RIS Revista Internacional de Sociología Rev. int. sociol. 1988-429X 0034-9712 Consejo Superior de Investigaciones Científicas España 10.3989/ris.2024.82.4.1299 ris.2024.82.4.1299 Artículos Misperceptions and affective polarization: Evidence from Spain Percepciones erróneas y polarización afectiva: evidencias desde España https://orcid.org/0000-0003-2600-8839 Crespo Martínez Ismael icrespo@um.es Formal analysis Conceptualization Investigation Methodology Writing - original draft Writing - review Universidad de Murcia (UM). España
Universidad de Murcia España
https://orcid.org/0000-0003-2247-6846 Mora Rodríguez Alberto alberto.mora@um.es Formal analysis Investigation Methodology Writing - original draft Writing - review Universidad de Murcia (UM). España
Universidad de Murcia España
https://orcid.org/0000-0001-9491-4839 Rojo Martínez José Miguel Formal analysis Investigation Methodology Writing - original draft Writing - review Universidad de Murcia (UM). España
Universidad de Murcia España
(corresponding author) josemiguel.rojo@um.es 30 12 2024 30 12 2024 82 4 e260 Recibido 28 11 2023 Aceptado 06 09 2024 Publicado 04 02 2025 © 2024 CSIC 2024 CSIC https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International (CC BY 4.0) License. ABSTRACT

What effect do misperceptions have on our feelings towards a party? Research on the impact of misperceptions on intergroup hostility is becoming increasingly important in affective polarization studies in the United States. However, little attention has been paid to this issue in Europe. We use data from the Second National Survey of Political Polarization in Spain conducted by the Murcian Public Opinion Research Centre (CEMOP) in 2022 to show that the highest levels of perceptual deviation in Spain occur concerning what people think about Vox voters. Furthermore, it is found that the higher the misperceptions, the greater the feelings of rejection towards right-wing parties, something that does not happen with left-wing parties. Our study has important implications for understanding the origins of affective polarization and the role that partisan stereotypes play in this process.

 RESUMEN

¿Qué efecto tienen las percepciones erróneas en nuestros sentimientos hacia un partido? La investigación sobre el impacto de las percepciones erróneas en la hostilidad intergrupal está adquiriendo cada vez mayor importancia dentro de los estudios sobre polarización afectiva en el ámbito norteamericano. Sin embargo, apenas se ha atendido a esta cuestión en Europa. Usamos datos de la II Encuesta Nacional de Polarización Política en España llevada a cabo por el CEMOP en 2022 para evidenciar que los niveles más altos de desviación perceptiva en el país se producen respecto a lo que piensan el conjunto de electores sobre los votantes de Vox. Además, se comprueba que, cuanto mayores son las percepciones erróneas, mayores son los sentimientos de rechazo hacia los partidos de la derecha, algo que no sucede con las formaciones de izquierdas. Nuestro estudio tiene importantes implicaciones para el entendimiento de los orígenes de la polarización afectiva y el papel que juegan los estereotipos partidistas en este proceso.

 biases intergroup attitudes party stereotyping social identity feelings sesgos actitudes intergrupales estereotipos partidistas identidad social sentimientos Autonomous Community of the Region of Murcia 21876/PI/2022 Ministry of Universities of the Spanish Government FPU20/01033 This paper was made possible by the project financed by the Autonomous Community of the Region of Murcia through the call for aid to projects for the development of scientific and technical research by competitive groups, included in the Regional Programme for the Promotion of Scientific and Technical Research of Excellence (Action Plan 2022) of the Seneca Foundation, Science and Technology Agency of the Region of Murcia. Ref.: 21876/PI/2022. José Miguel Rojo Martínez thanks the Ministry of Universities of the Spanish Government for funding his FPU grant (State programme to develop, attract and retain talent). Ref.: FPU20/01033. Introduction

The party system in Spain has changed dramatically in the last decade. The founding of new parties with a considerable percentage of votes at both ends of the ideological spectrum has coincided with escalating conflict, the breaking down of important cross-cutting consensuses and increasing difficulties in reaching legislative agreements or forming a government. Furthermore, in 2018, the entry of Vox into institutions brought a close to the “Iberian exception” -the limited success that far-right parties seemed to have in Spain and Portugal, compared to the rest of Europe (Alonso and Rovira Kaltwasser, 2015; Turnbull-Dugarte, Rama and Santana, 2020; Heyne and Manucci 2021)-. Spain went from having an imperfect two-party system with one-party governments in which the two main national parties portrayed a centripetal dynamic to a multi-party system with four large nationwide parties that need to create coalitions to access power. This situation has replaced the centripetal two-party system with a polarized two-bloc logic.

This political climate is sparking social and academic debate surrounding the so-called “affective polarization”. We have observed how the disagreements about partisan competition have escalated, becoming an affective discord that generates hostility between the voters of the parties that make up both ideological blocs. Most literature on affective polarization is limited to underlining the scope of policy or ideological dissent as the source of this emotional tension both at elite and mass levels or the role that political identities as new social identity forms (Iyengar et al. 2019) play in certain attitudes or behaviours, thereby recovering the Michigan model’s classic approach but with an affective turn (Rivera and Lagares, 2022).

However, this article intends to provide a view hitherto unexplored in Spain’s case on the misperceptions generated by partisan biases. The attribution of mental states (i.e. what I believe others believe) appear in electors’ minds when judging the position or way of thinking of a party’s voters. These attribution of beliefs cause an overestimation of the actual level of polarization, artificially alienating opponents. If, when we think of others, the image we conjure up matches the partisan stereotype, we run the risk of accepting that all voters of a party resemble that stereotype, which is the group’s most distinctive identifier and which contrasts the most with other groups. In short, we will explore the psychological grounds of affective polarization, focusing on the cognitive consequences of identities rather than the identities themselves. Similarly, by paying attention to cognitive biases, we prove that people’s beliefs are not as divergent as they think.

 Misperceptions

Political parties have settled in our minds as social tribes (Iyengar, Sood and Lelkes 2012; Whitt et al. 2021) thanks to their growing ability to classify internal political and social identities, which by overlapping, increase the feeling of distance between each group’s members (Mason 2018; Harteveld 2021). Identity perceptions of political tribes are based on heuristic reasoning, which implies the biased representativeness of certain categories within the group (Ahler and Sood 2022).

We often believe that others seem more like their party’s stereotype than they actually are, which affects how we judge their political opinions or their personal traits. This phenomenon has been named “perception gap” (Yudkin et al. 2019) or is more commonly known as “misperceptions” (Chambers et al. 2006; Ahler 2014; Garrett et al. 2019; Orr and Huber 2021). Both concepts are closely related to the literature about false polarization (Wilson et al. 2020; Fernbach and Van Boven 2022) and second-order beliefs (Mildenberg and Tingley 2017) because attributing thoughts or ways of being to others from a stereotype-based estimation produces a greater sense of estrangement, which can potentially magnify the actual degree of the disagreement (Chambers et al. 2006).

Misperceptions can be analyzed using two closely related theoretical perspectives: the psychological-cognitive and the psychological-social. From a cognitive perspective, our estimates about our out-groups’ positions are distorted, considering traits that more commonly match with the stereotype generated for each category and which create an illusion of validity as part of a representativeness heuristic (Tversky and Kahneman 1974). By believing that a trait is more likely to be within one category than another, and that, in accordance with this likelihood, it can predict affiliation, we assume that the trait is dominant in the category (Ahler and Sood 2022). To this end, misperceptions would be the result of information-processing and meaning-generation strategies that promote animosity processes because they excessively represent the most extreme traits or stances, which are precisely those that maximise the distinction-intellect utility.

Political stereotyping -the mental image that appears when we think about members of a partisan out-group- can refer to the presumption of ideological or issue positions, personality traits or sociodemographic characteristics and is often related to the distinctive profile portrayed by the elites of each group (Flynn, Nyhan, and Reifler 2017; Rothschild et al. 2019). These types of cognitive structures influence voters’ environment processing (Rahn 1993) and end up conditioning their attitudes (Myers 2023). While serving to simplify or classify, they also generate affective hostility spirals, especially when the stereotype refers to the assumption of personality traits (Rothschild et al. 2019). For years, it has been proven that stereotypes do not only influence our judgements but also our positions and behaviours (Biernat 2003).

Partisan stereotypes generate an illusion of symbolic representativeness that is confused with descriptive representativeness. They simplify intergroup perception, reduce the likelihood of individual differentiation, generate estimation deviations when attributing certain positions or characteristics to the out-group and, as a result, increase the sense of polarization (Sherman, Nelson and Ross, 2003; Levendusky and Malhotra 2016; Ahler and Sood 2018). Even if the stereotype does not refer to personal traits or attributes, presupposing stances in certain issues implies a reflection on the moral attitude of the subject and indirect reasoning on their way of being. Clifford (2020) highlights the relationship between establishing moral stereotypes and the view held about each group’s ideology. When moral judgements are held that assert that the Democrats are more compassionate and Republicans are tougher, it is done so based on the ideological position that each party upholds (Clifford 2020).

From a social psychology perspective, stereotypes are not only the individual’s cognitive strategies, they are also mechanisms that make the in-group united, keep one’s own identity by alienating the out-group and justify the moral superiority of our people, something which elevates our self-esteem (Rubin and Hewstone 1998). They, therefore, maximise the sense of belonging.

Studies on social identity have recurrently questioned the influence categorization has on intergroup discrimination (Tajfel 1981; Bourhis, Sachdev and Gagnon 1996). To what extent does classifying an individual as a voter of another party lead to a rivalry dynamic that promotes in-group favouritism biases? This is a pivotal question for affective polarization studies, which try to explain the origin of hostility between subjects for political reasons (Rudolph and Hetherington 2021). Under the minimal group paradigm (Tajfel et al. 1971), simple categorization would be a sufficient condition for a discriminatory attitude, whether there is an actual conflict or not (Bourhis et al. 1996), given that the group’s aim is always to achieve a level of positive distinction. It could be argued that by categorizing voters, a system of discriminatory intergroup perception is generated without there necessarily being objective elements for conflict. By incorrectly assigning political positions, we are looking for evidence that supports our attitudes of discrimination and rejection given the impossible nature of supporting the competition using rational arguments.

Considering this, categorization underpins “a discrimination in favour of the intergroup in the categorization condition”, according to social identity theorists Deschamps and Devos (1996: 51). When the perception refers to a central category in forming the internal collective identity, in-group favouritism bias magnifies the differences, not only to maintain the representation system’s coherence but to make us feel assured that our group has a privileged position compared to others.

 Misperceptions and affective polarization

Previous research has highlighted the tendency of many citizens to think that the voter profile of a party they see on social media or in the press corresponds to the typical (average) voter of that party, which causes a greater degree of intergroup animosity because certain traits are wrongly assumed to be representative of the members of a group (Druckman et al. 2022). The largest perception gap can, therefore, be expected to be related to higher consumption of some media types (Yudkin et al. 2019). In reality, a party’s average voter does not resemble the image that comes to mind when we try to think about how they are and what type of things members from these groups defend. The average voter’s reality is a lot more complex and nuanced than what political and media discourse reflects. In line with this logic, several studies suggest that implementing interventions to correct misperceptions could reduce affective hostility (Lees and Cikara 2020; Druckman et al. 2022; Hartman et al. 2022).

A recent study confirms this hypothesis: By correcting misperceptions that reduce the sense of threat or opposition from the out-group to issues that are relevant to an individual, cold feelings towards rival parties, and in turn, affective polarization are reduced (Voelkel et al. 2023). This would show the causal connection between misperceptions and affective polarization (the onset of our affective judgements is inexact). Other research has linked correcting perception deviations to reduced support for partisan violence (Mernyk et al. 2022). All in all, false beliefs aligned with motivated reasoning do not tend to disappear easily. They are proof of the predispositions that are not easily changed because of the cognitive resistance that exposure to information contrary to our beliefs would generate and which would weaken identity markers (Nyhan and Reifler 2010).

Even by recognizing the limitations of these intervention proposals, it is evident that misperceptions contribute negative associations about the out-group and uphold an exaggerated account on the levels of polarization, in that simply disclosing the true extent of disagreements could improve intergroup relations (Ruggeri et al. 2021). In addition to dissemination, there are other intervention methods to correct stereotype-based perceptions. The two best-known are probably contact and dialogue between individuals from opposing groups as a way of understanding each other better and empathising with their positions and opinions (Levendusky and Stecula 2021; Miles and Shinew 2022; Thomsen and Thomsen 2022) and the displays of collaboration between the elite (Huddy and Yair 2021; Bassan-Nygate and Weiss 2022).

Based on this evidence, the growing climate of affective polarization in most of the world’s democracies would not originate only in ideological disagreements or the consolidation of political identities as social identities. Perception biases arising from intergroup conflict also appear, amplified by media and political communication strategies, as a way of explaining the cause of this phenomenon. This pigeonholing underpins discriminatory judgements that alienate the adversary to positively single out one’s own group (Spears and Otten 2013).

 Hypothesis

Based on other findings on the importance of misperceptions in affective attitudes, we tried to answer the following questions:

RQ1. Which voter groups trigger more misperceptions in Spain?

RQ2. How are these misperceptions towards different voter groups distributed according to individuals’ political identities, media consumption and other sociodemographic variables?

RQ3. Do misperceptions about a party increase rejection towards that party? This question is particularly relevant for the study of affective polarization, as the increase in out-group rejection is the main driver of this phenomenon (Iyengar et al., 2019).

While the first RQ is more descriptive, the latter two pose a relationship between variables, which needs to be tested. Accordingly, we formulate the following two hypotheses:

H1 (related to RQ2). Misperceptions about a group of voters are highest among the members of its out-groups (whether partisan or ideological).

This hypothesis is linked to the theory of intergroup accentuation and group bias. Previous studies have shown that ‘misperceptions about in-groups are substantially smaller than those about out-groups’ (Bursztyn and Yang, 2022; Ahler and Sood, 2018). We make fewer errors in estimating the position of members of our own group. This is the result of the principle of intergroup accentuation: the tendency to increase the contrast between categories, maximising the distinction between oneself and the members of the out-group (Hewstone and Greenland, 2000). However, more evidence is still needed to support these earlier findings.

H2 (related to RQ3). The greater the level of misperceptions about what voters of a party think, the stronger the feelings of rejection towards that party. Misperceptions influence affective attitudes, alienating us from the out-group.

The second hypothesis represents the main thrust of the article. While for the United States it has been shown that misperceptions about rivals can increase negative feelings towards them and consequently affective polarization (Druckman et al., 2022; Hartman et al., 2022; Voelkel et al., 2023), there is no evidence that this also occurs in other countries. Specifically, there are no data in Spain to corroborate this relationship. Misperceptions increase rejection of members of the outgroup because they attribute to them motivations, positions or ideas that we find particularly reprehensible.

 Data and Methods

To move forward in answering the research questions and checking the hypotheses, we used data from the “Second National Survey of Political Polarization” conducted by the Murcian Public Opinion Research Centre (CEMOP) from 25 April until 18 May 20221. In the cited study, we asked the respondents to express their feelings towards the main Spanish political parties (Partido Socialista Obrero Español [PSOE], Partido Popular [PP], Vox and Unidas Podemos [UP]) using a scale of 0 to 10, in which 0 represented strong feelings of “aversion and rejection” and 10 strong feelings of “sympathy and membership”.

Similarly, to measure the presence of misperceptions, we asked the respondents to express their opinion on a list of four policy issues, and later asked them in what position they would place voters of PSOE, PP, Vox and UP on the same topics. By cross-referencing the position that the individuals placed themselves regarding the issues with the variable “vote+sympathy”,2 we obtained a close approximation to the actual average of what the voter groups think about some of the political agenda items. This average is compared with the estimated views that each individual assigned to each group (the difference between both will determine the level of misperceptions). We used the following scales and topics3 to understand the self-positioning and estimation about each group’s positions:

(Where would you place yourself? /Where would you place voters of PSOE / PP / Vox / UP?)

Feminism (FEM). On a scale of 0 to 10 in which 0 is “women continue being discriminated against and that is why we need to keep promoting feminist policies that favour women” and 10 is “rather than defending equality, feminism is used to attack men”.

Ecologism (ECO). On a scale of 0 to 10 in which 0 is “priority must be given to protecting the environment, even if it means slower economic growth and losing some jobs” and 10 is “economic growth and job creation must be top priority, even if the environment were to suffer to certain extent”.

Gender violence (GENVIO) On a scale of 0 to 10 in which 0 is “gender violence is a very serious problem and the State should increase resources to combat it” and 10 is “gender violence does not exist and the State should worry about other more important issues”.

Immigration (IMMIGRA) On a scale of 0 to 10 in which 0 is “the State should allow all types of immigration into our country” and 10 is “the State should close the borders and not permit any type of immigration”.

Following the calculation formula proposed by Rojo, Crespo and Mora (2023), which is based on the coefficient of variation (CV), we obtained a measure of perceptual deviation for each individual as the result -in absolute value- of subtracting the average that the group actually has on that topic ( X ¯ IP ) from a person’s estimate positioning of a group on a topic ( X ¯ IP ). The larger the resulting difference in the numerator, the larger the level of misperceptions (MP). This calculation allows us to aggregate the deviations produced on a party in all of the topics (N). The deviation formula also includes the estimations that an individual makes on their own group.

M P = ∑ i = 1 N | X IP − X ¯ IP | N

Throughout the research, we have worked with another series of variables that serve as control mechanisms for the main relationships being compared: Individual’s ideology4; sex; age; level of education (university/non-university); positive partisan identity (PPID) regarding those parties that are different to the party positioned in the dependent variable.5 Lastly, we looked at the frequency with which political information is consumed via different media sources (television, social media and digital newspapers)6. You can find more information about how all the variables were encoded and their descriptive and frequency statistics in Annex 1.

 Results and discussion The distribution of misperceptions

Misperceptions are particularly notable when considering the positions of radical-right voters and sympathisers. We observed the largest gap between the attributed positioning and the actual positions about the topics for this group (see Table 1, RQ1). The misperceptions indicator about Vox ( X ¯ MP = 3.382, sd = 1.310) contrasts with results obtained from other parties, which are somewhat similar yet all smaller than those for Vox (PSOE X ¯ MP = 2.030, sd = .967; PP X ¯ MP = 2.349, sd = 1.225; UP X ¯ MP = 2.236, sd = 1.187). Remember that the higher the indicator, the greater the differences between the positions and the attribution.

Scope of misperceptions about what voters of the different parties think in Spain (aggregate gap by party and by topic for each party).
X̅ SD N (by list)
PSOE aggregate MP 2.030 .967 982
PP aggregate MP 2.349 1.225 993
Vox aggregate MP 3.382 1.310 995
UP aggregate MP 2.236 1.187 999
PSOE FEM MP 2.071 1.495 1071
PP FEM MP 1.964 1.529 1073
Vox FEM MP 3.212 2.012 1086
UP FEM MP 2.656 1.783 1093
PSOE ECO MP 1.810 1.474 1045
PP ECO MP 2.104 1.609 1052
Vox ECO MP 3.122 1.668 1044
UP ECO MP 2.240 1.628 1054
PSOE GENVIO MP 2.336 1.779 1076
PP GENVIO MP 3.340 1.954 1076
Vox GENVIO MP 4.372 1.989 1073
UP GENVIO MP 1.948 2.092 1073
PSOE IMMIGRA MP 1.984 1.472 1076
PP IMMIGRA MP 1.993 1.474 1083
Vox IMMIGRA MP 2.840 1.616 1090
UP IMMIGRA MP 2.182 1.564 1076
Source: Created by the authors with data from CEMOP’s Second National Survey of Political Polarization.

Analysing which topics produce the largest gaps between what Vox voters and sympathisers think and what others believe they think, we find bigger distortions when recognizing gender-based violence as a public priority ( X ¯ MP = 4.372, sd = 1.989). Spaniards think that Vox voters are less sensitive towards gender-based violence than they claim to be, something which is also prominent in the case of PP voters. The results show that this morally charged topic is recognized as an element of intergroup differentiation when the out-groups of Vox, and in general of the right bloc, want to make comparisons.

Positions on gender-based violence are seemingly not a desired identity marker by Vox members, unlike immigration, as the party’s position on the latter is not so far from its voters’ position and does not contrast with other groups’ perceptions. For Vox, immigration is the topic with the smallest gap between perception and actual positioning ( X ¯ MP = 2.840, sd = 1.616). The actual positions on immigration are better adjusted to the average voter’s reality and generate fewer problems when being expressed as markers of belonging -forecast of lower impact on the group’s status and self-esteem- than those referring not only to gender-based violence but also to feminist policies in general ( X ¯ MP = 3.212, sd = 2.012) or ecologism ( X ¯ MP = 3.122, sd = 1.668). These findings call for us to reflect on the topics that underpin the identification process with Vox.

Once we identified the existence of particular discrepancies when comparing what Vox voters and sympathisers think and how each of the respondents imagine them, we could look more closely at the distribution of average misperceptions about the four parties according to the different variables (see Table 2).

At a sociodemographic level, the leftist parties (PSOE and UP) generate more misperceptions among people without university education and Vox does so among people with university education. Regarding sex, significant differences were only noted for PP (women have more misperceptions about this party). However, the size of the effect is reduced. Aside from the sociodemographic distribution, we were interested in the different perception deviation averages referring to each of the partisan groups’ identification (positive partisanship) or ideologies (H1). To this end, we will explore if the misperception phenomenon is - as theory suggests - an effect of intergroup comparison and conflict.

We obtained significant evidence (see Table 2) that misperceptions about a group are always higher among out-group members. For example, those who are strongly identified with PSOE have an average of 4.073 (sd = 1.135) over Vox, while it is 3.271 (sd = 1.301) for the rest of the sample. For those who identify with Vox, their level of misperceptions towards PSOE is 2.463 (sd = .992), while it is 1.969 (sd = .952) for the rest of the sample. The more removed our group is from the object of perception, the more erroneous the estimation of its members is and the more homogenization and stereotypical generalization tend to occur. This explains why those who identify with UP have the highest average score on Vox ( X ¯ MP = 4.115, sd = .911). On the contrary, similar groups have lower estimation errors. The perception gap on Vox is smaller for PP sympathisers and voters ( X ¯ MP = 2.775, sd = 1.278) than for the rest of the sample ( X ¯ MP = 3.471, sd = 1.291). This also happens when considering Vox voters’ perception of PP voters and sympathisers. The effect sizes in all cases are high. These results are coherent with those obtained from other psychosocial studies on representation patterns, social projection, the forming of judgements and the tendencies to exaggerate intergroup differences (Mullen, Brown and Smith 1992; Moore-Berg et al. 2020; Rojo et al. 2023).

Similarly, those who feel attached to a group tend to be those who have the least misperceptions about that group, which they know best and which they do not judge based on unrepresentative stereotypes that favour homogenization. Those who positively identified with Vox have a misperceptions indicator of 2.488 (sd = 1.168) on their party, being 3.522 (sd = 1.277) for those non-positive identified. Although group membership implies a more precise and less extreme representation ability (Linville et al. 1989), those who identify positively with Vox are - in comparison with those who identify positively with other parties- those who deviate most in estimating their peers’ positions. Partisan loyalty towards the far-right could be explained by a multi-layered thematic framework, making it difficult for those who are part of the group to understand their own identity.

Significant mean differences in misperceptions according to positive partisan identity (PPID), sex, age, level of education and ideological polarization.
X̅ (SD) PSOE aggregate MP Test X̅ (SD) PP aggregate MP Test X̅ (SD) VOX aggregate MP Test X̅ (SD) UP aggregate MP Test
Sex Male 1.988 (.955) ns 2.255 (1.155) ** d=-.188 3.340 (1.309) ns 2.190 (1.157) ns
Female 2.072 (.983) 2.443 (1.286) 3.423 (1.310) 2.283 (1.215)
Age 18-30 1.878 (.998) *** n2=.036 2.406 (1.123) ns 3.372 (1.154) ** n2=.010 1.985 (1.247) *** n2=.034
31-44 1.806 (.877) 2.260 (1.154) 3.283 (1.131) 1.959 (.967)
45-64 2.077 (.906) 2.319 (1.267) 3.304 (1.317) 2.408 (1.237)
65 and over 2.293 (1.066) 2.452 (1.292) 3.604 (1.525) 2.428 (1.203)
Education University 1.898 (.913) *** d=.241 2.357 (1.205) ns 3.484 (1.254) ** d=-.175 2.130 (1.187) *** d=.202
No university 2.139 (1.002) 2.342 (1.241) 3.309 (1.352) 2.333 (1.180)
PSOE PPID Yes 2.209 (1.273) * d=-.208 3.044 (1.434) *** d=-.804 4.073 (1.135) *** d=-.802 2.354 (1.423) ns d=-.696
No 2.001 (.906) 2.240 (1.153) 3.271 (1.301) 2.220 (1.146)
PP PPID Yes 2.348 (1.064) *** d=-.367 2.0230 (1.07782) *** d=.376 2.775 (1.278) *** d=.696 2.508 (1.174) *** d=-.313
No 1.981 (.942) 2.399 (1.239) 3.471 (1.291) 2.195 (1.178)
VOX PPID Yes 2.463 (.992) *** d=-.494 2.017 (1.096) *** d=.385 2.488 (1.168) *** d=-1.034 2.503 (1.053) *** d=-.309
No 1.969 (.952) 2.403 (1.238) 3.522 (1.277) 2.193 (1.186)
UP PPID Yes 1.875 (1.058) * d=.172 3.384 (1.268) *** d=-1,153 4.115 (.911) *** d=-.822 2.043 (1.458) * d=.214
No 2.048 (.955) 2.229 (1.164) 3.293 (1.322) 2.257 (1.143)
Ideology Leftist 1.858 (.911) *** d=-.504 2.814 (1.247) *** d=.847 3.939 (1.019) *** d=1.219 1.983 (1.170) *** d=-.539
Right 2.362 (1.015) 1.967 (1.030) 2.719 (1.305) 2.522 (1.126)

Bivariate analysis: T test for dichotomous variables and ANOVA test for multiple comparison analysis (including Cohen’s d test for t tests and eta squared - n2 - for ANOVA, to measure the effect size).

p<0.01

p<0.05

p<0.1

not statistically significant.

 Source: Created by the authors.

The mean differences in the levels of misperceptions are related to partisan group identities, confirming the first part of H1. However, it is also appropriate to consider the relationship between this phenomenon and the individual’s ideology. An individual’s self-positioning on the left or on the right could be considered as another relevant form of group political identity, changing the party’s marker for the ideological bloc’s marker. It would not be correct to limit the affective reference groups to parties (Hobolt et al. 2021), especially in multi-party systems in which ideological blocs frequently act as elements that articulate political relations (Hagevi 2015).

All MP averages obtained significant differences with a high effect size (> 0.5) when comparing individuals from the left and from the right. For example, misperceptions about Vox among leftists are higher than those obtained for those on the right (Vox X ¯ MP leftists = 3.939, sd = 1.019; Vox X ¯ MP right-wingers = 2.719, sd = 1.305, t = 12.947, p < 0.01). There are also fewer misperceptions about UP and PSOE among leftist respondents than those from the right (UP X ¯ MP leftists = 1.983, sd = 1.170; UP X ¯ MP right-wingers = 2.522, sd = 1.126, t = -5.659, p < 0.01; PSOE X ¯ MP leftists= 1.858, sd=.911; PSOE X ¯ MP right-wingers = 2.362, sd = 1.015, t = -6.291, p < 0.01). The differences for PP follow the same direction as those for Vox.

What seems to be more connected to the mean difference of perceptual deviation -ideology or partisan feelings? As closely aligned and intertwined identities, ideology and partisanship seem to be related with similar cognitive effects, being the only variables with significant differences and a considerable effect. However, the comparison between ideology categories shows mean differences with a more relevant effect (especially for misperceptions about Vox and PP).

The bivariate analyses and the above descriptive statement act by way of introduction. We will now test multiple linear regression models that control the influence of different variables to find out how misperceptions affect the individuals’ feelings towards a party (H2).

 The effect misperceptions have on our feelings towards a party

We present a multiple linear regression model ( Y i = β 0 + β 1 X i + β k X i k + ε ) that allows us to predict feelings towards different parties considering the level of misperceptions towards people who support or sympathise with those party. We presume that if misperceptions explain higher levels of rejection or membership, then they increase the difference in feelings between partisan groups and, therefore, affective polarization (they affectively unite or alienate us from the out-groups).

Although the coming sections focus on analysing the case of Vox -given its differential nature shown by the results in the last section- Annex 2 includes the results from the same regression applied to feelings towards PSOE, PP and UP to compare how the model works. We control the relation proposed by H2 including a series of covariables that, by keeping them constant, will allow us to precisely measure the influence of misperceptions: positive partisan identification compared to other parties (measuring the conflict impact between group identities), ideology, misperceptions about other parties, the level of consuming different media and sociodemographic variables like sex, age and level of education. The first two control variables (positive partisan identity and ideology) represent alternative explanations to the one that we are focusing on in this study -that of misperceptions- on the assumption that the phenomenon of affective polarization cannot be addressed based on a single relationship between variables. The formulation of the linear regression model equation to check how misperceptions affect feelings towards a party is:

Y i ( Feelings VOX ) = β 0 + β 1 ( PPI D P P ) + β 2 ( PPI D PSOE ) + β 3 ( PPI D UP ) + β 4 ( Ideology ) + β 6 ( Sex ) + β 7 ( Age ) + β 8 ( Level of education ) + β 9 ( Level of TV consumption ) + β 10 ( Level of social media consumption ) + β 11 ( Level of digital newspaper consumption ) + β 12 ( Vox MP ) + β 13 ( PSOE MP ) + β 14 ( PP MP ) + β 15 ( UP MP ) + ε

Based on the results shown in Table 3, we observed that the level of perceptual deviation on what Vox voters really think is -after ideology (cognitive shortcut par excellence for the formation of our political judgements)- the variable with the highest level of influence when explaining the position on the sympathy/rejection scale for that party (β = -.222, p<0.01). On the contrary, contributing to the partisan identity within the designed model is limited, given that none of the PPID variables are significant.

The higher the perceptual deviations about Vox voters, the more the individual will move towards positions of affective rejection towards this party. Misperceptions are proven to significantly affect our attitudes. With misperceptions appearing to be variables that explain feelings towards Vox, we must ask ourselves to what extent is affective polarization influenced by group-based psychological processes, enriching classic theories (ideology, intergroup rivalry). The images that are portrayed within us about a group’s beliefs seem to become a relevant starting point for our sentimental reactions. Likewise, we found that misperceptions about PSOE support membership with Vox (β = .061, p<0.01). This upholds the theoretical assumption that these biases respond to a need for in-group reinforcement but similarly, it highlights the idea that these two parties are engaged in an intense opposing relationship to the extent that the “us vs. them” dynamic could be occurring more between PSOE and Vox than between the alternative left (UP) and Vox.

For a long time, we have focused on the impact of ideology (Rogowski and Sutherland 2016; Webster and Abramowitz 2017) and political identity as a form of social identity (Iyengar et al. 2012) on affective expressions. The first theory outlines that a greater worldview divide creates an emotional detachment. From an identity approach, identifying with a group can be expected to -as a minimum condition, in line with Tajfel- cause negative feelings towards external groups, especially in a scenario of competition for scarce resources (power). The Vox model confirms that ideology plays an important role but that group identity is influenced by the perception biases it generates.

Our research has contributed to the debate on how populist radical-right parties’ ability to generate affective polarization is related to the stereotyping process that we have observed people who vote for this type of parties suffer (at least in Spain); a process that is influencing affective assessments of citizens as a whole. Misperceptions about Vox voters, which are particularly notable, increase a sense of false polarization that distorts the complex reasoning of these individuals and their motives, attributing to them positions that coincide with the identity prototype or with positions of the elite but do not necessarily translate to the reality of an ordinary voter. The scant attention paid to Europe in the literature regarding the relationship between misperceptions and affective polarization should now be reconsidered with these findings, which highlight how powerful stereotypes are in generating negative feelings towards a party.

As we have argued, identity variables (ideology or partisanship), representative of last century’s political behaviour theories, need to be supported by a novel explanation related to perception biases. Misperceptions can be as important as other variables, extensively reviewed in the literature, and moreover, they open an interesting path for implementing intervention projects that reduce the levels of interpartisan hostility. This finding replicates what social psychologists warned us about from studying intergroup relationships, even since the seminal works of Sherif et al. (1988): Rather than being due to objective reasons, conflict is the result of a discriminatory belief structure that looks for the in-group’s positive distinction to maintain status and protect the individual. However, what happens with the rest of the parties? (See Annex 2). Applying the same regression model, we observe that the influence of misperceptions is not similar in all cases. Misperceptions do not influence feelings towards PSOE (nor those about PSOE or the other groups). Nevertheless, when generating an explanatory model about affective attitudes towards this party, sex, ideology (once again) and identifying with some groups - particularly Vox and UP - are relevant. Positive identity with Vox reduces positive feelings towards PSOE (β = -.179, p<0.01), which is something that does not occur with positive identity regarding PP. Once again, the “us vs. them” dynamic is reflected between the social democracy and the radical right in Spain.

Multiple linear regression analysis of the effect that misperceptions on feelings towards Vox.<xref ref-type="fn" rid="fn-7-e261">1</xref>
Feelings towards Vox
B (E) β p VIF
Positive partisan identity (PPID) (ref.: no PPID)
PSOE PPID -.440 (.239) -.048 1.192
PP PPID -.107 (.282) -.011 1.526
UP PPID -.209 (.250) -.022 1.189
Ideolog. (ref.: right) -4.719 (.251) -.642 *** 2.091
Sex (ref.: male) -.145 (.170) -.021 1.072
Level of education (ref.: no university studies) -.167 (.172) -.024 1.098
Age .009 (.006) .042 1.286
Political news consumption
TV .019 (.059) .008 1.160
Social media -.142 (.053) -.069 *** 1.202
Digital newspapers .072 (.053) .035 1.199
PSOE MP .215 (.106) .061 ** 1.620
PP MP -.124 (.079) -.045 1.444
Vox MP -.613 (.082) -.222 *** 1.594
UP MP -.063 (.088) -.021 1.590
Constant 7.661 (.455) ***
R2 .689
Adjusted R 2 .681
N (according to the list) 571

p<0.01

p<0.05

Numbers in parentheses are standard errors.

If misperceptions are not important for explaining the feelings towards PSOE, in the case of PP (see Table 8, Annex 2), the model results are similar to those obtained for Vox. Higher levels of misperceptions about PP increase rejection towards this party (β = -.264, p < 0.01), with a slightly greater impact than that observed for Vox. Misperceptions also become the most significant predictive variable, second only to ideology. By conducting a regression analysis on the feelings towards the two right-wing parties (see Tables 4 and 5), only including misperceptions about Vox and PP and ideology in the model, we found that ideology further reduces the explanatory contribution of misperceptions in Vox’s case. We noted that ideology moderates the effects of perception biases within the regression models, understanding part of its explanatory ability, which is more pronounced in the extreme party than the centrist party in the bloc. Attitudes towards more extreme parties may be more dependent on structural elements such as ideology, which would reduce the ability of misperception-correction interventions to reduce animosity towards these groups.

Multiple linear regression analysis to check the explanatory ability compared between misperceptions and ideology about feelings towards Vox
Feelings towards Vox
Model 1 Model 2
B (E) β p B (E) β p
Vox MP -1.293 (.069) -.514 *** -.601 (.070) -.218 ***
Ideology (ref.: right) -5.115 (.187) -.699 ***
Constant 6.908 (.249) *** 7.871 (.234) ***
R 2 .265 .676
Adjusted R2 .264 .675
N according to the list 990 628

p<0.01,

p<0.05.

Numbers in parentheses are standard errors.

 Multiple linear regression analysis to check the explanatory ability compared between misperceptions and ideology about feelings towards PP.
Feelings towards PP
Model 1 Model 2
B (E) β p B (E) β p
PP Pe -1.070 (.070) -.440 *** -.649 (.073) -.258 ***
Ideology (ref.: right) -3.929 (.191) -.600 ***
Constant 6.375 (.184) *** 7.991 (.205) ***
R 2 .194 .527
Adjusted R2 .193 .525
N according to the list 987 626

p<0.01,

p<0.05.

Numbers in parentheses are standard errors.

Briefly, it would seem that rightist parties suffer the affective consequences of misperceptions in Spain the most. This is consolidated when confirming that (see Table 9, Annex 2) -in the model about feelings towards UP- misperceptions about this party’s voters do not significantly affect feelings towards the party but misperceptions towards PP and Vox do. The more deviations there are regarding the opinions of rightist party voters, the greater the membership to UP (β MP PP = .134, p<0.01; β Pe Vox = .100, p<0.01). This allows us to extend the scope of H2: misperceptions about a party do not only act by generating rejection towards this party, when they are produced for out-groups, they can act as a positive affiliation mechanism to the in-group.

We confirm that the extent of the biases and their effects on affective attitudes is not bidirectionally homogeneous in the intergroup conflict, in such a way that leftist voters in Spain uphold a considerable stereotypical view about rightist sympathisers. These misjudgements of what others are like may be another major driver of recent political hostility, coupled with a sort of artificially created moral superiority.

 Conclusions, study limitations and future research

Throughout this research, we have proven that the highest level of misperceptions among the Spanish electorate is produced around what citizens think about radical-right voters. Vox voters are the most stereotyped and those who are furthest removed from their profile set by society as a whole, especially regarding feminist-related issues. The lack of commitment of these individuals regarding gender-based violence is somewhat exaggerated and their values about gender equality and ecologism are underestimated. However, there seem to be fewer disparities regarding immigration. These findings open up a new debate on the central topics that justify far-right identification in Spain and the coherence of positions between elites and base voters while re-emphasizing the non-proportional nature of different parties’ contributing to polarization dynamics.

We have also shown that, in all cases, misperceptions about a group of voters are higher among its out-groups. The further away we are from a group, the more errors we make in attributing a mental state. The groups diverge from the comparison process, generating artificial divisions, assuming that the actual position of a party or its elites is homogeneously shared by its voters. This comparison seems dominated by a heuristic way of thinking and motivated reasoning that serves to improve the environment’s processing capacity but also maintain each group’s status and strengthen its identity. It seems clear that misperceptions are biases caused by belonging to an “us” compared to a “them” both as a cognitive strategy and as a strategy to reinforce in-group affiliation and its members’ self-esteem.

The main focus of this research’s contribution is to underline how -in the case of the two right bloc parties (Vox and PP)- misperceptions about rightist voters are the most important variable when predicting affective attitudes towards them, second only to ideology. In both cases, the higher the levels of misperceptions about these parties, the greater the levels of rejection towards them. This can increase affective polarization by leading to a greater spread in the evaluation of the in-group and out-groups. However, the fact that this does not occur with any of the two parties from the left bloc (PSOE and UP) shows us that, although the phenomenon of differential cognitive biases by affiliation group (see Table 2) is cross-cutting, the intensity is not the same for all groups and it affects some groups more than others in affective terms.

Similarly, we present three substantive findings to analyse the political dynamics in Spain: (1) the importance of gender issues in the construction of the cultural, emotional and identity battle, (2) the constitution of PSOE as Vox’s leading antagonist -contrary to what the spatial interpretation of polarization could have predicted- and (3) the strategic use of rhetoric favouring the sense of false polarization and making it difficult for ordinary voters -who do not resemble their perceived stereotype as much- to understand each other, even if some discourses strive to continually remind us of the unbreakable boundaries between groups.

Likewise, it is important to mention this study’s main limitation: social desirability could alter respondents’ replies when stating their position regarding certain issues, spuriously increasing the level of misperceptions. Despite answers being anonymous and using a relatively impersonal survey method (CATI), two types of biases could be involved in the responses observed: a) that the respondents adapted their beliefs to what is understood as being more socially acceptable, therefore recognizing that their group’s status about that issue is illegitimate (Hinke and Taylor 1996) and wanting to respond to this situation by transforming their postures to safeguard their personal image -which would be an interesting finding in itself-; b) that a meta-perception calculation was carried out when responding, attempting to portray a positive impression that protects the group’s status, changing the actual belief, not to protect themselves from being judged but to protect the group’s social position. Other issues will also need to be considered to see if they produce the same results even if they are areas of different social or political symbolism. We can accept that the origin of such misperceptions may be related to the nature of the issues, the symbolic ownership, the social hegemony of each party on each issue, or the importance that parties give to different issues (which may affect the cohesion of their electorate), but this does not detract from the fact that misperceptions occur, and misperceptions have affective effects.

Lastly, future studies will have to overcome other limitations of this study, such as developing experimental designs that allow us to clarify further the direction of the causal relationship between misperceptions and affective attitudes. This could be argued, contrary to what is stated here, that these are a posteriori judgements to justify well-established affective attitudes. Additionally, any future studies should further examine the origins of these misperceptions, paying special attention to the role of the elites, social media and the press. In particular, it is relevant to see what role communication strategies play in reinforcing misperceptions (top-down approach). It is possible to argue that this perception gap is installed by the leaders in order to artificially increase the differentiation with respect to other parties. Polarization and the exaggeration of party differences could respond to electoral strategies to enhance competitive advantage.

 Data availability statement

You can consult and download the database of the survey used in this link: https://www.cemopmurcia.es/estudios/ii-encuesta-nacional-de-polarizacion-politica/.

 Funding acknowledgement statement

This paper was made possible by the project financed by the Autonomous Community of the Region of Murcia through the call for aid to projects for the development of scientific and technical research by competitive groups, included in the Regional Programme for the Promotion of Scientific and Technical Research of Excellence (Action Plan 2022) of the Seneca Foundation, Science and Technology Agency of the Region of Murcia. Ref.: 21876/PI/2022. José Miguel Rojo Martínez thanks the Ministry of Universities of the Spanish Government for funding his FPU grant (State programme to develop, attract and retain talent), Ref.: FPU20/01033.

 Conflict of interest statement

The authors of this article declare that they have no financial, professional or personal

conflicts of interest that could have inappropriately influenced this work.

 Authorship contribution statement

Ismael Crespo Martínez: Conceptualization, Formal analysis, Investigation, Methodology, Writing - original draft, Writing - review.

Alberto Mora Rodríguez: Formal analysis, Investigation, Methodology, Writing - original draft, Writing - review.

José Miguel Rojo Martínez: Formal analysis, Investigation, Methodology, Writing - original draft, Writing - review.

 References Ahler, Douglas J. 2014. “Self-Fulfilling Misperceptions of Public Polarization”, The Journal of Politics 76(3): 607-620. 10.1017/S0022381614000085. Ahler, Douglas. J. and Gaurav Sood, G. 2022. “Typecast: A routine mental shortcut causes party stereotyping”, Political Behavior 45: 1581-1607. 10.1007/s11109-022-09780-8. Ahler, Douglas. J. and Gaurav Sood . 2018. “The Parties in Our Heads: Misperceptions about Party Composition and Their Consequences”, The Journal of Politics 80(3): 964-981. 10.1086/697253. Alonso, Sonia and Cristóbal Rovira Kaltwasser . 2015. “Spain: No Country for the Populist Radical Right?” South European Society and Politics 20(1): 21-45. 10.1080/13608746.2014.985448. Bassan-Nygate, Lotem and Chagai M. Weiss . 2022. “Party Competition and Cooperation Shape Affective Polarization: Evidence from Natural and Survey Experiments in Israel”, Comparative Political Studies 55(2): 287-318. 10.1177/00104140211024283. Biernat, Monica . 2003. “Toward a Broader View of Social Stereotyping, American Psychologist 58(12): 1019-1027. 10.1037/0003-066X.58.12.1019. Bourhis, Richard Y. , Itesh Sachdev and Andréane Gagnon . 1996. “Las matrices de Tajfel como un instrument para realizar investigación intergrupal”. P.p. 61-103 In Identidad social. Aproximaciones psicosociales a los grupos y a las relaciones entre grupos, Ed. By J. Francisco Morales , Darío Páez , Jean-Claude Deschamps and Stephen Worchel . Valencia: Promolibro. Bursztyn, Leonardo and David Y. Yang . 2022. “Misperceptions about others”. Annual Review of Economics, 14, 425-452. 10.1146/annurev-economics-051520-023322. Chambers, John. R. , Robert S. Baron and Mary L. Inman . 2006. “Misperceptions in Intergroup Conflict: Disagreeing About What We Disagree About”, Psychological Science 17(11): 38-45. 10.1111/j.1467-9280.2005.01662.x. Clifford, Scott . 2020. “Compassionate Democrats and Tough Republicans: How Ideology Shapes Partisan Stereotypes”, Political Behavior 42: 1269-1293. 10.1007/s11109-019-09542-z. Crespo Martínez, Ismael and Alberto Mora Rodríguez . 2022. “El auge de la extrema derecha en Europa: el caso de Vox en la Región de Murcia”, Política y Sociedad 59(3): e75974. 10.5209/poso.75974. Deschamps, Jean-Claude and Thierry Devos . 1996. “Relaciones entre identidad social e identidad personal”. P.p. 39-57 In Identidad social. Aproximaciones psicosociales a los grupos y a las relaciones entre grupos, Ed. By J. Francisco Morales , Darío Páez , Jean-Claude Deschamps and Stephen Worchel . Valencia: Promolibro. Druckman, James N. , Klar, Samara , Krupnikov, Yanna , Levendusky, Matthew and Ryan, John B. 2022. “(Mis) estimating affective polarization”, The Journal of Politics 84(2), 1106-1117. 10.1086/715603. Fernbach, Philip. M. and Leaf Van Boven . 2022. “False Polarization: Cognitive Mechanisms and Potential Solutions”, Current Opinion in Psychology 43: 1-6. 10.1016/j.copsyc.2021.06.005. Ferreira, Carles . 2019. “VOX como representante de la derecha radical en España: un estudio sobre su ideología”, Revista Española de Ciencia Política 51(51): 73-98. 10.21308/recp.51.03. Flynn, D. J. , Brendan Nyhan and Jason Reifler . 2017. “The Nature and Origins of Misperceptions: Understanding False and Unsupported Beliefs About Politics”, Political Psychology 38(51): 127-150. 10.1111/pops.12394. Garrett, R. Kelly , Jacob A. Long and Min Seon Jeong . 2019. “From Partisan Media to Misperception: Affective Polarization as Mediator”, Journal of Communication 69(5): 490-512. 10.1093/joc/jqz028. Hagevi, Magnus . 2015 “Bloc identification in multi-party systems: the case of the Swedish two-bloc system”, West European Politics 38(1): 73-92. 10.1080/01402382.2014.911480. Harteveld, Eelco . 2021. “Ticking all the boxes? A comparative study of social sorting and affective polarization”, Electoral Studies 72: 1-11. 10.1016/j.electstud.2021.102337. Hartman, Rachel , Will Blakey , Jake Womick , Chris Bail , Eli J. Finkel , Hahrie Han , 2022. “Interventions to reduce partisan animosity”, Nature Human Behavior 6: 1194-1205. 10.1038/s41562-022-01442-3. Hewstone, Miles and Katy Greenland . 2000. “Intergroup conflict”. International Journal of Psychology: 35(2), 136-144. 10.1080/002075900399439. Heyne, Lea and Luca Manucci . (2021). “A new exceptionalism? Comparing the populist radical right electorate in Portugal and Spain”, Political research exchange 3(1): 1-27. 10.1080/2474736X.2021.1989985. Hinkle, Steve W. and Laurie A. Taylor 1996. “Identidad social y aspectos de la creatividad social: cambios a nuevas dimensiones de comparación intergrupal”. P.p. 199-221 In Identidad social. Aproximaciones psicosociales a los grupos y a las relaciones entre grupos, Ed. By J. Francisco Morales , Darío Páez , Jean-Claude Deschamps and Stephen Worchel . Valencia: Promolibro. Hobolt, Sara B. , Thomas J. Leeper and James Tilley . 2021. “Divided by the vote: Affective polarization in the wake of the Brexit referendum”, British Journal of Political Science 51(4): 1476-1493. 10.1017/S0007123420000125. Huddy, Leonie and Omer Yair . 2021. “Reducing Affective Polarization: Warm Group Relations or Policy Compromise?”, Political Psychology 42(2): 291-309. 10.1111/pops.12699. Iyengar, Shanto , Gaurav Sood and Yphtach Lelkes . 2012. “Affect Not Ideology: A Social Identity Perspective on Polarization”, Public Opinion Quarterly 76(3): 405-431. 10.1093/poq/nfs038. Iyengar, Shanto , Yphtach Lelkes , Matthew Levendusky , Neil Malhotra and Sean J. Westwood . 2019. “The Origins and Consequences of Affective Polarization in the United States”, Annual Review of Political Science 22(1): 129-146. 10.1146/annurev-polisci-051117-073034. Lees, Jeffrey and Mina Cikara . 2021. “Understanding and Combating Misperceived Polarization”, Philosophical Transactions of the Royal Society B 376: 1-8. 10.1098/rstb.2020.0143. Levendusky, Matthew S. and Dominik A. Stecula . 2021. We Need To Talk. How Cross-Party Dialogue Reduces Affective Polarization. Cambridge: Cambridge University Press. Levendusky, Matthew. S. and Neil Malhotra . 2016. “(Mis)perceptions of Partisan Polarization in the American Public”, Public Opinion Quarterly 80(S1): 378-391. 10.1093/poq/nfv045. Linville, Patricia W. 1982. “The complexity-extremity effect and age-based stereotyping”, Journal of Personality and Social Psychology 42(2): 193-211. 10.1037/0022-3514.42.2.193. Mason, Lilliana . 2018. Uncivil agreement: how politics became our identity. Chicago: The University of Chicago Press. Mernyk, Joseph. S. , Sophia L. Pink , James N. Druckman and Robb Willer . 2022. “Correcting Inaccurate Metaperceptions Reduces Americans’ Support For Partisan Violence”, PNAS 119(16): 1-9. 10.1073/pnas.2116851119. Mildenberger, Matto and Dustin Tingley . 2019. “Beliefs about Climate Beliefs: The Importance of Second-Order Opinions for Climate Politics”, British Journal of Political Science 49(4): 1279-1307. 10.1017/S0007123417000321. Miles, Joseph R. , and Hannah J. Shinew . 2022. “A Breakdown (and rebuilding) of Intergroup Dialogue, Group Dynamics: Theory, Research, and Practice 26(3): 274-287. 10.1037/gdn0000190. Moore-Berg, Samantha L. , Lee-Or Ankori-Karlinsky , Boaz Hameiri and Emile Bruneau . 2020. Exaggerated meta-perceptions predict intergroup hostility between American political partisans, PNAS 117(26): 14864-14872. 10.1073/pnas.2001263117. Moore, Christopher . 1995. El proceso de mediación. Métodos prácticos para la resolución de conflictos. Buenos Aires: Ediciones Granica. Mullen, Brian , Rupert Brown and Colleen Smith . 1992. “Ingroup bias as a function of salience, relevance, and status: An integration”, European journal of social psychology 22(2): 103-122. 10.1002/ejsp.2420220202. Myers, C. Daniel . 2023. “Issues, Groups, or Idiots? Comparing Theories of Partisan Stereotypes”, Public Opinion Quarterly 87(3): 635-66. 10.1093/poq/nfad038. Nyhan, Brendan and Jason Reifler . 2010. “When Corrections Fail: The Persistence of Political Misperceptions”, Political Behavior 32: 303-330. 10.1007/s11109-010-9112-2. Orr, Lilla V. and Gregory A. Huber . 2021. “Measuring Misperceptions: Limits of Party-Specific Stereotype Reports”, Public Opinion Quartely 85(4): 1076-1091. 10.1093/poq/nfab062. Rahn, Wendy M. 1993. “The Role of Partisan Stereotypes in Information Processing about Political Candidates”, American Journal of Political Science 37(2): 472-496. 10.2307/2111381. Rivera Otero, Xosé Manuel , Lagares Díez, N. 2022. “Emociones y política: Introducción”, Revista Española de Ciencia Política 58: 13-18. Rogowski, Jon C. and Joseph L. Sutherland . 2016. “How ideology fuels affective polarization”, Political behavior 38: 485-508. 10.1007/s11109-015-9323-7. Rojo Martínez, José Miguel , Ismael Crespo Martínez and Alberto Mora Rodríguez . 2023. “Culture Wars, Perception Gap and Affective Polarization: An Approach from the Spanish Case”, OBETS. Revista de Ciencias Sociales 18(1): 79-96. 10.14198/obets.21976. Rothschild, Jacob. E. , Adam J. Howat , Richard M. Shafranek and Ethan C. Busby . 2019. “Pigeonholing Partisans: Stereotypes of Party Supporters and Partisan Polarization”, Political Behavior 41: 423-443. 10.1007/s11109-018-9457-5. Rubin, Mark and Miles Hewstone . 1998. “Social Identity Theory's Self-Esteem Hypothesis: A Review and Some Suggestions for Clarification”, Personality and Social Psychology Review 2(1): 40-62. 10.1207/s15327957pspr0201_3. Rudolph, Thomas J. and Marc J. Hetherington . 2021. “Affective Polarization in Political and Nonpolitical Settings”, International Journal of Public Opinion Research 33(3): 591-606. 10.1093/ijpor/edaa040. Ruggeri, Kai , Bojana Većkalov , Lana Bojanić , Thomas L. Andersen , Sarah Ashcroft-Jones et al . 2021. “The general fault in our fault lines”, Nature Human Behavior 5: 1369-1380. 10.1038/s41562-021-01092-x. Sherif, Muzafer , O.J. Harvey , B. Jack White , William R. Hood and Carolyn W. Sherif . 1988. The Robbers Cave Experiment. Pennsylvania: First Wesleyan Edition. Sherman, David K. , Leif D. Nelson y Lee D. Ross . 2003. “Naï Realism and Affirmative Action: Adversaries are More Similar Than They Think”, Basic and Applied Social Psychology 25(4): 275-289. 10.1207/S15324834BASP2504_2. Spears, Russell and Sabine Otten . 2013. “Discrimination: Revisiting Tajfel´s minimal group studies”. P.p. 164-182 In Social Psychology. Revisiting the classic studies. Ed. By Joanne R. Smith and S. Alexander Haslam . London: Sage. Tajfel, Henri , Michael Billig , Robert P. Bundy and Claude Flament . 1971. “Social categorization and intergroup behaviour”, European Journal of Social Psychology 1(2): 149-178. 10.1002/ejsp.2420010202. Tajfel, Henri . 1981. Human groups and social categories. Cambridge: Cambridge University Press. Thomsen Jens Peter Frølund and Anna Håland Thomsen . 2022. “Intergroup Contact Reduces Affective Polarization But Not Strong Party Identifiers”, Scandinavian Political Studies 46: 241-263. 10.1111/1467-9477.12242. Turnbull-Dugarte, Stuart J. , José Rama and Andrés Santana . 2020. “The Baskerville's dog suddenly started barking: voting for VOX in the 2019 Spanish general elections”. Political Research Exchange 2(1): 1781543. 10.1080/2474736X.2020.1781543. Tversky, Amos and Kahneman, Daniel . 1974. “Judgment under Uncertainty: Heuristics and Biases”, Science 18: 1124-1131. 10.1126/science.185.4157.1124. Webster, Steven W. and Alan I. Abramowitz . 2017. “The ideological foundations of affective polarization in the US electorate”, American Politics Research 45(4): 621-647. 10.1177/1532673X17703132. Whitt, Sam , Alisandra Yanus , Brian McDonald , John Graeber , Mark Setzler , Gordon Ballingrud , y Martin Kifer . 2021. “Tribalism in America: Behavioral Experiments on Affective Polarization in the Trump Era”, Journal of Experimental Political Science 8(3): 247-259. 10.1017/XPS.2020.29. Wilson, Anne. E. , Victoria A. Parker and Matthew Feinberg . 2020. “Polarization in the Contemporary Political and Media Landscape”, Current Opinion in Behavioral Sciences 34: 223-228. 10.1016/j.cobeha.2020.07.005. Yudkin, Daniel A. , Stephen Hawkins and Tim Dixon . (2019). The perception gap: How false impressions are pulling Americans apart.New York: More in Common. https://osf.io/preprints/psyarxiv/r3h5q/ ANNEX 1 Summary of frequencies and descriptive statistics on variables considered in the study.<xref ref-type="table-fn" rid="twf-16-e261">*</xref>.
Variable Descriptive frequencies
Sex Male = 594, 48.1% Female = 642, 51.9%
Age 18-30= 176, 14.2% 31-44 = 290, 23.5% 45-64 = 450, 36.4% 65 and + = 320, 25.9% X̅ = 50.12 SD = 16.176 Range = [18-93]
Level of education (dummy). University studies = 513, 42.0% No university studies = 709, 58.0%
Ideology (eliminating the central categories, recoding of the self-placement scale as a dichotomous variable). Left (1-4) = 473, 38,3% Right (7-10) = 260, 21,0% Missing (5-6): 503, 40,7%
VOX sympathy/rejection feeling scale (dependent variable). Scale from 0 to 10 where 0 is “dislike and rejection” and 10 is “sympathy and adhesion”. X̅ = 2.49 SD = 3.253 Range = [0-10] N = 1212
PPID PSOE (recoding from the feeling scale, taking scores 8-9-10 as indicative of adherence). Yes PPID = 151, 12.5% No PPID = 1060, 87.5%
PPID PP (recoding from the feeling scale, taking scores 8-9-10 as indicative of adherence). Yes PPID = 150, 12.4% No PPID = 1063, 87.6%
PPID VOX (recoding from the feeling scale, taking scores 8-9-10 as indicative of adherence). Yes PPID = 150, 12.1% No PPID = 1062, 87.6%
PPID UP (recoding from the feeling scale, taking scores 8-9-10 as indicative of adherence). Yes PPID = 109, 9.0% No PPID = 1099, 91.0%
TV consumption (scale 1-5, where 1 is “every day” and 5 is “almost never or never”). Every day = 63,4% Three to four days per week = 16,7% Almost never or never = 19,9%
Digital newspapers consumption (scale 1-5, where 1 is “every day” and 5 is “almost never or never”). Every day = 35,7% Three to four days per week = 10,9% One to two days per week = 11,4% Less frequently = 6.8%. Almost never or never = 35.1%.
Social media consumption (scale 1-5, where 1 is “every day” and 5 is “almost never or never”). Every day = 38,0% Three to four days per week = 10,2% One to two days per week = 11,1% Less frequently = 6,4% Almost never or never = 34,3%
Source: Created by the authors with data from CEMOP’s Second National Survey of Political Polarization.

The frequencies relative to MP are shown in Table 1.

 ANNEX 2 Multiple linear regression analysis of the effect that misperceptions on feelings towards PSOE<xref ref-type="fn" rid="fn-8-e261">7</xref>.
Feelings towards PSOE
B (E) β p VIF
Positive partisan identity (PPID) (ref.: no PPID)
PPID PP .038 (.336) .005 1.526
PPID VOX -1.458 (.336) -.179 *** 1.583
PPID UP 1.488 (.293) .179 *** 1.149
Ideology (ref.: right) 2.706 (.320) .429 *** 2.380
Sex (ref.: male) .458 (.203) .077 ** 1.076
Level of education (ref.: no university studies) .115 (.205) .019 1.097
Age .020 (.007) .108 *** 1.250
Political news consumption
TV -.114 (.071) -.057 1.160
Social media .024 (.064) .014 1.205
Digital newspapers .001 (.064) .001 1.200
MP PSOE -.221 (.126) -.073 1.631
MP PP -.107 (.094) -.045 1.444
MP VOX .089 (.099) .038 1.627
MP UP .175 (.105) .069 1.586
Constant 1.603 (.558) ***
R2 .400
Adjusted R2 .385
N (according to the list) 571

p<0.01,

p<0.05.

Numbers in parentheses are standard errors.

 Multiple linear regression analysis of the effect of misperceptions on feelings towards PP<xref ref-type="fn" rid="fn-9-e261">8</xref>.
Feelings towards PP
B (E) β p VIF
Positive partisan identity (PPID) (ref.: no PPID)
PPID PSOE .573 (.257) .069 ** 1.193
PPID VOX .210 (.304) .025 1.585
PPID UP -.145 (.269) -.017 1.188
Ideology (ref.: right) -3.585 (.266) -.545 *** 2.019
Sex (ref.: male) .143 (.184) .023 1.077
Level of education (ref.: no university studies) .108 (.185) .017 1.097
Age .010 (.006) .055 1.268
Political news consumption
TV -.174 (.064) -.083 *** 1.161
Social media .003 (.057) .001 1.204
Digital newspapers -.014 (.057) -.008 1.201
MP PSOE .282 (.114) .090 ** 1.630
MP PP -.653 (-.085) -.264 *** 1.444
MP VOX -.145 (.090) -.059 1.629
MP UP -.086 (.095) -.032 1.590
Constant 7.529 (.502) ***
R2 .550
Adjusted R2 .539
N (according to the list) 571

p<0.01,

p<0.0.5

Numbers in parentheses are standard errors.

 Multiple linear regression analysis of the effect that misperceptions on feelings towards UP<xref ref-type="fn" rid="fn-10-e261">10</xref>.
Feelings towards UP
B (E) β p VIF
Positive partisan identity (PPID) (ref.: no PPID)
PPID PSOE 1.501 (.276) .174 *** 1.153
PPID PP -.587 (.332) -.065 1.525
PPID VOX -.567 (.332) -.064 1.585
Ideology (ref.: right) 2.757 (.319) .401 *** 2.431
Sex (ref.: male) -.031 (.201) -.005 1.079
Level of education (ref.: no university studies) .311 (.202) .048 1.097
Age -.009 (.007) -.044 1.290
Political news consumption
TV -.018 (.070) -.008 1.161
Social media -.056 (.062) -.029 1.187
Digital newspapers .031 (.063) .016 1.194
MP PSOE -.272 (.124) -.083 *** 1.620
MP PP .347 (.091) .134 *** 1.396
MP VOX .258 (.098) .100 *** 1.627
MP UP -.106 (.104) -.038 1.590
Constant 1.311 (.557) ***
R2 .506
Adjusted R2 .493
N (according to the list) 571

p<0.01,

p<0.05.

Numbers in parentheses are standard errors.

The survey was conducted via a computer-assisted telephone interviewing (CATI) system to a representative sample of the Spanish population of both sexes over 18 years of age. The survey sample included 1,236 cases. The sampling error is ± 2.8% for a confidence level of 95.5% (two sigmas) and P = Q. We used a stratified multi-stage sampling procedure with the final selection of individuals according to sex, age and strata created by crossing the 17 Spanish autonomous communities (plus Ceuta and Melilla) with various habitat sizes.

The variable “vote+sympathy” is widely used in Spain to identify group affiliation and generate vote-forecasting models. This variable includes those individuals who state their direct intention to vote for a party and those who have not decided who to vote for or simply prefer not to state it but who do consider themselves as sympathisers with a party. This variable can increase the number of cases to be included in the analysis via the mentioned aggregate.

Given that 3 of the 4 issues can be considered symbolic property of left-wing parties (Rojo et al., 2023), we analyse the standard deviation in the actual position on each issue held by each group of voters to conclude whether the perception gap may be caused by a greater heterogeneity or ambiguity of right-wing voters towards issues related to feminism, gender violence and environmentalism as opposed to the greater internal coherence of left-wing voters around these issues. X ¯ FEM PSOE = 3.90, sd = 2.730; X ¯ FEM PP = 5.62, sd = 2.706; X ¯ FEM Vox = 7.12, sd = 3.135; X ¯ FEM UP = 2.71, sd = 2.878; X ¯ ECO PSOE = 4.01, sd = 3.003; X ¯ ECO PP = 5.05, sd = 2.689; X ¯ ECO Vox = 5.41, sd = 2.914; X ¯ ECO UP = 2.68, sd = 2.796; X ¯ GENVIO PSOE = 1.14, sd = 2.118; X ¯ GENVIO PP = 2.31, sd = 2.845; X ¯ GENVIO Vox = 3.590, sd = 3.275; X ¯ GENVIO UP = .75, sd = 1.434; X ¯ INMIGRA PSOE = 3.81, sd = 2.643; X ¯ INMIGRA PP = 5.61, sd = 2.363; X ¯ INMIGRA Vox = 6.96, sd = 2.410; X ¯ INMIGRA UP = 2.57, sd = 2.257. In general, the levels of standard deviation are not too dissimilar. In fact, it is not on all issues that right-wing voters have the highest standard deviation.

This variable is obtained by recoding the 1-10 ideological self-positioning scale as a dichotomous variable for the regression models, in which “1” groups individuals positioned on the left (1-4) of the 1-10 ideological self-positioning scale and “0” groups together those positioned on the right (7-10). Central positions are considered missing values (5-6). Recoding of the initial scale occurs in order to capture what is really intended to be captured. In the case of ideology, the starting point is whether identification with one or another ideology (left or right) influences feelings towards a party. This is not a question of the intensity of ideology within each of these spaces (greater or lesser extremism). We cannot forget that the concept of ideology redirects to qualitative labels that are treated quantitatively through a scale of 1-10, which is certainly artificial. Our purpose is to distinguish and compare the effects of identity and the spatial location of individuals. Therefore, we need to separate and dichotomously confront those individuals located on the left (1-4) with those located on the right (7-10).

This variable is the result of dichotomously recoding the feeling thermometer of sympathy/rejection towards parties and positioning those individuals who show a sound level of sympathy to a party in “1” (8, 9 10), which lets us consider them as positively identified, and the others in “0”. It is not a matter of considering how an individual moves on the feeling thermometer as a predictive element. It is not a question of scale or intensities. The variable is a condition, which is either present or absent. We need to differentiate those individuals who have strong feelings of attachment and sympathy, and who therefore show a clear positive partisanship, from those who do not.

The frequency that political information is consumed via different media sources was measured on a scale of 1-5 with 1 being “every day”, 2 being “three or four days a week”, 3 being “one or two days”, 4 being “not frequently” and 5 being “almost never or never”.

The model explains the 68.1% variance of the scale of feelings towards Vox (adjusted R2 = .681), meaning that the model’s goodness of fit and predictive ability are considered optimum. Moreover, we conclude that the equation is significant - the explanatory variables have a joint and linear influence on DV - (F = 88.901 p<0.01). We include a multicollinearity diagnosis confirming that the values referring to the variance inflation factor (VIF) are never greater than five (5) for any variable. Lastly, we observe that the highest Cook’s distance value for the model (residuals) is .067 with an average of .002, meaning that we rule out issues relating to the model caused by an excessive influence of atypical residuals.

The model explains the 38.5% of the variance of the scale of feelings towards PSOE (adjusted R2 =.385), meaning that the model’s goodness of fit and predictive ability are considered optimum. Moreover, we conclude that the equation is significant - the explanatory variables have a joint and linear influence on DV (F= 26.454 p<0.01).We include a multicollinearity diagnosis confirming that the values referring to the variance inflation factor (VIF) are never greater than 5 for any variable. Lastly, we note that the highest value of the Cook’s distance measure for the model (residual statistic) is .054 with a mean of .002, so we rule out issues relating to the model caused by an excessive influence of atypical residuals.

The model explains the 53.9% of the variance of the scale of feelings towards PP R2 =.539), meaning that the model’s goodness of fit and predictive ability are considered optimum. Moreover, we conclude that the equation is significant - the explanatory variables have a joint and linear influence on DV (F= 48.512 p<0.01). We include a multicollinearity diagnosis confirming that the values referring to the variance inflation factor (VIF) are never greater than 5 for any variable. Lastly, we note that the highest value of the Cook’s distance measure for the model (residual statistic) is .047 with a mean of .002, so we rule out issues relating to the model caused by an excessive influence of atypical residuals.

The model explains the 49.3% of the variance of the scale of feelings towards PP (adjusted R2 =.493), meaning that the model’s goodness of fit and predictive ability are considered optimum. Moreover, we conclude that the equation is significant - the explanatory variables have a joint and linear influence on DV (F= 40.619 p<0.01). We include a multicollinearity diagnosis confirming that the values referring to the variance inflation factor (VIF) are never greater than 5 for any variable. Lastly, we note that the highest value of the Cook’s distance measure for the model (residual statistic) is .053 with a mean of .002, so we rule out issues relating to the model caused by an excessive influence of atypical residuals.