This is an open access article distributed under the terms of the Creative Commons
Attribution License (When real estate mogul Donald Trump announced his campaign for Republican presidential nomination in June 2015, it seemed like a “gift to late nights from the comedy gods” (Phillips, 2015). Despite public derision, he actively joined the race and raised substantial media attention by insulting practically everyone around him—starting with his closest competitors. He discredited their appearance (“Look at that face! Would anyone vote for that?” - Fiorina), their achievements (“He was an ok doctor” - Carson), and their intelligence (“can’t negotiate his way out of a paper bag” - Jeb Bush; “reminds me of a spoiled brat without a properly functioning brain” - Paul; “idiot” - Graham; see “Trump’s most insulting jabs”, 2015). However, while the media published pieces on “why no-one should take Donald Trump seriously” (Cillizza, 2015), he became the most supported Republican candidate. Political analysts claimed that his support would vanish in the wake of the real elections. Rather, he became the Republican presidential candidate, leaving his competitors and the world perplexed. Trump’s success thus raised a potent question: Would people vote for a president who based his political communication on disrespect? Are all norms of social conduct forgotten in the face of political competition?
In the present article, we suggest that disrespectful candidate behavior can indeed lead to success—if it meets the right receivers. Some voters may consider it necessary to behave aggressively and disrespectfully in the political arena, while others may feel that such behavior violates social norms and is illegitimate, irrespective of the domain. We expect this underlying attitude towards disrespect in politics to alter voters’ judgments toward observed disrespectful behavior and, by extension, their decisions at the polls.
To develop a more nuanced view on these dynamics, our study adds three elements to the extant literature. First, we go beyond a general perspective on person perception and examine consequences of disrespect in politics the two basic dimensions of social judgment: communion and agency (e.g., Abele & Wojciszke, 2007; Cislak & Wojciszke, 2008; Fiske, Cuddy, & Glick, 2007; Fiske, Cuddy, Glick, & Xu, 2002). Second, we investigate the behavioral consequences of these social judgments not only in terms of voting intention, but also in terms of voting decision. To accomplish this latter goal, we collected real world data in the course of an actual election. Lastly, we develop an instrument that taps general attitude towards disrespect in politics, which enables us to unfold the above dynamics. Figure 1 provides a summary of the hypothesized relationships.
Figure 1
The hypothesized relationships.
In sum, we aim to contribute to the fields of both social psychology and political science by a) further investigating how voters’ underlying attitude towards disrespect in politics affects the relation between politicians’ perceived respectfulness and voters’ social judgment, and b) showing that voting intentions and actual voting are affected through these social judgments. Our findings reveal interesting perspectives, both theoretical and practical, on the phenomenon of disrespect in politics. In particular, our study further illustrates that voters’ varied reactions to disrespectful behavior stem from the different mindsets underlying their judgments. This helps to illuminate why politicians perpetually engage in negative communicative practices: Just like Donald Trump, they may assume that many voters approve of disrespectful behavior in the political arena.
Defining Respect and Disrespect
In social psychological research just as in daily life, several notions of respect are used, resulting in a heterogeneous body of respect research (Simon, 2007; Van Quaquebeke, Henrich, & Eckloff, 2007). Darwall (1977) offered a dualist approach to respect, differentiating recognition respect and appraisal respect. Recognition respect means treating another person as equal in worth and dignity and consciously dealing with his or her needs and opinions. Simon (2007) and Van Quaquebeke and colleagues (2007) note that the term “respect” – derived from the Latin word “respicere” which means look back or consider – refers to the Kantian concept of ‘Achtung’ or ‘respect for persons’ (Kant, 2012): Individuals should always be treated as ends in themselves due to their human dignity. Thus, to respect someone implies the willingness to seriously consider his or her perspective (Simon, 2007). Accordingly, it exceeds the mere tolerance of a person’s presence (Van Quaquebeke et al., 2007).
In contrast, appraisal respect is granted when another person has notable abilities, knowledge or achievement that the subject evaluates positively or admires (Simon, 2007). This kind of respect is based on individual differences rather than equality. It is not, in a moral sense, owed to every person, but it must be deserved.
This dualism echoes Huo, Binning, and Molina’s (2010) finding that links feeling respected to status and inclusion. Likewise, Spears and colleagues (2005) differentiated between competence-based and liking-based respect. In contrast, Wojciszke and colleagues (2009) used an appraisal-based concept of respect, defining it as “asking the other for opinion on important issues, voting for the other, and respecting the other’s knowledge” (p. 974). In line with this, Fiske and colleagues (1999) linked being respected to the attribution of competence.
However, in this paper, we consider this notion of recognition respect as most appropriate for our purposes, as we focus on how politicians treat each other in political debates. In the heat of such debates, politicians tend to forget “to give proper weight to the fact that [their opponents] are persons” (Darwall, 1977, p. 39). In order to ‘win’ a debate against the opponent, they decide to personally attack their opponents, not take them seriously, and discredit them as persons, (i.e., their integrity, reliability, or intelligence). In Argumentation Theory, this is called an ad hominem (‘to the person’) argument, i.e., a “direct personal attack in which the opposite party is represented as stupid, dishonest, unreliable, or negative in some other way” (van Eemeren & Grootendorst, 2004, p. 177). The opponent’s arguments are devalued by discrediting the opponents themselves. The opponent is not acknowledged as equal but degraded. It is this kind of arguing which we refer to as disrespect: personal attacks that treat a counterpart as inferior in worth and dignity. Respectful behavior, in contrast, displays the intention to consider someone else‘s perspective (Simon, 2007). It can be defined as “not humiliating [others], taking their views seriously, and providing reasons for one’s own view” (Lalljee, Tam, Hewstone, Laham, & Lee, 2009, p. 667).
Importantly, approaching disrespectful behavior through ad hominem arguments implies to a focus on verbal forms of disrespect. Undoubtedly, nonverbal behavior such as ignoring (Blincoe & Harris, 2011), rolling one’s eyes, shaking one’s head, etc. (Mutz & Reeves, 2005) can express disrespect as well. However, it is the verbal form of disrespect which we consider as most relevant for the political context, especially for political debates: political opponents attack each other with words, and media multiply these words in their coverage.
Disrespect in the Light of Flexible Social Norms
A person behaving disrespectfully towards others is often evaluated negatively—not only from the target, but from bystanders as well (e.g., Martijn, Spears, Van der Pligt, & Jakobs, 1992; Skowronski & Carlston, 1987). In the political context, research on negative campaigning shows that disrespectful communication can cause damage to attackers’ reputations (e.g., Brooks & Geer, 2007; Fridkin & Kenney, 2004; Mutz & Reeves, 2005), decreasing their likeability (Carraro & Castelli, 2010; Fridkin & Kenney, 2004), perceived fairness (Brooks & Geer, 2007), and their audiences’ positive affect (Fridkin & Kenney, 2004; Roese & Sande, 1993). This is most likely due to the fact that disrespectful behavior violates social norms regarding how to treat each other.
Granted, some studies have revealed that disrespect can also have positive consequences in terms of conformity with the attacker (Carraro & Castelli, 2010) or intention to vote for the attacker (Kaid, 1997; Roddy & Garramone, 1988). However, many other studies found no effect at all (see Lau, Sigelman, & Rovner, 2007, for an overview). Such mixed results most likely point to the presence of moderators. In this sense, variables such as ideological closeness between politician and observer (Houston & Doan, 1999) or newspaper consumption (Faber, Tims, & Schmitt, 1993) have been identified as potential moderators that are able to buffer the negative effects of disrespect. Taking a more respect-specific approach, Fridkin and Kenney (2011) investigated how the voters’ “tolerance of negative messages” (p. 312) in political commercials altered political candidate evaluation, showing that voters’ underlying attitudes play an important role in the conclusions drawn from disrespectful messages. With this paper, we build on their approach and extend it by examining the effects of disrespectful candidate behavior on voting intention and actual voting decision. In doing so, we provide an integrated model of the proposed relationships and thus also advance Fridkin and Kenney’s approach methodologically. Furthermore, we introduce a scale that taps attitude towards disrespect in political debates in general, thus going beyond the specific context of negative commercials. The instrument we developed can also be applied to cultures less shaped by negative campaigning (e.g., European countries with multi-party systems; see Walter & van der Brug, 2013).
There are three reasons why we argue that disrespect in political communication may not always engender negative evaluations. First, research shows that people find disrespect to be entertaining (Mutz & Reeves, 2005). Beyond all norms and credibility, this could be a very simple reason behind voters’ approval of disrespect in the political context. For instance, London’s Mayor Boris Johnson is “better known for his supposed gaffes–and subsequent apologies–than anything he had achieved as an MP [Member of Parliament]” (Wheeler, 2008).
Second, some suggest that voters want politicians to be authentic instead of scripted (Louden & McCauliff, 2004). Authenticity is linked to a politician’s honesty and credibility and can be signaled by emotional expression (Hareli et al., 2009; Louden & McCauliff, 2004). However, emotional expression alone may not sway an audience; to be authentic it must be appropriate to a time and place. This seems somewhat intuitive: Skilled orators in ancient Greece “already understood that the person who is not able to get angry in the right way about the right things at the right time will be seen as a fool” (Vaes, Paladino, Castelli, Leyens, & Giovanazzi, 2003, p. 1022). In this sense, lashing out against an opponent, and losing one’s respectful countenance in the process, could be seen as an authentic expression of a candidate’s inner state, and therefore a legitimate reaction towards an opponent’s standpoint.
Third, people perceive others’ behavior differently depending on the situation in which said behavior is embedded (Reeder & Brewer, 1979; Snyder & Frankel, 1976): “One basic assumption guiding the attribution process is the belief that persons will vary their behaviors, if possible, in order to best meet environmental contingencies and demands” (Reeder & Brewer, 1979, p. 64). In the case of politics, voters may be cognizant of the fact that the rationale of the “political game” is to stand apart and outrival the opponent. Accordingly, politicians are usually expected to be strong and courageous leaders (Kinder, Peters, Abelson, & Fiske, 1980), as well as energetic and aggressive (Trent, Trent, Mongeau, & Short-Thompson, 1997). Based on this assumption, voters might shift the social standards downwards (see Biernat & Manis, 1994) in order to accommodate politicians’ interpersonal behavior. Thus, disrespectful behavior can be seen as the political actor’s effort to meet expectations rather than violate social norms.
We rely on these theoretical explanations when we suggest that voters’ general attitude of approval toward disrespect may buffer or reverse negative evaluations of candidates’ disrespectful behavior. This, in turn, may further impact their voting intention and decisions. In the next section we differentiate why we suppose such an effect for the two central dimensions of social judgment, communion and agency.
How Attitude Towards Disrespect Relates to Social Judgment and Voting
Two dimensions have been consistently found to underlie people’s social judgment of others (e.g., Abele & Wojciszke, 2007; Cislak & Wojciszke, 2008; Fiske, Cuddy, & Glick, 2007; Fiske, Cuddy, Glick, & Xu, 2002): communion (or warmth, morality, expressiveness, social relatedness) and agency (or competence, instrumentality, power; Abele & Wojciszke, 2013; Judd, James-Hawkins, Yzerbyt, & Kashima, 2005). Communion taps an emotional, interpersonal orientation, while agency taps a sense of assertiveness and power (Conway, Pizzamiglio, & Mount, 1996). Both dimensions are essential for the evaluations of politicians (e.g., Caprara, Barbaranelli, & Zimbardo, 2002; Cwalina & Falkowski, 2016; Funk, 1996; Kinder, Peters, Abelson, & Fiske, 1980; McCurley & Mondak, 1995), and we thus hypothesize that disrespectful behavior will affect both communion and agency.
Concerning the communion dimension, we generally expect disrespectful candidate behavior to produce a negative effect, in line with numerous research results: When speaking and acting disrespectfully, attackers are often met with fewer voters feeling positive about them (Fridkin & Kenney, 2004), voters perceiving them as less warm (Carraro & Castelli, 2010), and less-positive overall evaluations (Haddock & Zanna, 1997). These negative consequences on the interpersonal level can be explained by disrespect violating social norms and expectations of “good” conduct. However, as outlined above, we propose that any such effect is ultimately shaped by voters’ attitude towards disrespect in the political arena. Specifically, voters who apply standard social norms to the political context, and thus disapprove of disrespect in politics, should provide negative communion ratings of such behavior. In contrast, voters who perceive disrespect in political debates as normal, and thus apply less restrictive social norms to the political context, are expected to derive less-negative communion judgments from observed disrespect. Formally:
H1: The more negative an observer’s attitude towards disrespect in politics, the more a disrespectful candidate’s behavior will translate into lower communion ratings.
Meanwhile, we expect disrespectful candidate behavior to exert a different effect on agency. Carraro and Castelli (2010) found that disrespectful candidate behavior increased participants’ spontaneous conformity towards the disrespectful person, indicating an openness to influence that is likely related to perceived strength and power. Similarly, verbal aggression has been linked to attributions of toughness and coolness (Rodkin, Farmer, Pearl, & Van Acker, 2000), as well as assertiveness and competitiveness (Buss & Perry, 1992). In line with this, we propose that disrespectful candidate behavior can increase voters’ perceptions of agency. However, this effect should be exclusive to voters who approve of disrespect in politics. Reversely, we expect those voters to interpret not being disrespectful as a sign of weakness and inferiority and thus as less agentic. In contrast, we do not expect voters who disapprove of disrespect in politics to positively relate disrespectful behavior with agency. Formally:
H2: The more positive an observer’s attitude towards disrespect in politics, the more a disrespectful candidate’s behavior will translate into higher agency ratings.
To demonstrate the practical reach of our proposal, we also factor in behavioral intention and actual behavior. Research has shown that voting behavior is potentially affected by both communion and agency judgments: Cwalina and Falkowski (2016) found that both dimensions essentially influence a political candidate’s image and whether voters will support this candidate in an election. Similarly, Pfau and colleagues (1993) found that voting intention was much more affected by communion and agency than by issues, and McCurley and Mondak (1995) showed that agency and communion both affect the vote choice.
Accordingly, we expect higher communion and agency ratings to engender a higher intention to vote for the respective candidate (Cislak & Wojciszke, 2008; Wojciszke, 2005; Wojciszke & Abele, 2008), and to translate into an actual voting decision (Gaffney & Blaylock, 2010). Given this link, we propose that disrespect indirectly affects voting intention through social judgment (mediation). As the relation between disrespect and social judgment is supposed to be moderated, we proposed the moderation effects in terms of the indirect effects (moderated mediation model).
H3: The more negative an observer’s attitude towards disrespect in politics, the more a disrespectful candidate’s behavior will decrease voting intention/ voting decision for the candidate due to decreased communion ratings.
H4: The more positive an observer’s attitude towards disrespect in politics, the more a disrespectful candidate’s behavior will increase voting intention/ voting decision for the candidate due to increased agency ratings.
Present Research
To test our hypotheses, we developed a scale aimed at tapping people’s attitude towards disrespect, and then tested its reliability and criterion validity. Afterwards, we conducted two studies to test our hypotheses. In an experimental vignette study, we provided participants with a description of a fictitious politician who reportedly behaved respectfully or disrespectfully, and then measured their social judgments, intention to vote, and attitude towards disrespect in politics. Afterward, we sought to validate our findings in the field by investigating the actual behavioral impact of our hypothesized relations. Therefore, prior to the elections for German parliament in 2013, we asked participants to rate the two top candidates for chancellorship, Angela Merkel and Peer Steinbrück, with regard to their displayed respect towards the opponent and their respective communion and agency; we then measured participants’ attitude towards disrespect. One week after the election, the same participants reported their voting decisions.
Pretests of the Scale “Attitude Towards Disrespect in Politics”
For our operationalization of attitude towards disrespect in politics, we drew on the theoretical notion that people’s higher acceptance of disrespect in politics can stem from perceiving such attacks as normal, legitimate, and perhaps even desirable. From that, we developed nine items that tapped participants’ attitude towards disrespect in politics. Responses were made on a 5-point Likert scale ranging from completely disagree (1) to completely agree (5). We pretested these items in a survey (N = 329; see Mölders, Van Quaquebeke, & Paladino, 2017) in order to assess their reliability and validity. We recruited the sample through a combination of social network platforms and the online recruiting platform of a German distance university. Participants’ ages ranged between 18 and 71 years (M = 32.26, SD = 11.05), and 66% were female. We removed two items after performing an analysis of internal consistency (α = .67). For the remaining seven items, Cronbach’s Alpha was α = .82. These final seven items are reported in the Appendix.
In order to establish validity, we checked the results against participants’ self-reports of conflict approach (Goldstein, 1999), moral awareness (Reynolds, 2006), and political interest. Participants’ openness towards conflicts might lower their sensitivity towards disrespect in debates; thus, we expected that a higher willingness to be involved in conflicts would correspond with a more positive attitude towards disrespect (positive relation). In contrast, as disrespect can be seen as immoral behavior, we expected participants’ moral awareness (i.e., acknowledgement “that a situation contains moral content and legitimately can be considered from a moral point of view”; Reynolds, 2006, p. 266) to align with a more negative attitude towards disrespect (negative relation). Finally, as the display of disrespectful candidate behavior is quite common in politics (Sobieraj & Berry, 2011), we expected participants’ interest in politics to be weakly linked to a more positive attitude towards disrespect as a matter of habituation and/or necessary predisposition. Our predicted correlations were accurate: Participants’ attitude towards disrespect in politics was positively related to their conflict approach (r = .34**), negatively related to their moral awareness (r = -.31**), and positively related to their political interest (r = .20**). Moreover, participants’ attitude towards disrespect was related to their age (r = .24**), but not to their gender (r = -.09).
The scale was then included in an additional survey study (N = 273; participants aged between 19 and 81 years, M = 44.45, SD = 14.08; 53% male; see Mölders et al., 2017), where it demonstrated consistently good internal reliability (T1: α = .82; T2: α = .87) and test-retest reliability (rtt = .79**) when presented with a time-lag of one week.
Study 1
Sample
The experimental vignette study was conducted online. Participants were recruited via a professional German online panel (Respondi AG) in exchange for panel credits. The first page of the study was visited by 308 persons. Of these, 261 finished the questionnaire without significant pauses. Participants who were younger than eighteen (and were thus not allowed to vote) and those who incorrectly answered a control item (“Please select the middle option here”) were automatically screened out and thus did not enter into our dataset. Research supports the use of such instructional manipulation checks to enhance data quality (see Oppenheimer, Meyvis, & Davidenko, 2009; Paolacci, Chandler, & Ipeirotis, 2010). Moreover, we removed 14 participants who did not indicate their gender and 18 participants who did not agree with the use of their data. In the end, our analysis included data from 229 participants, 48% of whom were female, their ages ranging from 18 to 75 years (M = 42.25, SD = 15.16). Asked for their political orientation, 18% reported to usually support the Christian Democrats (CDU or CSU), 23% the Social Democrats, 10% the Green Party, 4% the Liberals, 9% the Left Party, 4% the Alternative for Germany, and 8% others, while 7% declined to indicate their favorite party and 18% reported that they did not support any specific party.
Procedure
Manipulation
We employed a one-factor, between-subjects design. Participants were randomly assigned to one of two conditions (respectful vs. disrespectful) and presented with a vignette about a fictitious candidate who was said to be a party member for 25 years, a member of German parliament for 10 years, and an expert on economic issues. Additionally, we randomly manipulated the candidate’s party (Christian vs. Social Democrats / two main parties in Germany), gender (male vs. female), and name (Johann/a Schulz vs. Paul/a Kunze) for control reasons. Subsequently, we manipulated respectful vs. disrespectful communication towards the opponent by presenting eight sentences (either respectful or disrespectful) that this candidate was said to have used in a TV debate (e.g., “I don’t completely understand your point of view” or “I can see your point, but I still disagree” vs. “You haven’t the faintest idea of what you’re doing” or “This is hypocrisy–there’s no other name for that”). These statements were obtained from actual political talk-shows and debates or, in case of some respectful statements, were newly created (please see Mölders et al., 2017 for the complete selection rationale). As a pretest, we had 58 participants rate 56 sentences on an 11-point-scale and picked eight sentences for our vignettes that were clearly rated as disrespectful (< M - 1 SD) and eight complementary statements that were rated as respectful (> M + 1 SD).
Measures
After participants provided their demographic data and read the vignettes, we asked them to rate the candidate’s communion and agency using items from the Personal Attributes Questionnaire (PAQ; Spence & Helmreich, 1978) that were validated by Ward and colleagues (2006). Responses were given on a 5-point scale that juxtaposed an adjective at one end with its opposite or negation at the other. Communion was measured by six items (e.g., not at all kind – very kind; very cold in relations to others – very warm in relations to others; α = .93); agency, by seven items (e.g., very passive – very active; not at all competitive – very competitive; α = .83).
Then, we measured intention to vote using three items (“Would you vote for the candidate if he or she ran for office in an election?”; “Do you think this politician is good for Germany?”; “Would you recommend voting for this politician to others?”; α = .93). Responses to each item were made on 5-point scales (no (1) – rather not – maybe – probably yes – yes (5)).
Afterwards, we informed the participants that the candidate had been fictitious and that the person’s gender, party affiliation, and statements had been randomly selected. We then asked them to indicate their identification with the Christian Democratic Union (CDU) and the Social Democratic Party (SPD) on a Venn diagram (Schubert & Otten, 2002), and to report which party they usually supported as well as their education. Finally, participants’ attitude towards disrespect in politics was measured by the scale we developed (α = .83), and participants indicated their educational attainment. The survey concluded by asking participants for the permission to use their data, thanking them, and revealing the purpose of the study.
Method
To analyze moderation and conditional indirect effects, we used the SPSS macro PROCESS provided by Hayes (comprehensively explained in Hayes, 2013). We standardized the moderating variables before entering them into the regression. For dichotomous variables, we report unstandardized regression coefficients in line with Hayes’ (2013) recommendations.
Results
Main Effects and Interactions
The politician’s name was marginally predictive of communion ratings, such that “Johann/a” was perceived as slightly more communal than “Paul/a” (b = 0.22, t(227) = -1.75, p < .09). The female politician received slightly higher communion ratings than the male politician (b = 0.21, t(227) = p < .10), while the politician’s party did not affect communion ratings (p = .50). Neither participants’ agency ratings nor their voting intentions were related to politician’s name, gender, or party. Thus, we used all three variables as controls in the following analysis.
First, communion and agency ratings were not correlated (r = .02, p = .81). Meanwhile, the candidate's respectfulness was positively related to participants’ communion judgments (b = 1.12, t(227) = 11.18, p < .001, R2 = .35) and negatively related to their agency judgments (b = -.42, t(227) = -4.68, p < .001, R2 = .08). Participants’ attitude towards disrespect in politics predicted their communion judgments of the candidate (β = .18, t(227) = 2.72, p < .01, R2 = .03), but not their agency judgments (β = .06, t(227) = 0.89, p = .37) and, moreover, moderated the effect of disrespectful candidate behavior on participants’ communion judgments (b = 0.25, t(226) = 2.01, p < .05): In line with H1, simple slope analysis revealed that the negative effect of disrespectful candidate behavior on communion ratings became stronger as participants reported a more negative attitude towards disrespect (Table 2; Figure 2).
Table 1
Regression-Based Moderations and Moderated Mediation in Study 1
| Communion
|
Agency
|
Voting intention
|
|||||||
|---|---|---|---|---|---|---|---|---|---|
| Predictor | R2 | b | SE | R2 | b | SE | R2 | b | SE |
| Politician’s gendera | .18† | .10 | -.02 | .09 | -.03 | .10 | |||
| Politician’s partyb | .01 | .10 | .01 | .09 | -.12 | .10 | |||
| Politician’s namec | -.23** | .10 | .11 | .09 | .06 | .10 | |||
| R2 after Step 1 | 0.02 | .00 | .00 | ||||||
| Candidated | 1.11*** | .10 | -.41*** | .09 | -.20 | .13 | |||
| Attitude | .30*** | .08 | .04 | .08 | .25** | .08 | |||
| R2 after Step 2 | 0.39 | .08 | .08 | ||||||
| Candidate*Attitude | -.20* | .10 | -.18† | .09 | -.30** | .10 | |||
| R2 after Step 3 | 0.40 | .09 | .13 | ||||||
| Communion | .80*** | .07 | |||||||
| Agency | .38*** | .08 | |||||||
| R2 after Step 4 | .53 | ||||||||
Note. Full model reported.
a0 = male, 1 = female; b0 = CDU, 1 = SPD; c0 = Johann/a, 1 = Paul/a; d0 = disrespectful, 1 = respectful.
†p < .10. *p < .05. **p < .01. ***p < .001.
The interaction effect of candidate’s respect and voters’ attitude towards disrespect in politics on voters’ agency ratings of the candidate was marginally significant (b = 0.22, t(226) = 1.92, p < .06). In line with H2, simple slope analysis revealed that the effect of disrespectful candidate behavior on participants’ agency judgments became more positive as participants reported a more positive underlying attitude towards disrespect (Table 2; Figure 2). This tendency was weaker for participants who disapproved of disrespect in politics. All effects held when we controlled for candidate’s gender, name, and party (see Table 1).
Table 2
Simple Slope Analysis of the Interaction Effect of Disrespectful vs. Respectful Candidate Behaviora and Voters’ Attitude Towards Disrespect in Politics on the Two Dimensions of Social Judgment in Study 1
| Attitude towards disrespect | DV: Communion
|
DV: Agency
|
||||
|---|---|---|---|---|---|---|
| b | SE | p | b | SE | p | |
| Negative (-1 SD) | 1.33 | 0.14 | .000 | -0.25 | 0.13 | .05 |
| Positive (+1 SD) | 0.89 | 0.14 | .000 | -.58 | 0.12 | .000 |
a0 = disrespectful candidate behavior, 1 = respectful candidate behavior.
Figure 2
Simple slopes for the interaction effect of disrespectful vs. respectful candidate behavior and participants’ attitude towards disrespect on their communion and agency judgments of the candidate in Study 1 (N = 229).
Conditional Indirect Effects
Participants’ voting intention in favor of the candidate was predicted by both their communion (β = .66, t(226) = 14.38, p < .001) and agency judgments (β = .29, t(226) = 6.21, p < .001; overall R2 = .51). Furthermore, their voting intention was predicted by the candidate’s respect (b = 0.54, t(227) = 3.88, p < .001, R2 = .06). Meanwhile, the more positive the participants’ attitude towards disrespect, the greater was their willingness to vote for the presented candidate (β = .17, t(227) = 2.55, p < .05; R2 = .02).
We found a negative indirect effect of disrespectful candidate behavior on participants’ voting intention through communion that varied with different levels of attitude towards disrespect (Table 3): It was strongest when participants’ attitude towards disrespect was negative, and it was weakest when their attitude towards disrespect was positive, thus supporting H3. In contrast, disrespectful candidate behavior had a positive indirect effect on participants’ voting intention in favor of the candidate through increased agency judgments; this effect was stronger when attitude towards disrespect was positive and weaker when it was negative (Table 3), thus supporting H4. These results also held when we controlled for participants’ party identification, gender, age, and education.
Table 3
Conditional Indirect Effect of Disrespectful vs. Respectful Candidate Behaviora on Voting Intention Dependent on Voters’ Attitude Towards Disrespect in Politics in Study 1
| Attitude towards disrespect | Med: Communion
|
Med: Agency
|
||||
|---|---|---|---|---|---|---|
| Effect | Boot SE | 95% BCI | Effect | Boot SE | 95% BCI | |
| Negative (-1 SD) | 1.10 | 0.15 | [0.82, 1.43] | -0.09 | 0.05 | [-0.22, -0.003] |
| Positive (+1 SD) | 0.77 | 0.12 | [0.53, 1.02] | -0.23 | 0.07 | [-0.39, -0.11] |
Note. Number of bootstrap samples = 20.000. BCI = Bias corrected limits of the confidence interval for the indirect effect.
a0 = disrespectful candidate behavior, 1 = respectful candidate behavior.
Discussion
As predicted, the experiment showed that the consequence of disrespectful candidate behavior in political debates not only differs between dimensions of social judgment, but also depends on voters’ general attitude towards disrespect in politics. The described disrespectful behavior decreased participants’ communion ratings and, subsequently, their voting intention in favor of the respective candidate. The effect became stronger as participants’ disapproval of disrespect in politics increased. In contrast, disrespectful candidate behavior increased agency ratings and subsequent voting intention in favor of the candidate, but only for participants who approved of disrespect in political debates. Seemingly, voters with a negative attitude towards disrespect in politics do not consider a politician’s degree of respect toward an opponent as significantly informative of his or her agentic qualities.
Meanwhile, a positive attitude towards disrespect was related to generally higher communion ratings and a higher intention to vote for the presented politician. However, this effect did not apply to the agency dimension, countering the idea that a more positive attitude towards disrespect produces a generally more favorable evaluation of disrespectful politicians. Another interesting side finding was the marginal differences in the communion ratings arising from the politician’s name. In the future, psycholinguists might unravel whether name length or name-related stereotypes might play a systematic role in the judgment of politicians.
In this experimental setting, the disrespectful and respectful candidates were fictitious and described in only basic terms. In order to demonstrate constructive replication and external validity, we conducted a second study that investigated disrespect, social judgment, and actual voting decisions in the context of two, real political candidates.
Study 2
Procedure
Sample
Participants were recruited via a professional German online panel (Research now) in exchange for credit within the panel system. They were invited to the survey about one week before (T1) and after (T2) the German parliament election of 2013. At T1, participants rated the two chancellor candidates for the two biggest German parties, Angela Merkel for the Christian Democrats (CDU/CSU) and Peer Steinbrück for the Social Democrats (SPD). In the course of the electoral race, challenger Steinbrück had raised a lot of attention by being disrespectful towards opponents, media, and even partisans, ultimately symbolized by a picture of him giving the middle finger to his critics during an interview (Denkler, 2013). Incumbent Merkel, in contrast, was known for her unexcited, somber attitude. As our later analysis shows, they nicely paralleled our disrespectful vs. respectful candidates in the experimental design. At T2, the same participants were asked for their voting decision. We matched the two datasets by a unique but anonymous code that the panel provider assigned to the participants.
The survey’s welcome page at T1 was seen by 470 participants, of which 362 finished the questionnaire without significant pauses and agreed with the use of their data. We excluded 25 participants who did not pass our instructional manipulation check (Oppenheimer et al., 2009; Paolacci et al., 2010). Of the remaining participants, 268 also took part at T2, of which 67 did not indicate their voting decision (not at all or “no comment”), one did not indicate gender, and one did not indicate education. In the end, we analyzed the data of 199 participants. Their ages ranged from 19 to 85 years (M = 44.92, SD = 15.19), and 48% were female. Asked for the party they had voted for, 30% indicated the Christian Democrats (CDU or CSU; actual parliament election results: 34%), 21% the Social Democrats (actual election: 26%), 12% the Green Party (actual election: 8%), 4% the Liberals (actual election: 4.8%), 15% the Left Party (actual election: 9%), 7% the Alternative for Germany (Alternative für Deutschland; actual election: 4.7%), and 13% others (actual election: 6%).
Measures
At the beginning of the questionnaire, participants were asked whether they were familiar with Angela Merkel and Peer Steinbrück. This question was intended as a screen-out; however, all participants indicated familiarity with these candidates. All participants then evaluated both candidates, Merkel and Steinbrück, in random order. First, they rated the candidate’s respect towards the political opponent on six items developed according to the definitions of respect proposed by Simon (2007) and Van Quaquebeke and colleagues (2007; recognizes his/her political opponent as equal in worth; attacks his/her political opponent on the issue level rather than the person level; his/her arguments sometimes aim below the belt (r); treats his/her political opponent with respect; does not try to enhance his/her status at the expense of others; uses an appropriate tone in debates; α = .87). Participants then judged the candidates’ communion and agency and indicated their attitude towards disrespect in politics with the same measures used in Study 1. Furthermore, we asked participants to supply their identification with the candidates’ parties (CDU and SPD) using a Venn diagram, as well as their partisanship, education, age, and gender. At T2, participants reported which party they actually voted for. Afterwards, they were thanked and debriefed about the purpose of the study.
Method
Voting decision was converted into two dichotomous outcome variables: Voting decision in favor of Merkel’s CDU (yes/no) and in favor of Steinbrück’s SPD (yes/no). The dataset was then restructured: Each participant was doubled so there were two datasets per participant with one row per rated candidate, and created a dummy variable for the candidate. As this restructuring caused clusters within the data, we used the robust standard error for clustered data (Rogers, 1993) for our analyses. It is based on the Huber-White Sandwich estimator and is robust to within-cluster correlations (Wooldridge, 2002), thereby providing a better estimate of the true variance (Williams, 2000). Degrees of freedom were set according to the number of participants. Analyses were conducted in Stata and logistic regression was applied to account for the dichotomous outcome variable. In our analysis, we used participants’ gender, age, and education as control variables.
Results
Ratings of Merkel and Steinbrück
In order to validate our approach of treating Steinbrück as a disrespectful candidate and Merkel as a respectful one, we tested for significant differences in their respect ratings (T-test for matched samples). We found that participants rated Steinbrück as significantly less respectful (M = 3.11, SD = .80) than Merkel (M = 3.73, SD = .79; t(198) = 6.77, p < .001). Furthermore, they rated Merkel as significantly more communal (M = 3.17, SD = .96) than Steinbrück (M = 2.92, SD = .87; t(198) = 2.46, p < .02). The difference concerning agency was just marginally significant (t(198) = 1.83, p < .07), with Steinbrück receiving higher ratings (M = 3.75, SD = 0.86) than Merkel (M = 3.63, SD = 0.64). Table 4 presents the regression-based predictions of participants’ communion and agency ratings, as well as their voting decisions.
Table 4
Regression-Based Moderations and Moderated Mediation in Study 2
| Communion
|
Agency
|
Voting decision
|
|||||||
|---|---|---|---|---|---|---|---|---|---|
| R2 | b | RSE | R2 | b | RSE | Wald χ2 | b | RSE | |
| Participants’ gendera | 0.05 | 0.08 | 0.09 | 0.07 | -0.17 | 0.27 | |||
| Participants’ age | -0.07† | 0.04 | -0.03 | 0.03 | 0.32 | 0.13* | |||
| Participants’ education | -0.01 | 0.04 | -0.02 | 0.03 | -0.01 | 0.13 | |||
| R2 after Step 1 | 0.01 | 0.01 | 5.91 | ||||||
| Candidateb | 1.99*** | 0.46 | 0.52 | 0.32 | 0.71 | 1.18 | |||
| Attitude | 0.38*** | 0.08 | 0.16* | 0.06 | 0.25 | 0.25 | |||
| R2 after Step 2 | 0.03 | 0.02 | 12.20* | ||||||
| Candidate*Attitude | -0.54*** | 0.15 | -0.20* | 0.10 | -0.14 | 0.35 | |||
| R2 after Step 3 | 0.09 | 0.04 | 13.84* | ||||||
| Communion | 0.87*** | 0.18 | |||||||
| Agency | 0.49† | 0.26 | |||||||
| R2 after Step 4 | 54.74*** | ||||||||
Note. Full model reported, controlled for 199 clusters. RSE = robust standard error.
a0 = male, 1 = female; b0 = Steinbrück (disrespectful), 1 = Merkel (respectful).
†p < .10. *p < .05. **p < .01. ***p < .001.
Interactions
Participants’ attitude towards disrespect was marginally related to their agency ratings of the candidates (β = .09, t(195) = 1,72, p < .10), but not to their communion ratings. There was a significant interaction between candidate and attitude towards disrespect that predicted communion, b = .54, t(195) = 3.76, p < .001. As hypothesized, the effect was strongest for voters with a negative attitude towards disrespect in politics (H1). For voters with a positive attitude towards disrespect, communion ratings did not differ between the two candidates (Table 5). Likewise, candidate and attitude towards disrespect interacted in predicting agency, b = .20, t(195) = 2.01, p < .05. As in the experiment, the effect was strongest for voters with a positive attitude towards disrespect in politics, supporting H2. It did not occur for voters with a negative attitude towards disrespect in politics (Table 5). Figure 3 shows the interaction effects of candidate and attitude towards disrespect on participants’ communion and agency judgments.
Table 5
Simple Slope Analysis of the Interaction Effect of Disrespectful (Steinbrück) vs. Respectful (Merkel) Candidate Behaviora and Voters’ Attitude Towards Disrespect in Politics on the Two Dimensions of Social Judgment in Study 2
| Attitude towards disrespect | DV: Communion
|
DV: Agency
|
||||
|---|---|---|---|---|---|---|
| b | SE | p | b | SE | p | |
| Negative (-1 SD) | 0.71 | 0.13 | .000 | .05 | .09 | .62 |
| Positive (+1 SD) | -.20 | 0.13 | .12 | -.29 | .09 | .002 |
a1 = Steinbrück, 2 = Merkel.
Figure 3
Simple slopes for the interaction effect of disrespectful vs. respectful candidate behavior and participants’ attitude towards disrespect on their communion and agency judgments of the candidate in Study 2 (N = 199).
Conditional Indirect Effects
Both participants’ communion and agency judgments of the candidate predicted their voting decision, with b = .88, SE = .17, p < .001 for communion, and b = .48, SE = .25, p < .06 for agency. For participants with a negative attitude towards disrespect in politics, the effect of the candidate on voting decision was mediated through their communion judgments: Those participants provided lower communion judgments for Steinbrück and therefore were less likely to vote for him. In contrast, when participants’ attitude towards disrespect was positive, their communion judgments of the candidate did not mediate the relation between candidate and voting decision. Thus, H3 was overall supported (Table 6). For agency, we found the expected reverse pattern: Participants with a positive attitude towards disrespect perceived Steinbrück as more agentic than Merkel and, therefore, were more likely to vote for him. In contrast, for participants with a negative attitude toward disrespect, their agency judgments did not mediate the relationship between candidate and voting, providing overall support for H4 (Table 6).
Table 6
Conditional Indirect Effect of Disrespectful (Steinbrück) vs. Respectful (Merkel) Candidate Behaviora on Voting Decision Dependent on Attitude Towards Disrespect in Study 2
| Attitude towards disrespect | Med: Communion
|
Med: Agency
|
||||
|---|---|---|---|---|---|---|
| Effect | Boot SE | 95% BCI | Effect | Boot SE | 95% BCI | |
| Negative (-1 SD) | 0.58 | 0.16 | [0.31, 0.94] | 0.02 | 0.06 | [-0.06, 0.18] |
| Positive (+1 SD) | -0.16 | 0.12 | [-0.42, 0.06] | -0.14 | 0.10 | [-0.40, -0.01] |
Note. Number of bootstrap samples = 5.000. BCI = Bias corrected limits of the confidence interval for the indirect effect. Both mediators were included simultaneously into the analysis. a1 = Steinbrück, 2 = Merkel.
Discussion
As in Study 1, the candidate and attitude towards disrespect interacted in predicting participants’ communion and agency judgments of the candidates, and, through this, their actual voting decision. The main difference in this study stemmed from a stronger accentuation: For participants with a positive attitude towards disrespect in politics, their communion judgments did not at all mediate the relationship between candidate and voting decision, while for participants with a negative attitude towards disrespect in politics, their agency judgments did not at all mediate that relationship.
In contrast to Study 1, Study 2 showed that a more positive attitude towards disrespect aligned with generally higher agency ratings, while the actual voting decision and communion ratings were not affected by this disposition. Seemingly, there is no systematic and stable influence of attitude towards disrespect per se on social judgments and decision-making.
General Discussion
In the presented studies, we found that voters’ attitude towards disrespect in politics affected their candidate evaluations, producing different patterns for the social judgment dimensions of communion and agency. Specifically, the link between disrespect and agency was stronger for voters with a positive attitude towards disrespect, while the (negative) link between disrespect and communion was stronger for voters with a negative attitude towards disrespect. This means that under certain conditions, disrespect indeed confers a positive sense of strength and assertiveness, but it can also lead to backlash on the interpersonal dimension. Through these social judgments, disrespect also affected voting intention and actual voting decision.
The results are in line with findings from a previous set of studies in which we showed that the importance that voters attach to being moral themselves interferes with the judgments they draw from perceived disrespect (Mölders et al., 2017). With the present study, we shifted our focus to voters’ explicit attitudes towards disrespect in politics in order to better understand the specific norms that guide political judgments and decision making, replicating research by Fridkin and Kenney (2011) and extending it to the context of political debates in general, as well as voting decisions as an outcome. The findings corroborate the conclusion that individual factors interfere with the judgments drawn from disrespect and should thus be taken into account when investigating the effects of disrespectful candidate behavior.
Our results support the importance of including both social judgment dimensions when investigating the effects of negative campaigning. However, the strength of such effects and whether they occur at all depends on voters’ underlying attitude towards disrespect in the political arena. In that context, voters may apply a different set of standards to politicians than what they use to guide everyday life. At this point, with regard to the underlying psychological processes, we speculate that voters contrast observed disrespectful behavior against their individual standards. In this sense, the answer to our question of whether a politician could possibly win elections by being disrespectful cannot go beyond a measured “It depends” – on the voter’s attitude, on the one hand, and the focal judgment dimension, on the other.
From a social psychological perspective, the concept of attitude towards disrespect can be linked to Cialdini and colleagues’ (1991) distinction of injunctive vs. descriptive norms. Voters with a positive attitude towards disrespect in politics possibly think of disrespect as meeting descriptive norms (i.e., norms that describe appropriate and effective behavior in a given group) for politicians. In contrast, voters with a negative attitude towards disrespect in politics possibly interpret disrespect as violating injunctive norms (i.e., norms on how one generally ought to behave) of being considerate and taking others seriously independent of the situation. To solidify this link, future research could collect justifications for/against disrespect in politics and content-analyze those responses in the light of injunctive vs. descriptive norms. Additionally, it may be useful to explore the relation between ethical predispositions and attitude towards disrespect: a positive attitude towards disrespect should be related to the position of utilitarianism, which focuses on utility and goal-orientation. In contrast, a deontological position focuses on the means themselves, i.e., whether they are morally sound, and could thus be related to a negative attitude towards disrespect. Exploring that link would also contribute to an understanding of whether voters’ attitude towards disrespect in politics is specific to that context or generalizes to others. We would hypothesize that the explicit power-relatedness and competitiveness of politics create a unique setting that shifts norms downward. To test this, future experimental research might, for example, introduce a political context vs. non-political context condition to our design. Furthermore, there is high practical value in understanding whether and how attitude towards disrespect in politics can be manipulated. We would speculate, for example, that a competitiveness prime could activate a more positive attitude towards disrespect, while a moral or cooperation prime could engender a more negative attitude towards disrespect. Likewise, expert statements on the effectiveness vs. non-effectiveness of disrespect in politics could change voters’ attitude. All this could help to unravel the nature of people’s attitude towards disrespect, as well as how to deal with it.
In our line of argumentation, we suppose that the negative effects of disrespectful attacks are due to an observed violation of social norms. However, there are other approaches to explain backlash effects: For instance, research has shown that communicators often acquire the valence they ascribe to others (Transference of Attitudes Recursively (TAR); Gawronski & Walther, 2008). Additionally, on an implicit level, communicators can simply become associated with the traits they describe in others (Spontaneous Trait Transference (STT); Skowronski, Carlston, Mae, & Crawford, 1998). For future research, it might be interesting to test our moderation effect using these research paradigms in order to challenge our norm-oriented rationale. For example, priming a positive attitude towards disrespect (e.g., using a word completion task) could also affect implicit evaluations of a communicator (STT).
Another interesting extension of our research might consider the role that political cynicism (see Thorson, Ognianova, Coyle, & Denton, 2000) plays in the relationships we uncovered. Political cynicism is defined as a pessimistic and definitive belief about the integrity of political actors (Dancey, 2012). We would speculate that voters with a negative attitude towards disrespect in politics who often observe disrespect will more likely develop political cynicism, while for voters with a positive attitude towards disrespect, there should be no such effect. Additionally, it would be interesting to test how voters who are already high in political cynicism react to disrespectful messages. On the one hand, they might support extremely disrespectful candidates (as Donald Trump) because they annoy the political establishment. On the other hand, cynical individuals have been shown to interpret ambiguous information about politicians in a rather negative way (Dancey, 2012). As a third possibility, voters high in political cynicism may not differentiate between respectful and disrespectful candidates due to their generally negative attitude towards politicians.
Lastly, future research should examine how a voter’s attitude towards disrespect relates to consequences faced by the attacked politicians, which we did not consider in our studies. Research has shown that negative attacks potentially damage the voters’ attitude towards the target (e.g., Homer & Batra, 1994). Possibly, voters who disapprove of disrespect will devaluate the attacked target less than people with a positive attitude towards disrespect. However, we can only speculate here.
Limitations
The two studies both have their unique weaknesses, though we strived to mitigate these through complementary designs. In Study 1, we presented fictitious candidates that were minimally described outside of their (dis)respectful behavior in order to eliminate any potentially disturbing information or participants’ previous experience. While our results lacked significant generalizability, the experimental design allowed us to draw causal conclusions. In Study 2, the correlational design prevented such conclusions, even though we collected data on voting decision at a different time. For example, voters who favored Steinbrück may have adapted their ratings of him in order to justify a decision they were already prepared to make (e.g., Bem, 1967; Festinger, 1957). Likewise, SPD voters might have, prior to participating, developed a more positive attitude towards disrespect in order to be able to support Steinbrück as a candidate. However, we accounted for such effects by using party identification as a control variable. In sum, even though Study 2 is embedded in a real political context with all its potential confounds, we still replicated the findings from Study 1, thus providing support of their external validity.
In Study 2, we simplified the political context by focusing only on the top candidates of the two main parties, even though a considerable number of other (though smaller) parties were part of the race, too. This may evoke the impression that we converted a multi-party system to a bi-partisan system. Importantly, we did not force our participants to choose between the two candidates we presented, but asked them for their actual voting decision so that other parties were considered as well. Additionally, we did not ask participants to rate specific disrespectful incidences between Steinbrück and Merkel, but the candidates’ general respect towards any political opponents. Thus, even though we only presented these two candidates, we maintained the actual variety of choice. Therefore, we are confident that our results provide a valid insight into the multi-party German political system. As our findings are similar to those provided by Fridkin and Kenney (2011) for a bi-partisan system, we are optimistic that they generalize to other political contexts.
Granted, there is a known difference in the use of negative campaigns and messages between multi-party and bi-partisan systems. For the former, there is not only a pragmatic need to maintain coalition options, but also a recognition that voters do not only choose the “least evil” of two options (e.g., Walter & van der Brug, 2013). Rather, they can choose a third, fourth or fifth option. Given that the basic social psychological mechanisms of backlash and adherence to norms function independently from the surrounding party system, potential differences could occur in terms of the baseline attitude towards disrespect. Specifically, we suppose that the attitude towards disrespect might be more positive in political cultures with more negative campaigning due to habituation. It might thus be an interesting extension of our research to compare attitude towards disrespect in different (political) cultures.
Finally, another weakness of our study is that we do not know how much information participants used to derive their judgments (i.e., information intensity). For future studies, this might be a very interesting point to include. Possibly, one striking observation related to respect or disrespect could potentially be enough to influence judgments. Furthermore, we would suppose that perceived respect vs. disrespect plays a greater role when voters have less information on other dimensions such as issue competence. However, we believe that perceived (dis)respect will ultimately produce an effect on most voters, however informed, as this factor provides insight on the candidate.
Implications for Practice
Successful political communication is demanding. Different voters have different standards and different needs, which makes it difficult to choose the right strategy in the right situation. Disrespectful behavior is rated as being strong and assertive by some, which is probably the reason why politicians still engage in disrespectful attacks. And yet, others see such actions as unfair violations of social norms. At the same time, being “too” respectful can promote a sense of weakness.
Given the variance and complexity of people’s reactions to disrespect, our advice towards politicians is threefold. First, know the target audience. Unless there are good reasons to assume that they embrace disrespect in politics (e.g., are politically active people), you should not engage in disrespectful attacks. Even when you know that your audience approves of disrespectful behavior, be aware that your behavior is no longer limited to a certain time and space, but can quickly spread to a wider audience through social media platforms. That makes it even more important to consider the different effects your behavior may have on different kinds of audiences.
Second, when you treat your political opponents with respect, consciously provide other cues of your agency (e.g., be energetic, lively, decisive, talk about your successes). On the other hand, if you choose a disrespectful strategy, try to provide cues of your communion in what you say (e.g., understanding for third parties, politeness towards the facilitator, etc.).
Third, if your opponent behaves disrespectfully, try to appeal to your audience’s attitude towards disrespect by questioning the appropriateness of his or her behavior; in other words, try to invoke their injunctive norms of conduct.
In the German race for chancellor of 2013, Peer Steinbrück likely considered disrespectful behavior to be an effective way of standing out and getting attention. However, it did not help him win the election. It is possible that the negative effects on the communion dimension ultimately outweighed the positive effects on the agency dimension. In contrast, Donald Trump, whose campaign was dominated by disrespectful behavior and statements (much more than Steinbrück would ever have dared and wanted), was elected President of the United States. Seemingly, some factors altered voters’ acceptance of disrespect – such as, for example, dissatisfaction with the political establishment or the controversial opponent Hillary Clinton. We are confident that future research will shed further light on Trump’s unexpected success.