It's not the what, it's the who: the importance of the political leader in getting the message across

Ismael Crespo Martínez

University of Murcia. Spain. 

icrespo@um.es                  

 

Alberto Mora Rodríguez

University of Murcia. Spain. 

alberto.mora@um.es

 

José Miguel Rojo Martínez

University of Murcia. Spain. 

josemiguel.rojo@um.es

How to cite this article / Standard reference:

Crespo Martínez, Ismael; Mora Rodríguez, Alberto & Rojo Martínez, José Miguel (2025). It's not the what, it's the who: the importance of the political leader in getting the message across. Revista Latina de Comunicación Social, 83, 1-23. https://www.doi.org/10.4185/RLCS-2025-2354

Date Received: 25/10/2024
Acceptance Date: 21/12/2024
Date of Publication: 20/02/2025

 ABSTRACT 

Introduction: In a context of increasing personalization of politics, the influence of leaders on communicative processes is increasingly prominent. However, we do not have much evidence comparing the effects of the senders and of the topics of the message on the reactions of audiences. Therefore, the present research seeks to test whether the levels of support or rejection of a message are influenced by who its sender is. Methodology: A survey experiment was designed with a total of 719 participants from the panel of the Centro de Estudios Murciano de Opinión Pública (CEMOP, for its acronym in Spanish). Data collection began on July 9, 2023 and ended on July 21, 2023, in the middle of the electoral campaign for the general elections held in Spain. The participants were randomly divided into four groups and had to show their levels of support or rejection to a message presented in the form of a text-vignette, changing the sender depending on the group. Results: The findings point to the existence of a singular leadership, that of Santiago Abascal, which has the capacity to influence the responses of individuals even when they do not emit positional messages. Abascal's presence as a sender influences the responses, regardless of the issue at hand or the attitudes and characteristics of the participants. A "negative personalism" effect is observed, according to which certain individuals only need to know who is defending something to reject a message. Discussion: Political leaders can introduce significant biases that condition citizens' responses to the messages issued, regardless of the specific thematic content of the messages. This allows us to reflect on the relationships between leadership and polarization, as well as to attend to the conditioning factors of the success of communication processes. Conclusions: The results of this research have important implications for understanding the positions of voters in the face of the messages.

Keywords: political leadership; message; political communication; Spain; elections; survey experiment. 

1. INTRODUCTION  

Every communication process involves a message articulated around a specific topic or content, a sender who is responsible for transmitting the message to some receivers who learn about it through different media or channels and react to it. The communicative act, as defined by Lasswell (1948, p. 37), answers the questions "who says what, through what channel, to whom, and with what effect". Based on this general framework, classical studies of political communication have examined each of the constituent elements of the communicative act through different theories.

Studies on the influence of leaders have asked, in relation to the expansion of mass media, to what extent the person responsible for transmitting the message can be relevant to the success of the communicative act (Crespo et al., 2011) and have even spoken of the leader as a message (Ortega Ruiz & Castillo, 2009). On the other hand, there are many studies dedicated to the characterization of the message, considering the nature of the topics referred to, the discursive and persuasive techniques used for its elaboration, the inclusion of these messages in the media and political agendas, or the supports and channels used for its dissemination (Norris et al., 2003). All elements of the communication process influence in some way its reception and the response that individuals develop.

The current concern about the rise of political polarization around the world is accompanied by a reflection on the specific contribution of postmodern communicative dynamics to this climate. Reflections on the role of social networks and algorithmic biases, new forms of access to information, the heightened tone of elite speech, populist rhetoric demonizing rival groups, negative campaign strategies, or the alteration of the partisan supply are common (Sood & Iyengar, 2016; Gervais, 2019; Sarrión Andaluz & Rodríguez-Gordo, 2021; Törnberg et al, 2021; Renström et al., 2023; Romero-Rodríguez et al., 2023; Rodríguez-Virgili et al., 2022; Rojo-Martínez et al., 2023; Torcal, 2023; Martin & Nai, 2024). The impact of social networks on communication processes deserves special attention, as these platforms have become one of the main access points to political messages. According to Pacchioni-Hurtado & Antezana-Corrieri (2023), social networks promote polarization because they facilitate the spread of fake news, create echo chambers that isolate users and reinforce their prejudices, and at the same time prioritize sensitive content through recommendation algorithms. However, we do not have much evidence that simultaneously compares the effects of senders and issues on public reactions. This question is relevant because it allows us to understand to what extent it is the elites who become a factor of polarization of the masses, or in other words, the power of people to arouse passions (positive and negative) regardless of their ideas or proposals. This will be the main focus of our research.

In this article, we will focus on one of the specific elements of the communicative process and we will show the results of a survey experiment developed during the campaign of the general elections of July 2023 in Spain. In order to determine the capacity of political leaders to generate rejection processes towards the messages they issue, we will observe how the reactions provoked by the same topic can differ depending on the sender who enunciates it. Even when a leader formulates a seemingly uncontroversial message with which everyone could agree, can he generate remarkable rejection processes that reveal the biases of the audience and explain a climate of strong polarization towards the elites? We assume that rejection of a message simply because of who is expressing it is a clear sign of polarization, revealing cognitive biases stemming from intergroup conflict and the political identities of the recipients.

As mentioned above, the experiment whose data are presented below took place during the Spanish electoral campaign for the July 23, 2023, elections, a campaign characterized by a climate of strong competition. After the first coalition government in recent democratic history between the Spanish Socialist Workers' Party (PSOE, for its acronym in Spanish) and Unidas Podemos, the municipal and regional elections of May 2023 brought with them an apparent change of cycle and other coalition pacts, that of the Popular Party (PP) with Vox. In this scenario, a dynamic of "us vs them" emerged among the ideological blocs. The right wing focused its entire strategy on the demonization of the figure of Sánchez under the slogan "repeal Sanchismo" (Gil Orantos, 2023). The president appeared as the epitome of the extensive and controversial list of legislative changes promoted by the progressive majority, which the PP-Vox coalition would reverse if it reached Moncloa. The left, for its part, fueled fears of the consequences of this repeal in various areas of social and identity significance, such as women's rights and the fight against gender violence, the rights of the LGTBI collective, freedom of expression in the cultural field, or economic aid to the most vulnerable. The need to stop the "reactionary wave" in order not to roll back rights was the great campaign framework of the PSOE and its allies (Merino, 2023), who also repeatedly highlighted the good unemployment figures after the labor reform promoted by the leader of Sumar, Yolanda Díaz.

1.1. Theoretical approaches

When we ask about the origins of political polarization, it is common to find studies that point to the elites of political parties as responsible for this situation (Druckman et al., 2013). In an era of increasing personalization of politics, understood as "the prominence that political leaders have acquired in relation to their parties" (Rebolledo, 2017, p. 48), leaders increasingly influence the attitudes of voters, even in parliamentary systems (Van Biezen & Hopkin, 2007; Bittner, 2021). In particular, the discursive polarization of elites has been shown to spill over to citizens, leading to greater affective polarization (Skytte, 2021). There is also evidence that we are increasingly polarized in relation to leaders. According to Miller (2023, p. 37), in the last decade in Spain the difference between the feelings evoked by ingroup leaders and those evoked by outgroup leaders has increased. The high personalization of politics makes leaders both the source and the object of polarization. More precisely, the concept of the personalization of politics implies the assumption that these leaders become priority senders of messages, referents for themselves of the political offer and protagonists of the competition, all trying to respond to the media logic imposed first by television and now by social networks (Garzia, 2014; Parmelee et al., 2023).

Candidate-centered politics, to follow Wattenberg's (2011) terminology, makes the individual characteristics of leaders decisive (Caprara, 2007; McGregor, 2018). This, in turn, encourages more emotional speeches and positions that depend more on the affection a person arouses in us than on the ideas or proposals he or she defends (Rudi, 2014; Wahl-Jorgensen, 2019). It is not strange to think that personalization rewards the emotional aspect over the rational one, since by giving a face to ideas, the tendencies of affinity or incompatibility present in any relationship between people appear (Lovatt, 2020). Thus, the news increasingly focuses on leaders and less on the organizations to which they belong, which gives the political arena a less substantial character and makes our position on certain issues mediated by sympathy or rejection of the leader who defends them (Van Aelst et al., 2012). Recent research shows that when a news item is starred by an impersonal source, such as a political party, it generates less affective polarization than when the information appears to be associated with a specific leader (Thijssen et al., 2024). The biases activated by leaders are related to the identity function they play: leaders construct identities, foster feelings of belonging, and act as prototypes for members of a group (Hogg et al., 2012). This identity function also means that within the dynamics of intergroup conflict, rival leaders become the primary object of attack or disqualification, as they are the main representatives of the outgroup (Fuentes-Rodríguez, 2023).

In parallel, other studies emphasize the influence of issues on polarization. The concept of issue salience, previously used to explain the impact of certain issues on the evaluation of presidential performance or on the behavior of legislators (Epstein & Segal, 2000), is now revived to show how voters who attach more importance to cultural issues are more polarized (Han, 2022). So-called culture war issues related to identity and morality seem to elicit more polarized responses than other issues related to the economy (Mouw & Sobel, 2001; Gidron et al., 2020). Accepting that issues matter is not new. Studies of political behavior have repeatedly assessed the influence of issues and their symbolic ownership, in terms of credibility for their management, in explaining voting (Budge, 1982; Bélanger & Meguid, 2008). Similarly, the concept of "niche parties" has been proposed to describe those formations that focus their political action on a dominant issue (Garrido et al., 2022). At the same time, the analysis of public opinion has always tended to distinguish between transversal or valence issues, such as economic growth, which are capable of generating consensus —and which are becoming more relevant as the ideological polarization of the electorate decreases and partisan polarization increases (Green, 2007)— and positional or distributive issues of a marked ideological nature, which are more controversial, such as abortion or taxes (Costa Lobo, 2013). 

However, there are also authors who believe that the basis of contemporary political polarization is not thematic, but emotional, in such a way that voters agree on a good part of the issues but maintain positions of hostility towards groups with which they do not share identity for spurious reasons (Mason, 2013; Mason, 2015). Thus, there would be no concrete policies justifying affective hostility, but a psychological process of rejection and distancing from rival groups, generated by the consolidation of partisanship as a form of social identity (Iyengar et al., 2012). The thematic disagreement would be a posteriori attempt to rationalize the affective rejection of a group nature. Thus, it would be evident that the conflict between these partisan social groups is mediated by prejudices and cognitive biases that create an unreal sense of competition (Torcal, 2023). According to this vision, the issues would be mere excuses that shape or justify the tribal attitudes that exist in the background (Reiljan & Ryan, 2021).

The question of the influence of issues on the polarization of the mass’s stems from the growing tendency of voters to take positions in all political debates that are consistent with those defended by their membership group, thus reducing independent thinking about them (Garner & Palmer, 2011). In addition, one's own opinion on an issue is often shaped by the influence of the group once it has expressed its position through its leaders (Cohen, 2003). Even preferences can change depending on who the individual defending an action believes: the same idea presented by one's own leader will receive support, but if that action is presented as the proposal of another group, it will be rejected, using identity-based motivational reasoning (Mullinix, 2016). As issues become more easily categorized within partisan competition, positions on policies become identity signals that reproduce intergroup competition. It doesn't matter what is being talked about, but who is saying it.

2. OBJECTIVES AND METHODOLOGY 

Through the present research, we seek to answer two fundamental questions: a) Does the sender influence the levels of support or rejection in front of a message, beyond the nature of the stated issue? b) Is this the case for all types of senders? These questions take on particular importance in the current context of debate about the rise of hyper-leadership and the increasing articulation of politics around the personality of leaders. The aim is to determine whether it is possible for the sender to introduce a bias such that the positions expressed by the receivers depend more on his person than on what he says, something that would also come to contradict the hypotheses about the influence of issues on polarization. With these questions we want to know to what extent the effect of a basic element of the communicative process, such as the sender, can influence voters' responses. These questions are formalized in the following research hypothesis:

Research Hypothesis: In the field of political communication, the levels of support or rejection of a message are influenced by who the sender is.

This hypothesis implicitly combines theories of leader influence and issue influence. To test it, a survey experiment was designed in which a total of 719 people from the panel of the Centro de Estudios Murciano de Opinión Pública (CEMOP) participated. All of them were citizens of the Region of Murcia over the age of 18. This is not a representative sample, since its purpose is not statistical inference, but causal inference from an experimental design. The use of the experimental method in studies of persuasion, propaganda and media has been recurrent, among other things, to analyze the predisposing factors —for example, the credibility of the communicator— that condition the effects of communication strategies and public attitudes (Hovland et al., 2017).

Data collection through a self-administered CAWI system began on July 9 and ended on July 21, in the middle of the campaign. The messages treated as reactive in the experiment correspond to real issues discussed during the campaign. Interviewees were randomly divided into four groups to make causal inferences based on the stimuli presented in each of these subsamples. All participants in the experiment had an equal chance of being in one of the groups, and therefore the stimulus they were exposed to depended only on a random question. 183 people participated in G1, 160 in G2, 177 in G3, and 199 in G4. The four randomly formed groups do not show significant differences between them, neither according to the ideology of their members, nor according to their self-reported vote in the November 2019 elections, nor according to their level of education, three variables that could decisively influence the results observed between the groups, thus disrupting the influence of the stimulus (see ANNEX 1). Since the groups do not differ in their composition with respect to the previously mentioned variables, the results of the dependent variable must be understood as caused by the introduced stimuli. These stimuli consisted of short textual vignettes in the middle of a questionnaire with routine questions about current political events and other items intended for the socio-demographic classification of the interviewee.

The participants of group 1 read two texts about the opinions of the leader of Vox and the leader of the PSOE on different issues, one of a transversal nature (the defense of full employment) and another of a positional or ideological nature, corresponding to a prominent area of the culture war (the laws of transgender persons), which during the legislative term led to a notable confrontation between the two ideological blocs and even within the parties of each bloc. The position of each leader on the issue used is the same as the one he/she defended. Then they were asked about their degree of agreement with the message they had just read:

The first stimulus said (P10): Vox candidate Santiago Abascal has stated that he wants to "guarantee full employment" in the country. How much do you agree with this proposal? Consider a scale from 0 to 10, where 0 means "strongly disagree" and 10 means "strongly agree". Santiago Abascal was presented as a defender of a message with little initial capacity to polarize.

The second stimulus said (P11): The PSOE candidate, Pedro Sánchez, has stated that he wants to "protect the rights of the transgender collective". How much do you agree with this proposal? Consider a scale from 0 to 10, where 0 means " strongly disagree" and 10 means "strongly agree". It showed that Pedro Sánchez was dealing with an issue that had been an element of strong controversy during the months leading up to the campaign.

The participants of group 2 read two texts with messages from the leader of the PP, Alberto Núñez Feijóo, and from the leader of Sumar, Yolanda Díaz. The messages had exactly the same wording as those seen by the participants of group 1, only the sender was changed: 

P10: The PP candidate, Alberto Núñez Feijóo, stated that he wants to "guarantee full employment" in the country.

P11: The Sumar candidate, Yolanda Díaz, stated that she wants to "protect the laws related to the transgender collective”.

The participants of group 3 read two texts about the opinions of the leader of Vox, Santiago Abascal, and the leader of Sumar, Yolanda Díaz. As can be seen, all possible combinations are made between the main national political leaders. In addition, each leader is issuing a message on both a positional and a transversal issue:

P10: The candidate of Vox, Santiago Abascal, has declared that he wants to "change the laws regarding the transgender collective".

P11: The candidate of Sumar, Yolanda Díaz, has stated that she wants to "guarantee full employment" in the country.

Finally, the participants of group 4 read two texts about the opinions of the leader of the PP, Alberto Núñez Feijóo, and the leader of the PSOE, Pedro Sánchez: 

P10: The PP candidate, Alberto Núñez Feijóo, has stated that he wants to "change the laws regarding the transgender collective". 

P11: The PSOE candidate, Pedro Sánchez, has stated that he wants to "guarantee full employment" in the country.

The summary of the research design is shown in Table 1. It ensures that each group was exposed to the two types of messages and to two senders of different ideology. In total, the entire sample saw the two messages used in conjunction with the four senders, creating all possible leader pairings. The survey methodology, the dataset with all the records, as well as the questionnaires used can be accessed at the following link: https://www.cemopmurcia.es/experimento-mediante-encuesta-cawi/ 

Table 1. Summary of research design.

Leader	Messages	Group	Group and message Identification
Pedro Sánchez	Positional	G1	X1
Pedro Sánchez	Transversal	G4	X2
Alberto Núñez Feijóo	Positional	G4	X3
Alberto Núñez Feijóo	Transversal	G2	X4
Santiago Abascal	Positional	G3	X5
Santiago Abascal	Transversal	G1	X6
Yolanda Díaz	Positional	G2	X7
Yolanda Díaz	Transversal	G3	X8

Source: Own elaboration.

The empirical analysis strategy followed combines, on the one hand, bivariate analysis tests in which we first observe intergroup differences in the level of support/rejection of the same message depending on who is the sender of the message. This allows us to estimate what is known in experimental methodology as the "average effect of the experimental treatment on the treated" (ATT) (Druckman et al., 2011).  However, assuming that in reality the support or rejection of the message may be conditioned by the previous political attitudes of the participants and that this deserves a more robust and conclusive statistical approach, we develop two linear regression models (OLS) that allow us to control not only for the effect of each explanatory variable on the outcome, but also for the interaction that may potentially occur between the independent variables. We consider the following set of variables:

Table 2. Variables used.

Type of variable	Name, characteristics values.
Dependent Variable (DV)	0-10 scale of support/rejection to a message/specific measure (P10 and P11 of the questionnaire). The respondent is presented with a message containing a measure proposed by a sender X (with the wording described above). The respondent is then asked:  How much do you agree with this proposal? Consider a scale from 0 to 10, where 0 means "strongly disagree" and 10 means "strongly agree".
Explanatory Variable (emitting stimulus)	It is formulated as a dummy variable that identifies the type of stimulus. To explain the results of P10, it takes the value 0 if the person read a text with Alberto Núñez Feijóo as a stimulus and the value 1 if the person read a text with Santiago Abascal as a stimulus.  To explain the results of P11, it takes the value 0 if the person read a text with Yolanda Díaz as a stimulus and the value 1 if the person read a text with Pedro Sánchez as a stimulus.
Control variable 1	This is formulated as a dummy variable indicating the type of message (positional or transversal). To explain the results of P10 and P11, it takes the value 0 if the person read a text with a transversal message as a stimulus and the value 1 if the person read a text with a positional message.
Control variable 2	Feelings towards the four leaders: Pedro Sánchez, Alberto Núñez Feijóo, Santiago Abascal and Yolanda Díaz. Scale 0-10, where 0 means "dislike and rejection" and 10 "sympathy and adhesion".
Control variable 3	Ideological self-positioning scale 1-10, where the value 1 is "extreme left" and the value 10 is "extreme right".
Control variable 4	Level of studies (dummy). It takes the value 0 if the interviewee has not studied and the value 1 if he/she does.
Control variable 5	Gender (dummy). It takes the value 0 if the interviewee is a man and the value 1 if the interviewee is a woman.
Control variable 6	Age. Uncoded metric variable.

Source: Own elaboration.

3. RESULTS

The analysis of the "average effect of the experimental treatment on the treated" (E[Y_ij1 − Y_ij0|T_k = 1]) applied to our research objective requires to check whether those who read a message attributed to leader X (Y_ij1) showed different levels of support/rejection compared to those who read the same message but attributed to leader Z (Y_ij0). To do this, it is necessary to compare the support/rejection averages in the following messages and for the following leaders: 

1. Guarantee full employment: P10-G1 (Santiago Abascal) vs. P10-G2 (Alberto Núñez Feijóo) and P11-G3 (Yolanda Díaz) vs. P11-G4 (Pedro Sanchez). We always compare leaders from the same ideological bloc to avoid basic biases of ideological predisposition that interfere with the effect of the stimulus.

Table 3. Difference in average on the support/rejection scale: message on full employment (right wing leaders).

Experimental group	Leader	Message	Average	DT	N
1	Santiago Abascal	Full employment	5,58	3,82	183
2	Alberto Núñez Feijóo	Full employment	7,38	3,21	160

Source: Own elaboration.

The T-test indicates that the average difference in the scale of support/rejection of the "full employment" message between Santiago Abascal and Alberto Núñez Feijóo is statistically significant (t(341) = 4.681, p = .000). Considering that both groups are made up of similar persons and that the theme used maintained the same wording, we initially sensed an influence of the sender on the positioning of the citizens, even when the message is constructed on such an uncontroversial issue. The differences in the standard deviations also show the greater polarization generated around Santiago Abascal. If we look more closely at the data obtained with another approach (percentages in sections of the scale), we can observe more precisely what is happening. The percentage of persons who are in positions of strong rejection (0-2) on the scale when it is Abascal who proposes full employment reaches a high 30.1%. This percentage is only 11.9% when it is Feijóo who proposes the same. These results show a certain emotional disposition of the electorate, since it is not the content of the communication (the proposals) that is judged, but who defends it. The issues can serve as excuses to justify ex-post attitudes of identity origin, questioning the full validity of the hypotheses of issue salience. At the same time, the importance of the sender for the effectiveness of the communicative process is verified, from the moment that his/her own presence can condition the reception of the message, creating an initial barrier as a preventive animosity.

Table 4. Difference in average on the support/rejection scale: message on full employment (left-wing leaders).

Experimental group	Leader	Message	Average	DT	N
3	Yolanda Díaz	Full employment	6,76	3,48	177
4	Pedro Sánchez	Full employment	6,99	3,48	199

Source: Own elaboration.

In the case of leftist leaders, no significant differences in averages are observed depending on who sends out the message (t(374)= -.662, p= .509). Looking again at the responses according to a percentage distribution of frequencies, 18.6% of the participants express a strong rejection of the full employment proposal when it is sent out by Yolanda Díaz, and the same percentage is obtained in Group 4 when the one sending out the message is Pedro Sánchez. In the absence of the "sender effect", the rejection rates could be due to cognitive shortcuts activated by the ideology shared by both senders.

The tests conducted so far suggest the existence of a singular leadership that could influence the responses of persons even when it does not issue positional messages on which a high degree of consensus would be expected. Although this should be verified through multivariate analysis models, it could be argued that this evidence may not occur in other messages in the same sense. Next, we examine the differences that occur in the second of the messages used, the positions of each ideological bloc on the laws related to the transgender collective.

2. To protect or change the Trans Laws: P10-G3 (Santiago Abascal) vs P10-G4 (Alberto Núñez Feijóo) and P11-G1 (Pedro Sánchez) vs P11-G2 (Yolanda Díaz).

Table 5. Average difference on the support/rejection scale: message about laws targeting transgender people. (right-wing leaders)

Experimental group	Leader	Message	Average	DT	N
3	Santiago Abascal	Trans law	3,19	3,87	177
4	Alberto Núñez Feijóo	Trans law	4,20	3,89	199

Source: Own elaboration.

Table 6. Average difference on the support/rejection scale: message about laws targeting transgender people. (left-wing leaders)

Experimental group	Leader	Message	Average	DT	N
1	Pedro Sánchez	Trans law	6,39	3,67	183
2	Yolanda Díaz	Trans law	5,70	3,72	160

Source: Own elaboration.

The differences between the two senders on the right are again statistically, but in this case only at the 95% confidence level (t(374) = 2.504, p = .013). This suggests that although the sender bias persists in messages with positional content, its differential effect is smaller (the bias introduced by message content is stronger). This conclusion is reached intuitively by comparing the difference of averages in Table 5 (1.01) and that in Table 3 (1.80). And in the case of the left-wing senders, are there again no differences between them? The same trend is confirmed (t(341)= -1.735, p= .084). Whether the sender is Pedro Sánchez or Yolanda Díaz does not produce significantly different effects.

In summary, we can conclude that the same pattern emerges when using two types of unrelated topics. On the right, who says things matters, on the left it doesn't. Even assuming positions of cultural struggle that can be installed in a framework of social desirability (we assume that the social-moral norm is more on the side of protecting a minority than of modifying a law that affects it), when Alberto Núñez Feijóo makes a proposal, it causes less rejection than when Abascal makes the same proposal, although it is formulated in the same terms. The defense of certain already sensitive positions becomes even less effective when the standard-bearer is the leader of Vox. These first impressions need more robust statistical verification through a linear regression model. Despite the fact that our groups do not differ significantly in their composition, attending to the aggregate level of results in an experiment that captures individual attitudes. In the manifestations of support/rejection of a person, not only the stimulus may intervene, but also the person's own ideas or characteristics, an element outside the experiment that needs to be controlled.

The equation of the linear regression model we propose is as follows: 

Yi (scores on the support/rejection scale) = β0 + β1 (Stimulus-sender) + β2 (Stimulus-type of message) + β3 (ideological self-placement) + β4 (Feelings towards Pedro Sánchez) + β5 (Feelings towards Alberto Núñez Feijóo) + β6 (Feelings towards Santiago Abascal) + β7 (Feelings towards Yolanda Díaz) + β8 (Educational level) + β9 (Sex) + β10 (Age) + β11 (Sender × Type of message) + ε

The results of the regression models indicate that our research hypothesis is fulfilled in the case of right-wing senders, which confirms what has already been established in the bivariate analysis. The presence of Santiago Abascal as a sender increases the levels of rejection of the message, controlling this effect by a wide range of variables that could distort it (the individual's ideology, their initial feelings towards the leaders, the nature of the topic on which the messages were about or sociodemographic characteristics such as gender, age and education level). Likewise, we observed that the interaction between leader and topic is not significant in the models applied to the senders on the right (Table 7), which reinforces the fact that the sender effect acts with independence and explanatory capacity by itself.

Table 7. Full model coefficients - Support/rejection scale (right-wing senders).

Predictor		Estimator		EE		t		p
Constant		5.18458		0.69296		7.4818		< .001
Sender		-0.90560		0.33999		-2.6636		0.008
Type of message		-2.58247		0.32260		-8.0051		< .001
Ideological self-placement		0.17861		0.06840		2.6113		0.009
Feelings towards Pedro Sánchez		-0.19833		0.04555		-4.3536		< .001
Feelings towards Yolanda Díaz		-0.02709		0.04412		-0.6140		0.539
Feelings towards Alberto Núñez Feijóo		0.25242		0.04130		6.1120		< .001
Feelings towards Santiago Abascal		0.29770		0.04350		6.8445		< .001
Level of studies		0.10191		0.25535		0.3991		0.690
Gender		-0.40194		0.23583		-1.7044		0.089
Age		-0.00377		0.00733		-0.5144		0.607
Sender × Type of message		0.03209		0.46119		0.0696		0.945

Source: Own elaboration.

Note: R²=0.474. Collinearity control with the VIF test (for no variable >5).

Communication strategies should not overlook the influence that senders can exert, among other elements, to persuade or dissuade audiences. With certain audiences, there are those who add and value while others detract from it. The case of Santiago Abascal is paradigmatic when it comes to understanding these processes and can be explained in large part by the tendency toward personalistic leadership of populist parties (Akkerman et al., 2014; Mudde & Kaltwasser, 2017). The centrality of his figure in Vox's communication strategies makes him the perfect prototype of the political proposal he represents, as if he himself embodies it. Thus, for those who reject that group, Abascal's presence is a sufficient signal to refuse to support the message, no matter how much individuals may agree with it. We have agreed to call this evidence "negative personalism": it is enough for me to know who is defending something to be against it, a kind of magnified ad hominem bias. We will have an opportunity to explore the implications of this concept, including psychological theories of social identity, in the discussion section.

On the other hand, the regression results in Table 7 show that the type of message (positional or transversal content) influences support or rejection. When the topic is positional, rejection increases, which is not surprising. What is relevant and deserves further analysis is that this happens for right-wing leaders, but not for left-wing leaders (see Table 8). This could be due, first, on the one hand, to the different ways in which they approach the issue. In the case of the Right, the positional issue of the culture war was raised from a framework of regression in acquired rights, which can generate defensive emotional reactions. The moral implications of this issue provoke important changes in interviewees' predispositions toward the messages, demonstrating the potential success of progressive framing. As we know, the leader of Vox arouses high levels of strong rejection even when he addresses a non-positional issue (30.1%, M=5.58). However, the levels of strong rejection with Abascal's message increases (+29.2 points) when his persona is combined with a positional issue. If we observe the changes that occur with respect to Santiago Abascal for each issue according to the self-reported vote of the participants in the experiment (Figure 1), it is found that this sender begins to arouse rejection in a relevant percentage of PP (25%) and significantly increases the rejection of socialist voters when it deals with a positional issue. When it comes to full employment, the majority of PSOE voters did not reject the messages of the leader of the radical right (38.8%), which does happen when a message critical of the laws on the transgender collective is issued (79.5%). The leader of Vox alienates voters from mainstream parties when he engages in the culture war. Of particular interest is the effect of the binomial positional Abascal-issue on PP voters, since the treatment of cultural issues unsettles a segment of PP supporters —possibly those with more socially liberal views— and agitates the cohesion of the bloc.

Figure 1: Changes in the opinions aroused by Abascal's messages depending on the message and according to the interviewees' self-reported vote.

Source: Own elaboration.

Similarly, Alberto Núñez Feijóo is the leader who experiences the greatest changes depending on the type of topic on which he focuses his messages (M=7.38 when he talks about full employment to M=4.20 when he talks about the laws related to the transgender collective). When the popular leader talks about this cultural issue, the levels of strong rejection (0-2) with his message increase by 34.3 points. The leadership of Núñez Feijóo is particularly sensitive to the influence of these issues. Meanwhile, he is the leader who performs best in terms of support when he focuses on a non-cultural issue with little conflict. From this result it can be concluded that the dominant strategy for the current president of the Popular Party is a communication far from issues of high moral charge, since there are no major tendencies of "negative personalism" around him. Looking at this question according to the self-reported vote of the participants, the data inform us of three prominent effects in the case of Núñez Feijóo: when talking about cultural issues, the rejection of the socialist voters quadruples (from 17.9% to 78.8%), the levels of agreement among his own voters decreases (from 94.4% to 73%), but the support among the supporters of Vox increases (from 53.3% to 70.4%). The strategy of choosing culturally charged issues for Núñez Feijóo brings him closer to the radical right voter, but distances him from the socialist voter (it hinders the conversion of the moderate socialist vote to articulate a majority from the center) and from a part of his electorate that could coincide with a profile of "soft vote" sensitive to certain positions or behaviors.

Figure 2: Changes in the opinions aroused by Feijóo's messages depending on the message and according to the interviewees' self-reported vote.

Source: Own elaboration.

Overall, two conclusions of high strategic value can be observed in the reaction to the messages of the right. The figure of Abascal, no matter what he says, can provoke the rejection of a large part of the population, a clear example of the mechanisms of intergroup emotional reactivity warned of by Torcal (2023, p. 41). In turn, the figure of Núñez Feijóo, for whom this does not happen, sees support for his messages weaken in all sectors of the voters, except in Vox, when he addresses positional issues with polarizing potential. In the case of Núñez Feijóo, the success of his communication depends much more on the right choice of contents. For Santiago Abascal, this is less relevant, since the emotional disconnection with a part of the voter may be so high that a change in the thematic agenda may have no effect (there is no incentive to abandon the culture war).  

Finally, the results presented in Table 8, which reflect the explanatory model for the support/rejection of the messages issued by leftist leaders, indicate that in this case the issues are not significant (the culture war does not increase the rejection because it is carried out from positions linked to a general moral sense) and, more clearly, neither do the senders. The positions towards the messages of Pedro Sánchez and Yolanda Díaz are strongly influenced by the ideology of the individual and the previous attitudes of the individual towards different leaders. This could be an indication that the margin of success of political communication to be developed by the left-wing parties will be much less suited to strategic moves linked to the construction of personalist leaderships or to the choice of certain "star" themes or messages. Dependence on the ideology and the deep feeling of the individual creates a scenario much less sensitive to conjunctural changes or temporary choices on the specific content and carrier of the message.

Table 8. Full Model Coefficients - Support/rejection scale (left-wing senders).

Predictor		Estimator		EE		t		p
Constant		5.80226		0.60236		9.633		< .001
Sender		0.44600		0.26672		1.672		0.095
Type of message		-0.55397		0.28437		-1.948		0.052
Ideological self-location		-0.27993		0.05824		-4.807		< .001
Feelings towards Pedro Sánchez		0.45975		0.03879		11.854		< .001
Feelings towards Yolanda Díaz		0.02349		0.03756		0.625		0.532
Feelings towards Alberto Núñez Feijóo		0.05019		0.03516		1.427		0.154
Feelings towards Santiago Abascal		-0.18544		0.03703		-5.007		< .001
Level of Studies		0.34756		0.21741		1.599		0.110
Gender		0.17554		0.20079		0.874		0.382
Age		0.00741		0.00624		1.186		0.236
Sender × Type of message		-0.30041		0.39457		-0.761		0.447

Source: Own elaboration.

Note: R2=0.521. Collinearity control with the VIF test (for no variable >5).

4. DISCUSSION: THE IMPORTANCE OF THE WHO

The results of our research have implications on three areas of study. On the one hand, they allow us to reflect on the role that elites play in polarization processes and the role that leaders assume in intergroup conflict relations (Lachat, 2015). Then, it is also necessary to reflect on the factors that condition the construction of messages and the lessons that strategists of electoral campaigns can draw from them. Finally, and considering the scope and context in which the research was carried out, we can generate some hypotheses about the communicative and affective factors that influenced the outcome of the last general elections in Spain.

Regarding the first question, the leadership of Santiago Abascal fulfills an identity-referential function for his group that is not comparable to that of the other senders. Therefore, his person captures the intergroup biases of rejection towards Vox when there is a high overlap of identity between sender and group, activating cognitive springs of hostility in the receivers towards the characteristics and ideas of this group with the mere presence of its maximum representative. In other words, Abascal activates many implications without the need for them to appear explicitly, proving that he himself acts as a message and that his figure is intimately linked to the process of categorizing Vox's identity. The intertwining of personalism and populism makes this type of leader a conditioning factor in the electoral success of their parties (Michel et al., 2020), to the extent that adherence to the party is mediated by adherence to the leadership, but also a source of singular rejection. This is not so common in the more established mainstream parties, such as the PP and PSOE, where the identity category represented by the general brand of the party seems to overcome the contingency of the leader, even if there are occasional personalist drifts in them. As mentioned above, we propose the concept of "negative personalism" to explain the bias introduced by certain leaders, which can function as a cognitive shortcut to facilitate the elaboration of a position contrary to a measure by the simple observation of who defends it. We have tended to focus discussions of the effects of leadership and the personalization of politics on the catalytic role of the sender (to the point of becoming a voting factor in itself) in an era of partisan dealignment and the dissolution of strong feelings of belonging (Michel et al., 2020).

In the same way, we again provide evidence of what has already been warned by authors such as Harteveld et al. (2022): The radical right has an extraordinary capacity to provoke rejection. Its appearance in the public sphere, with positions that challenge the status quo and question important social consensus, provokes a singularity that contributes to more emotional responses on the part of left-wing voters (who do not act in this way with the traditional right). We observe this assumption because the leader of the radical right is censored even in the most innocuous communicative scenarios, and this cannot be understood without assuming how this name excites the recipient when read. Audiences cannot dissociate a message, however positive it may be, from its sender's prior history, i.e., messages are always interpreted in light of prior events, meanings, and experiences outside the immediate moment. Certainly, when the participants read "Santiago Abascal," they began to think about their previous opinions about this person or about things that this person defended or represented, which clouded their judgment about the second part of the sentence: Abascal wanted to defend something reasonably good, full employment. At the same time, certain social norms or desirability biases may have acted, making it difficult to support that message by whoever was defending it.

The predisposition to support Abascal's message may be mediated not by a prior opinion about that leader or affection for what he represents, but by a reflection on the social valuation of that fact. Supporting a stigmatized actor may be perceived as socially undesirable, and the individual may therefore prefer not to be disruptive and do what is expected of him. Social norms interfere with communication processes and reduce support for certain groups. Finally, it is undeniable that the experimental treatment X₆ (Abascal issues a non-positional message) was a clear test of cognitive dissonance.

Many participants had their schemas changed when they were exposed to a leader, they disliked defending something good, which could potentially weaken their negative beliefs about that person and create contradictions. The resolution of the dissonance situation, to overcome the discomfort involved, could take place through two motivated reasoning strategies (Festinger, 1957). The first consists in "completing": when the message was read, suspicions about the hidden intentions or the meaning of this simple text were quickly aroused by paying attention to the sender, completing the available information with elements outside the experiment but present in the individual's memory and learning. For example, the message was good, but the means to achieve it could be assumed to be bad. The second strategy is simply to change the actual object of the positioning. Although the question was directed to the individual's position on the policy proposal ("What is your degree of agreement with this proposal?"), the attitude may have been expressed toward the sender, which was much easier to resolve in a congruent way, reducing the tension of the contradiction. This leads to further reflection on the difficult interrelationship that arises when designing a question in which a sender, a message and a subject coincide, even in non-experimental surveys, since doubt remains as to what the reaction is really being produced towards.   

Returning to the debate on the implications for political communication, if we put the scenarios described above in a more abstract way and try to apply them to the communicative situations typical of campaigns, we should accept that leaders cannot be separated from the image that has long been installed about them in the minds of voters, nor from the social norms that operate towards their group. A naive turn to soften the image of leaders with a strong negative charge may not be a completely profitable strategy, because the reservations will remain in an important part of the public (whatever they say) and along the way an inconsistency will have been created that damages their positioning.

5. CONCLUSIONS

We began this research by asking about the contribution of one of the main elements of the communication process (the sender) to the positioning of citizens in the face of different messages. Focusing our interest on the effects of the sender makes sense in a political moment such as the present, when the emphasis is on the rise of personalization and hyperleadership. Our research hypothesis proposed to evaluate the causal impact of the sender on the level of support or rejection of a message, controlling for different factors such as the topic treated or the ideology of the individual. To verify this, a survey experiment was conducted in the context of the July 2023 election campaign in Spain, in which a total of 719 people participated, randomly divided into four groups that were shown different textual stimuli.

The bivariate analysis showed us how, faced with the same issue, different means of support were produced among the leaders of the right. And this happened both in a message of transversal content and in one of positional nature linked to a culture war. On the left, however, the sender did not cause any changes. Since these results could be influenced by multiple variables that introduce distortions at the individual level of response, two linear regression models were formulated for the senders of each block. These models confirmed to us that Abascal was a stimulus with the capacity to generate rejection processes regardless of the topic it said, that is, polarization scenarios are not always explained by the relevance of the issues. Our hypothesis was confirmed for the sender who corresponded, precisely, to a more personalistic leadership and with a higher level of influence on the development, public presence and stereotypical construction of his group, which is consistent. Under the concept of "negative personalism", we attempted to explain how leadership not only has a positive influence on reinforcing group has membership but can also become a powerful signal for rejecting certain proposals or motivating defensive reactions in outgroups. Thus, we have achieved our research goal. In addition, it allows us to complete the hypotheses about the salience of the issues by considering another important component of the communicative process. At the same time, it was found that the nature of the issue was also relevant on the right, which led us to conclude, through various descriptive analyses, that the culture war (at least using the issue of trans laws as a stimulus) does not seem to penalize the left in terms of reactive polarization. On the other hand, it is clear that if the Popular Party wishes to grow towards the center and weaken the PSOE's support bases, the influence of the issues is a key factor that could jeopardize this objective.

Our research is not without limitations. The capacity for causal inference allowed by the experimental design comes at the cost of reducing the external validity of its results and of assuming a certainly artificial environment that, in this case, introduces some specific issues and not others. In this sense, it would be advisable to replicate the experiment designed with a representative sample of the Spanish population and confirm whether the same conclusions are obtained. Likewise, it would be useful to check whether the same results are obtained when other types of issues and messages are selected, maintaining the distinction between their positional or transversal nature. It could also be asked whether the image of the parties influences the reactions generated by the leader, i.e., whether when people react against or in favor of the leader's message, it is because of their rejection or support for him or her, or because of his or her status as a representative of a group (party). The comparison of public reactions according to the type of sender (group vs. individual) is a line of research to be developed in the future. All in all, the contribution made makes it possible to open up new paths, which are synthesized in the hypotheses generated from the empirical evidence presented. In the two months that elapsed between the early call for elections and the election day, certain leaderships, issues and messages helped to generate affective reactions that resolved the battle between "abolishing sanchismo" and "stopping the reactionary wave" in favor of the second framework. We contribute reflections to a still very early debate on the campaign experienced in Spain in July 2023, as well as to the literature on polarization, countering the attention shown to the thematic relevance and proposing to delve into the effect of leaders.

6. REFERENCES

Akkerman, A., Mudde, C. y Zaslove, A. (2014). How populist are the people? Measuring populist attitudes in voters. Comparative political studies, 47(9), 1324-1353. https://doi.org/10.1177/0010414013512600  

Bélanger, É. y Meguid, B. M. (2008). Issue salience, issue ownership, and issue-based vote choice, Electoral studies, 27(3), 477-491. https://doi.org/10.1016/j.electstud.2008.01.001 

Bittner, A. (2021). The personalization of politics in Anglo-American democracies, Frontiers in Political Science, 3(70). https://doi.org/10.3389/fpos.2021.660607 

Budge, I. (1982). Electoral volatility: Issue effects and basic change in 23 post-war democracies. Electoral Studies, 1(2), 147-168. https://doi.org/10.1016/0261-3794(82)90001-4

Caprara, G. V. (2007). The personalization of modern politics. European Review, 15(2), 151-164. https://doi.org/10.1017/S1062798707000178

Cohen, G. L. (2003). Party over policy: The dominating impact of group influence on political beliefs, Journal of personality and social psychology, 85(5), 808-822. https://doi.org/10.1037/0022-3514.85.5.808 

Costa Lobo, M. (2013). Dimensions of the economic vote: Valence, positional and patrimony voting in Portugal's 2011 elections, Electoral Studies, 32(3), 460-464. https://doi.org/10.1016/j.electstud.2013.05.021 

Crespo, I., Garrido, A., Carletta I. y Riorda, M. (2011). Manual de Comunicación Política y estrategias de campaña. Candidatos, medios y electores en una nueva era. Biblos.

Druckman, J. N., Peterson, E. y Slothuus, R. (2013). How elite partisan polarization affects public opinion formation, American political science review, 107(1), 57-79. https://doi.org/10.1017/s0003055412000500 

Druckman, J. N., Green, D. P., Kuklinski J. N. y Lupia, A. (2011). Experiments: An Introduction to Core Concepts. En J. N. Druckman, D. P. Green, J. H. Kuklinski y A. Lupia (2011). Cambridge Handbook of Experimental Political Science (pp. 15-27). Cambridge University Press. 

Epstein, L. y Segal, J. A. (2000). Measuring issue salience, American Journal of Political Science, 44(1), 66-83. https://doi.org/10.2307/2669293 

Fuentes-Rodríguez, C. (2023). Adversarial Attacks in Spanish Politics: Direct and Indirect Forms. En O. Felmand (Ed.), Political Debasement: Incivility, Contempt, and Humiliation in Parliamentary and Public Discourse (pp. 31-51). Springer Nature Singapore.  

Garner, A. y Palmer, H. (2011). Polarization and issue consistency over time. Political Behavior, 33, 225-246. https://doi.org/10.1007/s11109-010-9136-7

Garrido Rubia, A., Martínez Rodríguez, A. y Mora-Rodríguez, A. (2022). Partidos spin-off, partidos nicho y derecha radical: El caso de Vox en España, Revista Internacional de Sociología, 80(2), e207-e207. https://doi.org/10.3989/ris.2022.80.2.21.10 

Garzia, D. (2014). Personalization of politics and electoral change. Springer.

Gervais, B. T. (2019). Rousing the partisan combatant: Elite incivility, anger, and antideliberative attitudes. Political Psychology, 40(3), 637-655. https://doi.org/10.1111/pops.12532

Gidron, N., Adams, J. y Horne, W. (2020). American affective polarization in comparative perspective. Cambridge University Press. https://doi.org/10.1017/9781108914123

Gil Orantos, B. (2023, 15 junio). Derogar el «sanchismo», según PP y Vox: un objetivo, distintas vías. EFE Noticias. https://efe.com/espana/2023-06-16/derogar-sanchismo-pp-vox-objetivo/

Green, J. (2007). When voters and parties agree: Valence issues and party competition, Political Studies, 55(3), 629-655. https://doi.org/10.1111/j.1467-9248.2007.00671.x 

Han, K. J. (2023). Issue salience and affective polarization, Journal of Elections, Public Opinion and Parties, 1-19. https://doi.org/10.1080/17457289.2023.2277429 

Harteveld, E., Mendoza, P. y Rooduijn, M. (2022). Affective polarization and the populist radical right: creating the hating? Government and Opposition, 57(4), 703-727. https://doi.org/10.1017/gov.2021.31 

Hogg, M. A., van Knippenberg, D. y Rast III, D. E. (2012). The social identity theory of leadership: Theoretical origins, research findings, and conceptual developments, European Review of Social Psychology, 23(1), 258-304. https://doi.org/10.1080/10463283.2012.741134 

Hovland, C. I., Lumsdaine, A. A. y Sheffield, F.D. (2017). Experiments on Mass Communication. Princenton University Press. 

Iyengar, S., Sood, G. y Lelkes, Y. (2012). Affect, not ideology: A social identity perspective on polarization. Public opinion quarterly, 76(3), 405-431. https://doi.org/10.1093/poq/nfs038  

Lachat, R. (2015). Leader effects and party polarization. En M. Costa Lobo y J. Curtice (Eds.), Personality Politics? The Role of Leader Evaluations in Democratic Elections (pp. 105-126). Oxford University Press. 

Lasswell, H.D. (1948). The Structure and Function of Communication in Society. En L. Brynson (1948). The communication of ideas (pp.37-53). Harper & Brothers. 

Lovatt, M. D. (2020). Personalised Storytelling in British Political Rhetoric: Performance, Emotions, and Argument [Tesis de doctorado]. University of Leeds.

Martin, D. y Nai, A. (2024). Deepening the rift: Negative campaigning fosters affective polarization in multiparty elections, Electoral Studies, 87, e102745. https://doi.org/10.1016/j.electstud.2024.102745 

Mason, L. (2013). The rise of uncivil agreement: Issue versus behavioral polarization in the American electorate. American Behavioral Scientist, 57(1), 140-159. https://doi.org/10.1177/0002764212463363

Mason, L. (2015). “I disrespectfully agree”: The differential effects of partisan sorting on social and issue polarization, American journal of political science, 59(1), 128-145. https://doi.org/10.1111/ajps.12089 

McGregor, S. C. (2018). Personalization, social media, and voting: Effects of candidate self-personalization on vote intention, New media & society, 20(3), 1139-1160. https://doi.org/10.1177/1461444816686103 

Merino, J. C. (2023, 30 mayo). El Gobierno justifica el adelanto electoral del 23-J para “combatir la ola reaccionaria”. La Vanguardia. https://www.lavanguardia.com/politica/20230530/9006932/gobierno-justifica-adelanto-electoral-combatir-ola-reaccionaria

Michel, E., Garzia, D., Ferreira da Silva, F. y De Angelis, A. (2020). Leader effects and voting for the populist radical right in Western Europe, Swiss Political Science Review, 26(3), 273-295. https://doi.org/10.1111/spsr.12404

Miller, L. (2023). Polarizados. La política que nos divide. Deusto. 

Mouw, T. y Sobel, M. E. (2001). Culture wars and opinion polarization: the case of abortion, American Journal of Sociology, 106(4), 913-943. https://doi.org/10.1086/320294 

Mudde, C. y Kaltwasser, C. R. (2017). Populism: A very short introduction. Oxford University Press. https://doi.org/10.1093/actrade/9780190234874.001.0001

Mullinix, K. J. (2016). Partisanship and preference formation: Competing motivations, elite polarization, and issue importance, Political Behavior, 38, 383-411. https://doi.org/10.1007/s11109-015-9318-4 

Norris, P., Curtice, J., Sanders, D., Scammel, M. y Semetko, H. A. (2003). On message. Communicating the campaign. SAGE.

Ortega Ruiz, M. y Luque Castillo, F. J. (2009). Obama es el mensaje: Una interpretación macluhaniana del nuevo liderazgo presidencial estadounidense, Psicología Política, 39, 19-38.

Pacchioni-Hurtado, V. y Antezana-Corrieri, M. E. (2023). Miradas comunicacionales de la polarización política en entornos digitales, una revisión bibliográfica. Cuadernos del centro de estudios de diseño y comunicación, 194, 35-52.  https://doi.org/10.18682/cdc.vi194.9620

Parmelee, J. H., Perkins, S. C. y Beasley, B. (2023). Personalization of politicians on Instagram: what Generation Z wants to see in political posts, Information, Communication & Society, 26(9), 1773-1788. https://doi.org/10.1080/1369118x.2022.2027500

Rebolledo, M. (2017). La personalización de la política: una propuesta de definición para su estudio sistemático. Revista de Comunicación, 16(2), 147-176. https://revistadecomunicacion.com/article/view/991 

Reiljan, A. y Ryan, A. (2021). Ideological tripolarization, partisan tribalism and institutional trust: The foundations of affective polarization in the Swedish multiparty system, Scandinavian Political Studies, 44(2), 195-219. https://doi.org/10.1111/1467-9477.12194 

Renström, E. A., Bäck, H. y Carroll, R. (2023). Threats, emotions, and affective polarization. Political Psychology, 44(6), 1337-1366. https://doi.org/10.1111/pops.12899 

Rodríguez-Virgili, J., Portilla-Manjón, I. y Sierra-Iso, A. (2022). Cuarenta años de polarización ideológica en España. Revista Empresa Y Humanismo, 25(2), 75-103. https://doi.org/10.15581/015.XXV.2.75-103

Rojo-Martínez, J. M., Crespo-Martínez, I. y Mora-Rodríguez, A. (2023). Intergroup Emotional Dynamics: An Analysis of the Characteristics of Affectively Polarised Spanish Voters, Revista Española de Investigaciones Sociológicas (REIS), 184(184), 205-242.

Romero-Rodríguez, L. M., Castillo-Abdul, B. y Cuesta-Valiño, P. (2023). The process of the transfer of hate speech to demonization and social polarization, Politics and Governance, 11(2), 109-113. https://doi.org/10.17645/pag.v11i2.6663 

Rudi, T. (2014). Emotions towards leaders and voting behaviour. En M. Costa Lobo y J. Curtice (Eds.), Personality Politics? The Role of Leader Evaluations in Democratic Elections (pp. 215-240). Oxford University Press.

Sarrión Andaluz, J. y Rodríguez-Gordo, C. (2021). Polarización social y control político: algunas consecuencias de la inteligencia artificial y las redes sociales para la razón práctica. Cuadernos Salmantinos De filosofía, 48, 157-185. https://doi.org/10.36576/summa.144497 

Skytte, R. (2021). Dimensions of elite partisan polarization: Disentangling the effects of incivility and issue polarization, British Journal of Political Science, 51(4), 1457-1475. https://doi.org/10.1017/s0007123419000760 

Sood, G. y Iyengar, S. (2016). Coming to dislike your opponents: the polarizing impact of political campaigns, SSRN Electronic Journal. https://doi.org/10.2139/ssrn.2840225 

Thijssen, P., van Dijk, R. y van Erkel, P. (2024). Exploring the effect of personalized voting on affective polarization: Prototypical leadership and campaign effects, Acta Política, 1-27. https://doi.org/10.1057/s41269-023-00319-1 

Torcal, M. (2023). De votantes a hooligans. La polarización política en España. Catarata.    

Törnberg, P., Andersson, C., Lindgren, K. y Banisch, S. (2021). Modeling the emergence of affective polarization in the social media society, PLoS One, 16(10), e0258259. https://doi.org/10.1371/journal.pone.0258259 

Van Aelst, P., Sheafer, T. y Stanyer, J. (2012). The personalization of mediated political communication: A review of concepts, operationalizations and key findings, Journalism, 13(2), 203-220. https://doi.org/10.1177/1464884911427802 

Van Biezen, I. y Hopkin, J. (2007). The presidentialization of Spanish Democracy: Sources of Prime Ministerial Power in Post-Franco Spain. En T. Poguntke y P. Webb. (Eds.), The presidentialization of politics: A comparative study of modern democracies (pp. 107-127). Oxford University Press.

Wahl-Jorgensen, K. (2019). Emotions, media and politics. John Wiley & Sons.

Wattenberg, M. P. (2011). US party leaders: Exploring the meaning of candidate-centred politics. En K. Aarts, A. Blais y H. Schmitt (Eds.), Political leaders and democratic elections (pp. 76-90). Oxford University Press. https://doi.org/10.1093/acprof:osobl/9780199259007.003.0005 

 

AUTHOR CONTRIBUTIONS, FUNDING AND ACKNOWLEDGEMENTS

Authors' contributions:

Conceptualization: Crespo Martínez, Ismael. Software: Mora Rodríguez, Alberto; Rojo Martínez, José Miguel. Validation: Mora Rodríguez, Alberto. Formal analysis: Crespo Martínez, Ismael; Mora Rodríguez, Alberto; Rojo Martínez, José Miguel. Data Curation: Rojo Martínez, José Miguel. Writing-Preparation of the original draft: Rojo Martínez, José Miguel. Writing-Revision and Editing: Crespo Martínez, Ismael; Mora Rodríguez, Alberto. Visualization: Rojo Martínez, José Miguel. Supervision: Crespo Martínez, Ismael. Project management: Crespo Martínez, Ismael. All authors have read and accepted the published version of the manuscript: Crespo Martínez, Ismael; Mora Rodríguez, Alberto; Rojo Martínez, José Miguel. 

Funding: This research has been funded by the Autonomous Community of the Region of Murcia through the Call for Project Grants for the Development of Scientific and Technical Research by Competitive Groups of the Seneca Foundation, Science and Technology Agency of the Region of Murcia, Project 21876/PI/22 "Affective polarization in the Region of Murcia. A study of its causes" (2023-2025). The author, José Miguel Rojo, is grateful for the funding received for his FPU Predoctoral Contract from the Spanish Government through the Programa Estatal de Promoción del Talento y su Empleabilidad (Ref. FPU20/01033).

Conflict of Interest: There is no conflict of interest.

AUTHORS: 

Ismael Crespo Martínez 

University of Murcia.

Professor of Political Science. Director of the Department of Political Science, Public Finance and Social Anthropology at the University of Murcia. Co-director of the Centro de Estudios Murciano de Opinión Pública (CEMOP) and Director of the magazine Más Poder Local. PI of the Transfer Group LABCOM (“Laboratorio de Comunicación Política”) at the University of Murcia. PI of the Research Project “POLARIZA” of the Seneca Foundation, Science and Technology Agency of the Region of Murcia (Project Ref.: 21876/PI/22).

icrespo@um.es   

Índice H: 18

Orcid ID: https://orcid.org/0000-0003-2600-8839

Google Scholar: https://scholar.google.es/citations?user=FvMZX9wAAAAJ&hl=es&oi=ao 

 

Alberto Mora Rodríguez

University of Murcia.

PhD in Political Science and Public Administration from the University of Murcia with Extraordinary Doctorate Award. Specialist in Applied Social Research and Data Analysis at the Center for Sociological Research (CIS). Permanent Professor of Political Science at the University of Murcia. Technical Director of the Centro de Estudios Murciano de Opinión Pública (CEMOP). Editor of the magazine Más Poder Local.

alberto.mora@um.es  

Índice H: 11

Orcid ID: https://orcid.org/0000-0003-2247-6846 

Google Scholar: https://scholar.google.es/citations?user=iGeFOvQAAAAJ&hl=es&oi=ao 

José Miguel Rojo Martínez

University of Murcia.

FPU Predoctoral Contract of the Ministry of Science, Innovation and Universities. Graduate in Political Science and Public Management with Final Degree Award, Graduate in Sociology and Master in Applied Political Analysis. He was Graduate Fellow at the Center for Sociological Research (CIS, for its acronym in Spanish) and Visiting Researcher at Stanford University (Spring 2023).

josemiguel.rojo@um.es   

Índice H: 9

Orcid ID: https://orcid.org/0000-0001-9491-4839 

Google Scholar: https://scholar.google.es/citations?user=3sMJEYcAAAAJ&hl=es

 

ANNEX 1

Testing the assumptions of intergroup homogeneity

i) Absence of significant differences in the ideological self-placement average between groups. One-factor ANOVA test: F= 0.0286, gl1= 3, p=0.993. G1: 5.22 (N=175, dt=2.12); G2: 5.25 (N=157, dt=2.13); G3: 5.22 (N=169, dt=2.13); G4: 5.19 (N=193, dt=2.23).

Figure 3: Ideological self-placement averages between groups

Source: Own elaboration.

2. Absence of significant differences in the voting recall (10N 2019 general elections) of participants in the different groups. Chi-square test: N=615, χ2 valor= 15.9, gl=18, p=0.596.

Table 9. Percentage of group composition in accordance with voting recall.

Voting recall	Experimental group
%	1 %	2 %	3 %	4 %
PP	21,1	26,9	19,6	22,0
PSOE	30,4	29,1	27,3	29,1
Citizens	8,1	20,4	14,0	12,4
Podemos-IU	20,5	17,9	18,9	17,6
Vox	12,4	11,2	12,6	13,0
MP-Equo	6,2	3,0	4,9	4,6
Others	1,2	1,5	2,8	1,5
Total N	100 161	100 134	100 143	100 177

Source: Own elaboration.

3. Absence of significant differences in the level of education of the participants in the different groups. Chi-square test: N=719, χ2 value= 6.26, gl=3, p=0.099. Percentage with university education G1= 69.4%; Percentage with university education G2= 66.3%. Percentage with university education G3= 73.4%. Percentage with university education G4= 77.4%.

Figure 4: Percentages of group composition by level of education.

Source: Own elaboration.

Related articles:

Bailén, M. Q. (2024). Los mensajes populistas en España ante la guerra de Ucrania. Los discursos de Vox y Unidas Podemos. Más Poder Local, 56, 84-104. https://doi.org/10.56151/maspoderlocal.151  

Montiel, L. I. T. (2024). Técnicas persuasivas empleadas en la comunicación política en la ciudad de Pilar, año 2024. Ciencia Latina Revista Científica Multidisciplinar, 8(5), 5353-5365. https://doi.org/10.37811/cl_rcm.v8i5.13976 

Panche Rodríguez, O. B., Alegría Varona, G. R., & Graus Cortez, L. E. (2024). Habilidades de negociación y manejo de conflictos en líderes comunitarios del Callao, Perú, 2024. European Public & Social Innovation Review, 10, 1-17. https://doi.org/10.31637/epsir-2025-464

Pop, A. I., Marín Pérez, B., & Pérez-Sánchez, J. (2023). Liderazgo y personalización de la política. Sánchez y Casado en las elecciones generales de 2019: el papel constructor de las televisiones generalistas al bipartidismo. Revista de Ciencias de la Comunicación e Información, 28, 56-76. https://doi.org/10.35742/rcci.2023.28.e282 

de Sola Pueyo, J., & Pascual Santisteve, A. (2024). Estrategia en Twitter de los grandes partidos políticos españoles: PSOE y PP en las autonómicas de 2023. Vivat Academia, 157, 1-24. https://doi.org/10.15178/va.2024.157.e1541