This study used a hybrid approach. Quantitative techniques were first used to make a relevant selection in order to limit the scope of the qualitative analysis to the comment threads which had the greatest potential for political discourse associated with Messi. The comments were then assessed in depth in order to identify the use of hate speech and also predictably complex content patterns within a potentially valuable but also manageable volume of text.

The Cartodio database was used to apply web scraping techniques to download more than one million news items, with their corresponding comments sections, from the newspapers 20minutos, ABC, AS, El Mundo, El País, La Vanguardia, Marca and Mundo Deportivo for the period 2019-2020. The database did not include all the news items from 2019: it excluded the whole year of Mundo Deportivo and the first eight months of El País, because these newspapers moved the comments to an external server that web scraping could not access. Nevertheless, it was possible to expand the data collection from ABC and El Mundo to 2021, which extended the study with these two newspapers to include the end of Messi's time at F.C. Barcelona. A broad shortlist was made from this database of all available news items in which the word ‘Messi’ appeared in the headline. This search resulted in 2,639 news items and almost 250,000 comments to work on (Figure 1).

Figure 1. News items and comments that included ‘Messi’ in the headline

Source: Developed by the authors.

Subsequently, six comment threads were selected from the Madrid newspapers and another six from the press published in Barcelona that had the highest prevalence of terms that would be potentially useful for our analysis. The terms sought were *catalanuf*, *cataluf*, *democracia*, *democratic*, *democrátic*, *independ*, *preso*, *procés*, *prosés*, *prucés*, *prprusés*, *polític*, *politic* (Tables 1 and 2). The search (and therefore, this list), which only included news items from Marca, El Mundo, La Vanguardia and Mundo Deportivo, may have returned false positives (any use of the adverb ‘independently’ [independientemente] would be selected by the filter, for example) and/or false negatives (any comment of a political nature that did not use any of these terms would have been excluded). However, this was not a relevant limitation, as the quantitative phase simply identified comment threads where the volume of discussion on the topic of the procés and the pro-independence movement were significant. Once the threads with the most political discussion were located, the qualitative analysis of the comments automatically eliminated the impact of Type I and Type II errors inherent in the quantitative phase.

3. Results

3.1. Political comments by type of newspaper and location

The quantitative selection of the 12 most important threads in Madrid and Barcelona, respectively, showed considerable differences between the comments sections in news articles published in each of the two cities. The incidence of political discussions about the procés and the Catalan pro-independence movement was overwhelmingly higher among newspaper users in Madrid compared to Barcelona: 95% compared to 5% (Figure 2). The Madrid-based newspapers published a greater volume of news about Messi in the period analysed: more than four times as much. While this was partly due to the greater number of news publications in Madrid, if this effect is eliminated by basing the ratio on averages per newspaper (Table 3), the imbalance is still remarkable: almost double. In Madrid, there was also a higher volume of comments on news items about Messi, regardless of whether politics was discussed or not. This higher average volume was found when analysed by news item, by newspaper and also overall. Proportions ranged between 6.1 and 1 for general comments and from 7.1 and 1 for specifically political comments (Table 3).

Figure 2. Frequency of political terms related to the procés and the Catalan pro-independence movement found in comments

Source: Developed by the authors.

Of the total number of comments in news articles on Lionel Messi that were related to the procés and the pro-independence movement, 71% were found in the sports press, compared to 29% in the general circulation press (Figure 2).

The news item about Messi with the highest number of political comments was the one related to the Sant Jordi Cross award that the player received. This was for a specific reason: as the audience applauded when slogans were shouted in support of the pro-independence prisoners, Messi did not do so. El Mundo did not highlight Messi's attitude to the applause in the headline and simply published a reflection on what happened in a few descriptive lines within the information: ‘he was also the protagonist during the applause and chants for the release of the prisoners of the procés when, unlike those present, he crossed his hands and did not join in other people’s gesture’ (Cabezas, 2019). In the 157 comments, the presence of pro-independence opinion makers was almost non-existent. This absence of a counterbalance marked the comment thread. There was speculation in the comments as to why Messi did not applaud and what it meant. The vast majority of comments conveyed the belief that Messi's gesture was against pro-independence interests. Most users felt that Messi was ‘neutral on the issue of independence’ (#130) or that he was ‘neither in favour nor against’ (#123). However, Messi's neutral position did not mean that the commentators took the same view: ‘he strongly dislikes these displays that destroy coexistence’ (#116; Cabezas, 2019).

When the same news item was published in Marca, the dynamics in the comments changed: they involved conflict. Firstly, there was mass participation, with a total of 1,299 comments. Secondly, there was a general tone that saw Messi's gesture in a positive light, as reflected in the 51 times the word ‘great’ was used to describe the player. Thirdly, as Marca is a sports newspaper, it has no ideological affiliation as such, although it is more identified with the city of Madrid and specifically with Real Madrid, which means that there were some pro-independence comments. Regarding the type of political comments, Marca users could be clearly divided into two groups: those who were in favour of independence and those who were against independence. The former were considerably fewer were distributed over the entire duration of the thread (almost two days). All these comments were posted by eight users. Some were more active, whereas others participated only very occasionally and their presence in the discussions was fleeting. Two of them even overlapped with each other and participated simultaneously, although they tended to take turns, which allowed for a sustained tension between positions throughout the thread. Those who were against independence believed that Messi did not clap because he was surfeited: ‘Messi, like many Catalans and non-Catalans living in Catalonia, is totally fed up with this issue. They’re riff-raff!’ (#59; Navarro, 2019) or ‘Good for him for sending them to hell with that gesture’ (#813). Another user saw it more as a test of his personality, but closely associated with their reading of what was happening in Catalonia: ‘they won’t brainwash you’ (#193). Others did not dare to go too far: ‘many Catalans who see the awkward attitude of their idol will identify with it’ (#525). And there were those who took the opportunity to criticise him: ‘what happened to him is that he got scared, like he did at Anfield’ (#193; Navarro, 2019).

The views of pro-independence supporters were remarkably different. Some assessed his gesture as neutral (‘he wants to stay out of it’ (#986) or ‘they seem to suggest that Messi, instead of applauding, was giving the finger’ (#986). Take a good look at the video, he didn’t clap but he showed respect’ (#922)); whereas there were some who chose to justify his decision by arguing that it was based on fear of what television channels could do to him if he had applauded: ‘[they would have been] throwing bombs and fake news against Messi for more than 10 years and demanding that he went to prison for sedition... On the side of the baddies, Messi knows that nothing happens’ (#907; Navarro, 2019).

3.2. Hateful insults and rhetoric

The presence of moderators who deleted messages was detected In Mundo Deportivo, La Vanguardia and Marca. In El Mundo there was no trace of this activity: all messages were numbered consecutively. This absence of a moderator translated into a higher number of disparaging comments, especially in news item number 4 (Table 1), where expressions such as ‘sufre separata rata’ (suffer, you separatist rat) and ‘AmiMeMola, muérete!’ (ILikeIt, just drop dead!) were published. Even so, the analysis as a whole revealed relatively few direct insults or derogatory words considering the volume of material reviewed. This type of language was found in comments referring to Messi, who was described as ‘an autistic dwarf’, ‘semi-literate’, ‘illiterate’ and ‘borderline’ (Cabezas, 2021a), and was accused of being ‘traitor and collaborator of the year’ (#1059; Navarro, 2019). Similarly derogatory language was directed at other users who posted in the threads, including ‘Sois paletos (...)’ (you are all rednecks), #117; Cabezas, 2019)), and also referring to all those in favour of independence: ‘clowns’ (Ruipérez, 2020); ‘(...) he has not let himself be manipulated by that “bunch” of “dimwits, bumpkins and troglodytes” (Cabezas, 2019); ‘(...) this way we will prevent spawns like Junqueras and Torra from coming out (#246; Navarro, 2019), or ‘there is no field without a cricket and no idiot without a yellow ribbon’ (#934; Navarro, 2019). Being fed up with this political issue also led to comments that criticised both sides: ‘I say that it’s equally stupid when independence supporters bring politics into football as when someone like you (who is not pro-independence) does the same (Mundo Deportivo, 2020).

Sometimes, the response to these hateful terms and rhetoric was a degree of self-restraint by the commentators themselves by appealing to equanimity, as in ‘don’t feed the Troll’ (Martinez, 2020), by helping to neutralise a participant, which occasionally also resorted to verbal violence. This was done by creating a kind of boomerang effect, as the original post was used to disparage its originator. A user (Cento) who had a history of aggressive comments in the same thread, wrote a hateful message: ‘I will continue until there is not even the smell of anything Catalan left in Spain. Everything Catalan out of Spain’ (#440; Cabezas, 2021a). This message resulted in four aggressive comments of rejection. One user reminded him that ‘anything Catalan, or all raving lunatics like you’ (#444). Another participant elaborated the argument further in the post, but also disapproved of the comment: ‘It must be very sad to live with such moral decay; believe me, I am sorry for you and for those who have to endure being near you, best regards’ (#484). Cento complained about the treatment he received and the ‘insults’: ‘Curious, is it moral decay to want the best for my country? I think Catalans are ruining coexistence in my country. You are the one displaying moral decay as you try to prevent me from expressing my opinion through insults’ (#497). In all this discursive process, which lasted more than six hours, one user supported Cento and suggested a boycott of Catalan products and ‘a fair divorce settlement for them to get the hell out [of Spain] straight away’ (#451). However, the last messages about the conversation, those that held a certain higher ground in the creation of final meaning because they were not answered or questioned by anyone, reinforced the rejection of his hateful message: ‘not all Catalans are like Puigdemont and his gang, but it seems that you are too short-sighted to understand that’ (#501) and ‘Crackpot!’ (#504).

The same user (Cento) had written a similar message months before, also in El Mundo: ‘What the Catalans have to do is to leave Spanish competitions, or better still, all Catalans should leave Spain; it's very hard to put up with these people on a daily basis. For a Spain free of Catalans’ (#9; Cabezas, 2021b). Again, three users attacked this message in less than an hour and reminded Cento that not all Catalans were in favour of independence: ‘I very much regret your statements and, even more, your lack of solidarity with the Catalans who have views like mine, there are a few millions of us’ (#72).

3.3. Ways of reading a thread: consequences

In order to understand the influence that hate speech messages can have on threads, it is important to discuss the two reading patterns identified among users. On the one hand, there were some who replied to commentators who issued their views close to the time when they decided to participate. This way of interacting with the thread means that they were not influenced by virtual conversations or comments that came long before: they did not read in a narrative way, from the beginning to the end. Therefore, users who chose to stay in the thread for a long time or leave and come back later to continue participating had a greater influence than others who posted and read on an ad hoc basis. A user participated in the same thread at three different points in time times with an interval of more than six hours to introduce the same idea:

‘Puigdemont is roaming around freely in Europe. Why?’ (#103. 10:57)
‘Josu Ternera, a fugitive from justice, lived in hiding and didn’t appear in court. Puchy [Puigdemont] appeared before Belgian and German courts; the civil guard is not watching him, as far as anyone knows, and everyone knows where he lives’ (#231, 11:44).
‘Why was Josu Ternera extradited and Puchy was not? What's going on?’ (#823. 17:02; Marca, 2019)

Within this reading dynamic, the more an idea appears, the more influential it is, because it can reach more people without them reading other messages that refute or qualify theirs. In contrast, other users read the thread narratively, going back to comments that were posted a long time ago, even as far as the beginning. This gave them the perspective of several or many virtual conversations, not just one or two. Moreover, despite the fact that a comment was a long way into the thread, and there was little or no chance of continuing the conversation, many users chose to leave a reply to support or counter someone else's opinion; and those messages were also read, despite the disconnection between those creating the debate. In all the threads there were some conversations that were hours or days apart: comment 190 of one thread replied to number 7 and did so almost a day later (Cabezas, 2021c); number 660 replied to number 28, which was made more than 5 hours earlier (Navarro, 2019); the user who wrote comment 436 to reply to number 2 did not notice the almost 12-hour interval between messages (Cabezas, 2021b), etc. Users continued to reply to messages more out of the need to express themselves or to have the upper hand in the ultimate meaning of the conversation than to communicate with the author of the original message.

Further evidence that the comment thread was often a narrative in itself, which users wanted to follow from the beginning, could be found in the top-rated comments. Systematically, the users whose post contents were the most widely accepted were the ones who participated first; thus, they gathered support not only from readers who read and wrote at the same time as they did, but also from those who came later and started reading from the beginning. Of the six comment threads analysed from the Madrid press, four of them were above or close to 500 messages (Marca, #1299 and #737; El Mundo, #534 and #488) and two were around 200 (El Mundo, #157 and #193). Only 3% of those who contributed beyond post 50 in these threads were among the most supported. If a user wrote one of the first 10 posts, they were more likely to be selected as one of the most strongly backed-up comments: 48% of the most widely supported posts were among the first ten who participated; and 77% were among the first 25 to take part.

4. Discussion

The press in Madrid paid more attention to Lionel Messi in the period analysed than the press in Catalonia: a higher volume of comments in news items about Messi were found in Madrid, regardless of whether they linked to politics or not. Of the total number of comments related to the procés and the pro-independence movement, 71% of the contributions were found in the sports newspapers, compared to 29% in general newspapers (RQ1). The sports press featured a greater number of political comments, as well as more conflicts and debates compared to the general press due to its audience's diverse ideologies and territorial sensitivities, contrasting with the more homogeneous ideological alignment of the general press audience. Once the differences due to the greater number of Madrid-based news publications studied had been discounted, the average number of political comments in the news was 7.1 to 1 in favour of Madrid users, as opposed to the abstention of those who posted messages in the Catalan press. Barcelona newspapers and users, whether out of self-censorship or sense of exhaustion with the issue, or out of a simple desire not to mix heterogeneous topics, refrained from making political references to Messi, or from using the news and comments to analyse or discuss political issues such as Catalan independence, and did not mention Messi in political discussions. On the other hand, the Madrid media did not shy away from political issues in some news reports, and users of these newspapers resorted to the comments section to engage in discussion on politics and the pro-independence movement in relation to Messi (RQ1). Beyond some occasional insults, the comments did not target Messi, but they did use some actions or events in his life to undermine the legitimacy or authority of pro-independence views and at the same time introduce words or arguments associated with hate speech.

In the Madrid press, newspapers had a decisive influence on the type of users who posted messages. The rhetorical dynamics and the tone of the comments did not change depending on whether a newspaper was a general or a sports one, but there were differences in the variety of the reflections displayed in the threads. In the reports in El Mundo, ideologically linked to the centre-right, pro-independence users were almost non-existent, which meant that virtual conversations were almost monologues, lacked major conflicts or substantial differences, while Marca, a sports newspaper with no ideological affiliation, had overwhelmingly more political discussions due to conflicts between users who were for and against the independence of Catalonia from Spain. Although the former were fewer in number, their presence was enough to generate discussion throughout the thread (RQ2).

In the sample analysed, online discussions were always incomplete; there was no closed narrative; there were no conclusions and the participants who posted the comments were not willing to understand or assimilate new arguments; on the contrary, they remained anchored to their views, although they did look for rhetorical formulas to gain an advantage and refute the arguments of their opponents. In threads with a higher number of comments there was more conflict and also more hostility (Ksiazek, 2018). Readers who accessed the comments had two reading patterns: they engaged on a casual basis (reading only the comments that were close to the time when they decided to participate) or as a narrative (reading from the first comment). The number of arguments or rebuttals a user could read conditioned their construction of meaning about the information and tolerance of hate speech. Although the measures taken by the media reduce direct insults, it is really complicated to prevent uncivil discourse, because ‘the richness of the language, the context of interpretation and even in some cases unintentional biases cause comments that would be part of hate speech to remain undetected’ (Paz-Rebollo et al., 2021, p. 6). Still, the more a reader follows any given thread, the more likely they are to notice the real impact that hate speech has on successive debates and how most participants react to the tone and substance of those comments. There was a boomerang effect in the threads: while hateful messages might always have had some support, they were more often rejected by the majority. If such comments were properly neutralised, their presence could lead to a greater and deeper rejection of the ideas they conveyed than if they did not appear in the thread.

Threads contained several virtual conversations running simultaneously, sometimes more than 50 over the course of the entire thread if it was long enough. The presence of uncivil comments, clearly identifiable as hate content, caused anger (Ksiazek et al. 2015) but only in part because, despite the presence of hateful messages, some conversations continued long enough to show a majority rejection of such content by number of speakers and rhetorical quality (RQ3). It is true that certain responses also revealed hateful content in the way they were expressed; this seems inevitable: ‘We cannot expect everyone to act rationally online’ (Lee, 2016, p. 1). But a comment with clear signs of hate speech may not be negative (Naab et al., 2020, p. 13) if it appears as defeated by a counterargument or simply by the cancellation of its response through silence or the joint action of several users in the dynamics of the conversation (RQ3).

The analysis of virtual conversations and the fact that the most valued comments were systematically the first ones in the thread (77%) suggests that the narrative reading of the thread is widespread, although more studies would be necessary to assess how many readers chose to read a thread as a narrative, and how many briefly stepped in and out. It would also be important to study to what extent the presence of comments with hate speech traits was disruptive or, on the contrary, users took a degree of aggressiveness in virtual conversations for granted and set boundaries when it became excessive and self-regulated, confirming the boomerang effect. It would be interesting to analyse under what circumstances and for what types of messages users were more likely to cancel hate speech comments themselves.

Contribution by authors

José Cabeza: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Project administration, Resources, Supervision, Validation, Writing - Original draft, and Writing - review & editing.

Raúl Casado: Conceptualization, Data curation, Formal analysis, Methodology, Resources, Software, Supervision, Validation, Writing - Original draft and Writing - Review & editing.

Manuel Gómez: Conceptualization, Data curation, Formal analysis, Investigation, Validation, Writing - Original draft and Writing - Review & editing.

All authors have read and agree with the published version of the manuscript.

Conflicts of interest

The authors declare that they have no conflicts of interest.

Funding

This research is part of research project PID2019-105613GB-C31, entitled ‘Cartografía de los discursos del odio en España desde la Comunicación’, funded by the Ministry of Economy and Competitiveness (Spain).

References

Barnes, Renee (2015). Understanding the Affective Investment Produced Through Commenting on Australian Alternative Journalism Website New Matilda. New Media & Society, 17(5), 810–826. https://doi.org/10.1177/1461444813511039.

Beckert, Johannes & Ziegele, Marc (2020). The Effects of Personality Traits and Situational Factors on the Deliberativeness and Civility of User Comments on News Websites. International Journal of Communication, 14, 3924–3945. https://tinyurl.com/47f3b49r

Cabezas, Francisco (16 de mayo de 2019). Messi recibe la Creu de Sant Jordi por los servicios prestados a Cataluña. El Mundo. https://tinyurl.com/y9tuty4v

Cabezas, Francisco (5 de agosto de 2021a). Leo Messi cierra su etapa en el Barcelona. El Mundo. https://tinyurl.com/2s372kuc

Cabezas, Francisco (31 de enero de 2021b). Messi y el Barça anuncian acciones legales contra EL MUNDO tras desvelar el mayor contrato de la historia del deporte. El Mundo. https://tinyurl.com/bduatpt4

Cabezas, Francisco (7 de agosto de 2021c). Messi, una cuestión de vida o muerte en el Barça: La supervivencia pasaba por su salida. El Mundo. https://tinyurl.com/y9bhv83t

Coe, Kevi; Kenski, Kate & Rains, Stephen (2014). Online and uncivil? Patterns and determinants of incivility in newspaper website comments. Journal of Communication, 64(4), 658-679. https://doi.org/10.1111/jcom.12104

Dohle, Marco (2018). Recipients’ Assessment of Journalistic Quality. Digital Journalism, 6(5), 563–582. https://doi.org/10.1080/21670811.2017.1388748.

Ervedosa, Clara (2018). ‘Elective affinities’: interdiscursive dynamics between football, the economy and nationalism in Germany. National Identities, 20(1), 41-62. https://doi.org/10.1080/14608944.2018.1530647

Gal, Noam (2019). Ironic humor on social media as participatory boundary work. New Media & Society, 21(3), 729–749. https://doi.org/10.1177/1461444818805719

Gómez García, Salvador & Ballesteros-Herencia, Carlos (2019). Deporte en el éter: las emisiones deportivas de Radio Nacional de España en los inicios del franquismo. Materiales para la Historia del Deporte, 19, 106-116.

Hwang, Hyunseo; Kim, Youngju & Kim, Yeojin. (2018). Influence of discussion incivility on deliberation: An examination of the mediating role of moral indignation. Communication Research, 45(2), 213-240. https://doi.org/10.1177/0093650215616861

Juventeny, Silvia (2017). Much ‘more than a club’: Football Club Barcelona’s contribution to the rise of a national consciousness in Catalonia (2003–2014). Soccer & Society, 10(1), 103-122. https://doi.org/10.1080/14660970.2016.1267624

Ksiazek, Thomas (2018). Commenting on the News. Explaining the degree and quality of user comments on news websites. Journalism Studies, 19(5), 650-673. https://doi.org/10.1080/1461670X.2016.1209977

Ksiazek, Thomas; Peer, Limor & Zivic, Andrew (2015). Discussing the News: Civility and Hos-tility in User Comments. Digital Journalism, 3(6), 850–870. https://doi.org/10.1080/21670811.2014.972079

Kümpel, Anna-Sophie & Springer, Nina. (2016). Commenting Quality: Effects of User Comments on Perceptions of Journalistic Quality. Studies in Communication and Media, 5(3), 353–366. https://doi.org/10.5771/2192-4007-2016-3-353

Lee, Eun-Ju (18 de abril de 2016). Despite Flaws, Comments are Good for Public Discourse. The New York Times. https://tinyurl.com/3eb7u3e4

Lee, Eun-Ju; Jang Yoon-Jae & Chung, Myojung (2020). When and How User Comments Affect News Readers’ Personal Opinion: Perceived Public Opinion and Perceived News Position as Mediators. Digital Journalism, 9(1), 42-63. https://doi.org/10.1080/21670811.2020.1837638

Martínez, Juan-Bautista (4 de septiembre de 2020). Messi se queda a disgusto en el Barça y cumplirá su contrato. La Vanguardia. https://tinyurl.com/4m39rzt2

Naab, Teresa, Heinbach; Dominique, Ziegele Marc & Grasberger, Marie-Theres (2020). Comments and Credibility: How Critical User Comments Decrease Perceived News Article Credibility. Journalism Studies, 21(6), 783-801. https://doi.org/10.1080/1461670X.2020.1724181

Navarro, Cristina (11 de septiembre de 2019). Cardoner: Messi puede entrar y salir del club cuando crea oportuno. Marca. https://tinyurl.com/2swbjcky

Newman, Nic; Fletcher, Richard; Kalogeropoulos, Antonis; Levy, David & Nielsen, Rasmus-Kleis (2017). Reuters Institute digital news report 2017. Oxford: Reuters Institute for the Study of Journalism. https://doi.org/10.60625/risj-5srx-vt08

Paz-Rebollo, María-Antonia; Cáceres-Zapatero, María-Dolores & Martín-Sánchez, Isabel (2021). Suscripción a la prensa digital como contención a los discursos de odio. Profesional de la información, 30(6), 1-13. https://doi.org/10.3145/epi.2021.nov.13

Paz-Rebollo, María-Antonia; Montero-Díaz, Julio & Moreno-Delgado, Alicia (2020). Hate speech: a systematized review. Sage open, 10(4). https://doi.org/10.1177/2158244020973022

Purcell, Kristen; Rainie, Lee; Mitchell, Amy; Rosenstiel, Tom & Olmstead, Kenny (2010). Understanding the Participatory News Consumer. Pew Research Internet Project. https://tinyurl.com/53w8rr2m

Redacción Marca (17 de mayo de 2019). El gesto de Leo Messi mientras todos coreaban Libertad, presos políticos. Marca.com. https://www.marca.com/futbol/barcelona/2019/05/17/5cde62f4ca47417d7d8b45c9.html

Redacción Mundo Deportivo (3 de noviembre de 2020). El City, listo para volver a la carga por Messi en enero. Mundodeportivo.com.https://www.mundodeportivo.com/futbol/fc-barcelona/20201103/49194840038/messi-manchester-city-precontrato-enero.html#:~:text=Después%20de%20marearle%20el%20pasado,capitán%20azulgrana%20puede%20empezar%20a

Redacción Sport (18 de diciembre de 2022). ¿Cuántos títulos tiene Lionel Messi? Sport.es. https://www.sport.es/es/noticias/barca/palmares-titulos-leo-messi-barcelona-argentina-7365831

Rojas-Torrijos, José-Luis & Guerrero-Caballero, Miguel (2021). ¿Un periodismo deportivo apolítico?: Estudio de las referencias al ‘procés’ catalán en la prensa deportiva digital en España. Estudios sobre el Mensaje Periodístico, 27(3), 947-960. https://doi.org/10.5209/esmp.72554

Rowe, Ian (2015). Deliberation 2.0: Comparing the deliberative quality of online news user comments across platforms. Journal of Broadcasting & Electronic Media, 59(4), 539–555. https://doi.org/10.1080/08838151.2015.1093482

Ruipérez, Carles (27 de agosto de 2020). La FIFA ampara a Leo Messi. La Vanguardia. https://tinyurl.com/2byavzpn

Santana, Arthur (2016). Controlling the Conversation. The availability of commenting forums in online newspapers. Journalism Studies, 17(2), 141-158. https://doi.org/10.1080/1461670X.2014.972076

Waddell, T. Franklin (2020). The Authentic (And Angry) Audience. Digital Journalism, 8(2), 249-266. https://doi.org/10.1080/21670811.2018.1490656

Wang, Sai (2020). The Influence of Anonymity and Incivility on Perceptions of User Comments on News Websites. Mass Communication and Society, 23(6), 912-936. https://doi.org/10.1080/15205436.2020.1784950

Wang, Sai (2021). Moderating uncivil user comments by humans or machines? The effects of moderation agent on perceptions of bias and credibility in news content. Digital journalism, 9(1), 64-83. https://doi.org/10.1080/21670811.2020.1851279

Weber, Patrick; Prochazka, Fabian & Schweiger, Wolfgang (2017). Why user comments affect the perceived quality of journalistic content: The role of judgment processes. Journal of Media Psychology, 31, 24-34. https://doi.org/10.1027/1864-1105/a000217

Zamith, Rodrigo & Lewis, Seth (2014). From public spaces to public sphere: Rethinking systems for reader comments on online news sites. Digital journalism, 2(4), 558-574. https://doi.org/10.1080/21670811.2014.882066

ⁱ User comments are reproduced during the article in response to their original wording, including typos, onomatopoeias, or misspellings