Research on selective media exposure in Spain: a critical review of its findings, application phases, and blind spots

This paper is the first review of selective exposure studies in Spain, offering comprehensive insights into the diverse findings and analytical strategies employed over almost three decades by studies that explore the politically oriented media consumption of Spanish audiences. The article divides the evolution of this research line in Spain into two main phases –an initial phase (1995-2016) and a consolidation phase (2017-present)– according to the methodological approaches used during each period. Despite the challenges posed by working with secondary data and the lack of experimental designs, we note a robust initiation of this research line in Spain, accompanied by increasing methodological sophistication and diversification and a substantial accumulation of evidence on how Spanish audiences selectively consume like-minded news media. The paper also identifies the blind spots of selective exposure research in Spain, such as the scarcity of primary data sources, which overlooks phenomena such as online selective exposure and selective avoidance, the need to supplement multivariate analyses with other techniques that allow the direction of causal influence between variables to be established, and the significance of assessing the role of public media as facilitators or inhibitors of cross-cutting exposure.


Introduction: origin and relevance of selective exposure
Selective exposure is a classic theory in communication studies, positing that individuals exhibit a preference to consume media and content that aligns with their pre-existing viewpoints, while shunning content that contradicts their established beliefs.Its origin goes back to the theories of Mass Communication Research on limited media effects (Lazarsfeld;Berelson;Gaudet, 1948;Klapper, 1960).
"Actual exposure to partisan propaganda does not parallel availability.Availability plus predispositions determine exposure -and predispositions lead people to select communications which are congenial; which support their prior position" (Lazarsfeld;Berelson;Gaudet, 1948, p. 89).
Within the framework of the two-step flow of communication, where the impact of media is influenced by social groups, interpersonal communication, and opinion leaders, as noted by Klapper (1960), selective exposure emerged as a viable rationale for the limited effects observed, subsequently evolving into a foundational element of the reinforcement thesis.Even in laboratory situations (Cooper;Jahoda, 1947;Postman;Bruner;McGinnies, 1948;Mills;Aronson;Robinson, 1959), the social psychology research conducted during the 1940s and 1950s revealed that individuals exhibited a strong inclination to avoid, misinterpret, or swiftly forget messages that contradicted their pre-existing beliefs, indicating selective tendencies in exposure, perception, and memory processes.
The Selective Exposure Theory refers to the presence of any systematic bias in the composition of the audience (Sears;Freedman, 1967, p. 165), which explains the discrepancy between availability and effective consumption.Despite the frequent association of selective exposure with confirmation bias and its predominant examination in the context of politically-oriented media consumption, it is essential to understand that the selection of media content can be influenced not only by political biases, but also by factors such as sociodemographic and racial attributes (Knobloch-Westerwick, 2015a, 2015b).Nevertheless, it remains evident that selective exposure guided by political beliefs is unequivocally the most extensively documented phenomenon (Hart et al., 2009;Stroud, 2011).
The phenomenon of selective exposure has been researched through two primary methodological approaches (Feldman et al., 2013).
-The first and most common approach involves proving a correlation phenomenon, i.e., determining the ideological point of view of the message and then comparing it with that of the audience, usually through the combination of content analysis and survey or panel (Stroud, 2008).-The second, less common, approach is focused on demonstrating an influence process through experimental designs that prompt individuals to choose between different types of media messages -ideologically neutral, aligned with their political preferences, or openly contrary (Iyengar;Hahn, 2009;Knobloch-Westerwick et al., 2003).
In general, studies based on correlation analysis have yielded more empirical evidence than those based on experiments, whose results are not so unanimous.In their well-known literature review, Sears and Freedman (1967) argued that correlation alone is insufficient to assert that political leanings genuinely motivate selective exposure.The correspondence only proves the existence of de facto selectivity.Nonetheless, subsequent meta-analyses have demonstrated that experimental studies conducted between 1956 and 1996 also support the existence of selective exposure (D'Alessio; Allen, 2002).
Decades after the first conceptualisations by Lazarsfeld, Berelson and Gaudet (1948), the profound transformations in media environments, such as the emergence of social networks and digital media, the multiplication of media supply, and fragmentation and hybridisation of consumption, have sparked renewed interest in this line of research in the context of new informational environments (Stroud, 2018;Camaj, 2019;Humanes, 2019), and as a result, it is now one of the eight most extensively utilised perspectives in international communication studies (Bryant;Miron, 2004).Its contemporary relevance is further evidenced in the publication of monographs in high-impact journals over the past decade, such as: -Communication methods and measures (2013) https://www.tandfonline.com/toc/hcms20/7/3-40 -Media and communication (2019) https://www.cogitatiopress.com/mediaandcommunication/issue/view/130 In Spain, there has also been a robust initiation of research in this field, with numerous studies exploring the impact of political leanings on consumer preferences (Martín-Llaguno; Berganza, 2001; Fraile; Meilán, 2012; Humanes, 2014; Humanes; Mellado, 2017; Ramírez-Dueñas; Vinuesa-Tejero, 2020).Indeed, the expansion of media supply presents a widened opportunity structure for selective exposure to information (Skovsgaard;Shehata;Strömbäck, 2016), consequently increasing the likelihood of selective exposure (Fischer;Schulz-Hardt;Frey, 2008;Iyengar;Hahn, 2009).
In Spain, there has also been a robust initiation of research in this field, with numerous studies exploring the impact of political leanings on consumer preferences What are the underlying reasons that lead individuals to prefer media and content that align with their existing beliefs?Selective exposure has found support from various theories in cognitive psychology, which have served as the foundation for understanding this phenomenon (Stroud, 2011).According to the theory of cognitive dissonance, individuals have an inherent tendency to seek psychic-cognitive coherence, striving to maintain internal consistency in their beliefs.As a result, they systematically prefer messages that are congruent with their existing beliefs while avoiding content that induces dissonance (Festinger, 1957).
Subsequent theories, such as motivated reasoning (Kunda, 1990) and epistemic motivations (Kruglanski, 1989), centre on the underlying motivations that drive information-seeking behaviours, which, depending on the circumstances, may be influenced by the desire for accuracy or directionality, meaning the inclination to confirm pre-existing beliefs (Hart et al., 2009).Finally, the concept of "cognitive misers" (Kahneman;Slovic;Tversky, 1982) describes how individuals employ heuristics to optimise decision-making processes and how party identification and ideology play a particularly significant role as drivers of media consumption due to their personal relevance for individuals (Donsbach, 1991).
From a normative perspective, the significance of selective exposure can be understood in light of democracies' historical aim to provide individuals with exposure to a certain diversity (Sampedro, 2000).This exposure is intended to foster social cohesion and political tolerance among citizens (Mutz, 2002).However, the evolving media landscape, characterised by digital technologies and the vast array of user choices in current media environments, has shifted the traditional function of media in producing public space (Sunstein, 2003;Precht, 2010).
In this sense, there are well-founded arguments regarding the effects of politically-oriented media consumption on democracy.On one hand, we know that consuming political information increases citizens' political knowledge, increases their sense of political efficacy, their interest in public affairs, and fosters their political participation (Cantijoch;Jorba;San-Martín, 2008;Meilán, 2010).But, on the other hand, selective exposure to information has an impact on the perception of the public issues deemed relevant (Stroud, 2011).It also increases the intensity of political attitudes and their cognitive accessibility (Knobloch-Westerwick, 2012;Matthes;Schmuck, 2017), and contributes to processes of political polarisation (Stroud, 2010) and affective polarisation (Levendusky, 2013).
This phenomenon occurs because political opinions are underpinned by emotions (Stroud, 2011).Furthermore, political ideologies not only offer individuals a way of perceiving the world, but also grant them entry into an affective community (Arias-Maldonado, 2016, p. 100).Indeed, the growing prominence of affects in shaping subjectivity and political mobilisation (Arias-Maldonado, 2016), along with the emergence of identity politics, cannot be fully comprehended without considering the role of selective exposure in promoting and reinforcing identities (Knobloch-Westerwick; Meng, 2011).Because of the above, researching how various affective communities are exposed to information sources in hybrid media environments (Chadwick, 2013) becomes essential.This involves exploring how "identification with politically relevant groups, including political parties and national, ethnic, linguistic, or gender groups" (Huddy, 2013) influences the consumption of information, making it an imperative undertaking in communication studies.

Selection of the corpus
Carrying out a literature review requires defining the criteria used to gather the bibliographic corpus.In this particular review, the focus is on studies that: -investigate politically-oriented media consumption patterns of audiences, and -prioritise Spanish audiences as the primary focus of analysis and reference population.This entails research on both online and offline consumption, as well as studies focused on one or multiple types of media.In essence, the review considers research whose primary objective is to explore the relationship between ideological/political predispositions and media consumption patterns among Spanish audiences.This decision implies encompassing studies conducted by both foreign academics and researchers from the fields of political and social sciences.In disciplinary terms, the review incorporates studies not only from the field of communication but also from the political and social sciences (Fraile;Meilán, 2012;Ramírez-Dueñas;Vinuesa-Tejero, 2020;Cardenal et al., 2019a, etc.).The authors' academic background matters less; what's crucial is how their work contributes to our understanding of how Spanish audiences make selective choices, considering almost three decades of empirical research.The interdisciplinary nature of the research subject, which involves exploring the relationship between political beliefs and media consumption, along with its multidisciplinary empirical development both within and outside of Spain, necessitates this approach.
However, this approach excludes international comparative studies, even those that involve Spain as part of their analysis.These studies are invaluable for testing hypotheses in diverse contexts and advancing theoretical knowledge by identifying universal patterns and explaining differences between countries.However, their primary focus is not on specifically describing or explaining the behaviour of Spanish audiences.News consumption is, nevertheless, highly contextual (Feldman et al., 2013), meaning that the differences between media environments at the macro level significantly in-fluence individual decisions about media exposure (Clay;Barber;Shook, 2013;Skovsgaard;Shehata;Strömbäck, 2016;Steppat;Castro-Herrero;Esser, 2022).
Because of the above, this review includes those research articles and book chapters that meet the specified criteria.Although these publications may not be subject to the same review processes, we find it necessary to include them because the review spans a significant period, from the 1990s to the present, when journal publication was not as prevalent in Spanish academia (Fernández-Quijada;Masip, 2013), and when many of the initial works were published in the form of book chapters.Their quality should not be underestimated as a result.To avoid duplication, we excluded conference communications that subsequently resulted in publications in the form of articles and book chapters.With one exception: the work of Humanes and Mellado (2017), which was presented at an AIMCR conference, but was never published as an article or book chapter.
Sampling employed diverse strategies.First, a search was carried out in the bibliographic portal Dialnet using the Spanish terms "exposure", "selective" and "media", with a "social sciences" filter, following the strategy of other literature reviews (Ardèvol-Abreu, 2015).The search returned 39 documents, of which only 7 were relevant (Martín-Llaguno; Berganza-Conde, 2001; Humanes, 2014; 2016; Valera-Ordaz, 2018; Ramírez-Dueñas; Vinuesa-Tejero, 2020; Moreno-Moreno; Sanjurjo-San-Martín, 2020; Sintes-Olivella; Fondevila-Gascón; Berbel-Giménez, 2023).Secondly, a comprehensive bibliography of all these articles was thoroughly examined to identify additional works that met the specified criteria, following a snowball approach.Thirdly, recent publications were searched for on Google Scholar, focusing on authors who had previously conducted selective exposure studies.Finally, the corpus of papers analysed totalled 26 documents: seven book chapters, one communication at an international conference and 19 research articles.

The application of the selective exposure theory in Spain: emergence and consolidation
Compared to the extensive trajectory of the selective exposure theory in North America, starting from the initial theorisations by Lazarsfeld, Berelson and Gaudet (1948), Cooper and Jahoda (1947) or Postman, Bruner and McGinnes (1948), in Spain, its trajectory is significantly more recent.It is important to specify that the theory of selective exposure initially made its way to Spain in the 1980s and 1990s through informative works that discussed the theories of political Researching how various affective communities are exposed to information sources in hybrid media environments becomes essential and sociocultural effects of the media (Dader, 1983;Roda, 1990;Muñoz-López, 1993;Canel, 1998).However, it wasn't until the mid-1990s that the first empirical studies on the politically-oriented consumption of Spanish audiences were published.
Next, we will explore how and when the first empirical research on selective exposure emerged in Spain and how scientific production can be categorised into different phases around this line of research, paying attention to the dominant methodological strategies employed at each stage.
The empirical research on selective exposure in Spain can be categorised into two periods: a) an initial emerging phase , characterised by bivariate studies exploring how the political leanings on the left-right axis of audiences influence their consumption of various media types (primarily press and television, with some consideration for radio); and b) a second consolidation phase (2017-2023), characterised by a higher methodological sophistication, where bivariate studies gave way to multivariate analyses, the use of primary data sources advanced, and the predictors of media consumption studied became more diversified: positions on the left-right axis are complemented with variables associated with the centre-periphery cleavage, interest in politics, political polarisation processes, various sociodemographic attributes, and even the nature of the news (Table 2).The main characteristics and studies of each phase are described in greater detail below.
We note here that this categorisation is based on the methodological strategies dominant during each of the periods.This does not mean there are no other ways to categorise them, nor does it exclude the possibility of diagnosing development phases through methodological approaches that consider criteria such as the social and institutional history of the Spanish scientific community.

Emergency of research into selective exposure in Spain (1995-2016)
Pioneering studies on selective exposure in Spain were published in the late 1990s and early 2000s, and were the first bivariate explorations of the partisanship of Spanish press and television audiences during elections in the 1990s (Díez-Nicolás;Semetko, 1995;1999;Gunther;Montero;Wert, 2000;Martín-Llaguno;Berganza-Conde, 2001).Most of them used contingency tables and the chi-square statistic to analyse whether there is a statistical association between the ideology and/or vote of individuals, and their media consumption preferences (Fraile;Meilán;2012;Humanes, 2014).In addition to these studies, there is some isolated work on selective exposure by political elites, which also confirms the occurrence of politically-oriented media consumption among members of the Parliament of Catalonia (Lacasa-Mas; Jandura; Cano-Castells, 2014).
Gunther, Montero and Wert (2000) conducted a pioneering study analysing the political leanings of readers of major Spanish newspapers during the 1993 election.The research revealed significant differences in newspaper consumption based on electoral behaviour.They found that readers of El País in 1993 had a leftist ideological orientation and voted more often for PSOE and Izquierda Unida (IU), those of El Mundo were located on the centre-left and usually voted for Partido Popular (PP) and IU, and readers of ABC were centre-right and more often voted for PP.However, it is important to note that their findings were derived from a very small sample of readers (N=175) from the 1993 CNEP study, mainly due to the relatively low press consumption in Spain.As a result, several more years of research with larger samples from the Spanish population would be necessary to conclusively establish the partisanship of Spanish audiences.
The initial analyses on television's come from Díez- Nicolás and Semetko (1995;1999).They investigated the correlation between voting patterns and viewers' preferred television channel during the 1993 and 1996 elections.These studies were conducted when television had emerged as the primary source of information for the Spanish population, and private channels like Antena 3 and Telecinco had gained significant market influence.Their approach does not solely fit within the scope of selective exposure, as they examine the relationship between voting behaviour and media consumption in both directions.They explore whether television consumption influences voting choices or if the vote influences the selection of a particular television channel.Nevertheless, their findings indicate "a highly significant correlation between watching TVE1 and voting for PSOE, and watching Antena 3 and voting for PP" (Díez-Nicolás; Semetko, 1999, p. 167).
Their study's merit lies in conducting content analysis to examine the informational treatment provided by both channels towards the two parties, revealing a more favourable treatment of PSOE by TVE1 and a similar pattern by Antena 3 for PP.The theory of selective exposure initially made its way to Spain in the 1980s and 1990s through informative works that discussed the theories of political and sociocultural effects of the media Martín-Llaguno and Berganza (2001) investigated the relationship between vote recall and the consumption of print media and television during the 1996 general election.The study took place in a media landscape marked by the emergence of private television channels, namely Antena 3, Telecinco, and Canal+.Through a telephone survey involving over 300 individuals in Pamplona, the researchers observed significant differences in media preferences among different groups of voters.This outcome lends support to the hypothesis of selective exposure in Spain concerning both print media and television consumption.Similarly, the researchers confirmed that the influence of voting behaviour on preferences for the written press primarily occurs with national newspapers but not with regional publications.However, it's worth noting that these studies approached their interpretations with caution.While they establish a significant relationship between voting and media consumption, they refrain from determining the direction of influence: causal relationships can only be adequately addressed through experimental studies (Alcalde-Villacampa, 2003; Martín-Llaguno; Berganza-Conde, 2001; Díaz-Nicolás; Semetko, 1995; 1999).
The works of Fraile and Meilán (2012) and Humanes (2014) would significantly enhance and deepen the study of politically-motivated media consumption in Spain for two crucial reasons.
-Firstly, because they use data sources that allow analysis replication.In fact, both studies used post-election surveys, either from the Centre for Sociological Research (CIS) or from the Comparative National Elections Project 1 (CNEP), and Fraile and Meilán also incorporated a TNS survey conducted during the 2004 election.These surveys offered data from substantially larger and more representative samples of the Spanish population.-Secondly, because they analysed the connection between political leanings and media exposure, considering a wider range of media: television, press, and radio, and incorporating multiple time points of analysis.
The study conducted by Fraile and Meilán (2012) researched the association between ideological self-placement on the left-right axis and media consumption of press, radio, and television during the 1993 and 2004 general elections, as well as the 2009 European elections.Their findings offer substantial empirical evidence in support of the thesis of  3,Telecinco,Canal+), measured as an increased proportion of the audience who are ideologically aligned between 1993 and 2009.These results, however, rely on contingency tables and do not incorporate a specific statistical test to demonstrate these temporal differences.(2014; 2016) explored the selective exposure of Spaniards to television, radio, and the press, employing three indicators of political leanings -vote recall, closeness to political parties, and ideological self-positioning-during the 2008 and 2011 elections.Her results broadly confirm that politically-oriented media consumption occurs in Spain, in addition to two other major findings: 1) that the written press is the medium most conditioned by selective exposure when compared to radio and television, and 2) that the relationship between political beliefs and selective consumption is particularly robust for partisan identification variables: "vote recall and proximity to a party exhibit stronger associations than ideological leaning across all the three types of media" (Humanes, 2014, pp. 788-789).

The consolidation of the study of politically-oriented media consumption (2017-2023)
These 20 years of slow but steady development in empirical research on selective exposure in Spain is followed by clear consolidation, driven by four circumstances: -studies now display greater methodological sophistication primarily attributed to the incorporation of multivariate techniques; -the analysis of variables predicting media consumption becomes more diverse; -there is growing research into online selective exposure; and -studies become less reliant on secondary data, with a stronger emphasis on the design of ad hoc questionnaires and other data collection strategies.Here, the application of multivariate analysis techniques contributes significantly to the progress of this research field in Spain. 2 On one hand, it enables researchers to ensure that the correlation between political attitudes and media consumption is not influenced by other latent variables, such as sociodemographic attributes or interest in politics.On the other hand, it quantifies the impact of political leanings on exposure while accounting for the influence of other relevant variables, aligning with international studies (Castro-Herrero; Nir; It should be noted that this methodological advancement, which may include the use of qualitative techniques such as focus groups (Masip; Suau; Ruiz-Caballero, 2018), needs to be contextualised within the fact that Spanish studies on selective exposure have primarily relied on correlation analyses, and there are no experimental designs that provide individuals with various options that allow us to unambiguously determine their preference for ideologically related information (Feldman et al., 2013;Iyengar;Hahn, 2009;Knobloch-Westerwick;Meng, 2011).
As highlighted by Knobloch-Westerwick (2015a), selective exposure research relying on correlations, even when employing panel data and introducing multivariate controls, cannot completely rule out the possibility of other unknown factors influencing the observed patterns of covariance between political beliefs and media preferences.Only experimental research enables the manipulation of stimuli in controlled environments and permits making causal inferences between variables (Knobloch-Westerwick, 2015a, pp. 111-112).In Spain, however, there is only one study (Cardenal et al., 2019a) that uses real news events as quasi-experiments to study online information consumption.Nonetheless, there is a notable absence of laboratory designs or anything comparable to the well-established track record of experimental research in the US.On the other hand, Spanish correlation analyses are based on large and representative samples, enhancing their validity and enabling the generalisation of results to the general population (Clay; Barber; Shook, 2013, Knobloch-Westerwick, 2015a ).
Pioneering studies on selective exposure in Spain were published in the late 1990s and early 2000s, and were the first bivariate explorations of the partisanship of Spanish press and television audiences during elections in the 1990s Secondly, in recent years there has been a significant diversification of the independent variables studied in Spain.Thus, along with the classic positions on the left-right axis and partisan identification (Humanes; Mellado, 2017; Ramírez-Dueñas; Vinuesa-Tejero, 2020; Humanes; Valera-Ordaz, 2023), studies have revealed that media consumption in Catalonia is also influenced by factors such as national identity and positions on the centre-periphery axis (Valera-Ordaz, 2018; 2022; Sintes-Olivella; Fondevila-Gascón; Berbel-Giménez, 2023).Indeed, national identity, nationalism, and support for independence serve as robust predictors of media consumption in Catalonia.Individuals who feel more Catalan than Spanish and support independence display a significantly higher preference for regional and local media, while those who identify more nationally with Spain and oppose independence tend to consume state media to a much greater extent (Valera-Ordaz, 2018; Sintes-Olivella; Fondevila-Gascón; Berbel-Giménez, 2023).
Furthermore, research indicates that a climate of political polarisation contributes to the intensification of selective exposure.During the procés in Catalonia, for instance, the attitudes of Catalans towards independence progressively gained prominence as predictors of radio and television consumption.This occurred concurrently with the intensification of the conflict surrounding independence (Valera-Ordaz, 2022).The reasons behind this trend can be attributed to the continuous activation of these attitudes and the subsequent increase in their cognitive accessibility (Knobloch-Westerwick; Meng, 2011; Knobloch-Westerwick, 2012).This was further fuelled by the extensive media coverage of the conflict, leading to a notable polarisation of radio and television audiences (Martínez-Amat, 2020).In this sense, the Spanish context still offers unexplored cases of analysis, where multiple axes of political conflict coexist in regions such as the Basque Country, Galicia, and the Valencian Community.Investigating these cases further could deepen the understanding of how selective exposure, driven by identity-related factors, shapes media consumption patterns.
Other studies have raised questions about the significance of interest in politics as a predictor of selective exposure (Humanes; Mellado, 2017; Rodríguez-Virgili; Sierra; Serrano-Puche, 2022).These studies aim to ascertain whether selective exposure in the Spanish media system is predominantly driven by ideological and partisan factors or whether it is primarily guided by interest in politics, such that audiences are segmented based on their preference for information-oriented content versus entertainment-oriented content (Dahlgren, 2019 However, the data sources used by Spanish researchers do not provide detailed information on the specific contents consumed, they only capture the preferred media of the audiences.This limitation hinders a comprehensive evaluation of the extent of selective exposure guided by interest in politics.Despite this, certain studies do suggest that an interest in public affairs leads to increased exposure to channels like La Sexta regardless of political leaning.This may be due to La Sexta's higher presence of political programmes on its schedule compared to Antena 3 and Telecinco (Humanes; Valera-Ordaz, 2023).
Other studies suggest that political beliefs exert a direct effect-as postulated by selective exposure-but also an indirect effect-mediated by factors such as trust in the media and the perceived quality of journalistic brands-on media behaviour (Moreno-Moreno; Sanjurjo-San-Martín, 2020; Victoria- Mas et al., 2023).In this sense, findings have shown that ideology plays a role in shaping trust in the media, and it can even influence how individuals perceive the media as distributors of disinformation (Masip; Suau; Ruiz-Caballero, 2020b).Thus, those who align with left-wing positions have much more trust in media outlets such as La Sexta, El País, and Cuatro, while those who align with right-wing positions have more trust in El Mundo, Antena 3, and ABC.In fact, not many media sources are perceived as reliable by the entire Spanish population.And citizens of both the left and the right believe that non like-minded media sources are creators of disinformation (Masip; Suau; Ruiz-Caballero, 2020b).This polarised and agonistic perception of the Spanish media makes the Spanish context an ideal case study to investigate the role of ideological extremism and political antagonism as drivers of both exposure and selective avoidance.
Other studies highlight that during situations where there is greater need for cognitive guidance-such as the COVID-19 pandemic-information utility may outweigh confirmation bias as a driver of media consumption (Knobloch-Westerwick; Kleinman, 2012; Zoizner et al., 2022).The study carried out by Cardenal et al. (2019a) underscores that the nature of news impacts individuals' motivations when seeking information, concluding that selective exposure tends to increase when there is a surge in information consumption during significant news events (Cardenal et al., 2019a).
The application of multivariate analysis techniques contributes significantly to the progress of this research field in Spain. 2 On one hand, it enables researchers to ensure that the correlation between political attitudes and media consumption is not influenced by other latent variables, such as sociodemographic attributes or interest in politics.On the other hand, it quantifies the impact of political leanings on exposure while accounting for the influence of other relevant variables, aligning with international studies In addition to the diversification of independent variables, initial studies have been published analysing the extent of online selective exposure in Spain (Moreno-Moreno; Sanjurjo-San-Martín, 2020; Rodríguez-Virgili; Sierra; Serrano-Puche, 2022), albeit in a very limited way compared to the vast international literature.In general, the studies challenge the hypotheses of self-reinforcing information loops, echo chambers (Sunstein, 2003), and filter bubbles.They emphasise that there is significantly more cross-cutting exposure than expected (Masip;Suau;Ruiz-Caballero, 2018;Cardenal et al., 2019a;2019b).For example, Cardenal et al. (2019b) demonstrate, using web crawl data from 408 Spanish internet users, that while direct browsing on the internet does lead to an increase in selective exposure, the influence of Google in reducing selective exposure and promoting cross-cutting exposure is more significant.Along the same lines, other studies conclude that Spanish internet users take advantage of the ideological diversity of the media offered by the internet to display transversal consumption (Cardenal et al., 2019a).
Similar conclusions were found by Masip, Suau and Ruiz-Caballero (2018; 2020a).Their study, based on a survey of more than 6,000 digital media users and data from discussion groups, showed that, despite 50% of internet users subscribing to like-minded media, 33% also read news from opposing media to cross-reference and gain a better understanding of reality.Their findings further validate that incidental exposure to content shared by friends and acquaintances on social networks plays a crucial role in promoting cross-cutting consumption of political information (Masip; Suau; Ruiz-Caballero, 2018).
In other words, Spanish studies align with international findings that refute the existence of ideological insularity in online news consumption (Garrett, 2013), despite initial concerns (Doval-Avendaño; Badía-Lago, 2009).Instead, social networks serve as diversification engines, enabling individuals to discover new information (Barnidge; Peacock, 2019), and access a variety of news sources (Fletcher;Nielsen, 2018).
Finally, this stage of consolidation is marked by significant advancements in the use of primary sources, reflecting the growing capability of the research community to employ data collection strategies to explore selective exposure.

Limitations in the application of selective exposure in Spain
Despite this prolific trajectory, selective exposure research in Spain also faces certain limitations which will be outlined below.We will set out some of the blind spots and argue critically why it is essential to expand research in specific directions to develop a research agenda that empowers the academic community and gives it a deeper insight into selective exposure in Spain.
We will focus on five interrelated issues: -the challenges related to available data sources; -the limited research on online selective exposure; -the ongoing issue of endogeneity and the lack of experimental designs; -the necessity to explore selective avoidance through suitable techniques; and -the relatively understudied role of public media in fostering or inhibiting cross-cutting exposure.

Available data: reliance on secondary sources and scarcity of primary sources
Research on selective exposure in Spain has demonstrated a significant reliance on secondary data sources, primarily surveys and barometers conducted periodically by different research institutes and consortia.Table 3 systematically lists the data sources used in selective exposure studies in Spain, together with the type of source and the specific nature of the data.As is evident from the information presented, there is a notable scarcity of primary sources; only a limited In Spain, however, there is only one study (Cardenal et al., 2019a) , 2023).The significant advancement in using primary sources in recent studies on selective exposure can be seen.1995,1999) and TNS Demoscopia (Fraile;Meilán, 2012), have also been used sporadically.Such instances are less common and can be attributed to the direct affiliations of certain researchers with these companies, as the data they use remain inaccessible to the public.Contrarily, using studies from these companies is expensive and only within the reach of well-funded research groups (Masip; Suau; Ruiz-Caballero, 2018; 2020; Cardenal et al., 2019a;2019b).This is why a significant portion of the Spanish academic community has chosen to use freely accessible data generated by research institutes for statistical analysis.
Secondary sources, nevertheless, have certain benefits that have contributed to the robustness of the study of selective exposure in Spain.
-Firstly, its use safeguards against the numerous issues introduced by confirmation bias that exists in all scientific research.In essence, they provide greater assurance against researchers' inclination to confirm preconceived notions and unintentionally introduce biases in questionnaire design and data collection processes.-Secondly, surveys like CIS, CEO, or CNEP are built upon extensive and well-constructed population samples, thereby enhancing the potential for generalisations.-Finally, the media consumption questions found in these surveys, which are employed as dependent variables in the analyses, can be readily regarded as measures of habitual exposure rather than singular exposure.Because they typically inquire about interviewees' overall media preferences, this enhances the capacity to generalise the findings (Clay;Barber;Shook, 2013;Stroud, 2008).
Despite these advantages, working with secondary data also has obvious limitations.For instance, this type of survey, where interviewees are questioned about past behaviours, is more vulnerable to post hoc rationalisations compared to experiments where individual actions can be directly observed.Moreover, employing general surveys introduces limitations in the empirical investigation of selective exposure, as it confines researchers to test hypotheses within the predefined boundaries set by questionnaires.For example, the CIS had only included questions about preferred media (radio, television, and newspaper) since the 2008 election; it was only in the November 2019 post-election survey when a specific question about interviewees' favourite digital newspaper was introduced, which was not retained in subsequent surveys.This has led to a focus on traditional media consumption in most research, with limited exploration of online selective exposure.
The majority of electoral surveys of this nature do not collect specific media content and programmes consumed by individuals, nor the time they dedicate to them, This stage of consolidation is marked by significant advancements in the use of primary sources, reflecting the growing capability of the research community to employ data collection strategies to explore selective exposure.These developments are often facilitated through research project funding only preferred media.A lack of disaggregated content data makes evaluating the extent of selective exposure driven by political interest in Spain challenging, unlike research conducted in other countries (Skovsgaard;Shehata;Strömbäck., 2016).Nor do they account for the consumption of digital media (such as digital natives, social networks, etc.) or the deliberate avoidance of certain media due to ideological reasons.
The fact that Spanish researchers depend on general surveys carried out by other organisations has not facilitated an analysis of the phenomenon of selective exposure beyond the level of individuals' preferred media.As a result, there is a lack of understanding regarding the variables that explain why certain individuals prefer television news, others opt for morning magazines, and some choose radio talk shows as their sources of information.Similarly, the need to further diversify the independent variables remains unaddressed.To enable researchers to conduct in-depth analyses of how factors like the nature of news or trust in the media influence selective exposure, having access to data sources that include these types of questions is essential.
What is the situation of studies on audiences?Unlike surveys, audience studies do collect detailed data on content consumption, encompassing various media formats, types, duration, and frequency of consumption.
-Firstly, the issue lies in the expensive nature of access to this data, making it unavailable to researchers, as it is reserved for members only.For instance, gaining full access to the General Media Study conducted annually by the AIMC (Association for Communication Media Research), which involves a consortium of media, agencies, and advertisers, requires a minimum annual fee of 7,500 euros. 3-Secondly, the methodology of these studies is frequently opaque as they are designed primarily for commercial purposes rather than scientific research.Indeed, this is the case with measurements of television audiences conducted by Kantar Media using audimeters, and measurements of digital media audiences by Comscore, 4 whose methodologies are not made public (Zain, 2014).-Lastly, audience studies typically include sociodemographic variables such as sex and age, but they lack political variables like ideology and voting behaviour.This limitation makes these data sources unsuitable for selective exposure studies.As an example, the General Media Study conducted annually by the AIMC involves interviews with approximately 30,000 individuals, and while it is likely the most transparent audience study in Spain, providing both a summary of the questionnaire and a detailed technical file, 5 it records sociodemographic classification data but lacks variables related to the political positions of individuals.Establishing collaboration between the communication research community and the AIMC, possibly through some of its reference associations (such as AE-IC, SEP), could be highly beneficial, as such a collaboration could aim to include questions of sociological and political significance.This would make the General Media Study a very valuable data source, not only for the industry, but also for the scientific community.

Online selective exposure
The second important limitation pertains to the scarcity of studies exploring selective consumption of online information among Spanish audiences.Most of the research has primarily focused on analysing the consumption of traditional media.The lack of studies addressing this phenomenon contrasts with the considerable consumption of digital media and the significance of social networks as information sources for Spaniards.As an example, according to the 2022 Digital News Report, the internet is a source of information for 79% of Spaniards, social networks are regular sources of information for 56%, while television is for 59%, and print media is for only 26%.
Despite this data, there is a notable scarcity of studies analysing the phenomenon of online selective exposure in Spain (Cardenal et al., 2019a;2019b;Masip;Suau;Ruiz-Caballero, 2018;2020a).These studies cover aspects such as the consumption of digital media, exposure to content on social networks, or browsing facilitated by search engines like Google.
The reasons have been outlined above, directing us to the available data.Nonetheless, two research teams from Catalan universities stand out as notable exceptions, as their scientific output significantly contributes to our current understanding of the extent of the phenomenon in Spain.The team led by the late Pere Masip from Ramon Llull University, stemming from the communication field, and the team of Ana S. Cardenal from Universitat Oberta de Catalunya, specialising in political science, published groundbreaking studies on the phenomenon of echo chambers and filter bubbles.Their research utilised ad hoc online surveys (Cardenal et al., 2019a;2019b;Masip;Suau;Ruiz-Caballero, 2020a), discussion groups (Masip; Suau; Ruiz-Caballero, 2018), and even web tracking data from the Spanish internet user population (Cardenal et al., 2019a;2019b).et al. (2019a) used web browsing data tracking from a sample of 408 Spanish internet users and concluded that there are no echo chambers in the consumption of online political information.Their findings show that individuals use the diversity of online media to expose themselves to both like-minded and non-like-minded content.This differs from the US scenario where there is a consumption of centrist media (Gentzkow;Shapiro, 2011;Nelson;Webster, 2017), in Spain, the exposure is truly transversal, encompassing all kinds of media.Furthermore, they highlight that while direct consumption of digital media via the internet slightly increases selective exposure, the influence of Google in reducing

Cardenal
Employing general surveys introduces limitations in the empirical investigation of selective exposure, as it confines researchers to test hypotheses within the predefined boundaries set by questionnaires selective exposure is more significant (Cardenal et al., 2019b).In essence, Google not only prevents filter bubbles but also fosters cross-cutting exposure.Their findings also reveal that online selective exposure becomes more pronounced as the volume of information consumed increases during specific news events (Cardenal et al., 2019a).This emphasises the need to expand the study of selective exposure by considering the nature of the news, a task that remains pending in the Spanish academic realm.
Studies by Masip, Suau and Ruiz-Caballero (2018; 2020a) yield similar findings.Using a triangulated approach, incorporating a survey of over 6,000 digital media users and discussion groups, they found that while half of the users exclusively subscribe to like-minded media, 33% also engage with news from non-like-media outlets to gain a better understanding of the reality.Furthermore, their work reveals incidental exposure to diverse content shared by friends and acquaintances on social networks, depicting a cross-cutting consumption scenario that deviates from self-reinforcing information (Masip; Suau; Ruiz-Caballero, 2018).In fact, in a subsequent study, the authors observed a notable association between the intensive use of social networks and consumption of information not aligned with ideological beliefs, particularly on platforms such as Twitter (Masip; Suau; Ruiz-Caballero, 2020a).
Similarly, the comparative study by Rodríguez-Virgili, Sierra and Serrano-Puche (2022) highlights that a majority of German, British and Spanish consumers of digital media prefer unbiased news.However, their results also demonstrate that Spanish citizens are more inclined than their British or German counterparts to prefer like-minded media.Meanwhile, the study conducted by Victoria-Mas et al. (2023) identified the key dimensions of journalistic brands for readers of the Catalan digital press.It concluded that political congruency is one of the most significant aspects, indicating that brands are perceived as valuable when they align with users' political beliefs.
In summary, the limited Spanish studies on online selective exposure align with international research findings.They reveal that ideological insularity in online news consumption is less pronounced than anticipated (Garrett, 2013), influenced by factors like search engine mediation (Cardenal et al., 2019b), the role of social networks and weak social ties that foster diversification by facilitating exposure to diverse and plural information (Barnidge; Peacock, 2019; Fletcher; Nielsen, 2018).

The ongoing issue of endogeneity and the lack of experiments
Thirdly, researching selective exposure in Spain poses the problem of endogeneity, which refers to the challenge of establishing causal relationships between two phenomena-political beliefs and media consumption-where a single direction of influence cannot be definitively determined.Indeed, from a theoretical standpoint, it appears more plausible to view them as mutually influential phenomena.Political beliefs are not shaped in isolation from media exposure, and media consumption cannot be considered entirely impervious to the ideological orientations of individuals.
At the core of the endogeneity issue lies a crucial issue: Spanish studies rely solely on analysing correlations between political beliefs and media consumption.The vast majority use retrospective reports of media consumption through surveys, with direct observation measures, such as web tracking data, being used to a much lesser extent (Cardenal et al., 2019).There are currently no experimental designs that expose individuals to different types of media content (related, neutral, contradictory) to observe their behaviour, as seen in the American academic tradition, which has a long history of experimental studies (Iyengar;Hahn, 2009;Knobloch-Westerwick, 2003).
Experiments are the only way to rule out that the correlation between two phenomena is due to a latent variable.For this reason, when dealing with cases that involve endogeneity, estimating the parameters of the regressor variables may lack consistency, leading to potentially significant discrepancies between the estimated effects and the real ones.It is important to remember that in the history of empirical research on selective exposure, studies based on correlations have tended to yield stronger effects than those based on experiments (D'Alessio;Allen, 2002;Stroud, 2011).It is not a question of empirical research being hindered by methodological dogmatism; however, due to the lack of experimental designs, it is highly probable that the effects of selective exposure in Spain have been overestimated.
It is worth noting, however, that experimental designs are not exempt from challenges and limitations.Firstly, experimental designs carry the risk of overestimating variables that researchers can control within limited laboratory settings, whe-Establishing collaboration between the communication research community and the AIMC, possibly through some of its reference associations (such as AE-IC, SEP), could be highly beneficial, as such a collaboration could aim to include questions of sociological and political significance re subjects participate for a restricted period.Secondly, experimental designs require individuals select media in highly unrealistic choice contexts, often with a limited number of created media messages by researchers.These messages are categorised as consistent, neutral, or inconsistent, but there are always fewer options than what exists in the real world (Clay;Barber;Shook, 2013).The fact that these laboratory choice contexts do not resemble the real situation in which individuals make their decisions about media consumption constitutes a major limitation of the experimental approach (Knobloch-Westerwick, 2015a), since media selection is highly context-sensitive (Feldman et al., 2013).
Lastly, other techniques, while not as robust as randomised experiments, can help establish the direction of influence between political attitudes and media consumption.Panel studies and time series analyses, which involve monitoring a sample of individuals over time, enable the identification of behaviour patterns and changes in individuals' media consumption over time.These approaches present a valuable opportunity to advance selective exposure research in Spain.

Selective avoidance, a pending task
Research conducted by Ramírez-Dueñas and Vinuesa-Tejero (2020) provided confirmation of the influence of media consumption on voting behaviour during the 2015 and 2016 elections.In addition to discovering evidence of individuals selecting media outlets that align with their opinions, they emphasise "the limited openness of voters to engage with media that are not aligned, and even a growing inclination to reject media that are contrary to their political position (ideological or territorial)" (Ramírez-Dueñas; Vinuesa-Tejero, 2020, p. 23).
Humanes and Valera-Ordaz (2023) arrived at similar conclusions in their examination of the variables influencing media consumption among Spaniards over an 11-year period (from 2008 to 2019).They discovered that certain voter groups are more likely to consume related media, and voting is also linked to a notably reduced willingness to be exposed to non-related content.
"The findings of this study also suggest that certain media produce more polarisation because they simultaneously inspire exposure and selective avoidance by certain ideological groups (…) Although both selective exposure and avoidance occur simultaneously only at specific times, the existence of these media is likely to compromise exposure to diversity" (Humanes;Valera-Ordaz, 2023, p. 11).
This leads us to another limitation in the study of selective exposure in Spain: the scarcity of research on selective avoidance.In Spain, there are no studies exploring whether individuals intentionally avoid certain media due to political or ideological reasons.The absence of specific questions on explicit media rejection in the data sources used contributes to this limitation.Indeed, the CIS questionnaires used in the aforementioned analyses (Ramírez-Dueñas; Vinuesa-Tejero, 2020, 2021; Valera-Ordaz; Humanes, 2022; Humanes; Valera-Ordaz, 2023) only gather information on individuals' preferred media (e.g., "Which newspaper/radio station/television station do you prefer to follow political and electoral information during the campaign?").However, they do not capture data on comprehensive media consumption patterns, nor do they include specific inquiries related to media avoidance.
While selective exposure studies in Spain often highlight that certain ideological groups or voters tend to avoid consuming non-like-minded media, this interpretation is based on the negative sign coefficients observed in logistic regression models (Ramírez-Dueñas; Vinuesa-Tejero, 2020; Valera-Ordaz; Humanes, 2022; Humanes; Valera-Ordaz, 2023).However, it is important to note that just because PSOE voters preferentially consume El País, this does not necessarily imply that they avoid El Mundo, La Vanguardia or ABC.To determine which media individuals avoid consuming, one would need to know their media diets exhaustively or inquire about them in detail.
So, while the interpretation of the coefficients is technically valid (indicating that certain ideological groups are significantly less likely to be exposed to certain media), the fundamental issue lies in the design of the CIS post-election questionnaires and the nature of the questions they contain, which do not allow such inferences to be made.In fact, from a theoretical perspective, it is more likely that individuals simultaneously consume multiple media sources.For instance, they may read both a national and a regional newspaper or a newspaper and a television channel, as well as being exposed to messages and content when browsing the internet (Masip;Suau;Ruiz-Caballero, 2018;Cardenal et al., 2019a).Empirically validating the occurrence of selective avoidance necessitates experimental designs or, at the very least, correlational analysis that includes explicit questions about which media individuals intentionally avoid.
Therefore, the fact that some groups of voters have a lower probability of preferring media that are situated at their ideological antipodes does not constitute a general verification of the phenomenon of selective avoidance.As highlighted by certain researchers (Garrett, 2009;2013;Camaj, 2019), selective exposure does not necessarily entail systematic Despite this data, there is a notable scarcity of studies analysing the phenomenon of online selective exposure in Spain (Cardenal et al., 2019a;2019b;Masip;Suau;Ruiz-Caballero, 2018;2020a).These studies cover aspects such as the consumption of digital media, exposure to content on social networks, or browsing facilitated by search engines like Google avoidance of non-like-minded information; inferring the absence of cross-cutting and incidental exposure from variables that solely include a preferred media outlet is not possible.Ultimately, it is more plausible that the effects of selective exposure are offset by media diets that include the consumption of all types of media.
These studies do not diminish the significance of the aforementioned research; the findings indicating the lower likelihood of certain ideological groups consuming specific media (Ramírez-Dueñas; Vinuesa-Tejero, 2020, Humanes; Valera-Ordaz, 2023) prompt the research community to further explore how variables such as ideological extremism, political interest, (dis)trust in institutions, or partisan loyalty influence selective avoidance.To achieve this, questionnaires should be designed to include questions on avoidance and comprehensively capture interviewees' media diets.This is a particularly pertinent task for Spain, since the media are considerably politicised (Hallin;Mancini, 2004), its journalistic style tends to emphasise opinion over information, and professional autonomy is limited by the strong interdependence of the political and media systems (Humanes; Martínez-Nicolás; Saperas, 2013; Humanes; Roses, 2018).

Public media: facilitators of selective or cross-cutting exposure?
The quality of democracy is significantly influenced by cross-cutting exposure, which refers to citizens' encounters with diversity.Such exposure plays a crucial role in enhancing political tolerance (Mutz, 2002) and promoting consensus-building in complex societies (Mutz;Young, 2011).Although the recent proliferation of media options has made individual interest a significant driver of media consumption, research has shown that the structure of opportunities provided by each media system also impacts the likelihood of individuals being exposed to diversity (Skovsgaard; Shehata; Strömbäck, 2016).
For example, Castro-Herrero, Nir and Skovsgaard (2018) conducted a study to explore the influence of public service broadcasters' dominance in the media market of 27 European countries on individuals' cross-sectional exposure.They found that public media's tendency to offer a greater amount of political information contributes to increasing the likelihood of individuals encountering non-like-minded information.Their results indicate that there are fewer disparities in cross-cutting exposure between individuals with varying levels of political interest in countries with a robust public service.In contrast, in countries with a weaker public service, cross-cutting exposure tends to be more unequal.In the context of this study, Spain recorded mean values of cross-cutting exposure well below Scandinavian countries, but higher than those in Baltic countries.In a similar vein, Bos, Kruikemeier and De-Vreese (2016) demonstrated that Dutch public television mitigated citizens' selective exposure.As the only medium consumed by all ideological groups, it acted as a unifying force for social cohesion, despite the considerable levels of selective exposure observed in other public media.
Numerous studies have established that European public service broadcasters exhibit varying levels of popularity and politicisation (Brüggemann et al., 2014).In systems of polarised pluralism like Spain, public media tend to be politicised (Hallin;Mancini, 2004), and it is common for governments to use the state and regional broadcasting systems they administer to their advantage (Arriaza-Ibarra, 2013;Jones, 2007), leading to perceptions of partisanship by the public (Azurmendi; Muñoz-Saldaña; Ortega-Mohedano, 2018).Even the granting of licenses to private broadcasters has been used as a strategy by national and regional governments to favour specific media groups (Arriaza-Ibarra, 2013;Jones, 2007).
In Spain, there have been relatively few studies analysing the role of public broadcasting systems as facilitators or inhibitors of cross-cutting exposure, despite them often being criticised for their instrumentalisation.In the case of RTVE, several studies find patterns of selective exposure based on the party in government, but there has been no systematic review, only cross-sectional studies that analyse specific moments.
For example, Fraile and Meilán (2012) found that the ideological composition of Televisión Española audiences changed over time according to whichever party was in government."The most intriguing finding pertains to the composition of the public television audience, which shifted from predominantly left-wing in 1993 to right-wing in 2004, and then returned to being primarily left-wing in 2009, aligning with the governments in power during those periods" (Fraile;Meilán, 2012, p. 125).
The same trend was found for Radio Nacional de España (RNE), so that the majority ideology of its audiences changed with the changes of government between 1993 and 2009.
Subsequent studies have endorsed this trend.Thus, according to the 2010 Report on the journalistic profession in Spain (cited in Humanes, 2014, p. 788), the Spanish placed RNE and TVE on the centre-left in 2010 (4.7 and 4.5, respectively, on a scale from 1 to 10) after six years of Zapatero's socialist government.Furthermore, the bivariate analyses conducted by Humanes (2014) revealed statistically significant associations between voting for PSOE and feeling an affinity to this party, and the consumption of TVE in both 2008 and 2011 (Humanes, 2014, p. 791), although this pattern was not observed with the public channel La 2. In the case of RNE, there was also an association with voting for PSOE, but effects were smaller.
It is not a question of empirical research being hindered by methodological dogmatism; however, due to the lack of experimental designs, it is highly probable that the effects of selective exposure in Spain have been overestimated In the 2015 election, following four years of Rajoy's government, voting for the PP was associated with an increased likelihood of consuming TVE1 (Humanes; Valera-Ordaz, 2023), a trend that persisted in the 2016 election (Ramírez-Dueñas; Vinuesa-Tejero, 2020).Similarly, in April 2019, individuals located on the centre and on the right exhibited a higher probability of preferring TVE1 compared to those situated on the left (Valera-Ordaz; Requena; Pop, 2021).
Nevertheless, the pattern differed for RNE.In 2015 and 2016, the consumption of RNE was evenly distributed among groups of voters (Ramírez-Dueñas; Vinuesa-Tejero, 2020), and in April 2019, it also "has a transversal presence in the different segments of the ideological spectrum" (Valera-Ordaz; Requena; Pop, 2021, p. 167), and "it is not a reference for any specific political faction" (Valera-Ordaz; Requena; Pop, 2021, p. 171).In summary, exposure to the public television channel varies depending on the party in government, while public radio has been consistently consumed transversally by all Spaniards since 2015, regardless of their ideology and vote.
However, at a regional level, only the case of the Corporació Catalana de Mitjans Audivisuals (CCMA) has been investigated, even though some studies clearly indicate that, in general, regional television systems -such as Telemadrid and TV3are the least trusted media sources for the Spanish population in terms of reliability (Azurmendi; Muñoz-Saldaña; Ortega-Mohedano, 2018; Masip; Suau; Ruiz-Caballero, 2020).It is worth noting that most of the regional public media in Spain were established with the aim of reinforcing regional identities and promoting linguistic and cultural integration (Holtz-Bacha, 2015).In some cases, they were even conceived as tools for national construction (Guimerà, 2014;2018), with the CCMA serving as an exemplary case.
In this context, studies have shown that watching TV3 is associated with an increased likelihood of supporting Catalan independence (Serrano, 2013) and a feeling of national identification with Catalonia (Hierro-Hernández, 2012).Additionally, research has found that individuals who support independence are significantly more inclined to prefer Catalan public television over non-independence supporters (Valera-Ordaz, 2018; Sintes-Olivella; Fondevila-Gascón; Berbel-Giménez, 2023), and that selective exposure to TV3 and Catalunya Radio has grown substantially over the last few years (Valera-Ordaz, 2022).In summary, studies concur that in recent years, the Catalan public broadcasting system has been selectively consumed by individuals who support independence, identify exclusively as Catalan, and endorse nationalism.Conversely, it has been avoided by those who oppose independence and exclusively identify as Spanish.
Indeed, beyond the Catalan case, there is a lack of longitudinal or multivariate studies exploring the role of most regional broadcasting services (e.g., Telemadrid, Euskal Telebista, Corporación Radio e Televisión de Galicia, the old Canal Nou and the new À Punt, Radio y Televisión de Andalucía, etc.) as promoters or inhibitors of cross-cutting exposure.Despite being administered by the same party for extended periods, and being perceived as politicised by the public, research on their impact remains unexplored (Azurmendi; Muñoz-Saldaña; Ortega-Mohedano, 2018).Humanes (2014) highlights that in 2008 and 2011, there was a significant statistical association between voting for a specific party and consuming the regional public television channel in that area, such as Canal Sur (associated with PSOE), TV3 (CiU), Canal Nou (PP), Telemadrid (PP), TVG (BNG), and ETB1 (PNV and Amaiur).
These data provide a compelling invitation to conduct in-depth analyses of the role of public media.In this sense, the research community in Spain must make an effort to determine the extent to which these systems encourage cross-cutting exposure or are selectively consumed only by some ideological groups.After all, the idea of public service is to provide quality journalistic information to all citizens beyond their social, economic and political differences.This should mean that public television and radio stations serve the whole of society, not only the sectors related to the government that administers them.Conducting a comprehensive analysis of the factors that influence the consumption of public media in Spain, both at national and regional levels, and examining their trends over time, is an essential task in selective exposure studies.

Conclusions: an agenda for the future
Selective exposure theory remains highly relevant in international communication studies (Camaj, 2019;Humanes, 2019) due to the recent evolving information environments (Chadwick, 2013) and the increased availability of media options, enabling individuals to engage in politically-oriented and selective media consumption (Skovsgaard;Shehata;Strömbäck, 2016).
Over the past three decades, selective exposure research in Spain has witnessed significant growth and contributions from the fields of communication, political, and social sciences; with an initial emerging phase of bivariate studies (Gunther; Montero; Wert, 2000; Díez-Nicolás Semetko, 1999; Fraile; Meilán, 2012; Humanes, 2014), followed by a consolidated stage marked by methodological advancements, diversified independent variables, exploration of online selective exposure, and an increased commitment to primary data sources (Humanes;Mellado, 2017;Cardenal et al.,  Empirically validating the occurrence of selective avoidance necessitates experimental designs or, at the very least, correlational analysis that includes explicit questions about which media individuals intentionally avoid 2020; Rodríguez-Virgili; Sierra; Serrano-Puche, 2022; Humanes; Valera-Ordaz, 2023).Through numerous studies, it has been established with certainty that Spanish press, radio, and television audiences choose their media according to their political leaning.Additionally, there is provisional evidence suggesting that online information consumption exhibits less ideological insularity and more cross-cutting exposure than initially presumed (Masip;Suau;Ruiz-Caballero, 2020b;Cardenal et al., 2019b).
In this article we have argued that, despite the significant progress in empirical research of selective exposure in Spain, there are still some gaps in our understanding of the phenomenon.These limitations stem from the reliance on secondary data and the absence of experiments, which have hindered the further development of this research area in the country.
To ensure effective advancement in the study of selective exposure in the future, two critical steps must be taken.Firstly, researchers should prioritise primary data sources and secure adequate funding to collect original data with diversified variables of interest, using methods such as questionnaires, discussion groups, or experimental designs.Secondly, public research institutes should incorporate a wider range of questions related to media consumption, considering the increasingly complex contemporary informational environments.These measures will enhance the quality and scope of research in this field in Spain.Ultimately, we will only be able to adequately assess the influence of various variables (e.g., interest in politics, ideological extremism, nature of the news) as drivers of selective exposure, particularly in the context of media proliferation, by analysing disaggregated consumption data.
We emphasised the need for the research community in Spain to focus more on selective avoidance and online selective exposure phenomena.Presently, there is a lack of studies specifically addressing selective avoidance using appropriate techniques, partially due to limitations imposed by post-election surveys as data sources.However, some existing studies offer evidence of significantly lower consumption of certain media by specific ideological groups (Ramírez-Dueñas; Vinuesa-Tejero, 2020; Valera-Ordaz; Humanes, 2022; Humanes; Valera-Ordaz, 2023).
Online selective exposure has been explored in a preliminary and constrained manner in Spain through certain studies (Masip;Suau;Ruiz-Caballero, 2020b;Cardenal et al., 2019a;2019b).However, considering the significant influence of the internet and social networks as primary information sources for Spaniards, further comprehensive research is necessary.Continued research in this direction will strengthen initial findings and deepen our understanding of the factors influencing internet browsing patterns and digital media consumption.
We believe that the theory of selective exposure can be used comprehensively to evaluate the role of both national and regional public broadcasting services in the Spanish public sphere, aligning with international research (Bos; Kruikemeier; de Vreese, 2016; Castro-Herrero; Nir; Skovsgaard, 2018).Ultimately, democracies bear the responsibility of exposing citizens to quality journalistic information and fostering diversity, a task specifically entrusted to public media systems.
Advancing the understanding of politically-oriented media consumption will shed light on its relation to phenomena such as misinformation, populism, and political and affective polarisation.Recent studies indicate that restrictive media diets, characterised by exposure to a limited number of media with highly partisan content, can impact voting decisions and diminish tolerance towards other political parties (Ramírez-Dueñas; Vinuesa-Tejero, 2021).
In conclusion, we would like to note some of the limitations of this study and propose potential avenues for further research.Firstly, it is worth noting that certain studies on the selective exposure of Spanish audiences may not be captured in the results, as the Dialnet database might not encompass all scientific publications and this study has a broad timespan .Secondly, the growing number of international comparative studies that have not been included must be noted (there is an increasing integration of Spanish researchers into international teams financed by solid transnational research projects, such as Ana S. Cardenal, Laia Castro, and the researchers at the University of Navarra who collaborate with the Reuters Institute for the Study of Journalism).Future reviews should consider the social, institutional, and intellectual history of the Spanish scientific community; employing additional research techniques such as interviews with experts, the Delphi method, or discussion groups 6 that can provide insights into the direction of research on selective exposure in Spain.
Other noteworthy areas for advancing the analysis of selective exposure in Spain include the significant role played by the social and political sciences, evident in various publications (Díez-Nicolás;Semetko, 1995;1999;Fraile;Meilán, 2012;Ramírez-Dueñas;Vinuesa-Tejero, 2020;Cardenal et al., 2019a;2019b).This influence might extend to the adoption of analytical approaches and techniques in communication studies.Indeed, from the perspective of the sociology of science, exploring this line of work to understand the evolution of selective exposure studies in Spain would be of interest.
Additionally, analysing the impact of early empirical work on subsequent research as the consolidation phase progresses and more empirical evidence accumulates In Spain, there have been relatively few studies analysing the role of public broadcasting systems as facilitators or inhibitors of cross-cutting exposure, despite them often being criticised for their instrumentalisation This phase commenced in 2017 with the research conducted by Humanes and Mellado, which introduced the first multivariate analysis of the selective exposure of Spanish audiences, and it continues today.Firstly, studies increasingly apply multivariate techniques, such as binomial and multinomial logistic regression, to larger and more representative samples of the Spanish population through the use of CIS or Digital News Report surveys (Humanes; Mellado, 2017; Ramírez-Dueñas; Vinuesa-Tejero, 2020; Moreno-Moreno; Sanjurjo-San-Martín, 2020; Rodríguez-Virgili; Sierra; Serrano-Puche, 2022; Valera-Ordaz, 2022; Humanes; Valera-Ordaz, 2023).
These developments are often facilitated through research project funding.These methodologies involve designing and conducting online surveys targeting representative samples of the general population (Masip; Suau; Ruiz-Caballero, 2020b), internet users(Masip; Suau;Ruiz-Caballero, 2018;Cardenal et al., 2019a) and individuals registered with digital newspapers (Masip; Suau; Ruiz-Caballero, 2020a; Victoria-Mas et al.., 2023), carrying out discussion groups, and web crawling data to directly collect data from internet users on their media consumption patterns(Cardenal et al., 2019a;2019b).Thus, the study of selective exposure in Spain has become less reliant on general surveys and barometers conducted periodically by various research institutes and companies.These offer the advantage of being based on well-constructed and broadly representative samples of the population.However, they do not provide researchers with the ability to test hypotheses concerning the factors that underlie media consumption beyond the variables collected in the pre-designed questionnaires.It is worth clarifying, however, that secondary sources remain the most commonly used resource in selective exposure research in Spain (see, for example, Moreno-Moreno; Sanjurjo-San-Martín, 2020; Valera-Ordaz, 2018; 2022; Ramírez-Dueñas; Vinuesa-Tejero, 2020; Rodríguez-Virgili; Sierra; Serrano-Puche, 2022; Humanes; Valera-Ordaz, 2023; Sintes-Olivella; Fondevila-Gascón; Berbel-Giménez, 2023).

Table 3 .
Data sources used in selective exposure studies in Spain More recently, researchers from the University of Navarra have published studies based on data pertaining to Spain collected by the Digital News Report.This report is yearly compiled by the Reuters Institute for the Study of Journalism at the University of Oxford, who maintains a collaborative partnership with the University of Navarra.The group of researchers associated with the Center for Internet Studies and Digital Life, which includes Alfonso Vara-Miguel, Samuel Negredo, Avelino Amoedo-Casais, Elsa Moreno-Moreno and Jürg Kaufmann-Arqueta, compiled the aforementioned annual report spanning 2014 to today through an online survey carried out by YouGov on a sample of approximately 2,000 internet users(Vara-Miguel, 2022).Whilst confined to the Spanish online user population, this internet survey encompasses a range of questions on media consumption, sociodemographic attributes, and political factors, and has facilitated comparative analyses concerning individual motivations behind news exposure (Rodríguez-Virgili; Sierra; Serrano-Puche, 2022), and research exploring the influence of political leanings and journalistic branding on news consumption (Moreno-Moreno; Sanjurjo-San-Martín, 2020).Likewise, surveys of demoscopic companies, such as Sociological, Economic and Political Analysis (Díez-Nicolás; Semetko,