15_Bodrunova.indd International Review of Management and Marketing | Vol 5 • Special Issue • 2015 97 Special Issue for "Media as the Tool: Management of Social Processes" International Review of Management and Marketing ISSN: 2146-4405 available at http: www.econjournals.com International Review of Management and Marketing, 2015, 5(Special Issue) 97-104. Twitter-based Discourse on Migrants in Russia: The Case of 2013 Bashings in Biryulyovo Svetlana Sergeevna Bodrunova1*, Anna Alexandrovna Litvinenko2, Dmitry Petrovich Gavra3, Aleksandr Vasilievich Yakunin4 1Saint Petersburg State University, Universitetskay Naberejnay 7/9, St. Petersburg, 199034, Russia, 2Saint Petersburg State University, Universitetskay Naberejnay 7/9, St. Petersburg, 199034, Russia, 3Saint Petersburg State University, Universitetskay Naberejnay 7/9, St. Petersburg, 199034, Russia, 4Saint Petersburg State University, Universitetskay Naberejnay 7/9, St. Petersburg, 199034, Russia. *Email: svetlana.s.bodrunova@rambler.ru ABSTRACT According to a number of scholars, Twitter possesses big potential to become a “crossroads of discourses” due to its openness, de-hierarchization, and spontaneity (Miller, 2010; Shirky, 2008). At the same time, substantial criticism has risen towards political and deliberative effi cacy of Twitter (Fuchs, 2014). The authors aim at analyzing the features of the Twitter-based agenda setting within the hybrid media system in Russia (Chadwick, 2013; Bodrunova and Litvinenko, 2013a). The research question is whether the use of Twitter in the Russian socio-political context potentially leads to the formation of the “crossroads of opinions” or, in contrast, to closing-up of political discussion and to further fragmentation of public discourse. The research focuses on structural and content aspects of discussion on anti-migrant bashings in Biryulyovo (Moscow) that happened in October 2013. Our research methods include automated vocabulary-based web crawling, word frequency analysis, manual coding of tweets, and interpretation of statistical data. Preliminary results suggest an unexpectedly high level of mediatization of the discussion; the hypothesis about the “crossroads” nature of the discussion on the Russian Twitter seems to be proven, which makes this platform differ from the Russian Facebook where, according to another recent study (Bodrunova and Litvinenko, 2013b), political discussions are held mostly in closed-up communicative milieus, or “echo chambers” (Sunstein, 2007). Keywords: Twitter, Russia, Web Crawling, Social Media, Migrants, Echo Chambers JEL Classifi cations: O32, Z13 1. INTRODUCTION 1.1. Literature Review With the growth of Internet penetration in the world, media systems have been undergoing qualitative shifts in their shape, borders, and relations with outer society, including the political sphere. Among many conceptualizations of this process, one of the recently developed concepts is the concept of a hybrid media system (Chadwick, 2013) where technological changes are causally related to the reshaping of socio-political cleavages in the societies. Under hybridization of a media system, we understand the two trends: (1) Growing transformation of offl ine media into the so-called “convergent” media characterized by multiplatform production and multichannel delivery of content; (2) growth of the new segment of the media sphere, namely online-only professional media outlets and web 2.0 aggregated individual media (Bodrunova and Litvinenko, 2013b). Understood this way, media hybridization framework allows for posing questions on the quality of online public discussion with high effi cacy. Parameters that may show whether an online networking/communication platform, is, indeed, a place where competing or even hostile discourses meet are better developed as media hybridization parameters, as the concept unites traditional media research areas such as, e.g., agenda setting research, with political Internet studies. Media hybridization research today focuses on several important issues. Among those, one of the most important questions is what factors – national-bound or universal/global – infl uence online discussions more within the language-/nation-bound public spheres (Adam Bodrunova, et al.: Twitter-based Discourse on Migrants in Russia: The Case of 2013 Bashings in Biryulyovo International Review of Management and Marketing | Vol 5 • Special Issue • 201598 and Pfetsch, 2011). Under national-bound factors, scholars usually mean national context, including both technological factors (like Internet penetration or the structure of online/offl ine parallelism) and public sphere factors (deliberative traditions, representation of confl ict sides, local issues etc.). Cross-national or global factors are often bound to platform features: Presumably, Facebook, Twitter and other globally popular platform organize the online discussions in their own ways, thus infl uencing the deliberative quality of the online discourse. In particular, it is interesting to know what media today play the leading roles in formation of agendas (Bodrunova and Litvinenko, 2013a; Mancini and Mazzoni, 2013) and how intermedia agendas work (McCombs, 2005) – e.g. how the agenda fl ows are directed (from/to online media), whether there are boundaries for agenda spill-overs, and how coherent the agenda formation in different media is. Despite early studies showed no difference in online agenda setting with “old” media (Rogers et al., 1991; Lee, 2007), today is research brought a revival to long-established research on two-step fl ow of political information (Katz and Lazarsfeld, 1957) and “spiral of silence” in political discussions (Noelle-Neumann, 1984). But, being rooted in media and political research, media hybridization studies are still poorly connected to internet studies that perceive media in a more technical way (as platforms). In our research, we try to look at Twitter as a part of a hybrid media system but also address previous fi ndings of political research on Twitter as a communicative milieu. As it is widely known, in 1990s and early 2000s, many authors claimed that Internet penetration would lead to democratization via bigger citizen involvement and horizontalization of communication. This was believed to be especially true for transitive democracies, including Russia (Rohozinski, 2009; Kuchins, 2007). With the rise of micro blogging since 2006 (predominantly on Twitter), this democratization potential of Internet, in theory, was expected to increase even more: First, because of the growing informational openness in social networks, as Twitted had no options of closing posts from other users (Miller, 2010); second, due to the nature of the 140-digit-limited posts that seemed more engaging than on other, more monologue-oriented, social platforms. But recently optimism gave place to demarcation of optimists and pessimists, where the former still perceive Twitter as a capable catalyst of political discussion while the latter consider Twitter a noisy dump of de-politicized content oriented to sex, pop music, gaming, spam, and trivia (Fuchs, 2014). Research on online deliberative publics (Grönlund et al., 2014), as well as media exposure studies, have both shown signifi cant criticism towards political effi cacy of Twitter. Within these studies, there may be traced a critical line that tells of fragmentation of the audiences via encapsulation of online mini-discussions (Tewksbury, 2005), which clearly has political implications, as the discussions become less involving and representative. These closed-up online milieus are described as “public sphericules” (Gitlin, 1998), “echo chambers” or “enclaves” (Sunstein, 2007), or “filter bubbles” (Pariser, 2011); their formation prevents communication platforms from becoming “virtual fi replaces” or “crossroads of opinions.” Despite all existing criticism, Twitter is still perceived as more capable of forming the “crossroads of opinions” than longer-text-based platforms. It is particularly important whether Twitter-based discussions, presumably the most involving ones, are still inclined to form encapsulated “clouds.” Thus, the research that we present today has three parts. First, we look at the structure of the Twitter-based discussion in terms of its representative quality: Who takes part in the discussion, how the confl ict sides are represented, who are the “discussion infl uencers” (Vaccari, 2013) and to what extent they are institutionalized, and whether there are “echo chambers” that lower the potential of deliberative “opinion crossroads.” As to the infl uencers, according to American and Swedish studies, the biggest infl uencers are experts, professionals and established organizations (Ruthl 2012); but this may vary depending upon the national context (Vaccari et al. 2013). Second, we look (and partly present the results in this article) at the role of Twitter in intermedia agenda setting – that is, what media work better as news alert providers on Twitter and beyond, what directions (traditional to web 2.0 media or vice versa) agenda spill-overs take, and what relations media accounts in social networks form with other infl uencers. Here, one needs to add that, in the West, Twitter has been described as an “alert system,” a “springboard for news stories,” and a tool for reporting (Hermida, 2010; Mancini and Mazzoni, 2013; Vis, 2013). The third part of the research is dedicated to frame analysis of the discussion (De Vreesel 2005; Dimitrova and Strömbäck, 2012), which in future will allow us, among other, to describe the patterns of distribution of guilt, responsibility, and solidarity, as well as to compare the levels of hate speech, call for action, and presence of nationalist discourse in the discussions. 1.2. Russian Hybrid Media System as a Research Object Today, Russia constitutes a nearly perfect object for research upon political implications of media hybridization – due to at least four reasons. 1. Today, Russia is described as a fundamentally fragmented society (Zubarevich, 2011), with cosmopolitan post-industrial, town-based post-soviet, rural, and Caucasian/migrant “Russias” co-existing but not fully merged. The Russian media system seems to refl ect these divisions to a great extent, with post-Soviet media still occupying a big share in newspaper, TV, and radio markets (Vartanova, 2013). This implies that we need to assess how the media spheres of the “four Russias” are represented and interact in online discussions. 2. Russia has low online/offl ine media parallelism (Toepfl , 2011): In early 2000s, a big number of online-only media formed an alternative news arena. This implies we need to assess how the online/offl ine parallelism is effectively represented on Twitter. 3. Russia has already experienced formation of “echo chambers” in blogs, fi rst and foremost on Livejournal (Gorny, 2009). Thus, we expect Twitter sphere to have similar tendency. 4. In 2009-2013, Russia repeatedly became “the most socially networked country in the world” in terms of user engagement Bodrunova, et al.: Twitter-based Discourse on Migrants in Russia: The Case of 2013 Bashings in Biryulyovo International Review of Management and Marketing | Vol 5 • Special Issue • 2015 99 (Synthesio, 2013). This means that the role of social networking platforms in Russia is high and grows higher (but perhaps depends upon the “four Russias”). Thus, partial political modernization is complemented by an online-only-oriented media hybridization trajectory; the country is very likely to provoke closing-up of online communicative milieus on competing platforms. Our own earlier research fi ndings (Bodrunova and Litvinenko, 2013b) suggest that, in 2010s, online media, especially Facebook and media of specialized and alternative agenda, have played a “cultivational” role in forming a negative political consensus within the community of the “For fair elections” protesters of 2011-2012. We also tracked evidence of platform dependence in political engagement, as Facebook became an attraction for the ex-Livejournal community, the “meeting point” for the “creative class” and a clearly identifi able political community while, e.g., its Russian counterpart Vkontakte did not. What we observed was a platform-dependent and online-media- based formation of a post-industrial urban “public counter-sphere;” its divergence from the mainstream still deepens. By far, we had mixed data on the “crossroads” nature of the Russian Twitter. One research group (Greene, 2012) showed that, indeed, the Russian Twitter of 2012 could be perceived as “crossroads” in terms of presence of pro-establishment and oppositional clusters; pro-establishment networks, though, were better organized and more active. This fi nding was only partly supported by Berkman Center for Internet and Society at Harvard (Kelly et al., 2012) which, for 2010-2011, identifi ed topic-oriented clusters in the Russian Twitter. Thus, the questions on the “crossroads” stance of the Russian Twitter remain unanswered, while the relevance of Twitter in Russia grows, with rough estimation of usage as 5 million out of circa 78 million of monthly internet users aged over 12 (data by TNS). 1.3. The Biryulyovo Bashings Case in the Context of Ethnic Relations in Russia In the post-Soviet times, Russia has witnessed huge migration fl ows from the ex-Soviet states, including South ones with dominant Muslim population (Tajikistan, Uzbekistan, Kazakhstan), directed to urban areas in European Russia. It coincided with a comparable wave of resettling of North Caucasian people (that is, Russian citizens) to urban centers up north (Moscow, St. Petersburg, and others). This has created a distorted public perception of both streams as one, both immigrants from CIS countries and North Caucasus being labeled as “migrants” in media discourse and everyday speech. This infl ow has already provoked several major anti-migrant bashings like in Kondopoga (near St. Petersburg) in 2006 or in Moscow of 2010 and 2013. The latest one took place in the Moscow district of Biryulyovo in October 2013, being provoked by an alleged killing of a Russian youngster Egor Scherbakov by an immigrant Orkhan Zeinalov. Within several days, local dwellers partly destroyed a local warehouse where mostly migrants worked and called for a “popular gathering” to attract authorities’ attention. This case is ideal for our purposes, as it represents an “active” period of the (anti-) migrant discourse in the Russian Twitter. It combines a line of events (alleged Egor Scherbakov’s killing, Orkhan Zeinalov’s arrest, warehouse bashings, the people’s gathering, urged retirement of Moscow civil servants etc.) and the issues surrounding migrant life in Moscow (migrant crime, non- licensed trade, undercover dwelling, and lack of assimilation of the incoming migrant population). Thus, the case, in theory, may create a reference discussion point for further framing of migrant discourse on the Russian Twitter. 1.4. The Research Hypotheses As stated above, our research aims at describing structural features of the discussion, some aspects of intermedia agenda setting, and content features based on issue-oriented framing (De Vreese, 2005). For the fi rst stage of the project, we have developed four hypotheses on the structural and agenda-setting aspects. The general idea is to test whether the Russian Twitter replicates the trajectory of formation of an anti-establishment “echo chamber” analogue to Russian Facebook; this would mean that national context is more powerful than platform features. H1: Character of the discussion. The discussion will have an “explosive” character with a peak at the beginning and micro- peaks as the discussion developed. The discussion will be robust, as large number of users not connected otherwise would participate in it. H2: Presence of “opinion crossroads” and the role of media in it. Divergent opinions will be, indeed, found in the discussion on the whole, but the “crossroads” nature won’t show up, as the discussion will be polarized to the extent that clear “echo chambers” will be visible; these may be based upon political views, institutional or group belonging. The visible clusters will form around in-platform non-institutional infl uencers (who mostly organize the discussion by frequent tweeting and are the key reference authors for less active tweeters). Media accounts will play a big role in posting information but will not be themselves actively engaged into the discussion. H3: Replication of fragmentation of the media-based public sphere in Twitter. Online-only media will cast more impact than the hybrid media; among the former, anti-establishment and alternative-agenda media (Bodrunova and Litvinenko 2013b) will gain the most attention. H4: Involved media content. Linking to media content by the users will be online-centered, with web 2.0 media being most quoted, online-only media in the middle, and hybrid media least quoted. The data collected to test H4 also allows for meaningful interpretation of the discussion, that is, for exploring issue-oriented discussion frames. It is beyond the scope of this paper to present the results of this analysis; we would just note that the discussion was quite heated. Violent topics (the alleged murder, the bashings and police actions) were discussed 1, 2 times more than non- violent ones, negative emotional discourse dominated over rational commenting (but not over the tweets containing news, facts, and details); 15% of tweets put blame on someone; 10% contained nationalist speech; 11% had hate speech. Bodrunova, et al.: Twitter-based Discourse on Migrants in Russia: The Case of 2013 Bashings in Biryulyovo International Review of Management and Marketing | Vol 5 • Special Issue • 2015100 2. METHODS 2.1. Mixed Methodology of the Project To map the discussion on Biryulyovo bashings in October 2013, we used vocabulary-guided web crawling (Blekanov et al., 2012) and subsequent analysis of the obtained tweets, word frequency- based vocabularies, and web graphs. Hashtag collections selected by manual reading of over 1000 tweets have constituted the vocabularies for a specially designed web crawler robot. As the results of the crawling, the following items were received: (1) A web graph reconstructing the hashtag-based discussion; (2) lists of the most active users (who tweet and comment) and the most infl uential users (who get retweeted and commented); (3) a discussion vocabulary of high relevance and clarity; (4) collections of tweets for manual coding (analysis in progress). To assess the “crossroads” nature of the discussion, we looked at two data pieces – namely, the web graphs and the lists of active users; we were mostly looking for mediatization parameters. These data can show us: (1) Whether there is a real “crossroads of opinions” of various stakeholders in the Russian Twitter; (2) whether the existing cleavage between media of the “fi rst” and “second” Russias (mainstream/counter-sphere) is repeated. 2.2. The Work-in-progress Report The works of the project so far have included several steps. 1. We analyzed the Russian Twitter trending topics at Trendinalia. com to select the research period (October 1 to 31, 2013) and the relevant hashtags. The search for hashtags started on Trendinalia.com and was fi nalized via manual reading of over 1000 tweets per each period. 6 hashtags were selected for the fi rst-step crawling: #Biryulyovo (in Russian), #Biryulyovo (in English, with variations), #Zeinalov, #warehouse, #migrants, and #populargathering (all in Russian). The selection was later cross-validated by co-hashtagging analysis. 2. With the help of a specially developed web crawling robot, we have run primary hashtag-based web crawling and received the collection of tweets for the designated data (dates and hashtasg). As a result of crawling, 3734 users were found to be tweeting under the hashtags in October 2013; 10715 tweets were downloaded. 3. The web graph of the discussion was created (Figure 1). The graphs included not only the users who were discovered via hashtagging but also the users connected via reposting or commenting to each primary tweet collection. Thus, for Biryolyovo, the fi nal number of graph nodes (users) was 12040, over 3, 2 times more than initially. To our viewpoint, this reconstructs the discussion in a more accurate way, without losing meaningful inter-user connection. As our programming capacities allow for vocabulary-based web crawling, another option to create a more precise web graph is collecting data on tweets based on the discussion vocabulary. We have created such a vocabulary by gathering all tweets on Biryolyovo in a “bag of words,” stemming them, cleaning the stem dataset, ranging the stems by frequency, and manual clearing out the fi rst 6000 stems to get the list of the most relevant stems. After several stages of cleaning, 443 stems remained. In future, we will compare the hashtag-based and stem-based web graphs. Wу also used the vocabulary for semantic grouping of stems to see which lexical strings were especially relevant. 4. We visually analyzed the web graph and its 20 key nodes. The graph on Biryulyovo was analyzed for users’ classic betweenness, input capacity (the number of comments and reposts by the user), output capacity (the number of comments and reposts the user received), and overall capacity (input+output capacity). 5. We ranged users who posted for Biryulyovo to create 3 lists: By frequency of posts – “active users,” by received comments – “authoritative users,” and the combined list – “junction users.” 6. We assessed media accounts among the 72 most active users who tweeted on Biryulyovo 20 or more times in October 2013 to see who the most active users actually are, whether media are present in this list, and what media are winning the battle. 7. Selection of tweets for manual coding was performed. As we were interested fi rst of all in the active creators of the discussion and their discourse, the 3734 users of the initial Biryolyovo dataset were ranged according to quantity of their posts. To defi ne the threshold for sampling, we introduced the measure K = 20 tweets per user. K was defi ned as the mean for the “noisy” section of the graph (marked with a red dot with coordinates [20; 6]). 8. The 72 most active users altogether posted 2800 tweets. 1120 tweets were selected for further manual analysis by random choice of 40% of tweets from each of the 72 users to preserve their respective weight in the discussion; after cleaning this dataset from fully irrelevant, non-Russian (English-and Ukrainian-language) tweets, self-reposts, and spam, there remained 1014 tweets, a feasible dataset at the same time representing almost 10% of the full primary collection. Over 90.5% of the automatically formed dataset proved to be unique and relevant, and the number of irrelevant tweets was less than 3%. 9. The codebook was created and tested via Kappa measurement to analyze the fi nal dataset (all parameters fi nally showed Kappa over 0.6, mostly close to or higher than 0.7). We (with the help of 24 coders) manually triple-coded the 1014 Figure 1: The web graph representing 20 key “junction” user Bodrunova, et al.: Twitter-based Discourse on Migrants in Russia: The Case of 2013 Bashings in Biryulyovo International Review of Management and Marketing | Vol 5 • Special Issue • 2015 101 Biryulyovo tweets of the 72 most active users. Manual coding of tweets included 25 parameters, of which 6 are metadata, 5 discuss the status of author of the tweet, 3 are communicative features of the tweet (like presence of quotes or news/comment content), 2 on the topic and mentioned actors, 4 on discursive features (like hate speech), 4 on frames (guilt, responsibility, comparison to other confl icts, and spatial dimension of the confl ict), and the last one on media content links in a tweet. All non-agreed coding units were discussed within the working group and re-coded. By today, we have analyzed data for 673 tweets of the 1014; full results will be available by June 2015. 3. PRELIMINARY RESULTS AND DISCUSSION H1: Character of the discussion. The discussion on Biryolyovo was, as expected, a sustainable research object. There were enough users who participated in at least one dialogue (tweeted 4-6 times); it is over 120 users who tweeted 10 times or more, presumably following the case for days. It was not a short few-actor discussion; it did not stop after the bashings and created a reference point on the discussions on migrants. Being quite intense for 120 users or so, the discussion also had the “buzz” component, with over 3200 users involved within 20 days (October 11-31, as fi rst news on the case arrived on October 11); this was enough to bring the case to the trending topics. The discussion had a character of an explosion. This, fi rst of all, tells us again that the discussion chosen was robust enough for studying. Second, the Biryolyovo discussion did not fade rapidly but lasted for 2 weeks, as expected. But, contrary to expectations, the line of events did not infl uence its intensity in any visible manner: There were no micro-peaks on the time-series graph. This was also a focused discussion. This may be concluded out of the assessment of the automated vocabulary. Its confi guration shows that the 443 stems that may be considered discussion descriptors are situated not in the middle of the vocabulary after general lexicon of high frequency but in the upper part of the list. Other sources of texts concerning Biryulyovo could not produce ontology of such a high relevance; on Twitter, we have received a practically full semantic list describing the discussion. Probably, it is the format of Twitter as a platform (the 140 symbols limit) that makes users select the most relevant words for expression. If we look at the structure of the web graph of the discussion, we will see that it is characterized by weak connectedness and high evenness; that is, we see an even conglomerate of non-connected users with no “clouds” in it. This is contrary to what was expected, and this supports the vision that the “crossroads” potential of the Russian Twitter is higher than expected (we emphasize “potential,” as we still do not know whether we are seeing an “opinion crossroads” or one big “echo chamber”). Another side of the “crossroads” potential is the users with high betweenness centrality. Only 46 of the 12040 users had centrality of ≥0.001 (and two – of >0.12); thus, the “clouds” could have been interconnected did they appear, but the majority of discussants were just individual contributors. H2: Presence of “opinion crossroads” and the role of media in it. As stated above, there are no identifi able “clouds” in the discussion. The hypothesis of internal discussion milieus is, thus, wrong, and, as stated above, we witness either “one big cloud” where all the space is occupied by one echo chamber – or a real crossroads of opinions where any user can freely participate in any part of the discussion. It is hard to support either of the two explanations before we look at who the infl uencers are and what their positions are. To see which tweeters have a potential to become “junction users,” we have looked at their input/output capacity. The lists of most active users (with the highest number of tweets per user) and the most infl uential ones (with the highest overall capacity) do not fully correspond. Of top20 most infl uential users, only 5 (25%) are those who are within top20 tweeters. Thus, it is in rare cases only that a user can gain infl uence just by frequent tweeting, though this may be one of the key elements of the popularity-rising strategy. The overall user capacity depends to a much higher extent on output (on the user is commented and retweeted) than on frequency of their tweeting. This may be a sign that Twitter discussions can be more rational and content-oriented than expected. Also, there is a cleavage between those who tweet and those who retweet and comment without producing content themselves. On the whole, there is just one user (@BorisALV) who is the only real infl uencer who possesses all the necessary features (high input, output, and betweenness centrality). Second, in our dataset of 1014 tweets for coding, there were only 5 retweets found (cf. to 51 self-retweets), which tells that those who actively tweet do not retweet from other active users. Twitter in Russia may become not only “opinion crossroads” but also an “echo chamber;” but in this particular case the “crossroads” potential seems to outweigh the “echo chamber” one. A possible explanation may be quite simple. As the “bridging” users are not homogeneous, it could be that, on the Russian Twitter, there is not yet any stratum of “infl uencer commentators” through whom discussion clouds are connected; their place is partly occupied by media and trash accounts. As to the role of media in forming the “opinion crossroads,” we fi rst of all need to underline the unexpectedly high mediatization of the whole discussion. It shows in the structure of discussion vocabulary, hashtagging, and the lists of the most active and most infl uential users. News-related hashtags were very popular (#breakingnews, #news, #it is reported, #Russian Information Agency, #media, #lifenews [a publishing house], #Russia Today). Semantic analysis of stem has shown the stem group dedicated to media to be the 5th in size, being comparable to city descriptors, migrant-related lexicon, and descriptors of police activities, which is also surprising and proves that not only activities of the media themselves helped mediatize the discussion but the role of users in it was also high. But the most striking is that, of 72 most active tweeters, 27 are media or journalists (and 27 are ordinary users); of the top72 most infl uential ones, there are 21 media/journalist’s accounts; of the Bodrunova, et al.: Twitter-based Discourse on Migrants in Russia: The Case of 2013 Bashings in Biryulyovo International Review of Management and Marketing | Vol 5 • Special Issue • 2015102 top20 in each case, 7 most active and 9 most infl uential accounts belong to media (with 8 in each case belonging to ordinary users). Obviously, media are both very active and infl uential in this Twitter-based discussion. And, of the top72 most active users, it is only media (6 accounts, of both pro- and anti-establishment media outlets) that made their way into top20 infl uential users. Beside the conclusion on mediatization, this also disproves the hypothesis of power of non-institutional infl uencers. This, though, may be explained via the fact that our data collection was hashtag-based, which might have distorted the picture, especially for the “active” time when Biryulyovo provided news. And if media’s active tweeting is very understandable as a strategy, their unexpectedly high role as infl uencers cannot yet be explained in full. But the other part of our hypothesizing on media strategies on Twitter proves right: Media accounts do not engage in intense discussion, they tend just to tweet and self-retweet. Only 2 in top72 users with high input (that is, who are engaged into discussions) are media, and they are not the leading infl uencers. Thus, the results on media accounts seem to support the notion of “real opinion crossroads.” But the Russian Twitter in the case of Biryulyovo appears to be a defi nitive echo chamber in terms of representation of the confl ict sides. Migrants are practically absent from the discussion as well as NGOs. Just fi ve tweets of over 670 (less than 0.8%) were marked by the coders as belonging to a pro-migrant author. In most active tweeting, 7 nationalist accounts outperform 2 pro-migrant ones 2, 6 times (236 vs. 90 tweets). Perhaps this is why the blame was put to migrants and local and federal authorities, not to the local majority or business. H3: Replication of fragmentation of the media-based public sphere on Twitter. So far, analysis of pro-/anti-establishment supports the “real crossroads” version. Twitter represents a picture quite different from, e.g., the Russian Facebook where oppositional, anti-establishment (including business newspapers), and alternative-agenda media dominate. On Twitter, pro-government media outperform anti-establishment and independent news media in activity nearly 3 times, but independent media also have signifi cant presence and are able to form audiences without putting into Twitter the efforts comparable to those by pro- establishment media entities like Lifenews or Voice of Russia. Thus, Twitter, unlike Facebook, effectively gives voice to pro- and anti-establishment media. This may be explained by the fact that the average user profi le of the Russian Facebook user would differ from that on Twitter (though there is no available statistics on that): If Facebook has become an enclave for the Russian cultural and business elite, other social networks have the public with the features of mass audience, and Twitter is closer to this stratum. Media that are most active in case of Biryulyovo are mid- and low-market (Lifenews), state-funded (Voice of Russia), or depoliticized (Metro); oppositional Grani.ru or business paper Kommersant, both being leaders in their audience segments, were almost 7 times less active than Lifenews; the same goes for Radio Liberty and even both Moscow News outlets – they should have been most interested in the case being local media but neither of them is low-market-oriented. Thus, Twitter is, indeed, a crossroads for media representation, but by far pro-establishment and low-market media win the battle in the aspect of presence in the issue-oriented content, providing Twitter users with views not deprived of political slant. H4: Involved media content. These results for H2 and H3 are cross-validated if we take a look at what media content users link to. We expected dominance of web 2.0 content in linking, but the fi rst 673 tweets show that hybrid and online media are represented absolutely equally. It is also evident that the discussion is not encapsulated within Twitter, as references to it are rare. As to the “crossroads” thesis, the pro-/anti-establishment representation basically repeats the one in supporting our opinion on the alleged user profi le and media use patterns. 4. CONCLUSION In general, we had like to say that Twitter in Russia shows more potential to become a real “crossroads of opinions” than the Russian Facebook dominated by anti-establishment discourse or Vkontakte where political discussions are much less visible and, even when do exist, are community-oriented. The discussion on Biryulyovo became highly mediatized and politicized, which shows that media and political actors may fi nd an interested audience on the Russian Twitter. But by far pro-establishment media of mid- and low-market nature dominate this communicative milieu in terms of infl uence as well as in frequency of tweeting and linking, though independent media are also represented among the most active and most infl uential accounts. Hybrid media easily compete in Twitter with online-only media, outperforming them in all positions. Visualization of the web graph has demonstrated that media of various nature have really become central nodes of the discussion, and the mainstream part of the fragmented public sphere is winning the game. Our results combined with previous research (Bodrunova and Litvinenko, 2013b) show that, contrary to today is dominant opinion that the social context is the primary definer for how the online discussions develop in different countries, online communicative milieus in the same country may differ signifi cantly in how political cleavages are reconstructed, as well as in the level of involvement and representativeness of social groups. Our research adds to the evidence that platform features and especially the mean user profi le play an important role in the development of online discussions. The current article presents only partial results on structural and agenda-setting aspects of Twitter-based discussions in Russia; full results on these aspects will also be complemented by the results on framing of the issue, and we will get the full picture of how the infl uencers shape the discussion and who they are. Our research on mediatization aspects and agenda setting on Twitter within migrant-/racial-oriented discussions has already been expanded within the next research project at SPbU, “The role of Twitter in agenda formation in various socio-political contexts” (2014-2015). Within it, Twitter-based discussions in four countries (Russia, Germany, France, and the USA) will be studied in “confl ict” and “calm” month-long periods. As the “confl ict” cases, the Biryulyovo bashings, the killing and violence at Bodrunova, et al.: Twitter-based Discourse on Migrants in Russia: The Case of 2013 Bashings in Biryulyovo International Review of Management and Marketing | Vol 5 • Special Issue • 2015 103 Ferguson, USA, the killings at the editorial offi ce of Charlie Hebdo magazine, and the PEGIDA rallies in Germany were chosen. This research will allow for comparative assessment of both media and sociological features of the Twitter-based discussions and will provide evidence on whether national-bound or platform-oriented factors play a bigger role in formation of the discussion structures and in agenda setting on Twitter. Another extension of the current research is involvement of texts of hybrid (“offl ine-fi rst”) mainstream media in assessing whether explosive discussions on Twitter play a role in setting the agenda in mainstream media. This is planned for the years 2016-2017 for all the four countries. 5. ACKNOWLEDGMENTS The authors are grateful to the School of Journalism and Mass Communications, SPbU, for a research grant “Hybrid media systems and political agendas” (2013-2014, grant #4.23.2203.2013). The authors also express gratitude to Dr. Ivan S. Blekanov and Alexey I. Maximov for developing the web crawling software and conducting of the crawling procedures. The authors acknowledge the big role of student participants of the project as raters (coders) of the tweet samples and are grateful for their contribution to the overall outcomes of the research. The authors contributed equally to the study. REFERENCES Adam, S., Pfetsch, B. (2011), Media agenda building in online and offl ine media - Comparing issues and countries. Paper Presented at the 6th ECPR General Conference. Available from: http://ecpr.eu/ fi lestore/paperproposal/748a2f64-cb7f-4beb-903f-9f21173c21b6. pdf. [Last retrieved on 2015 Oct]. Blekanov, I.S., Sergeev, S.L., Martynenko, I.A. (2012), Constructing topic-oriented web crawlers with generalized core. Scientifi c and Research Bulletin of St. Petersburg State Politechnic University, 5(157), 9-15. Bodrunova, S.S., Litvinenko, A.A. (2013a), Hybridization of the media system in Russia: Technological and political aspects. In: Vartanova, E.L., editor. World of Media 2012. Moscow: Moscow State University Press. p37-50. Bodrunova, S.S., Litvinenko, A.A. (2013b), New media and the political protest: The formation of a public counter-sphere in Russia of 2008-12. In: Makarychev, A., Mommen, A., editors. Russia’s Changing Economic and Political Regimes: The Putin Years and Afterwards. London, England: Routledge. p29-66. Chadwick, A. (2013), The Hybrid Media System: Politics and P o w e r. Av a i l a b l e f r o m : h t t p : / / w w w. d x . d o i . o rg / 1 0 . 1 0 9 3 / acprof: oso/9780199759477.001.0001. De Vreese, C.H. (2005), News framing: Theory and typology. Information Design Journal Document Design, 13(1), 51-62. Dimitrova, D.V., Strömbäck, J. (2012), Election news in Sweden and the United States: A comparative study of sources and media frames. Journalism, 13(5), 604-619. Fuchs, C. (2014), Social Media: A critical introduction. Available from: http://www.dx.doi.org/10.4135/9781446270066. Gitlin, T. (1998), Public spheres or public sphericules. In: Liebes, T., Curran, J., editors. Media, Ritual and Identity. London, England: Routledge. p168-174. Gorny, E. (2009), Understanding the real impact of Russian blogs. Russian Analytical Digest, 69, 7-11. Greene, S.A. (2012), Twitter and the Russian Protest: Memes, Networks and Mobilization. Working Materials, Center for the Study of New Media & Society. Moscow New Economic School. Available from: http://www.newmediacenter.ru/2012/05/22/twitter-and-the-russian- street-memes-networks-mobilization/. [Last retrieved on 2015 Oct]. Grönlund, K., Bächtiger, A., Setälä, M. (2014), Deliberative Mini-Publics: Involving Citizens in the Democratic Process. London, England: ECPR Press. Hermida, A. (2010), Twittering the news. Journalism Practice, 4(3), 297-308. Katz, E., Lazarsfeld, P. (1957), The two-step fl ow of communication: An up-to-date report on a hypothesis. The Public Opinion Quarterly, 21(1), 61-78. Kelly, J., Barash, V., Alexanyan, K., Etling, B., Faris, R., Gasser, U., Palfrey, J. (2012), Mapping Russian Twitter. Publication #2012-3, Berkman Center for Internet & Society, Harvard University, Cambridge. Available from: http://www.cyber.law.harvard.edu/ publications/2012/mapping_russian_twitter. [Last retrieved on 2015 Oct]. Kuchins, A.C. (2007), Alternative Futures for Russia to 2017. A Report of the Russia and Eurasia Program. Washington, DC: Center for Strategic and International Studies. Available from: http://www. csis.org/fi les/media/csis/pubs/071214-russia_2017-web.pdf. [Last retrieved 2015 Oct]. Lee, J.K. (2007), The effect of the internet on homogeneity of the media agenda: A test of the fragmentation thesis. Journalism and Mass Communication Quarterly, 84(4), 745-760. Mancini, P., Mazzoni, M. (2013), Politicie social network: Un trampolino per i media mainstream: Un sistema ibrido (tutto) italiano. Paper Presented at SISP Annual Convent. Florence (Italy), 2013. Abstract. Available from: http://www.sisp.it/convegno/2013/sezioni/70/. [Last retrieved on 2015 Oct]. McCombs, M. (2005), A look at agenda-setting: Past, present and future. Journalism Studies, 6(4), 543-557. Miller, C.C. (2010), Why Twitter ’s C.E.O. Demoted Himself. The New York Times. Available from: http://www.nytimes. com/2010/10/31/technology/31ev.html?pagewanted=all&_r=0. [Last retrieved 2015 Sep]. Noelle-Neumann, E. (1984), The Spiral of Silence: A Theory of Public Opinion - Our Social Skin. Chicago: University of Chicago Press. Pariser, E. (2011), The Filter Bubble: What the Internet Is Hiding from You. London, New York: Penguin Press. Rogers, E.M., Dearing, J.W., Chang, S. (1991), AIDS in the 1980s: The Agenda-setting Process for a Public Issue. Vol. 126. Columbia, USA: Association for Education in Journalism and Mass Communication. Rohozinski, R. (1999), Mapping Russian cyberspace: Perspectives on democracy and the net. Discussion Paper #115. New York: United Nations Research Institute for Social Development (UNRISD). Available from: http://www.unpan1.un.org/intradoc/ groups/public/documents/untc/unpan015092.pdf. [Last retrieved on 2015 Oct]. Ruth, P. (2012), The linguistics of self-branding and micro-celebrity in Twitter: The role of hash tags. Discourse and Communication, 6(2), 181-201. Shirky, C. (2008), Here Comes Everybody: The Power of Organizing Without Organizations. London, England: Penguin Books. Sunstein, C.R. (2007), Republic 2.0. Princeton: Princeton University Press. Synthesio, (2013), 4 Fascinating Facts on the Social Media Landscape in Russia. Available from: http://www.synthesio.com/corporate/ en/2013/uncategorized/4-fascinating-facts-on-the-social-media- landscape-in-russia. Synthesio. [Last retrieved on Oct 2015]. Bodrunova, et al.: Twitter-based Discourse on Migrants in Russia: The Case of 2013 Bashings in Biryulyovo International Review of Management and Marketing | Vol 5 • Special Issue • 2015104 Tewksbury, D. (2005), The seeds of audience fragmentation: Specialization in the use of online news sites. Journal of Broadcasting and Electronic Media, 49(3), 332-348. Toepfl , F. (2011), Managing public outrage: Power, scandal, and new media in contemporary Russia. New Media and Society, 13(8), 1301-1319. Vaccari, C. (2013), Digital Politics in Western Democracies: A Comparative Study. Baltimore: Johns Hopkins University Press. Vaccari, С., Valeriani, A., Barbera, P., Bonneau, R., Jost, T.J., Nagler, J., Tucker, J. (2013), Social Media and Political Communication: A survey of Twitter users during the 2013 Italian general election. Paper presented at SISP Annual Convention. Florence (Italy), 2013. Abstract. Available from: http://www.sisp.it/convegno/2013/ sezioni/70/. [Last retrieved 2015 Oct]. Vartanova, E.L. (2013), Post-Soviet transformations of media and journalism. Moscow: MediaMir. Vis, F. (2013), Twitter as a reporting tool for breaking news. Digital Journalism, 1(1), 27-47. Zubarevich, N. (2011), Four Russias. Vedomosti. Available from: http:// www.vedomosti.ru/opinion/articles/2011/12/30/chetyre_rossii. [Last retrieved on 2015 Oct]. << /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles true /AutoRotatePages /All /Binding /Left /CalGrayProfile (Dot Gain 20%) /CalRGBProfile (sRGB IEC61966-2.1) /CalCMYKProfile (U.S. Web Coated \050SWOP\051 v2) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Warning /CompatibilityLevel 1.4 /CompressObjects /Tags /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages true /CreateJDFFile false /CreateJobTicket false /DefaultRenderingIntent /Default /DetectBlends true /DetectCurves 0.0000 /ColorConversionStrategy /LeaveColorUnchanged /DoThumbnails false /EmbedAllFonts true /EmbedOpenType false /ParseICCProfilesInComments true /EmbedJobOptions true /DSCReportingLevel 0 /EmitDSCWarnings false /EndPage -1 /ImageMemory 1048576 /LockDistillerParams false /MaxSubsetPct 100 /Optimize true /OPM 1 /ParseDSCComments true /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveDICMYKValues true /PreserveEPSInfo true /PreserveFlatness true /PreserveHalftoneInfo false /PreserveOPIComments false /PreserveOverprintSettings true /StartPage 1 /SubsetFonts true /TransferFunctionInfo /Apply /UCRandBGInfo /Preserve /UsePrologue false /ColorSettingsFile () /AlwaysEmbed [ true ] /NeverEmbed [ true ] /AntiAliasColorImages false /CropColorImages true /ColorImageMinResolution 300 /ColorImageMinResolutionPolicy /OK /DownsampleColorImages true /ColorImageDownsampleType /Bicubic /ColorImageResolution 300 /ColorImageDepth -1 /ColorImageMinDownsampleDepth 1 /ColorImageDownsampleThreshold 1.50000 /EncodeColorImages true /ColorImageFilter /DCTEncode /AutoFilterColorImages true /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /ColorImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000ColorACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000ColorImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasGrayImages false /CropGrayImages true /GrayImageMinResolution 300 /GrayImageMinResolutionPolicy /OK /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 300 /GrayImageDepth -1 /GrayImageMinDownsampleDepth 2 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages true /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /GrayImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000GrayACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000GrayImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasMonoImages false /CropMonoImages true /MonoImageMinResolution 1200 /MonoImageMinResolutionPolicy /OK /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 1200 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict << /K -1 >> /AllowPSXObjects false /CheckCompliance [ /None ] /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile () /PDFXOutputConditionIdentifier () /PDFXOutputCondition () /PDFXRegistryName () /PDFXTrapped /False /Description << /CHS /CHT /DAN /DEU /ESP /FRA /ITA /JPN /KOR /NLD (Gebruik deze instellingen om Adobe PDF-documenten te maken voor kwaliteitsafdrukken op desktopprinters en proofers. De gemaakte PDF-documenten kunnen worden geopend met Acrobat en Adobe Reader 5.0 en hoger.) /NOR /PTB /SUO /SVE /ENU (Use these settings to create Adobe PDF documents for quality printing on desktop printers and proofers. Created PDF documents can be opened with Acrobat and Adobe Reader 5.0 and later.) >> /Namespace [ (Adobe) (Common) (1.0) ] /OtherNamespaces [ << /AsReaderSpreads false /CropImagesToFrames true /ErrorControl /WarnAndContinue /FlattenerIgnoreSpreadOverrides false /IncludeGuidesGrids false /IncludeNonPrinting false /IncludeSlug false /Namespace [ (Adobe) (InDesign) (4.0) ] /OmitPlacedBitmaps false /OmitPlacedEPS false /OmitPlacedPDF false /SimulateOverprint /Legacy >> << /AddBleedMarks false /AddColorBars false /AddCropMarks false /AddPageInfo false /AddRegMarks false /ConvertColors /NoConversion /DestinationProfileName () /DestinationProfileSelector /NA /Downsample16BitImages true /FlattenerPreset << /PresetSelector /MediumResolution >> /FormElements false /GenerateStructure true /IncludeBookmarks false /IncludeHyperlinks false /IncludeInteractive false /IncludeLayers false /IncludeProfiles true /MultimediaHandling /UseObjectSettings /Namespace [ (Adobe) (CreativeSuite) (2.0) ] /PDFXOutputIntentProfileSelector /NA /PreserveEditing true /UntaggedCMYKHandling /LeaveUntagged /UntaggedRGBHandling /LeaveUntagged /UseDocumentBleed false >> ] >> setdistillerparams << /HWResolution [2400 2400] /PageSize [612.000 792.000] >> setpagedevice