conducted simple slopes analyses one standard deviation above (conservative) and below (liberals) the midpoint of the continuous ideology variable.In the Figures, the coefficients between the lines for same and opposing party congress people above the liberal and conservative labels represent the estimated emotion difference between same and opposing party congress people for liberals and conservatives respectively.
We also conducted a simple slopes analysis examining the slope of continuous political ideology for both same and opposing party congress people.In the Figures, coefficients above and below the lines for same and opposing party congress people represent the slope of continuous ideology among same and opposing party congress people respectively.Overall, this strategy allowed us to compare participants of equal levels of self-reported liberalism/conservatism on emotion ratings (thereby controlling for any sampling discrepancy).In addition, this analytic strategy allowed us to examine the manner in which participant ideology predicts emotions towards congresspersons both similar and dissimilar to participants from an ideological standpoint, an important distinction between previous studies and the present work.
We included the House versus Senate condition variable as a control in all analyses reported here (1 = House, -1 = Senate).Although we had no predictions regarding the House vs.Senate condition, we included it as a control since it was an experimental manipulation.No significant main effects or interactions emerged for House vs.
Senate condition, so we have omitted it from the results for the purpose of simplification.
The results of these analyses are depicted in Table 1 and Figure 1.The main effect of target politician was significant for each emotion (see the far-left column of Table 1 labeled "Target politician", a term we use to describe participants' emotion ratings towards ideologically similar versus dissimilar congresspersons), demonstrating that participants experienced more negative emotions (i.e.anger, contempt, disgust, and fear) towards ideologically dissimilar politicians, and experienced more positive emotion (happiness) towards ideologically similar politicians compared to dissimilar politicians.There were no main effects of participant ideology, indicating that neither liberals nor conservatives experienced greater degrees of any emotion overall (see the middle column in Table 1 labeled "Main effect: Participant's ideology).The main effects and interactions including the House vs.Senate condition across emotions were all non-significant.The only significant interaction that emerged was between target politician and participant ideology for the emotion contempt (see the right-hand column in Table 1, labeled "Interaction: Target politician x Participant's ideology").The interaction pattern suggested a larger gap between contempt ratings for dissimilar vs. similar politicians for liberals compared to conservatives.This pattern seemed to be driven by liberals feeling slightly more contempt towards dissimilar congresspersons and less contempt towards similar congresspersons on the part of liberals, compared to conservatives.We also ran a second set of analyses controlling for ideological distance discrepancy.In these analyses we included the ideological distance discrepancy variable as a predictor of emotions towards similar and dissimilar congresspersons.The inclusion of this variable had no effect on the initial results.All main effects of target politician held across emotions, no main effects of participant ideology were significant, and the target politician by participant ideology interaction for contempt remained.See Table A.1 in the Appendix for ANOVA tables that include ideological distance discrepancy.
Overall, the results of Study 1 suggest that the discrepancy between emotions towards similar and dissimilar congresspersons is largely symmetrical between liberals and conservatives.These results contrast with a robust body of literature linking conservatism with increased propensities towards experiencing fear, disgust, happiness, and negativity bias in general, and instead coincide more strongly with patterns that would be predicted by the ideological conflict hypothesis.In line with the ideological conflict hypothesis, the main effect of target politician demonstrated that people in general experienced more negative emotions regarding incongruent compared to congruent politicians.The one exception to this overall pattern of results was for contempt, in which liberals exhibited a larger gap than conservatives between ideologically congruent and incongruent politicians.
In addition, our analytic strategy allowed us to compare participants of equal levels of self-reported liberalism/conservatism, and controlling for ideological distance discrepancy (which could vary as a function of sampling bias) did not alter our results.Combined with the fact that distance discrepancy was not correlated with participant ideology, this suggests that the asymmetry finding for contempt was not driven by greater ideological extremity on the part of liberals or conservatives respectively, or by perceptions of greater ideological distance from oppositeparty congresspersons.

Study 2
Study 1 had several limitations, which we addressed in Study 2. For the purposes of the present studies, the inclusion of House versus Senate condition in the Study 1 survey was not particularly informative or relevant.Since House versus Senate condition had no significant influence on the results, we did not include it in the Study 2 survey.Instead, we simply referred to "[Democratic/Republican] congresspersons" in all of the measures.Another limitation of the Study 1 survey was a lack of attention check items included in the survey.For online samples, such as Mturk, participants may vary in how carefully they are reading the survey which could introduce random error that can impact our ability to detect differences.In Study 2, we added two attention check items.Also, Study 1's sample had a modest number of conservatives (n = 49 on the conservative end of the scale).As such, we recruited a larger number of participants in Study 2 in order to ensure a larger sampling of conservative participants.
Finally, Study 2 was conducted in order to see whether the pattern of results from Study 1 was stable over time, or whether any results were perhaps influenced by the time in which the data were collected.Study 1's data were collected between January and February of 2014, a few months after the government shutdown of 2013, in which House Republicans were widely perceived to be at fault, even by many conservatives (Woodward & Agiesta, 2014).The data for Study 2 was collected in May of 2015.Control of the senate had also shifted from Democrats to Republicans.

Method Participants
We recruited 329 Amazon M-Turkers.There were 163 women, 155 men, and eight people who did not report gender.The average participant age was 41.84, and ages ranged from 19 to 80. Participants were 81% White, 8% Black or African American, 4% Hispanic or Latino/a, 5% Asian or Pacific Islander, and 2% did not report their ethnicity.

Procedure
Participants completed items about political identification as the first step of the survey.Next, participants completed five survey sections -largely identical to those in Study 1. Once again, for the purposes of this research, only the emotions and demographic items were analyzed.The order of politician's party was randomized, the order of the survey sections was randomized, and the order of the items within each of these sections was randomized.After completing all questions pertaining to congresspersons from one political party, participants then answered the same questions again about congresspersons from the other political party.Finally, participants then filled out a demographics section.

Emotions -
The emotion measures for contempt, anger, disgust, happiness, and fear were identical to those used in Study 1, except that they used a 7-point scale (1 = not at all, 7 = very much).In keeping with Study 1, we recoded each pair of emotion ratings into congruent-ideology (liberals' emotions towards Democratic congresspersons, conservatives' emotions towards Republican ones) and incongruent-ideology emotion ratings (liberals' emotions towards Republican congresspersons, conservatives' emotions towards Democratic ones).
Congressional ideology and ideological distance discrepancy -In the same manner as Study 1, participants rated how liberal or conservative both Democratic and Republican congresspersons are, using a 7-point scale (1 = very liberal, 7 = very conservative).Ideological distance discrepancy was calculated in the same manner as Study 1.We first computed distance from ideologically-incongruent politicians via the absolute value of the difference between participants own liberal-conservative ideology and how liberal-conservative they rated their respective incongruent politicians.We next computed distance from ideologically-congruent politicians via the absolute value of the difference between participants' own ideology and how liberal-conservative they rated their respective congruent politicians.Finally, we computed ideological distance discrepancy via the difference between incongruent politician distance minus congruent politician distance.

Ideological Distance Discrepancy
In keeping with Study 1, we tested whether or not participant ideology was related to levels of ideological distance discrepancy.There was no relationship between the two r(327) = -.05,p = .34.

Emotions Towards Ideologically-Congruent Versus Ideologically-Incongruent Politicians
We used the same analytic strategy as in Study 1 to examine whether or not the differences between emotions depended on participants' ideology.The results are depicted in Table 2 and Figure 2.  As in Study 1, the main effect of target politician was significant across all emotions (see the far-left column of Table 2), in line with the ideological conflict interpretation suggesting people (liberals and conservatives) dislike ideologically dissimilar congresspersons to a greater extent than ideologically similar politicians.No significant main effects of participant ideology emerged across emotions except for contempt, which trended towards significance (see the middle column of Table 2) and indicated a larger level of contempt among conservatives overall (ŷ = 3.54) compared to liberals (ŷ = 3.26).
All interactions between target politician and participant ideology were significant with the exception of fear, which did not quite reach the cutoff for statistical significance (see the far-right column of Table 2).The pattern of effects suggests that liberals had a larger gap between emotion ratings for ideologically similar compared to dissimilar politicians (i.e., had greater degrees of emotional bias).However, the patterns of this bias varied across emotions.
In the case of anger, liberals-compared to conservatives-had more anger towards dissimilar congresspersons (although this pattern fell just short of significance).For contempt and disgust, liberals compared to conservatives showed less contempt and disgust towards ideologically similar congresspersons than conservatives, but no such pattern emerged for dissimilar congresspersons.For happiness, liberals experienced less happiness towards opposing party congresspersons than conservatives, but both groups showed a similar level of happiness towards like-minded congresspersons.
As in Study 1, we also controlled for ideological distance discrepancy in a second set of analyses.Again, the main effect for target politician held across all emotions, the pattern of effects and significance levels for the main effect of participant ideology held across emotions (with the exception of happiness with a reduction in p value from .11 to .09).The level of significance for all the participant ideology by target politician interaction remained the same across all emotions with the exception of happiness, with a change in p value from .04 to .06,suggesting that some of the effect of ideology on emotion may be a function of perceived ideological alignment.See Table A.2 in the Appendix for ANOVA tables that include ideological distance discrepancy.

General Discussion
A robust body of literature describes conservatives, compared to liberals, as more prone to experiencing fear, disgust, and happiness (e.g., Jost et al., 2003), which in turn have been found to predict conservatives' attitudes towards politically-relevant groups (e.g., Terrizzi et al., 2010).It logically follows that conservatives, compared to liberals, could potentially experience these emotions more strongly towards groups that represent a threat to their values (such as ideologically dissimilar congresspersons).In contrast, a growing body of work related to the ideological conflict hypothesis has found that liberals and conservatives may have equivalent levels of generalized negative affect towards ideologically dissimilar groups and/or generalized positive affect towards ideologically similar groups (Roscoe & Christiansen, 2010;van Prooijen et al., 2015).However, no studies had yet fully examined specific discrete emotions as separate variables within this context.
Across two studies, we provided a balanced test for these competing perspectives, and examined the extent to which anger, contempt, disgust, fear, and happiness towards ideologically similar and ideologically dissimilar politicians were driven by ideological differences in emotion, by ideological conflict effects, or by some mix of the two.In both studies, liberals and conservatives had greater magnitudes of negative affect (to a large degree) towards ideologically dissimilar congresspersons than they did towards ideologically similar ones, which held true across Second, it indicated that more than any other measured emotion, fear towards ideologically similar vs. dissimilar politicians might most closely correspond with the pattern of results that would be anticipated by the ideological conflict hypothesis (i.e., ideological symmetry).Finally, it further highlights the importance of assessing different discrete emotions as separate variables, as it indicated that not all negative emotions are equivalent or interchangeable within this context.

Contempt
Prior research has highlighted the unique characteristics of contempt as an especially toxic emotion in a political context (e.g., Tausch et al., 2011).Our inclusion of contempt was a unique contribution in and of itself, given that neither previous emotion studies in the context of the ideological conflict hypothesis nor any large-scale political surveys (e.g., the ANES) had yet assessed this emotion.Our results pertaining to contempt also further demonstrated that not all negative emotions are interchangeable in the context of ideological conflict: it was the only emotion that displayed an unambiguous and consistent ideological difference in bias across both studies.
The finding that liberals exhibited greater contempt bias than conservatives across both studies is novel and worthy of further exploration.It is not entirely clear what drove this finding.One possible explanation is that these results may reflect a general ideological difference, such that liberals may generally be more contemptuous than conservatives are towards ideologically threatening groups.However, another possible explanation is that these results could have been driven by conservatives being somewhat more contemptuous of ideologically similar politicians than liberals were, thus widening the contempt bias gap for liberals and narrowing it for conservatives.After all, there was an ideology main effect in Study 2 showing conservatives as higher on contempt overall.If conservatives were somewhat more contemptuous of ideologically similar politicians than liberals were, this may tie into a third possible explanation -participants' levels of contempt, and perhaps other emotions, towards congresspersons were predominately influenced by the political climate or political power dynamics at the time of data collection.

Political Climate
The possible influence of political climate at the time of data collection is difficult to account for, but likely played a role in our findings.Given that liberals only exhibited a contempt bias in Study 1, but exhibited anger, contempt, disgust, and happiness biases in Study 2, it may be the case that either current events or the overall political climate at the time of data collection may have led to an increase in emotional negativity towards Republican congresspersons across the ideological spectrum.This could account for a wider emotional bias gap for liberals and/or a narrower emotional bias gap for conservatives.Alternatively, the increase in overall negativity towards Republican congresspersons may have simply been associated with the duration of time the Republican-controlled congress had been in power.The list of potential reasons to feel negative emotions towards conservative members of congress may simply have gotten longer in the middle of their term (Study 2) than towards the beginning of their term (Study 1).

Political Power
The finding that liberals exhibited emotional bias across the majority of emotions in Study 2 but only exhibited bias for contempt in Study 1 could also be reflective of a shift in liberals' emotional response to ideologically dissimilar congresspersons due to a shift in political power between these two time periods.During Study 1, Republicans controlled the House while Democrats controlled the Senate.In Study 2, Republicans controlled both the

Future Research
While Studies 1 and 2 assessed multiple negative emotions (contempt, anger, disgust, and fear), only one positive emotion (happiness) was assessed.As such, there could be some ambiguity regarding what "happiness" represented in the eyes of participants.Although happiness was measured as an emotion, some participants may have instead construed happiness as a measure of approval/disapproval, which can be more nuanced and conditional than happiness as an emotion.Future studies could benefit from examining positive affect using more distinctly emotional names, such as pride or admiration.
The present studies' data were collected within a limited timeframe (2013)(2014)(2015), in which results may have been impacted by political power dynamics.Future research could examine similar questions in different political power arrangements, such as when Democrats have greater political power.Future research in a similar context could also examine the degree to which the ideological conflict hypothesis applies to emotion bias towards a wider variety of ideologically congruent versus incongruent targets.For instance, emotion biases in this context are likely not limited to politicians or political parties, but may also apply to groups more broadly seen as ideologically threatening to liberals versus conservatives (e.g., environmentalists versus religious fundamentalists).

Conclusion
Our studies provide several unique contributions to the literature.Although negative affect in general towards ideologically-incongruent groups may be equivalent between liberals and conservatives (e.g., van Prooijen et al., 2015), our results demonstrated that there are important distinctions to be made at the level of individual emotions, particularly regarding the under-studied emotion of contempt.
Our use of Democratic and Republican members of Congress as the target of our emotion measures also appears to be unique.Although numerous political science studies have been done on congressional approval (e.g., Pew, Gallup), and some research has assessed general negative and positive affect towards congress as a single entity (e.g., ANES feeling thermometers), our studies appear to be the first to assess multiple discrete emotions towards congressional Democrats versus Republicans.
Literature pertaining to the ideological conflict hypothesis contributed to the field of political psychology by showing that ideological similarity or dissimilarity is a primary consideration when it comes to expressing attitudinal bias.
That literature highlights multiple instances in which negative attitudes predominantly associated with conservatism in prior literature (e.g., prejudice and intolerance) may in fact be equivalent between liberals and conservatives when directed towards ideologically-incongruent groups on both sides of the political spectrum (vs.just incongruent for conservatives).Because there are consistent findings linking conservatism specifically with fear and disgust (but not anger or contempt), our analysis of these specific emotions (rather than just general affect) provides an important conceptual contribution to the ideological conflict literature, and adds important nuance to the ideology and emotions literature more broadly.The present studies show that the ideological conflict motive extends to even more fundamental reactions-our emotions-and that this motive may override general tendencies towards emotions experienced by ideological groups (e.g., conservatives).Finally, in showing that there are distinctions between individual emotions in this context (e.g., contempt), we hope to inspire additional research into the role of specific emotions within this new and exciting body of work.
Participants were instructed: "Please rate how strongly you feel each of the following emotions towards [Democratic/Republican] members of the [House of Representatives/Senate]."They rated five emotions:

Figure 1 .
Figure 1.Graphs for Study 1 results.Note.The Liberal and Conservative labels on the horizontal axis represent continuous political ideology.

Figure 2 .
Figure 2. Graphs for Study 2 results.Note.The Liberal and Conservative labels on the horizontal axis represent continuous political ideology.

Table 1
Study 1 ANOVA Statistics

Table 2
ANOVA Results for Study 2

Table A . 2
Study 2 ANOVA Results While Controlling for Ideological Distance Discrepancy