original scientific article UDC 316.755:004.738.5(460) received: 2014-06-01 WHEN THE MEDIUM IS ON THE MESSAGE: EXPLORING HATE IN MEDIAREADER INTERACTIONS IN SPANISH ONLINE SPORTS JOURNALISM Hibai LOPEZ-GONZALEZ University of Pompeu Fabra, Barcelona, Spain. Roc Boronat, 138. 08018. e-mail: hibai.lopez@upf.edu Frederic GUERRERO-SOLE University of Pompeu Fabra, Barcelona, Spain. Roc Boronat, 138. 08018. e-mail: frederic.guerrero@upf.edu ABSTRACT The aim of this article is to examine the way readers who participate in the comments section of online sports newspapers produce hate speech toward the medium. The research focuses on the Marca sport-based online community in Spain, in the context of the Real Madrid vs. FC Barcelona rivalry. The article explores the medium's quandary between allowing the free speech of his readers and therefore promoting the audience's engagement and the protection of its brand image, threatened by comments containing hate speech. Although only a small part of the readers addressed the medium through their comments, almost 75 per cent of them contained hate speech, accounting for a very violent discussion environment. Keywords: Participatory journalism, comments, sport, hate speech, hostile media, Spain. ko je medij sporočilo: raziskovanje sovrastva v interakciji med medijem in bralci v Španskem spletnem Športnem novinarstvu IZVLEČEK Namen tega članka je proučiti, kako bralci, ki pišejo komentarje pod prispevki v spletnem športnem časopisu Marca, tvorijo sovražni govor do medija. Raziskava se osredotoča na športno spletno skupnost v Španiji, in sicer v okviru rivalstva med Real Madridom in FC Barcelono. Članek raziskuje medijsko dilemo med svobodo govora svojih bralcev in spodbujanjem komentiranja ter zaščito blagovne znamke časopisa, ki jo ogrožajo komentarji s sovražnim govorom. Čeprav je le majhen del bralcev s svojimi komentarji nagovoril medij, je skoraj 75 odstotkov komentarjev vključevalo sovražni govor, ki ustvarja zelo nasilno razpravljalno okolje. Ključne besede: participatorno novinarstvo, komentarji, šport, sovražni govor, sovraštvo do medija, Španija. INTRODUCTION Online sports newspapers that host comments sections have to navigate between the Scylla and Charybdis of participatory journalism. On the one hand, a too zealous and conservative approach to the comments would reduce the number of user commenting and therefore the visits to the site and the money charged to the announcers. Also, the more the readers comment the deeper the involvement with the newspaper, making it a long-term investment in terms of loyalty and engagement with the medium. Nonetheless, on the other hand, a too permissive look at comments could damage the equity brand of the newspaper and result in a loss of trustworthiness (Di-akopoulos and Naaman, 2011). In addition, insults and hate speech attract the attention of some commentators but refrain others from entering the discussion. The opportunities that new technologies offer for reader participation in terms of interactivity and immediacy (Nielsen, 2013) need to be balanced with the popular perception that online comments hardly ever add to the debate, and what is more important, that very frequently impregnate the online discussion with derogatory language, hate and insults (Ruiz et al., 2010; Neu-rauter-Kessels, 2011; Ruiz et al., 2011). The dilemma of allowing or censoring certain opinions is especially notable in the case of the comments sections of online sports newspapers. Fans' involvement is an intrinsic component of sports spectacle and sports journalism, and the socializing and the community-building virtues that sports discussion entails cannot be undervalued (Lopez-Gonzalez et al., 2014). However, the passionate partisanship and sectarianism sometimes involved in sports chatter makes it convenient for the media to oversee it (Boyle, 2012). Therefore, this article aims at examining the relationship between news commentators and the online sports media hosting the discussion. We are specifically interested in the way fans produce hate speech toward a medium and how it manages that hatred. First, a quantitative analysis of comments was conducted, verifying if a media-reader relationship actually occurred. When that relationship existed, it was also quantitatively examined the media response to those comments of users addressing them. Second, a qualitative approach was followed to determine if user comments addressing the medium included hate speech. Finally and most importantly, those comments were examined to identify the main narratives articulating them. This work presents a local perspective and focuses on the Spanish context of football fandom, selecting as a case study one of the most important online sports community in Spain, Marca Community, hosted by Marca newspaper. Spanish sport landscape is dominated by football and is deeply divided into two contending sides: FC Barcelona supporters and Real Madrid supporters. We examined how the hate toward the medium took the shape of a fight between the two fan crowds and how Marca reacted to that hatred. THEORETICAL BACKGROUND Online comments, as a popular form of active audiences in the context of participatory journalism, has attracted the attention of a number of scholars in the last decade (Deuze, 2009; Domingo et al., 2008). Audience's comments challenge the traditional role of journalists as gatekeepers and force them to become 'gatewatchers' (Mitchelstein, 2011). The User-Generated Content (UGC) is often used by media outlets because it has become one of the most sought-after characteristics of online journals by the readers (Örnebring, 2008). In particular, newspapers have developed extensive sections of their Web pages based on UGC. But there is still relatively little discussion of the exact relationship between producing and consuming in these sections. What is being produced and what is being consumed? Does the blurring of the producer/consumer represent a real shift in power away from traditional media/news organizations, or is the rise of UGC just a way for newspapers to get content produced for free? Manosevitch argues that due to its embryonic state, research on online comments section is still 'primarily descriptive' (Manosevitch, 2011, 21) and emphasizes the need for more developed longitudinal studies on the area. In one of the most prolific efforts up to date the investigators examined and compared the participatory journalism cultures of 10 European countries and dedicated one chapter to users' comments as a privileged example of UGC (Reich, 2011). Among the methodologies employed we can identify a trend toward the ethno-methodological approaches. In-depth or semi-structured interviews were conducted in numerous researches (Chung, 2007; Paulussen and Ugille, 2008; Thurman, 2008). In others, interviews with practitioners were complemented with discourse analysis of the comments, either critical (Erjavec and Kovačič, 2012, 2013) or computed-medi-ated (Ruiz et al., 2011, 2010). Audience studies, such as online questionnaires, are more time-consuming and sometimes expensive and they have been very rare so far (Bergström, 2008). We seek to position our work between two gaps in the existing literature. One the one hand, while research on online sport fans' participation is extensive (Horn-moen, 2012; Özsoy, 2011; Ruddock et al., 2010; Steens-en, 2012) little or nothing has been done to date in the specific area of sports commenting under the news. On the other hand, the media-reader relationship allows the authors to explore the implications for media industry of fans participation. From a customer service perspective, hateful comments toward the medium can be understood as service failure and therefore addressed in order to recovery. Thus, the first goal of our study consists in verifying the existence of the aforementioned media- reader relationship. We focused on reader's messages referring to the medium, irrespective of the content of those messages. Therefore we propose the first research question: RQ1: Do readers address the medium in their comments? Since the early 2000s many interviewees from media outlets declared that the future of journalism lies in interactivity (Deuze et al., 2004). In the context of participatory journalism, media like to see themselves as promoters of audience-brand dialogue, which is why they encourage the participation of the consumers. This dialogue has been severely questioned by Domingo, who coined the expression 'the myth of interactivity' (2008) to refer to the ambiguous efforts of media companies to reach out consumers. In similar terms, other authors have explained that relationship as an 'interactive illusion' (Jönsson and Örnebring, 2011) variously described by terms such as interactivity and user-generated content, is frequently held up as a democracy-enhancing development. However, these concepts say little about the exact nature and character of media/audience relations. We wish to introduce a more detailed taxonomy of user-generated content. In both cases the underlying idea is that media only wants to incorporate readers or viewers in a superficial way while disempowering them in the long run. This pursued interaction with the user, if occurred as posed in RQ1, can backlash against the medium. One of the five problems envisioned by Reich when studying users' comments was their controversial content (Reich, 2011), a characteristic already posed by Singer and Ashman (2009). In Reich's work the editor of The National Post (Canada) confesses that 'the tone of your paper can really suffer from readers comments' (Reich, 2011). The specificity of those controversial comments in sports journalism resides in the fact that fans involved in sport discussions, and especially football in Spain and other football-centric countries, can easily turn heated comments into hate speech. In a football country like Scotland the Scottish Parliament introduced in 2012 'The Offensive Behaviour at Football and Threatening Communications Act' in order to tackle the traditional sectarianism and religious hatred derived from the rivalry between the two main clubs. There are two fundamental and original things about this unprecedented Act. The first one is its implication that football constitutes a species of its own, that the hatred involved in sports deserves a specific regulation beyond general legislation. The second thing is that the Act includes a whole section about hate speech on the Internet and it explicitly focuses on 'posted sectarian comments on the Internet' (Scottish Parliament, 2012). Hate speech has been defined as 'speech that denigrates persons on the basis of their race, ethnic origin, religion, gender, age, physical condition, disability, sexual orientation, and so forth (Leets, 2002, 342). We have adhered to a wider definition of hate speech ranging from discrimination based on national origin to sectarianism, as it is the case in Spanish football for the Madrid-Barcelona rivalry. Also, we included in our definition censorship allegations as some readers perceived censoring as a limitation to civil liberties. We aim to examine if the media-reader interaction is pervaded by hate speech. Our second research question then reads as follows: RQ2: To what extent do readers' opinions include hate speech toward the medium? In the event of hateful comments the role of the media remains unclear. The regulation on the comments posted online is still under construction and the responsibility of the hosting site is disputed. One of the last sentences made the Estonian web portal Delfi responsible for the offensive messages posted on its site. The European Court of Human Rights ruled that Delfi had editorial control over the comments section and that the filter was 'insufficient for preventing harm being caused to third parties' (European Court of Human Rights, 2013). The medium, hence, is bestowed with the task of finding the right equilibrium between the freedom of speech and the hate speech in the comments. Sometimes, though, fans from either side perceive the equilibrium has been broken and the medium is caught in the crossfire. How is it supposed to react in those occasions? RQ3: How does the medium moderate hate speech? Finally, we are determined to explore how this hate speech is constructed upon the messages. We must take into consideration that sporting narratives have been consistently characterized as 'narratives of conflict'. Sport has been considered a primordial communicative tool for gender discrimination (Billings and Eastman, 2002), racism (Van Sterkenburg et al., 2010) and nationalism (Alabarces et al., 2001; Knight, 2005). As a consequence some authors have pointed out that the main task of sport is the 'production of difference' (Rowe, 2003, 282), that is, that sport's primary focused is always the discovery of otherness and subsequently the mass production of narratives around this otherness. That is why Moragas reminded us that sport journalism is not there to give us a conciliatory account of the events but to 'emphasize the elements of crisis and contradiction' (1992, 15). In the light of this observation one might come to the conclusion that sport is the perfect breeding ground for the exacerbation of hate speech. We understand a narrative as 'a specific form of representation, one that accentuates the sequential or syntagmatic nature of meaning' (Knight, 2005). This means that a single narrative can potentially relate to numerous different comments by grouping them around a single storyline. The last question would be then: RQ4: What main narratives articulate the hateful discourse toward the medium? CONTEXTUALIZING HATE IN SPANISH SPORT In the last years in Europe's top sport leagues several black footballers have suffered racially abusive actions such as insults and monkey gestures. In 2004 the Spanish Football Federation was fined with £44,750 for racist chanting against the international English players Ashley Cole and Shaun Wright-Phillips (BBC, 2004). Two years later, in 2006, the Cameroonian FC Barcelona striker Samuel Eto'o threatened to leave the pitch during a game against Real Zaragoza in protest for the racist chants in the terraces (Lowe, 2006). To fight back those aggressions the Union of European Football Associations (UEFA) in collaboration with Football Against Racism in Europe (FARE) launched in 2008 the social responsibility programme called 'Respect'. In the commercial accompanying the campaign UEFA tries to address its concerns about gender, racial and religious discrimination as well as the improvement of the access to stadia of people with disabilities. Spain has been repeatedly warned by the International Olympic Committee for not being adamant enough on the fight against racism, that being supposedly one of the reasons why Madrid did not win the right to hold neither the 2016 nor the 2020 Olympic Games (Logothetis, 2009). In 2014, two new episodes of racism and religious discrimination gained public attention in Spain. In April, the Brazilian FC Barcelona football player Dani Alves was thrown a banana from the public accompanied with monkey gestures. Alves' response became viral as he decided to grab the banana from the pitch, peel it and eat it. Some weeks later, in the Euroleague basketball final between Real Madrid and Maccabi Tel-Aviv, several Jewish associations in Spain denounced that more than 17,500 Twitter users wrote anti-Semitic tweets (Levs, 2014) and remembered that Spain was the third country in Western Europe with the highest number of adults believing in anti-Semitic stereotypes (Anti-Defamation League, 2014). The Catalan-Spanish identity conflict In Spain, nevertheless, the most notorious 'production of difference' comes from the Real Madrid - FC Barcelona rivalry, known as El Clasico (Lopez-Gonzalez et al., 2012). These are the two most important clubs in the country and two of the most acknowledged all over the world. The relationship between these clubs is the tale of two cities. Madrid and Barcelona are the largest cities in Spain and their sporting rivalry dates back to 1902, their first match together, and continues for over a century until today. Each institution reaffirms proudly its symbolic heritage. Real Madrid has been awarded ten times with the European Champions League title. They represent the aristocracy of European football and are the wealthiest football club in the world (Deloitte, 2014), fact that enables them to attract the best-paid footballers available. FC Barcelona has constructed an image of creativity and passion, based on a very imaginative and forward way of playing the game. Madrid is the capital city of Spain and as such it represents the centrality, the status quo and the reactionary values of the country. Barcelona is the main city of Catalonia, an autonomous province of Spain seeking for independence from the rest of the country. The sport rivalry reflects the political struggle between the two territories and this conflict is permanently in the mind of the news commentators. Salvador has defined El Clasico as 'a game of identity and otherness' (Salvador, 2004, 64). Other authors have referred to Catalonia's situation in the Peninsula as 'a never-ending civil war, armed or metaphorical, against Spain' (Vazquez Montalban, 2006). For many years Real Madrid was perceived as the national team, supported by the dictator Franco, who favoured them to the detriment of FC Barcelona's interests (Leon Sol^s, 2003). However, there are some recent works that show a more nuanced picture of the history (Relano, 2012). CASE STUDY: MARCA For the purpose of this analysis we have selected the Marca online version' comments under the news as an online meeting point where FC Barcelona and Real Madrid supporters gather to discuss. The Marca Community is arguably one of the world largest online sports communities in Spanish. Almost 500,000 registered users participate in it and every month the news published on its website marca.com receive in the range of 800,000 to 1,200,000 comments. The community is named after the Marca daily, which is the most sold sport-based newspaper in Spain (EGM, 2013) and one of the most influential sports medium in the Spanish-speaking countries. Created in 1938, Marca permeates the everyday life of many sports fans and has the ability to dominate the Spanish cultural conversation about football. Although commentators remain anonymous in the Marca Community they must provide a verifiable email address and post under the same nickname all the time. This strategy allows Marca community manager to permanently block undesirable users and thus re-address the online conversation. Usually, comments under the news on Marca are irrespective of the content of the news story itself. Fans lead the thread of discussion towards non-related issues and turn the forum into a continuum of self-references, making it very hard to follow for a casual bystander. In the end this leads to an autonomous conversation out of control for the medium that finds itself unable to impose neither the topic of discussion nor the tone of it. Interaction moderation The Marca Community deals regularly with hate speech and offensive language. They have outsourced the management of the online discussion to Interactora, a user-generated content services company that delivers the same service to almost every digital media in Spain. The CEO and founder, Joan Llorach, explained to us in an interview the strategy followed by Interactora to restrain bad behaviour and build a healthy community. First, Marca sends Interactora the comments via the Content Management System (CMS), a process that wraps comments in small packages and delivers them. Second, Interacora filters those posted messages - it is a post-moderation tool, meaning that messages are first published and seconds later moderated - through a 19 steps tool called Moderation Technology Platform (MTP). Those steps include commonsensical measures such as deleting comments all written in capital letters as well as those blank. More sophisticated procedures encompass the use of a terminological black list with banned expressions and also a black list of users wherein commentators with a history of deleted messages are given special treatment. The MTP essentially categorizes users' comments in three groups: comments to be deleted, comments to remain published and comments yet to be determined. These undetermined comments are tagged by the robot to be passed to the human moderators, who read them and ultimately determine their convenience. Interactora employs around 40 moderators whenever Marca presumes a peak in the traffic volume at its website. It is very unusual for Marca to shut down the commenting on its website but this is exactly what happened for a brief period of time in March 2011. Eric Abidal, a FC Barcelona French player of Caribbean ancestry, was diagnosed with a liver tumour. Abidal happened to be black and also a convert to Islam. Javier Muina, Marca Community Manager at the time, was overwhelmed by the amount of racist and anti-Muslim comments spread on the forums, many of them wishing him dead by the cancer. Marca decided to disable the commenting option and removed all the comments on Abidal. METHODS In order to answer the research questions, we first exported the dataset provided by Marca to a customized database (Guerrero-Sole and Lopez-Gonzalez, 2013). The dataset consisted of all the comments posted to any of the news published by online news sport portal Marca from April 22nd to May 2nd, 2013. We selected the date because on those days took place the 2012-13 knockout phase of the UEFA Champions League football competition. On April 22nd and 23rd the four semi-finalists played the first leg of their matches; on April 31st and May 1st the second legs. We had a special interest in those confrontations because both Barcelona, who faced Bayern Munich, and Real Madrid, who played against Borussia Dortmund, were involved. Afterwards, we processed the dataset to identify those users that posted spam messag- es and removed all their comments from the database (103,495 messages). The resulting dataset consisted, then, of 209,584 posts. Of these messages, 30,195 were flagged as censored by the content manager. After cleansing the dataset, we used the search engine of the database to find whether Marca readers expressed their opinion about the medium (RQ1). To achieve our objective, we decided to perform a search by the keyword 'marca'. The search was performed in all the dataset, had the messages been censored or had not. The use of the keyword 'marca' as the only way to determine the media-reader interaction has its limitations as 'marca' stands also for the third person singular of the verb 'marcar', 'to score' in Spanish. That is why to complement the answer to RQ1, we randomly selected a sample of posts to check that the users actually made reference to the medium. We exported a list with the identifying codes of the messages to statistical software and executed the function 'Random samples of cases'. We randomly selected 200 of the messages containing the keyword 'marca' with a sampling error of 0.07. RQ2 was answered by examining those 200 comments and seeking hateful speech in them. To answer RQ3 and RQ4 we selected a second sample, slightly larger, this time only composed of comments representing hate speech. We selected 230 hate messages that included keywords and derivatives of dictatorship, censorship, fascism, Nazism, racism and inquisition. How did we choose these search terms? When answering RQ1 and RQ2 we familiarized with the contents of the messages. Based on the theoretical assumptions of the grounded theory (Strauss and Cor-bin, 1990), we inductively proposed some open coding categories. The open coding was a preliminary classification process of the material by which we could order the raw data into intelligible categories of analysis (Walker and Myrick, 2006). Once the categories were created, by means of the axial coding the items were regrouped in each category and create new categories that could form complete narratives articulating hate speech as posed in RQ4. In the field of the sports studies this method has been employed before to shed light into the narratives generated by sports journalism (Vincent et al., 2010; Lopez-Gonzalez et al., 2012). In online sports discussion we can find a precedent in the work of Kian et al. (2011) who specifically utilized the open coding and axial coding levels to make sense of the college football recruiting message board posts. RESULTS In Figure I we can observe a summary of the answers to RQ1 and RQ2. From the 209,585 comments composing the dataset only 12,316 contained the keyword 'marca'. Apparently, all those comments directly addressed the daily but on closer examination only half of them approximately actually referred to Marca Figure 1. Number and % of comments toward Marca containing hateful speech Slika 1: Število komentarjev, ki vključujejo sovražni govor do Marce newspaper. Thus, the final number of messages in the comments under the news addressing the medium was 2.84 per cent. However, among this small percentage of comments, almost every one of them contained hateful speech toward the medium (73.18 per cent). In answer to RQ3 we have to point out that comments section displayed a great amount of insults, many of them in disguise to avoid detection by the moderation systems. Abbreviations and misspellings were common techniques to elude them. Hateful commentators grew in sophistication and elaborated sometimes difficult formulas to compose messages that allowed the correct de-codification by other users but made them impossible to detect by the content managers in Marca. Hate speech in messages have to be concealed enough to pass undetected but at the same time it would fail the purpose if the target readers were not able to decode them, that is why hate speakers must balance their comments to avoid being blocked. Specifically, word derivations and irony proved to be very wise methods of camouflage of hatred. Non-existing derived words in conjunction with misspellings makes it very hard to the MTP to run the black list and therefore to identify forbidden words. Users are increasingly aware that robots run the moderation and that implies a whole different battle scenario. In the aggregate, among the 209,584 messages received by Marca, 30,195 of them were censured, amounting to a 14.40 per cent. This number includes all sorts of comments, hateful and non-hateful. In the 12,316 comments containing the word 'marca' we observed a slightly bigger rate of deletion, 15.41 per cent. However, if we come closer to examine the messages toward the medium identified by the researchers as containing hate speech, the percentage of censorship increased up to 54.34 per cent (125 deleted comments out of 230). The contextual information provided by the data of the first three research questions allowed us to expand our focus on RQ4. The axial coding showed two main axes around which the fans built their narratives articulating the hate speech toward Marca. We identified two fundamental questions that articulated the hate feelings: What should be talked about in Marca Community, and, who should be talking in Marca Community. The first question reflects what fans consider acceptable for the forum, and yet Marca censors; and also, on the contrary, what fans consider censorable and yet Marca leaves it unpunished. The second question creates a barrier between true Marca commentators and outsiders. The idea behind it seems to be that those who do not belong here should refrain from sharing their opinions here and Marca should moderate, and if necessary censor, those outsiders. Don't you dare to remove my comment! Some readers who wrote hate speech comments complain about Marca's arbitrariness to decide what is censorable and what is not. Usually, those readers re-enter the conversation after their post has been deleted to expose the injustice: 'Marca has censored my comment three times for expressing the same opinion as you but without insults'. Actually, readers with censored comments find in those the justification for the escalation of hate towards the medium. Very eloquently, a commentator warns Marca about the reasons behind his attitude: 'Moderator. I told you to f*** off because you deleted my innocuous comment. Best wishes, moderator-dictator'. A majority of readers with censored comments attacked Marca claiming the web infringed their rights to speak freely. They accused Marca of censoring them without any reason and some compared Marca's lack of respect for freedom of speech to the Spanish Inquisition. The Grand Inquisitor, Juan de Torquemada, a XV Century Spanish monk, was often cited. 'Here comes again the Holy Inquisition' or 'We are back to the years of Torquemada'. Commentators delve into the darkest ghosts of Spanish History in an attempt to shame the medium and make them re-evaluate the removed comments. Readers with previously deleted comments rejoined the discussion, sometimes from a different profile, and blamed Marca for doing so. The most common way of attacking Marca's alleged inclination towards censorship was calling them fascists: 'Marca fascist', and particularly connecting them to Franco's regime. Spain underwent 40 years of dictatorship until 1975. Under General Franco's rule citizens had no freedom of speech and media suffered prior restraint and the confiscation of publications. Franco was believed to support Real Madrid, the team from the capital city Madrid, and as a consequence Marca is perceived by FC Barcelona fans as the inheritor of Franco's will. This comparison also reveals a latent political belief underlying the online discussion. Barcelona fans see themselves as progressive, open-minded and modern; on the contrary, they perceive Madrid fans as centralist, conservative and loyalist. We can clearly observe the chained equation Marca = Real Madrid = Madrid = Spain = Francoism in the following example: 'This is what Madridism and its journalistically immoral pamphlet Marca look like. The Madridism has been morally schooled in the Franco's dictatorship hehehe'. In another comment that perfectly summarizes the terms of this equation a reader said: 'Marca people, as always, censoring. You removed a comment about Inaki Urdangar^n [alleged white collar thief, son-in-law of Juan Carlos I, the Spanish monarch] the other day; freedom of speech is not your thing. Paquito [Franco] died but his ashes are still with us. Long live Barga and long live Catalonia!'. In the message we can observe the two poles of the comparison. On the one hand, Marca is aligned with Spanish corruptors. They censor and manipulate to prevent the readers from learning the truth, just as Franco used to do during his years in power. On the other hand, the message concludes with a cry in favour of Barcelona, the team, and Catalonia, the nation. Here lies the moral of the commentator's intervention. If Marca's side is symbolically formed by dictatorship, censorship and lies, Barcelona's side denotes freedom of speech and truth. In the same fashion, this other comment 'Marca, you fascists! You delete comments against Madrid. Shame on youuuuuu!'explicitly blames Marca for taking sides and infers that the rules of participation that guide the deletion or publication of the comments are just a masquerade, the real reason being the ideology behind the comment. Paradoxically, commentators seem to value the freedom of speech when it comes to their own comments but meanwhile have favourable opinions on censorship regarding other's comments. 'I cannot understand why Marca does not censor this fella insulting me, calling me "ignorant" and staining the name of Real Madrid'. What transpires with this attitude is the reader's ultimate opinion on censorship. They believe some sort of censoring is needed although they criticize being censored. They share the idea that there should be a limit for free speech and admonish the content managers against the rampant racism hidden in El Clasico rivalry comments. 'I've been saying for quite some time now how is it possible to read on Marca such racist, xenophobic and disrespectful comments. Marca is looking forward to increase their audience and provoke people's desire to buy the newspaper. Marca, you should hang your head in shame'. Whose paper is this? A very determinant battlefield for the readers in Marca Community was the recognition that Marca was their community. As happens in the terraces in the stadium, fans develop a territorial sense and attack any trespasser that dares to challenge the boundaries of that territory. 'It is surprising that the culerdos [pejorative name for Barga fans] come here to protest and criticize Marca's front page, specially when Barcelona comic books [meaning Catalonian sports newspapers] shit all over Real Madrid every day. If you don't like it, go home!' The anonymity of the community wherein users are solely identified by a nickname often provokes a sort of witch-hunting among the members. The hunting of the foreigner, that who does not belong in the community, turns everyone into a suspect. The next three comments are good example of that. 'Marca could be as well be named Marga [implying it supports FC Barcelona] because there are envious catalufos [pejorative name for Catalans] all around here'; 'this small-time Catalonian paper of Marca'; 'In Marca there are more Barga fans and Catalonians than from any other team'. In a particularly aggressive comment, a reader tried to make Barcelona fans refrain from commenting on the community by saying 'I also find it illegal Catalonian Government's Nazi propaganda but do not come to Marca to say it'. The reader felt compelled to react against a perceived aggression. The rules of participation are annulled for a moment and are substituted by the laws of the land, stay off my property or I will feel entitled to shoot you, seems to be the thinking. This idea is shared by many other Marca Community members 'Catalonians from Sport [FC Barcelona supporting newspaper] and all those separatist dailies come to Marca The Madridistas don't need to go to Sport to say stupid things'. A number of Madrid fans openly equate Marca to Spain 'Live Spain and live Marca. Note: please abstain little Catalonians from commenting'. By highlighting the Spanish identity they succeed to exclude Catalan identity fans from the discussion. DISCUSSION Comments referring to the medium have proved to be rare among the habits of Marca readers, less than 3 per cent. Yet, the vast majority of those comments included hateful remarks and therefore have to be considered undesirable from Marca's point of view. Almost every comment toward the medium contained inappropriate language and criticism, but on top of that, almost 3 out of 4 comments were considered hate speech according to the proposed working definition. In terms of absolute numbers, Marca management and moderation teams had to deal with over 4,000 hateful comments in ten days that threaten their business and the equity brand of the company. In addition to that, what must be unequivocally troublesome for the company is the percentage of deletion of those messages. Only 54 per cent of them were tagged as inacceptable by the moderation platform and removed from the conversation, meaning that approximately 1,800 hateful comments passed through the filters and made it to the forum undetected. This relatively huge number of comments poses a threat to the readers' participation and might be harmful if the medium wants to promote a healthy discussion in its website. Also, given that virtually every comment referring to the medium was negative, they should re-evaluate what good comes from user commenting and the possible solutions to increase respectful media-reader dialogue (Ruiz et al., 2011; Lopez-Gonzalez and Guerrero-Sole, 2014). Media usually do not disclose their deletion rates so there is no easy way to compare statistics between different newspapers and actually determine if the hateful comments found in Marca were a huge or a small number. More often than not scholars perform comment analysis on the 'visible part' of the forum because the deleted comments have been withdrawn from there. To circumvent this obstacle, some scholars have interviewed journalists to estimate the number of comments deleted in the moderation process. Reich (2011) roughly calculated that The Guardian (UK) and The Washington Post (USA) had less than 10 per cent of deleted comments, and the Der Spiegel (Germany) between 10 and 20 per cent. In another research on the derogatory language in the comments under the news Ruiz et al. (2011) discovered that elpais.com in Spain contained such language in the 13.4 per cent of the news comments, the highest proportion in Europe only surpassed by Lemonde.fr (23 per cent). The deletion rates observed in Marca, 14.40 per cent for the whole dataset and 15.41 for comments containing the word 'marca', seem to be in line with the findings on other online newspapers. Deriving from the results of RQ4, we have found that supporters of Real Madrid and FC Barcelona indistinc-tively perceived Marca as a hostile media. Partisanship claims on both sides heated the debate and hate speech toward the medium derived greatly from this fact. The hostile media effect (Arpan and Raney, 2003; Lee, 2012), here only theoretically and not empirically approached, took place no matter which team the reader rooted for, and determined to a large extent the escalation of hate speech. Both Barcelona and Real Madrid fans found arguments to support the thesis that Marca was biased against them and articulated their narratives of hate accordingly. Barcelona fans hated Marca for supporting Madrid while Madrid fans hated Marca for not being aggressive enough against the enemy and required more censorship on them. According to this group of commentators, Marca should prevent Barcelona fans from ostentatiously discussing on a forum supposedly integrated by Madrid fans. Due to the lack of research at the specific topic of comments under the news in the area of online sports discussion this work can only be considered an exploratory effort into the field. Among the limitations faced by the investigators we could single out two. First, the difficulty of finding an operational definition of hate speech makes very challenging the identification of what is and what is not hate speech. In our dataset, we had to be constantly discriminating between insults and derogatory language and those expressions that went further and constituted actual hateful language. The boundaries between controversial, harmful and hate speech are not clear enough yet. Second, this study has tried to move between a quantitative and a qualitative approach to the case study in order to draw the broadest possible map of the situation. In the future, as the research area gains attention and funding, a more specific research would be desirable. CONCLUSION Although sport as competition provides by itself the basic elements for confrontation and rivalry, these elements seem to be insufficient to exert hate speech toward the enemy and the medium. In the heat of the discussion political issues and national stereotypes enter the stage to increase the level of humiliation. It is interesting to see this happen because it shows that commentators, unlike news stories by Marca, which never expand the sporting conflict into a political or social Catalonia-Spain metaphorical combat (Lopez-Gonzalez et al., 2012), are ready to go the extra mile and make use of whatever hurts the most the opposing side. Given this attitude, the medium has to be especially careful in the moderation of its community of fans in order to avoid harmful contents. This research has demonstrated that the moderation conducted by the medium does not prove to be useful to identify comments containing hate speech as observed by the poor connection between hateful comments and their deletion. Here the results suggest that online newspapers employing exclusively automatic moderation systems might be neglecting their responsibility of preventing hate expressions and could be facing some kind of liability in the near future. The sporting context offers a unique perspective on the area of participatory journalism. The myth or illusion of interaction (Domingo, 2008; Jönsson and Örnebring, 2011) seems to be all the more relevant for sports journalism, since fans perceive sports chatting as a natural space for opinion sharing and free expression, a space for escaping the constraints of everyday life. The media company needs to be extremely cautious about censoring comments as readers find this fact very irritating and a cause for abandoning the community. Managing user-generated content is not an inexpensive task at all but the reward is the engagement of the public with the brand. Hate speech toward Marca was obviously not desirable but in an online sports forum, wherein a certain amount of verbal violence seems to be the rule, it came with no consequences for the medium. Altogether, the analyst senses some irony and humour in the hateful comments, as if the liberating component was actually prevailing over the explicit content of it, even though the liberation came in the form of hate speech. quando il medium e' il messaggio: esplorare odio nelle interazioni tra media e lettori nel giornalismo sportivo spagnolo on-line Hibai LOPEZ-GONZALEZ Universitä Pompeu Fabra, Barcelona, Roc Boronat, 138. 08018, Spagna e-mail: hibai.lopez@upf.edu Frederic GUERRERO-SOLE Universitä Pompeu Fabra, Barcelona, Roc Boronat, 138. 08018, Spagna e-mail: frederic.guerrero@upf.edu RIASSUNTO Lo scopo di questo articolo e esaminare le modalita che portano i lettori che prendono parte alla sezione com-menti dei giornali sportivi on line a intervenire con discorsi di incitamento all'odio nei confronti della testata. La ricerca si basa sulla comunitä della versione on line del giornale sportivo Marca in Spagna, nell'ambito della rivalita Real Madrid vs Barcelona FC. L'articolo esamina il dilemma della testata tra lasciare libertä d'espressione ai suoi lettori e di conseguenza promuoverne la partecipazione oppure proteggere la propria identitä di marca, minacciata da commenti contenenti incitazioni all'odio. Sebbene solo una piccola parte dei lettori si rivolga alla testata nei loro commenti, circa il 75% di questi contiene incitazioni all'odio, creando un clima di discussione molto violento. Parole chiave: Giornalismo partecipativo, commenti, sport, incitamento all'odio, media ostili, Spagna. SOURCES AND BIBLIOGRAPHY Alabarces, P., Tomlinson, A., Young, C. (2001): Argentina versus England at the France '98 World Cup: narratives of nation and the mythologizing of the popular. Media, Culture & Society, 23, 5, 547-566. Anti-Defamation League (2014): The ADL GLOBAL 100: An Index of Anti-Semitism. Http://global100.adl. org/#country/spain (15. 7. 2014). Arpan, L. M., Raney, A. A. (2003). An experimental investigation of news source and the hostile media effect. Journalism and Mass Communication Quarterly, 80, 2, 265-281. BBC (2004): Spain fined for racist chanting. BBC. Http://news.bbc.co.uk/sport2/hi/football/internatio-nals/4110135.stm (21. 7. 2014). Bergström, A. (2008): The Reluctant Audience: Online Participation in the Swedish Journalistic Context. Westminster Papers in Communication and Culture, 5, 2, 60-80. Billings, A., and Eastman, S. T. (2002): Selective Representation of Gender, Ethnicity, and Nationality in American Television Coverage of the 2000 Summer Olympics. International Review for the Sociology of Sport, 37, 3-4, 351-370. Boyle, R. (2012): Social media sport? Journalism, public relations and sport. In: R. Krovel, Roksvold, T (Eds.), We Love to Hate Each Other: Mediated Football Fan Culture. Goteborg: Nordicom, 45-62. Chung, D. S. (2007): Profits and Perils: Online News Producers' Perceptions of Interactivity and Uses of Interactive Features. Convergence: The International Journal of Research into New Media Technologies, 13, 1, 43-61. Deloitte (2014): Football Money League. http:// www.deloitte.com/view/en_GB/uk/industries/sportsbu-sinessgroup/sports/football/deloitte-football-money-lea-gue/ (15. 7. 2014). Deuze, M. (2009). The Future of Citizen Journalism. In: S. Allan, Thorse, E. (Eds.), Citizen Journalism. Global Perspectives. New York: Peter Lang, 257-263. Deuze, M., Neuberger, C., Paulussen, S. (2004): Journalism education and online journalists in Belgium, Germany, and The Netherlands. Journalism Studies, 5, 1, 19-29. Diakopoulos, N., Naaman, M. (2011): Towards quality discourse in online news comments. In Proceedings of the ACM 2011 conference on Computer supported cooperative work - CSCW '11. New York, New York, USA: ACM Press,133. Domingo, D. (2008): Interactivity in the daily routines of online newsrooms: dealing with an uncomfortable myth. Journal of Computer-Mediated Communication, 13, 3, 680-704. Domingo, D., Quandt, T., Heinonen, A., Paulussen, S., Singer, J. B., Vujnovic, M. (2008): Participatory journalism practices in the media and beyond. Journalism Practice, 2, 3, 326-342. EGM (2013): Resumen general de resultados Estudio General de Medios: Febrero a Noviembre 2013. Erjavec, K., Kovačič, M. P. (2012): "You Don't Understand, This is a New War!" Analysis of Hate Speech in News Web Sites' Comments. Mass Communication and Society, 15, 6, 899-920. Erjavec, K., Kovačič, M. P. (2013): Abuse of online participatory Journalism in Slovenia: Offensive comments under News Items. Medijska Istrazivanja, 19, 2, 55-73. European Court of Human Rights (2013): Delfi AS vs Estonia http://hudoc.echr.coe.int/sites/fra/pages/search. aspx?i=001-126635#{"itemid":["001-126635"]} Guerrero-Sole, F., Lopez-Gonzalez, H. (2013): Lotus Notes como herramienta de analisis de comunidades virtuales: el caso de la Comunidad Marca. BiD: Textos Universitaris de Biblioteconomia I Documentacio, 30, June. Hornmoen, H. (2012): Battling for Belonging. How Club and Supporter Identities are Created in the Mediation of an Oslo Derby. In: We love to hate each other. Mediated football fan culture. Oslo: Nordicom, p. 323. Jönsson, A. M., Örnebring, H. (2011): User-Generated Content and the news. Empowerment of citizens or interactive illusion? Journalism Practice, 5, 2, 127-144. Kian, E. M., Clavio, G., Vincent, J., Shaw, S. D. (2011): Homophobic and sexist yet uncontested: examining football fan postings on Internet message boards. Journal of Homosexuality, 58, 5, 680-99. Knight, G. (2005): The Disappointment Games: Narratives of Olympic Failure in Canada and New Zealand. International Review for the Sociology of Sport, 40, 1, 25-51. Lee, E.-J. (2012): That's Not the Way It Is: How User-Generated Comments on the News Affect Perceived Media Bias. Journal of Computer-Mediated Communication, 18, 1, 32-45. Leets, L. (2002): Experiencing Hate Speech: Perceptions and Responses to Anti-Semitism and Antigay Speech. Journal of Social Issues, 58, 2, 341-361. Leon Sol^s, F. (2003): Negotiating Spain and Catalonia: Competing Narratives of National Identity. Bristol, Intellect. Levs, J. (2014): Groups demand punishment for anti-Semitic tweets. CNN. Http://edition.cnn. com/2 014/05/21/world/europe/spain-anti-semit-ic-tweets/ (19.07.2014). Logothetis, P. (2009): Racist fans may thwart Spain's goals. The New York Times. Http://www.nytimes. com/2009/02/10/sports/10iht-RACISM.1.20077003. html?_r=0 (18.07.2014). Lopez-Gonzalez, H., Guerrero-Sole, F., Haynes, R. (2012): Manufacturing conflict narratives in Real Madrid versus Barcelona football matches. International Review for the Sociology of Sport, ahead of print. Lopez-Gonzalez, H., Guerrero-Sole, F., Larrea, O. (2014): Community building in the digital age: Dynam- ics of online sports discussion. Communication & Society, 28, 3, 83-105. Lopez-Gonzalez, H., Guerrero-Sole, F. (2014): Discussion moderation in online media. The case of Marca newspaper. El Profesional de la Informacion, 23, 1, 51-57. Lowe, S. (2006): Eto'o on verge of walking off after racist abuse at Zaragoza. The Telegraph. Http://www. telegraph.co.uk/sport/football/european/2332860/Etoo-on-verge-of-walking-off-after-racist-abuse-at-Zaragoza. html (27. 6. 2014). Manosevitch, I. (2011): User generated content in the Israeli online journalism landscape. Israel Affairs, 17, 3, 422-444. Mitchelstein, E. (2011): Catharsis and Community: Divergent Motivations for Audience Participation in Online Newspapers and Blogs. International Journal of Communication, 5, 2014-2035. Moragas, M. (1992): Cultura, s^mbols i jocs ol^m-pics: La mediacio de la comunicacion. Barcelona, Centre d'investigacio de la comunicacio. Neurauter-Kessels, M. (2011): Im/polite reader responses on British online news sites. Journal of Politeness Research. Language, Behaviour, Culture, 7, 2, 187-214. Nielsen, C. E. (2013): Coproduction or cohabitation: Are anonymous online comments on newspaper websites shaping news content? New Media & Society, 16, 3, 470-487. Örnebring, H. (2008): The consumer as producer- of what? User-generated tabloid content in The Sun (UK) and Aftonbladet (Sweden). Journalism Studies, 9, 5, 771-785. Özsoy, S. (2011): Use of New Media by Turkish Fans in Sport Communication: Facebook and Twitter. Journal of Human Kinetics, 28, 165-176. Paulussen, S., Ugille, P. (2008): User Generated Content in the Newsroom: Professional and Organisational Constraints on Participatory Journalism. Westminster Papers in Communication and Culture, 5, 2, 24-41. Reich, Z. (2011): User Comments: The Transformation of Participatory Space. In J. B. Singer, A. Hermida, D. Domingo, A. Heinonen, S. Paulussen, T. Quandt, and M. Vujnovic (Eds.), Participatory Journalism: Guarding Open Gates at Online Newspapers (96-117). Oxford, UK: Wiley-Blackwell. Relano, A. (2012): Nacidos para incordiarse. Un siglo de agravios entre el Madrid y el Barga. Madrid, Mart^nez Roca. Rowe, D. (2003): Sport and the Repudiation of the Global. International Review for the Sociology of Sport, 38, 3, 281-294. Ruddock, A., Hutchins, B., Rowe, D. (2010): Contradictions in media sport culture: The reinscription of football supporter traditions through online media. European Journal of Cultural Studies, 13, 3, 323-339. Ruiz, C., Masip, P., Mico, J., D^az Noci, J., Domingo, D. (2010): Conversacion 2.0. y democracia: analisis de los comentarios de los lectores en la prensa digital cata-lana. Comunicacion Y Sociedad: Revista de La Facultad de Comunicacion, 23, 2, 7-39. Ruiz, C., Domingo, D., Mico, J. L., Diaz-Noci, J., Masip, P., Meso, K. (2011): Public Sphere 2.0? The Democratic Qualities of Citizen Debates in Online Newspapers. The International Journal of Press/Politics, 16, 4, 463-487. Salvador, J. (2004): Futbol, metafora d'una guerra freda. Tdx.cat, Universitat Rovira i Virgili. Scottish Parliament (2012): Offensive Behaviour at Football and Threatening Communications Http://www. scottish.parliament.uk/S4_Bills/Offensive Behaviour at Football and Threatening Communications (Scotland) Bill/SB11-48.pdf. Singer, J. B., and Ashman, I. (2009): "Comment Is Free, but Facts Are Sacred": User-generated Content and Ethical Constructs at the Guardian. Journal of Mass Media Ethics, 24, 1, 3-21. Steensen, S. (2012): Conversing the Fans. "Coverit-live" and the Social Function of Journalism. In: R. Kro-vel, Rokswold, T. (Eds.), We Love to Hate Each Other. Mediated Football Fan Culture. Goteborg, Nordicom, 48-51. Strauss, A., Corbin, J. M. (1990): Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory. Newbury Park, CA, Sage. Thurman, N. (2008): Forums for citizen journalists? Adoption of user generated content initiatives by online news media. New Media & Society, 10, 1, 139-157. Van Sterkenburg, J., Knoppers, A., De Leeuw, S. (2010): Race, ethnicity, and content analysis of the sports media: a critical reflection. Media, Culture & Society, 32, 5, 819-839. Vazquez Montalban, M. (2006): Futbol: Una religion en busca de un dios. Barcelona, DeBolsillo. Vincent, J., Kian, E. M., Pedersen, P. M., Kuntz, A., Hill, J. S. (2010): England expects: English newspapers' narratives about the English football team in the 2006 World Cup. International Review for the Sociology of Sport, 45, 2, 199-223. Walker, D., Myrick, F. (2006): Grounded theory: an exploration of process and procedure. Qualitative Health Research, 16, 4, 547-59.