Laughing your vote off: the impact of candidates’ humor on voters’ emotions and intentions

Both the Theory of Affective Intelligence and the Expectancy Violation Theory suggest emotions are crucial in supporting political candidates. However, despite the considerable existing evidence on this topic, the impact of humor-induced emotions has been less explored. In this study, we examine whether positive and negative emotions mediate the effects of politicians’ use of affiliative humor on their likeability and citizen’s intention to vote for them. To do this, we designed an experiment in the context of the Chilean presidential elections of 2021, where two candidates representing extreme opposing ideologies competed for the office of president: Gabriel Boric, from a coalition including the Communist Party, and José Antonio Kast from the extreme-right Republican Party. Both candidates used different campaign techniques, including humor that did not always have political content. The experiment was run prior to the second round of the election with a sample of Chilean voters two months before the election ( n = 1,033) and had four conditions: two experimental groups watched a video of a candidate using nonpolitical humor, and two control groups did not. Results indicate that humor boosted both candidates’ likeability and vote probability differently: for Boric, it increased positive emotions; for Kast, it reduced negative emotions. Perceptions of a candidate’s humor predicted higher likeability and declared probability of voting for him, independent of political stance. We discuss these outcomes through ITA and EVT frameworks and their significance for political communication.


Introduction
Leading up to elections, voters form impressions of politicians based not just on their policies, but also on the personal characteristics of candidates (e.g., Peacock et al., 2021).Everything from their general bodily mannerisms (Kramer et al., 2010) to whether they wear glasses (Fleischmann et al., 2019) to the pets they own (Robinson, 2021) can play a role in how people view them.This "political personalization" (Mitchell, 2000), or the process by which individual political figures become defined not solely by the parties and policies they represent but also as an individual person, contributes to a politician's likeability, and in turn plays a role in citizens' voting decisions beyond politician's stances and qualifications.This is especially true for candidate-centered electoral systems (Van Aelst et al., 2012;Balmas et al., 2014).
One such personal characteristic is their humor.Politicians have long used humor in various forms throughout their campaigns during debates and on their social media (Stewart, 2012), and many employ comedians to write humor into their speeches (Gardner, 1986).This is no coincidence; research shows we generally like funny people (Martin and Ford, 2018) and likability contributes to voting intentions (Meng and Davidson, 2020).In this way, the present study examines how seeing politicians use humor affects their likability and subsequent declared probability of voting for him, as well as potential mediators (positive and negative emotions) and moderators (political ideology, specific candidates) of this relationship, using an experimental design based on the 2021 Chilean political cycle.
Chuckling your way to the ballot box: humor, emotions, and politics Political personalization elicits emotional responses (Van Santen and Van Zoonen, 2010;McLaughlin and Macafee, 2019;Metz et al., 2020), which influence how members of the public think of them independently of other cognitive processes and identity considerations (Abelson et al., 1982;Brader et al., 2011;Otto, 2018).This has given emotions a central role in political communication research.Everything a politician says or does can elicit an emotional response, be it positive or negative, and those emotions serve as information in decision-making (Mutz, 2009) and attitude formation (Brader et al., 2011;Gabriel, 2020).
The Affective Intelligence Theory (AIT; Marcus et al., 2000) provides a framework for the study of emotions in political communication.According to the AIT, two "systems" of emotions contribute to the trust we place in politicians and our evaluation of political information (Marcus et al., 2011).The disposition system encompasses positive emotions (e.g., enthusiasm).This system reinforces existing attitudes, strengthens loyalty to one's political ingroup, promotes partisanship, and generally makes people less likely to change their existing perceptions.Meanwhile the surveillance system encompasses negative affect (e.g., anxiety, fear).This system lessens reliance on heuristics/patterns, increases attention to new information, and generally makes people more likely to consider a wide range of aspects when forming new perceptions or evaluating existing perceptions (Brader, 2005;Brader, 2006;Brader et al., 2011;Fridkin and Gershon, 2021).In the realm of politics, emotions like enthusiasm, anxiety, and fear can increase political participation and voting/volunteering behaviors much in the way that anger increases political mobilization or other positive emotions explain individual participation in collective political actions (Marcus et al., 2000;Brader, 2006;Valentino et al., 2011;Zúñiga et al., 2023).
Another way to understand the role of emotions in politics lies in the Expectancy Violations Theory (EVT; Burgoon, 1993;Johnston et al., 2015), which proposes that people evaluate and interpret deviations from expected behaviors based on what would be typical and appropriate.According to Lavine et al. (2012), voters rely on the heuristic of partisanship but can become ambivalent when the candidate from their political position is the cause of negative emotions.In other words, EVT does not only considers emotionslike the AIT -but also expectations, which would lead to different ways of explaining the role of positive and negative emotions.For example, Johnston et al. (2015) found that experiencing positive emotions (enthusiasm) toward a candidate from an opposite party suppressed partisan cue-taking (e.g., by decreasing party voting), and that the opposite happened when experiencing anxiety toward the other party's candidate.In this way, EVT differentiates from AIT in that it considers how people can get involved in more in-depth information processing not only when feeling anxiety toward and in-party candidate, but also when feeling enthusiasm toward an out-party candidate (Redlawsk and Pierce, 2017).
EVT has not only been studied considering emotions and partisanship, but also expectations regarding how people (and specifically, politicians) should behave.In an experiment where participants were exposed to the Twitter profile of a fictitious political candidate, Bullock and Hubner (2020) found that politicians' use of informal communications increased expectancy violation, which in turn lead to reduced credibility.In this case, expectancy violation was assessed with three items that asked participants the extent to which they considered that tweets were appropriate, how surprised they were by them, and how expected they were.
What would be the role of humor in the relation between politics and emotions?Humor as a mode of communication evokes emotional responses.Primarily, humor is known to elicit feelings of mirth, amusement, and exhilaration (each a facet of happiness or joy; Ruch, 1993;Ruch, 1998) as well as reduce feelings of anxiety, fear, and stress.For example, experimental research by Cann et al. (1999) found that viewing stand-up comedy after experiencing stress increases positive emotions and decreases negative emotions.Similar research mirrors these results while further showing the type of humor matters.Cann et al. (2016) found that people exposed to affiliative humor (i.e., humor intended to amuse others and strengthen relationships without ridiculing or otherwise devaluing/harming anyone) experience more positive emotions and fewer negative emotions than those viewing aggressive humor (i.e., humor that ridicules or otherwise devalues/ harms others, intended to establish one's superiority over them).
In the political realm, humor helps candidates with their public image.In addition to humor's ability to draw attention (Brader, 2006;Stewart, 2012), we tend to make positive stereotypical associations funniness with other positive traits or characteristics, like extraversion or intelligence (Decker, 1987;Kuiper and Leite, 2010).Because of people's tendency to generalize positive personality traits, candidates that invoke happiness or other positive emotions at some point, may be perceived more positively themselves (Sülflow and Maurer, 2019).Humor also humanizes politicians to citizens (Young and Lukk, 2017) and helps them connect with their audiences (Bippus, 2007), thus making people evaluate politicians more positively and more likely intend to vote for them (Baumgartner et al., 2015).Humor can also affect our evaluation of politicians as leaders, specifically the perception of certain "soft" characteristics, such as their warmth (Stewart et al., 2024), a quality that may deem more importance than other characteristics such as leadership or integrity when evaluating political leaders (Laustsen and Bor, 2017).
Interestingly, politicians can elicit the emotion of amusement in ways unrelated to politics (Brader and Marcus, 2013); that is, humor by politicians, whether or not it is focused on political issues, can foster positive emotions that affect citizens' evaluations of them.Lastly, humor-related facial expressions can be important.When candidates make humorous comments, they show signs of happiness and affiliative intention on their faces and are better evaluated by their audiences for it (Stewart, 2010), as long as those audiences do not disagree with the verbal sentiments accompanying those smiles  (Sülflow and Maurer, 2019).In a Chilean sample, Segovia (2021) showed that when candidates' faces display positive emotions, people were more likely to vote for them, regardless of political ideology.

The Chilean case and the 2021 presidential election
Political ideology plays a relevant role throughout this phenomenon, both in relation to the candidate and to the voters.In general, there appears to be no difference by ideology in the ways politicians use different kinds of humor (Stewart, 2012;Mendiburo-Seguel et al., 2022), though some evidence suggests conservatives in general (not specifically politicians) use humor for affiliative purposes less than liberals (Kfrerer et al., 2021).However, evidence suggests people do react more positively to political messaging that reaffirms their beliefs (Marquart et al., 2022) and more heavily rely on political heuristics when candidates behave according to what is expected of them as part of a certain party (Sniderman and Stiglitz, 2012).As EVT proposes, feeling unexpected emotions toward candidates, like positive emotions toward a candidate of an opposing ideology, likely leads to less reliance on heuristics and encourages more deliberation, while feeling emotions consistent with expectations would cause the opposite effect (Johnston et al., 2015).
There is little research on the effectiveness of a politician's humor as a function of that politician's ideology.In her experiment, Bippus (2007) observed that when people read different scenarios about either a Republican or a Democrat candidate making a humorous remark during a debate, the Democrat candidate's humor was more effective, regardless of whether the recipient was Democrat or Republican.Furthermore, Silvia et al. (2021) found that right-wing authoritarians (a distinct but somewhat overlapping concept with political conservativism) are generally considered less funny.Through the lens of EVT, these studies seem to suggest that left-wing politicians using humor do not violate expectations to the same extent as right-wing politicians using humor, and therefore receive less penalty to their credibility/support (Brader and Marcus, 2013;Bullock and Hubner, 2020) and perhaps even improved perceptions when that humor is affiliative (Baumgartner et al., 2015), though no research to date appears to have explicitly explored this notion.
Research on the role of political ideology of the recipient/audience in their receptiveness to humor is similarly inconclusive and even somewhat inconsistent.Some research suggests that political orientation is not related to enjoyment of humor (Kfrerer et al., 2021).However, Buie et al. (2022) found that conservatives tend to endorse stronger cavalier humor beliefs, which in turn leads them to have greater appreciation of humor overall.Yet still, other research suggests that, compared to liberals, conservatives have less appreciation of certain types of humor, namely irony and hyperbole, viewing such humor as less funny, smart, enjoyable, and interesting (Young et al., 2019).Considering EVT, this could suggest that both left-wing and right-wing recipients hold similar expectations about the humor use of politicians across party lines.In other words, humor is more expected from the left-wing than the right-wing.

The Chilean case and its 2021 presidential election
The Chilean case is interesting for studying the use of humor by political candidates for two reasons.On the one hand, Chile is situated in a region that is underrepresented in academic publications on this topic.This is significant, especially considering that many relationships between variables of interest for this study, such as conservatism/liberalism or the left/right ideological distinction, are not universal (Malka et al., 2019).Thus, it is plausible to consider that these regional differences also affect the use of humor in political contexts.
On the other hand, Chile is a country that demonstrates notable differences in its political system compared to other Latin American countries and has exhibited, at least since 2019, a trend toward greater ideological polarization compared to the past (Barreda and Ruiz, 2020).Before the social outburst of 2019, Chile's growing ideological polarization has led political discourses to move to extremes.There are different explanations for this, such as the rise of populist discourses, the apathy and subsequent demobilization of the political center (which led politicians to deviate from their usual political platforms to more extreme ones), or the problems with the binomial electoral system (that determined how Congress members were elected), which led to less possibility of agreements between opposing coalitions and the emergence of new ones (Fábrega et al., 2018).This polarization has led voters to hold highly extreme views and broadened the ideological distance between high and low socioeconomic levels, as well as between the generations born before 1955 and after 1980, making it even greater than the distance that existed at the beginning of the 1990s, shortly after the end of the Pinochet dictatorship (Lindh et al., 2019).
Chile's 2021 presidential election clearly displayed this polarization.The country's presidential election implies two possible rounds.If none of the candidates in the first round receive more than 50% of the votes, the two candidates with the most votes contend on a second round.In the first round of the 2021 presidential election, those majorities were José Antonio Kast of the Republican Party (the conservative Christian Social Front Alliance), who obtained a 28% of the votes, and Gabriel Boric of the socialist Social Convergence Party (the liberal Apruebo Dignidad alliance, along with the Communist party), with 26%.According to Benedikter and Zlosilo (2022), this result demonstrated the polarization in the country, with two opposing sides that were almost identical in strength.In the second round, Boric was elected president with 55% of the votes.

The present study
Research in political communication shows emotions, especially positive emotions, play a key role in the likeability and probability of voting for a candidate.But political psychology has largely failed to consider politicians' use of humor.This is a critical omission because humor is known to induce positive emotions like amusement and joy while reducing negative emotions like anxiety and nervousness (e.g., Cann et al., 1999).Thus, the present study attempts to understand the relationships between politicians' use of humor, its emotional effects, and the subsequent impacts on likeability and voting behavior, as well as examine the role of political ideology.In addition, we focus specifically on politicians' use of affiliative humor unrelated to political issues.Due to the increasing use of social networks and candidates' increasing participation in less formal interview formats, we believe investigating such humor may be of particular value and interest.Similarly, the existing evidence focuses on assessments of candidates' funniness, yet it does not address whether the mere use of humor could have effects.
This study uses an experimental design carried out on the 2021 Chilean political cycle, with data collected before the second round of the presidential election.Based on the literature reviewed above, we propose the following hypotheses: H1: Positive and negative emotions will mediate (positively and negatively, respectively) the relationship between politicians' use of affiliative humor and likeability.This will result in: H1.1: Higher likeability and subsequent higher probability of voting for the candidate when humor elicits more positive emotions.
H1.2: Higher likeability and subsequent higher probability of voting for the candidate when the humor decreases negative emotions.
Along with this, participants' political ideology and the candidate they evaluate (Boric or Kast) will moderate these proposed effects.Based both on the EVT and AIT, we propose that: H1.3: Participants in the humor condition will find a sameideology candidate more likable and report higher voting probability due to positive emotions.
H2: Regardless of political ideology, funnier candidates elicit more positive and fewer negative emotions, increasing likeability and voting probability.

Method Participants and design
We recruited 1,033 people from Chile (46.6% female, M age = 41.34,SDage = 13.27)two and a half weeks before the second round of the 2021 presidential election.Participants were contacted through a company that provides online panels.They all had to read and declare their agreement with an informed consent that had been approved by the Ethics Committee of the University of the first author.The vast majority (98.3%) were residents of one of the three most populated region of Chile (Metropolitan, Valparaíso, and Biobío).We randomly assigned participants to one of four between-subjects conditions based on a 2 (Humor: affiliative humor, no humor) × 2 (Candidate: Boric, Kast) design.Table 1 provides the detailed description of the four groups' demographics and other descriptive statistics.As anticipated from random assignment, we observed no differences in demographics between conditions with respect to gender (Fisher's exact test, p = 0.169), age [F(3, 1029) = 0.389, p = 0.761], region (Fisher's exact test, p = 0.283), interest in politics [F(3, 1029) = 1.666, p = 0.173], attitudes toward politicians in general [F(3, 1029) = 0.086, p = 0.968], or political ideology [F(3, 1029) = 0.414, p = 0.743].There were differences regarding funniness between Boric and Kast [t (439) = 6.892, p < 0.001], which could not be anticipated given the nature of the stimuli (there are not many publicly available videos of the candidates using non-aggressive humor).For this reason, we tested our second hypothesis separately for each candidate.

Procedure
We conducted an online experimental study with postmeasurement only and two control groups.Participants received a link to the study through a paid online survey panel provider.After granting informed consent, participants completed the first phase of the study, in which they reported demographics (gender, age, education, geographic area of residence in the country [region]) and responded to items gauging political attitudes and preferences.Specifically, we asked about their interest in politics ("How much interest would you say you have in politics?"; 1 = "no interest, " 10 = "total interest"), their attitudes toward politicians in general ("How much would you say you like politicians?"; 1 = "I do not like them at all, " 10 = "I absolutely like them"), and their political ideology ("In politics, it is normal to speak of left and right.In a scale where 0 is 'Left' and 10 is 'Right, ' where would you position yourself?";0 = "Left, " 10 = "Right").This last question is useful since Chile is a stable emerging presidential democracy with an institutionalized party system in which people can easily associate parties with a left-right axis and locate themselves around it.
After completing the first phase, participants were randomly assigned to one of four conditions.Two of these were control groups, in which they rated one of the candidates (Gabriel Boric or José Antonio Kast) without being exposed to any stimuli, which allowed us to obtain a top-of-mind association of candidates with emotions, mimicking traditional polls that address this topic.The other two were the experimental groups, in which participants first viewed a video of one of the candidates engaging in some kind of affiliative comical effort unrelated to political issues.
The two control groups evaluated the candidate of their condition immediately after responding to the items regarding political attitudes, while the two experimental groups did it after watching the video.Specifically, the four groups had to evaluate the extent to which they felt different emotions toward the candidate, likeability ("How much would you say you like Gabriel Boric / José Antonio Kast?"; 1 = "I do not like him at all, " 7 = "I absolutely like him"), and probability of voting for the candidate (or "vote probability"; "How probable would it be for you to vote for Gabriel Boric / José Antonio Kast in the next presidential elections?"; 1 = "Not probable at all, " 10 = "Very probable").Regarding the assessed emotions, we measured "worry" and "fear" as proxies for negative emotions, and "enthusiasm" and "joy" for positive emotions ("How much [emotion] would you say Gabriel Boric /José Antonio Kast causes you?"; 1 = "Nothing, " 5 = "A lot").The mean of the "worry" and "fear" items comprised our measure of negative emotions, while the mean of the "enthusiasm" and "joy" items comprised our measure of positive emotions.
Finally, in the case of participants in the experimental conditions, after viewing the video they were also asked to rate its funniness ("How funny did you find the video?" with possible responses ranging from 0 = "Not funny" to 10 = "Very funny").

Stimuli
The videos presented to the two experimental groups were each 1 min and 10 s in duration and depicted one of the candidates performing or participating in some kind of affiliative comical intent not related to politics.The length of the videos was mainly due to the difficulty in finding videos of the candidates in actions of this type and in which humor was exclusively affiliative (that is, with no other intention, such as aggression).In the case of Boric, we used an extract from an interview in which the candidate imitated Donald Duck and boxed with the interviewer.In the case of Kast, we used a compilation of clips from three videos from different sources.Two of them were obtained from Twitter: one in which Kast playfully dances to Gangnam Style and a second one in which he plays with a mime (Mimo Tuga).The third one was obtained from his TikTok account, and it presents him humorously fixing himself a drink.Both videos can be found at this link: http://bit.ly/3ERERmb.

Descriptive and bivariate analyses
We first compared the means of the dependent and mediating variables in the different conditions using ANOVAs and Games-Howell post-hoc tests (see Table 2).Within the two control conditions, there were no differences in participants' positive emotions toward either candidate.However, compared to Boric, participants reported more negative emotions and less likeability toward Kast, as well less probability of voting for him.Looking at the effect of humor for each candidate, Kast elicited significantly fewer negative emotions in his humor condition than in his control condition but no difference in positive emotions, while Boric elicited significantly more positive emotions in his humor condition than his control condition but no difference in negative emotions.Both candidates received higher ratings of likeability and voting probability in their humor condition relative to their control condition.
Table 3 depicts bivariate correlations between all continuous variables.Correlations were largely identical within the humor and control conditions.One exception is that the strength of negative emotion's correlations with likeability and voting probability is weaker in the humor conditions (r likeability = −0.25;r Voting Probability = −0.25)than in the control conditions (r likeability = −0.49;r Voting Probability = −0.47).Similarly, the strength of positive emotion's correlations with those same variables is somewhat greater in the humor condition (r likeability = 0.83; r Voting Probability = 0.81) than in the control condition (r likeability = 0.78; r Voting Probability = 0.77).

Emotions as mediators of the effect of humor on likeability and voting probability
We hypothesized that positive emotions (Hypothesis H1.1.)and negative emotions (Hypothesis H1.2.) would mediate the relationship between politicians' use of humor and their likeability, which would in turn yield higher probability of voting for the candidate, and that this mediation would be moderated by participants' political ideology and the candidate being evaluated.To test these hypotheses, we conducted a moderated moderated mediation analysis using a custom model built with the PROCESS macro (Hayes, 2018) for R (R Core Team, 2022).In this model, the condition (Humor = 0.5, Control = −0.5)was the independent variable, positive and negative emotions were secondstage parallel mediators, likeability was the third-stage mediator, probability of voting for the candidate was the dependent variable, and the two moderators were candidate (Boric = 0.5, Kast = −0.5)and political ideology.We used bias-corrected 95% confidence intervals with 10,000 bootstraps, and in order to deal with possible heteroscedasticity, we used the HC4 method.
Hypothesis 1.1 was partially supported.As shown in Table 4, according to the index of moderated moderated mediation (IMMM), there is no evidence that political ideology and candidate  interact together to influence the indirect effect of humor use on probability of voting through positive emotions and likeability.However, the indirect effect is moderated by the candidate alone (Index of moderated mediation = −0.82;95% BCCI [−1.39, −0.25]), which is exclusive for the case of Boric (b = 0.82, 95% BCCI [0.42,1.21]).In other words, regardless of the recipient's political ideology, Boric's use of humor improves the declared probability of voting for him by eliciting more positive emotions, which in turn bolsters his likeability.
In the case of negative emotions, there is evidence of a moderated moderated mediation, supporting Hypothesis 1.2.The index of conditional moderated mediation by political ideology was only significant in the case of Kast (−0.60; 95% BCCI [−0.10, −0.03]) and showed a positive indirect effect on liberals and centrists.That is, Kast's use of humor improves the declared probability of voting for him, specifically in liberal and centrist recipients, by decreasing the negative emotions he arouses, which in turn bolsters his likeability.

Direct and indirect effects of perceived funniness
Hypothesis 1.2 posited that viewing a candidate of the same political ideology as funnier would yield more positive emotions and fewer negative emotions, in turn leading to higher likeability and declared probability of voting for him.Also, viewing the candidate of the opposite ideology as funnier would cause the recipient to reassess that candidate and ultimately lead to the same effect.In this case, due to the differences in funniness evaluation between both candidates, we conducted a moderated mediation model separately for each of them, using funniness as the predictor variable.We again used biascorrected 95% confidence intervals with 10,000 bootstraps and the HC4 method.
Table 5 details all paths in the resulting models for both candidates.Concerning direct paths, funniness had a direct negative effect on probability of voting for Boric in the case of right-wing participants evaluating him.Additionally, there was a significant indirect effect of funniness through positive emotions and likeability in the case of both Boric and Kast, yielding increased declared probability of voting for them regardless of political ideology or candidate.
Looking at the effects (see Table 5), it is clear that positive emotions play the most important role in the relationship between funniness and probability of voting for a candidate.However, the indirect effect through likeability (that is, without considering positive or negative emotions) was also positive and significant in the case of both candidates, although not as strong as in the case of positive emotions.In sum, Hypothesis 2 is also partially supported.Those who considered the candidate funnier felt more positive emotions toward him, which in turn yielded higher likeability and declared probability of voting for him.

Discussion
Our first hypothesis aimed to compare the humor attempt with the top-of-mind images of the candidates, while the second focused on perceived funniness or the success or failure of that attempt.The latter is of particular importance, as failure to be considered funny when attempting to be so can have negative consequences on perceptions of competence, status, and likeability (Bitterly et al., 2017;Ji et al., 2023), and affect a person's public and private image (File and Schnurr, 2019).Given that the standard for politicians using humor may be higher than for other individuals (Bell, 2015), it was crucial to focus on this specific aspect.
The results of the present study partially supported these two hypotheses.Politicians' use of humor improved their likeability and declared probability of voting for them, and emotions mediated this relationship (in fact, emotions were essential in this relationship, as there was little or no evidence of a direct effect of humor on vote probability or of an indirect effect through likeability; see Table 4).This relationship worked differently between our two candidates and, to some extent, depended on the participants' political ideology.Boric's use of humor improved his likeability and vote probability by eliciting positive emotions in all participants, regardless of their political ideology, though the effect was stronger for more liberal participants.Conversely, Kast's use of humor improved his likeability and vote probability by decreasing participants' negative emotions toward him, specifically participants with more centrist or left-wing ideologies.Participants' evaluation of humor also played a key role in the emotions elicited (and subsequent likeability and vote probability): the funnier they deemed the candidate, the more positive emotions they felt toward him, regardless of their own ideology.
Our results also contribute to the EVT in the political arena by highlighting how emotions may not only have a direct role, but also an indirect one in people of contrary political ideologies of that of a given candidate.For centrists and conservatives, the observed direct effects of humor were negative in the case of Boric, but positive through the indirect path of positive emotions-likeability.This may have happened because joking breaks the expectation of a serious politician (Funk, 1997).In this case, expectancy violation would work over the stereotype of a politician, but it would have the opposite effect if that violation causes positive emotions.Similarly, in the case of Kast the indirect effects were also positive through negative emotions and likability in the case of centrists and liberals.It is possible that the contrast between the humorous behavior and the attributes typically associated with a rightwing senior politician might help soften any potential negative emotions among these citizens.On the other hand, and partially contrary to what could be proposed from the AIT, positive emotions only strengthened partisanship in the case of Boric, but not Kast.These results support the importance of expectancy violation in the appraisal of political candidates.
The present study broadens our understanding of the effects of political communication in general and politicians' use of humor more specifically for three reasons.First, it provides relevant evidence considering its high external validity and applicability of its findings (because of the use of real situations, candidates, and stimuli), and the use of a Chilean sample, which is especially valuable given that much of political humor's empirical research takes places in the United States context (Bippus, 2007;Stewart, 2012;Baumgartner et al., 2015).The fact that our findings generally align with expectations from previous (U.S.-based) research may suggest the observed phenomenon explored may not be culture-specific, or at least specific to cultures that overlap with both Chile and the U.S., though future research may attempt direct comparisons across cultures.
Second, this study also examined the effect of politicians' use of affiliative, unrelated to politics humor.Previous research has examined the effects of politicians' use of a variety of types of humor, such as self-deprecating humor (e.g., Baumgartner et al., 2015), which shares the affiliative intent but disparages the politician delivering the humor, risking damage to perceptions of their credibility (e.g., Hackman, 1988).Given that both self-deprecating humor and affiliative humor share similar goals (i.e., to build rapport/relationships; Cann et al., 2016), future research might compare politicians' use of these two forms of humor more directly to understand whether/how they differ in key outcomes such as eliciting favorable emotions and perceptions around credibility.
Third, and perhaps most critically, we established positive and negative emotions as key mediators in the relationship between politicians' humor and their likeability and vote probability.Positive emotions were the strongest of the two mediators, and both candidates could elicit those positive emotions if participants found them funny.This adds on to previous research on the benefits of inducing positive emotions through humor, such as drawing more attention (Fredrickson and Branigan, 2001;Fredrickson and Branigan, 2005) and activating brain regions associated with better retention of political information and more openness to opposing political views (Coronel et al., 2021).

Limitations and future directions
The results of our study should be understood considering some limitations.First, the effect sizes we observed are relatively small, especially when compared to the direct effects of political ideology or candidates' political affiliation.While this highlights the central role that affiliation and political ideology play in electoral decisions, the  fact that -regardless of the candidate-participants with more centrist political ideologies do appear susceptible to the use of humor, is very interesting.In highly polarized elections like the one we describe in our study, candidates must take every advantage they can to win the votes of those in the moderate/centrist ideologies if they wish to win elections, and our results suggest that the use of affiliative humor offers one such advantage.Second, there are limitations regarding the type of stimuli we used.Looking at mean funniness for each clip, it is possible that range restriction contributed to our results.Both clips were rated by participants as below-average, which means we are unable to detect potential rebound effects at higher levels of funniness.For example, if a candidate is viewed as extremely funny, it is possible that they will subsequently be considered "less serious, " which could elicit more negative emotions toward them for taking the election process too lightly.
Related to this, the clips we showed participants were viewed differently between candidates.The humor clip for Boric was viewed as funnier than the humor clip for Kast, so it is possible that if Kast's humor was perceived to be as funny as Boric's, then it would have elicited more positive emotions as well.However, considering that Boric finally won the elections by a 10% margin, it is possible that general attitudes toward him were in fact more positive, which could explain why his video was considered funnier.Our use of real stimuli from these candidates helps situate these findings in the real-world context, but future research might use more controlled stimuli to better understand the role of funniness in raising positive emotion and decreasing negative emotion in political audiences.
Our use of real candidates running for election bolsters the external validity of this study but comes with a crucial trade-off, which is that the candidates differ in more ways than just their party affiliations.They carried very different backgrounds with them, which could easily affect the evaluations made in each condition.For example, there is a significant age gap at the time of this study between Kast (53 years old) and Boric (33 years old).The candidates also have different presentational "styles" that may influence the effects of their humor.Boric has a very casual style, which includes openly displaying his tattoos, refusing to wear ties, or sporting a "mohawk" hairstyle during his time in Parliament, while Kast maintains a more traditional/ formal way of dressing and presenting himself.Because of this, participants may have had expectations of Boric being more casual and funny, and of Kast being more formal in his political communications.The humor clips used in this study helped meet the expectations of Boric's casual style but violated expectations of Kast's formal style.However, assessments of participants' prior attitudes showed that there were no differences between groups regarding relevant variables (such as age, gender, political ideology or attitude toward politicians).Future research should examine the moderating role of candidate qualities on humor's relationship with positive and negative emotions and subsequent likeability and voting intentions.
Despite these limitations, the present study provides further evidence that politicians' use of humor matters.Politicians themselves already seem to understand the value of humor, it is no coincidence that U.S. presidents have hired speech writers with comedy backgrounds to prepare more humorous speeches (Gardner, 1986).Previous research has demonstrated benefits of humor that would appeal to politicians, such as distracting their audience from the political content of their speeches (Becker and Waisanen, 2016) or deflect questions from reporters (Carpenter et al., 2019).Our results add to this body of research by showing that humor can generate emotions in citizens that correspond to improved likeability and ultimately the extra votes needed to get into office.
Finally, we hope to contribute to a broader understanding of America as a continent with a rich cultural diversity and various political frameworks, which challenges us to study these topics from an inclusive perspective.

TABLE 1
Descriptive statistics for each condition.

TABLE 2
ANOVAs and post-hoc tests (Games-Howell) for dependent and mediating variables between conditions.

TABLE 3
Correlation matrix for all continuous variables, grouped by condition and collapsed across candidate.

TABLE 4
Regression coefficients (standard errors), tests of conditional interaction of moderators and conditional effects of moderators for the use of humor (Hypothesis 1).

TABLE 5
Regression coefficients (standard errors), test of conditional interaction of moderators, and conditional effects of moderators for the funniness model (Hypothesis 2).