Twitter,YouTube,and Flickr as platforms of alternative journalism: The social media account of the 2010 Toronto G20 protests

Introduction

On 26 and 27 June 2010, the fourth meeting of the G20 heads of government took place in Toronto, Canada. The summit was focused on organizing a coordinated response to the ongoing world recession. As with many preceding international political and economic meetings, the summit was accompanied by protests, organized by among others the Canadian Labour Congress, Greenpeace, Oxfam, and the Ontario Coalition Against Poverty. In the days before the summit, demonstrations were already taking place in downtown Toronto. On 22 June, about 200 people from Toronto’s gay community gathered to draw attention to the rights of marginalized minority groups. Two days later, about 1,000 demonstrators took to the streets to protest Canada’s failure to sign the United Nation’s Declaration on the Rights of Indigenous Peoples. The biggest protests occurred on the afternoon of 26 June when a crowd of around 10,000 people marched through Toronto. These protests were met with extensive security efforts by the Canadian government, which reportedly deployed as many as 20,000 policemen. On 26 June, it came to a clash after a group of so-called ‘black bloc’ demonstrators broke away from the protest route, smashed the windows of a few stores, and set four police cruisers ablaze. Subsequently, during the evening and the following days, the police arrested over 1000 people (Austen, 2010; CBC News, 2010; Edwards, 2010; Jordan, 2010).

These events are interesting from a media theoretical perspective, as the Toronto Community Mobilization Network (TCMN), which coordinated and facilitated many of the protests, built heavily on major social media platforms for reporting on the demonstrations. The network urged protestors to ‘broadcast breaking news’ using Twitter, YouTube, or Flickr and recommended them to organize their reports with the hashtag ‘#g20report’. The tweets, videos and photos with this hashtag were subsequently aggregated in real-time on the website of the G20 Alternative Media Center.

Evidently, these attempts to appropriate commercial social media as platforms of alternative reporting take place against a long history of activist efforts to establish alternative channels of public communication. As activists have always had difficulty communicating their point of view through the mainstream press, they have tried to develop their own media (Atton, 2002; Atton and Hamilton, 2008; Couldry and Curran, 2003; Downing, 2001; Hackett and Carroll, 2006). Various theorists have suggested that social media indeed provide such powerful alternative platforms of public communication (Bruns, 2008; Castells, 2009; Jenkins, 2006; Shirky, 2008). Manuel Castells, for example, maintains that ‘as people (the so-called users) have appropriated new forms of communication, they have built their own systems of mass communication, via SMS, blogs, vlogs, podcasts, wikis, and the like’ (2009: 65). He calls this development the rise of ‘mass self-communication’. In a similar vein, Clay Shirky claims that now ‘everyone is a media outlet’. He even contends that in the current process of mass amateurization, the profession of journalism is disappearing, just as the ‘scribe’ has once disappeared (2008: 55–80).

These claims, however, raise a number of critical questions, which are examined in this article. First, what kind of account emerges when thousands of people collectively report on an event? Second, although a lot of data is available on numbers of users posting videos, or tweeting on particular events, it is not clear how different users contribute to a specific social media account. Can such an account be considered as a crowd effort, or is it determined by a relatively small number of users? Finally, and most importantly, it is not clear how activist social media accounts should be valued in relation to previous activist efforts to create their own platforms of public communication. To what extent do social media allow activists to realize the ideals of alternative journalism, which have been developed in response to mainstream reporting practices?

This article addresses these questions through a detailed examination of the use of Twitter, YouTube, and Flickr during the G20 protests in Toronto. The analysis focuses specifically on the 11,556 tweets, 222 videos, and 3,338 photos tagged #g20report, which were produced in the 12 days around the summit. While this is a relatively small set in comparison to the massive social media activity in the recent protests in the Middle East and North Africa, it does allow us to specifically examine social media reporting practices. Because the use of social media during the anti-government protests of the past years in Iran, Tunisia, Egypt, Libya, and Syria was largely uncoordinated, the results are also extremely heterogeneous and not necessarily concerned with reporting (Gaffney, 2010; Hwang et al., 2009; Lotan et al., 2011; Zuckerman, 2011). By contrast, the social media activity organized by the #g20report tag did have this particular focus.

To prepare the critical analysis of these social media reports, the next section reviews why activists and media theorists alike have been critical about mainstream protest reporting, and why they have, over and over, argued and strived for the development of strong platforms of alternative journalism. This review should provide insight into the ideals which have been associated with alternative journalism. The social media reporting activity during the G20 protests will be considered in the light of these ideals in the following sections.

Mainstream and alternative journalism

The idea that the mainstream media do not provide a balanced account of social protest has been developed especially in the period from the 1960s to the early 1990s. In this period, critical media theorists maintained that dominant political and economic actors can set the terms of the news agenda because journalists tend to identify these actors as ‘authoritative sources’. By contrast, marginal political actors, according to these theorists, are systematically labeled as deviant, or literally rendered speechless by the press (Becker, 1967; Bennett, 1988; Chomsky and Herman, 1988; Cohen and Young, 1981; Hall et al., 1978).

On the basis of this critical perspective, media researchers have argued that social activism is systematically signified by the mainstream press as illegitimate. For example, Todd Gitlin (1980: 271) has asserted, based on a study of the news coverage of the American anti-war movement during the 1960s, that ‘a demonstration is treated as a potential or actual disruption of legitimate order, not as a statement about the world’. As a result, media coverage is diverted away ‘from critical treatment of the institutional, systemic, and everyday workings of property and the state’. Halloran et al. (1970: 302) have argued that this particular treatment of demonstrations has a lot to do with mainstream journalism’s philosophy of objectivity and impartiality, which leads to a strong focus on the event itself. The authors maintain that by focusing on the event, some aspects of a story are more likely to emerge as newsworthy than others. Violence, which is highly visible, especially tends to become a central theme in demonstration reporting, whereas considerations of the protest issues, as well as background developments, and of the issues involved, tend to be excluded (see also Gitlin, 1980: 27; Murdock, 1981: 210).

Over the past two decades, these kinds of critical studies have been substantially qualified and taken to task for their ‘essentialist’ and ‘monolithic’ understanding of sources, which cannot account for the fact that previously marginal sources can under particular circumstances become secondary or even primary sources (Cottle, 2003, 2008; Deacon and Golding, 1994; Hansen, 1993; Manning, 2001; Schlesinger, 1990). Nevertheless, contemporary media theorists have remained rather pessimistic concerning the opportunities of activists to gain mainstream press attention for their issues. In her study on environmental activism and the news media, Alison Anderson (2003: 123), for example, comes to the conclusion that even though environmental groups, such as Greenpeace, often succeed in gaining campaign-related media coverage, it turns out to be much harder for these groups to secure legitimacy for their objectives. As Dieter Rucht (2004: 29) makes clear, this has a lot to do with how protest groups gain media attention, which is, inter alia, through conflict, spectacle, and newness. These phenomena have mainstream news value, but they are also prone to draw the attention away from the reasons for protest.

Alternative journalism, which includes the media of protest groups, dissidents, and fringe political organizations, can be seen as an attempt to counterbalance the dynamic of mainstream reporting (Atton and Hamilton, 2008: 1; Harcup, 2003: 371). Particularly important in providing such a counterbalance, according to Atton and Hamilton (2008: 86), has been to challenge the mainstream philosophy of objectivity, and of so-called ‘balanced reporting’. As discussed, this philosophy often leads journalists to disregard the reasons for protest, and instead to focus the attention on the spectacular, newsworthy aspects of an event. Atton and Hamilton stress that alternative media are, in contrast, ‘characterized by their explicitly partisan character’. ‘Skeptical of what counts as balance in the mainstream media, they [alternative journalists] seek to set up their own counterbalance.’ In doing so, these journalists, who typically have no professional qualifications, ideally not only devote attention to the issues ignored by the mainstream press, but they also seek ‘to invert the “hierarchy of access” to the news’. Atton and Hamilton maintain that this means ‘explicitly foregrounding the viewpoints of “ordinary people” (activists, protesters, local residents): citizens whose visibility in the mainstream media tends to be obscured by the presence of elite groups and individuals’.

Although media theorists have only recently started to consider alternative journalism more systematically, clear examples of this form of journalism can already be recognized in the 19th-century oppositional presses, which were often connected to specific social movements (Atton and Hamilton, 2008: 17). However, for appreciating the current use of activist social media as a form of alternative journalism, it is more relevant to look at a contemporary online example.

Particularly interesting in the context of this study is Indymedia, the Independent Media Center (IMC), which can rightfully make a claim for the title of main global online network for alternative journalism of the past decade. This network, which at its prime in 2006 featured over 150 media centers around the world, was created in November 1999 to cover the protests against the WTO summit in Seattle. The founders were motivated to establish Indymedia because they feared that these protests ‘would be poorly covered (if at all) by the mainstream, corporate media’ (Tarleton, 2000: 53). The explicit objective of Indymedia was to emphasize issues and include voices not featured in mainstream reporting (Tarleton, 2000; see also Atton, 2003; Coyer, 2005; Downing, 2003; Kidd, 2003; Platon and Deuze, 2003).

Research shows that Indymedia can be considered partly successful as an alternative journalism project. Various studies make clear that the network has, especially in its first years, provided a rich coverage from multiple perspectives of particular events, such as the 9/11 attacks, or the anti-war demonstrations of February 2003. In these instances, IMCs critically gatewatched output from both mainstream and alternative media, and at the same time provided first-hand witness accounts (Atton and Hamilton, 2008: 80; Wright, 2004: 80).

Yet, it also became clear that Indymedia’s reporting, despite the fact that open publishing is at the heart of this project, is still determined by a relatively select group of activists. First, since all Indymedia reporters are activists, and since eyewitness reporting is a key feature of this project, these alternative journalists effectively become ‘recorders of their own reality’ (Atton and Hamilton, 2008: 80). Second, Indymedia is certainly not free from editorial decision-making. Platon and Deuze found that ‘selection of content is common practice for IMC editorial teams’ (2003: 347). A core group of editors or technical staff decides, often on the basis of their personal involvement with the issues at hand, what content can stay on the newswire (2003: 349).

These observations correspond with the research of Atton and Wickenden (2005: 358) on the UK weekly free-sheet and online news site SchNEWS. They found that, in terms of its sourcing practices, ‘SchNEWS’s native reporting might be considered as a form of elite reporting’. More precisely, it presented ‘an alternative hierarchy of sources’, with a strong bias towards activists. Consequently, Atton and Wickenden come to the conclusion that SchNEWS ‘seems to exclude the diversity of voices that comprise a heterogeneous public’ (2005: 355, 368).

The crucial question is how the current social media protest-reporting practices match up. As activists are increasingly embracing social media as the preferred platforms of public communication, it is vital to find out whether these media are instrumental in fulfilling the ideals of alternative journalism. Does the use of Twitter and YouTube permit a large number of people to share their observations and points of view? Moreover, do the social media accounts focus the attention on the protestors’ issues, instead of on violence and spectacle?

Analyzing social media accounts

These questions have been addressed through the following steps. First, all the posts on Twitter, YouTube, and Flickr with the hashtag #g20report were collected for the 12-day period between 22 June and 3 July. Data was collected starting a few days before the actual G20 summit because there was already significant protest activity during these days. The collected data span until after the summit, as several demonstrations took place in reaction to the arrest of over a thousand protestors. Note that the resulting data set only concerns a subset, specifically focused on protest reporting, of the total social media activity on the G20. Thus, other G20-related hashtags, as well as blog posts and forum discussions about the protests and the summit, have not been included. Also left out of the analysis were the protest reports directly uploaded to the G20 Alternative Media Center, as these reports were limited in volume and not hosted by a major social media platform.

The #g20report tweets were collected with the Google Realtime Scraper. ¹ This produced a set of 11,556 tweets for the 12-day period. In turn, by searching #g20report in YouTube, 222 videos were harvested with the assistance of the Tubekit. ² Finally, the 3338 Flickr photos were collected manually through Flickr’s advanced search interface, which allows one to query for particular tags within a specific date range. The variance in the number of results immediately makes clear that there are large differences between the three platforms in terms of user activity, which is further explored in the following sections.

Second, the top results from the Twitter, YouTube, and Flickr sets were collected, using natively digital selection principles (Rogers, 2009). This analytical step builds on the idea that in social media the selection of relevant content does not take place before publication, as in the mainstream press, but after publication by users themselves. In other words, social media function on the basis of gatewatching and not gatekeeping (Bruns, 2008: 69–76; Shirky, 2008: 81–108). In this sense, not all social media content can be considered as relevant. Consequently, in reconstructing the activist social media account of the G20 protests, the examination has focused on those tweets, videos, and photos that were considered relevant by the users.

For Twitter this meant focusing on the top retweets (RT). As various authors have pointed out, retweeting is effectively passing ‘along interesting pieces of information’ (Cha et al., 2010: 3; see also: Boyd et al., 2010: 3). Thus, the most retweeted messages can be considered as the most relevant tweets. Given the real-time nature of Twitter, particular accounts could be reconstructed for each of the 10 days under analysis, making it also possible to show how the #g20report account progressed over time. This was done by selecting for each day the top 10 most retweeted messages. In addition to the retweeted messages, a separate selection was made of the top 10 most tweeted URLs per day. ³ We expanded all the shortened URLs, such as bit.ly, and counted 1651 unique URLs over the investigated 12 days. As Boyd et al. (2010: 2) point out, providing URLs to outside content has been one of the core practices on Twitter from the beginning. In this sense, these URLs can be understood as an essential part of the accounts emerging from the Twitter space.

For YouTube and Flickr, determining the results deemed most relevant by the users works somewhat differently. For these platforms the most relevant results were selected on the basis of numbers of views, and numbers of inlinks from other websites. The inlinks for the individual videos and photos were collected with the Yahoo Inlink Scraper. ⁴ Not surprisingly, these two user-attributed measures of relevance largely overlapped. Selected for further analysis were YouTube videos and Flickr photos which received more than 3,000 views and/or three inlinks from different sites. On the basis of these criteria, 46 videos were selected, and strikingly only one photo. The latter result is already an interesting finding, as it suggests that Flickr only played a minor role in the activist social media ecology. Apart from the one selected photo, which received 7,000 views and 14 inlinks, the other Flickr photos received only a few hundred views and no inlinks.

In the third step of the analysis, it was examined how many users, distinguished on the basis of account names, contributed to the #g20report account on each platform. For Twitter these numbers were subdivided for each day under analysis. This corresponds to the daily rhythm of the protests, and immediately made clear how a much larger number of users contributed to the account during the summit days than in the days before and after the summit. An important question was whether a small set of users or a large crowd determined the social media accounts. To address such question, the Pareto principle is often used. Following this principle, which states that 80 percent of the effects come from 20 percent of the causes, it was investigated what percentage of all posts on each platform came from the 20 percent most active users (Shirky, 2003). Subsequently, it was examined what percentage of the most relevant posts came from the 20 percent most active users.

Finally, all top retweets, photos, and videos were categorized through emergent coding (Stemler, 2001). The material was first examined and a checklist of features, consisting of keywords and characteristic images, was consolidated. Second, the related features were assembled into categories, which can be found in the coding manual in the Appendix. Subsequently, the relevant material was coded on the basis of these categories. This made it possible to make more systematic claims about the #g20report account, and whether it was primarily concerned with the issues of the G20 protests, or the accompanying spectacle.

Important to note is that, in contrast to much of the literature discussed above, this analysis starts from an individualist perspective. It treats users as individuals, and not as part of particular social or cultural groups. Correspondingly, the examined social media accounts are considered as the result of aggregate user activity, and not as the accounts of specific groups. Employing a natively digital approach, the analysis, in this respect, follows the medium: the examined social media platforms present, and also interact, with users as individuals, or aggregates of users. Moreover, since, these platforms are in principle open to everyone, and user identities cannot be verified, no substantive claims can be made concerning possible demographics that contribute to the #g20report accounts.

This conceptual focus should not simply be understood as instigated by the examined social platforms; it effectively corresponds with a much broader transformation of political activism, and politics in general. As a variety of contemporary theorists and researchers have argued, personal identity narratives are replacing collective social scripts as the basis for social organization. Individuals, especially those engaged in transnational activism, are likely to develop political and activist ties through shifting affinity networks based on these narratives, instead of through clearly defined social and cultural group identities (Beck and Beck-Gernsheim, 2002; Bennett and Segerberg, 2011; Castells, 1997; Della Porta, 2005; Giddens, 1991). This can also be observed in the case of the Toronto Community Mobilization Network, which does not claim to represent a specific group or class, but rather aims to connect broad categories of people, including ‘people of colour, indigenous peoples, women, the poor, the working class, queer and trans people’, for a wide array of causes, such as ‘Environmental and Climate Justice, Migrant Justice, and an End to War and Occupation, Income Equity and Community Control over Resources’ (TCMN, 2010).

Users

Considering the users who contributed to the #g20report accounts, the question is whether the adoption of the three social platforms effectively opened protest reporting to a large number of users. Research on Indymedia showed, as discussed, that the protest accounts on this platform were determined by relatively small groups. In a more general sense, countering optimistic claims by authors such as Shirky and Castells, various critics have pointed out that the percentage of users on most major web platforms who actively participate and contribute content is relatively small (Hackett and Carroll, 2006: 49–50; Lovink and Niederer, 2008: 11, 33; Van Dijck, 2009: 44). But is this also true in the case of the Toronto G20 protests, in which the TCMN explicitly called upon protesters to contribute content to the selected social platforms? In other words, do platforms such as Twitter and YouTube facilitate a crowd-sourcing of activist reporting, or is it still determined by a small number of users?

On first impression, the #g20report account indeed appears to have been constructed by rather large numbers of users. This was especially the case on Twitter. During the two days of the G20 summit on 26 and 27 June 2010, when the largest protests took place, more than 800 users (825 and 819 respectively) posted over 3000 tweets a day tagged #g20report. Over the investigated 12-day period a total of 2109 different users contributed tweets. The numbers of users on YouTube and Flickr are clearly less impressive. The 222 YouTube videos were uploaded by 65 different authors, whereas the 3,338 Flickr pictures were posted by only 46 different accounts. Nevertheless, these are still larger numbers than those comprising the kinds of editorial teams of more traditional alternative online media projects, such as Indymedia.

On closer inspection, however, it becomes clear that fewer users effectively determined the #g20report accounts. The data in Figure 1 show that the 20 percent most active users on all three platforms posted over 50 percent of all reports. Of course, this is still much more diverse than predicted by Pareto principle. There are, however, some important variations between the three platforms.

OPEN IN VIEWER

Figure 1.

Posts by top 20 percent of users, per platform

Flickr is clearly the most strongly dominated by a relatively small number of users. The 20 percent most active users posted 69 percent of the selected photos. The uneven distribution of user participation on this platform becomes even more evident when considering that the two most active users together posted about 40 percent of all pictures. In the light of these findings, and especially given that only one picture reached a wider public, it does not seem justified to characterize Flickr’s user participation as successfully crowd-sourced.

Some of the same conclusions can be drawn concerning the distribution of influence on YouTube. As on Flickr, the most active users accounted for a major part of the posts: the top 20 percent users uploaded 62 percent of all videos. Moreover, when the 46 most relevant videos are considered, there is one account that stands out, which contributed 13 (28%) of these videos. Yet, at the same time, what is also striking concerning the most relevant results is how many different users made a contribution to these results. The 46 most relevant videos came from 22 different authors. In fact, several of the top videos, receiving more than 50 inlinks and around 100,000 views, were uploaded by users who only posted one or two videos tagged #g20report. Thus, even though one account had a relatively large impact, the user influence on YouTube was already more distributed than on Flickr, and arguably also more so than on older alternative online platforms.

Examining user participation and influence is the most complex on Twitter. It is complex because #g20report activity changed dramatically from day to day on this platform. In the days preceding and following the summit, the number of users was fluctuating between 68 and 613. During the summit, when the largest protests took place, more than 800 users were using the hashtag. As Figure 2 shows, it was also during these two days that the 20 percent most active users contributed a larger share of the total number of tweets than on the other days. During the summit days, the 20 percent most active users produced over 70 percent of the tweets, approaching the Pareto distribution. The explanation for this is that the top users suddenly started to greatly increase their Twitter activity during the summit: from an average of five to six tweets per day to an average of about 14 tweets per day. The top three users on 26 June even posted 97, 92, and 88 tweets respectively. These figures suggest that the 20 percent top users, especially when it mattered, exerted a large influence over the #g20report account on Twitter.

OPEN IN VIEWER

Figure 2.

Tweets by top 20 percent of users, per day

Still, the question is whether the conclusion should be drawn that a relatively small number of users determined the Twitter accounts of the protests. On both summit days, the top 20 percent included more than 160 different users, which cannot be considered little. More fundamentally, examining the retweets, it is striking how ‘interactive’ the #g20report account is, especially when contrasted with other samples. In fact, analyzing a random sample of 720,000 tweets, Boyd et al. (2010: 3) concluded that ‘3 percent of the tweets are likely to be retweets’. By contrast, as Figure 3 indicates, for the #g20report hashtag, the retweets constituted more than 50 percent of all tweets on each of the investigated days. This suggests that #g20report can to some extent be characterized as a collective account of the protests. This impression is further reinforced by an examination of the individual tweets, which show that not only the most active users were retweeting, but also those who only posted one or two tweets. At the same time, it should be noted that retweeting is a rather lightweight form of participation, which demands a relatively small effort from the user.

OPEN IN VIEWER

Figure 3.

Number of tweets, per day

Overall, the activist social media reporting of the G20 protests was clearly more crowd-sourced than the reporting through older alternative online media. Yet, there were crucial differences between the three platforms. On Flickr, and to a lesser extent on YouTube, reporting was still determined by a relatively small number of users. On Twitter, protest reporting as well as content selection appears to have been much more of crowd-produced affair, especially owing to the practice of retweeting.

Social media protest accounts

Mainstream protest reporting has been extensively criticized for focusing too much on the spectacle and the occasional violence that accompanies protest activities, and for not paying sufficient attention to the reasons for protest. Historically, the objective of alternative journalism, as discussed, has been to counterbalance mainstream reporting by specifically devoting attention to the social, economic, and political issues underlying particular protests. The question is whether the use of Twitter, YouTube, and Flickr by the G20 protesters accomplished this objective. What kind of account emerges when protest reporting is crowd sourced?

Coding the most relevant results from Twitter, YouTube, and Flickr made one thing clear above all, the #g20report account was primarily about police activity. Given the overwhelming police presence and eventual police violence, this is of course not surprising. On all platforms, with some variations, it was documented in minute detail how the Toronto police were all around: patrolling the streets, checking IDs, and guarding the three-meter high security fence around the summit’s convention center. Moreover, the social media reports documented how the police ‘illegally searched’, detained, and arrested people. And finally, the reports discussed how the police were ‘beating people’, ‘breaking and stealing’, and ‘firing on the crowd’. Among this omnipresent police activity, the protestors’ own activities, such as the demonstrations for ‘Native Land Rights’, ‘Queer Liberation’, and the ‘Environment’, were largely lost. For those reporting through the #g20report, oppressive police activity became the main issue.

Figure 4 shows how overwhelmingly the police-oriented accounts figured among the most relevant results. Note that the percentages in this figure indicate the weighted share of the videos, RTs, and outlinks. The weighted share has been determined by calculating the relative spread of the videos and tweets. For the YouTube videos, the number of inlinks that each video receives are reflected in the proportion per account type. For the retweet and Twitter outlink pie charts, the number of retweets has been taken into account.

OPEN IN VIEWER

Figure 4.

Relative spread of account type per platform

The police were the most present in the YouTube videos, many of which were accompanied by titles and descriptions full of moral outrage. Of the ‘most relevant videos’, the weighted share of the police-activity oriented videos was 89 percent. Coded in this category were videos documenting the overall police presence, and the extensive security measures by the Canadian authorities. These videos typically showed images of roadblocks, the security fence, and of policemen lining up, or patrolling the streets. Moreover, also in this category are videos of people being stopped in the street, getting searched, or arrested. Finally, about half of the ‘police activity’ videos documented police violence, showing policemen suddenly coming in and grabbing protestors, chasing them away, or shooting rounds of tear gas. ⁵ Two of these violence-oriented videos, receiving more than 100 inlinks and more than 100,000 views, can be considered as the most prominent activist social media reports of the G20 protests.

Among the most relevant YouTube videos, it is hard to find the original reasons, or the issues of the protests: the weighted share of the videos coded as such was only 2 percent. One of these is a background news program from the Real News Network, which discusses the G20 summit itself, which is virtually absent from most of the other social media reports. The other issue-oriented videos contain images of demonstrations for peace, global justice, and against the G20 summit. ⁶ None of these videos received anywhere near the number of views or inlinks of the two top police-oriented videos. Thus, not only did few of the most relevant videos visualize the reasons for protest, the videos that did were certainly not among the most prominent reports as one can clearly see from Figure 4.

The one relevant Flickr photo directly corresponds with the police-oriented character of most of the top YouTube videos. This photo portrays a lone man holding a sign saying ‘Everything is O’, while a cordon of policemen can be seen in the background. ⁷ As the other photos did not receive any inlinks, and at most only a few hundred views, this picture was effectively the only prominent G20 protest report from Flickr.

Finally, on Twitter, the protest issues were also drowned out by reports focusing on police activities. Only 4 percent of the volume of retweets was concerned with the reasons for protest, mentioning among others Native Land Rights and queer activism. Retweets concerned with police activities constituted 57 percent of the top retweets. This is remarkable given that the top retweets can be considered the most collective account. Clearly, the overall police-oriented character of the social media reports was not due to the fact that only a few users posted these reports. The interactive nature of the Twitter retweet account certainly did not seem to generate more attention for the original protest issues.

Figure 4, however, also shows that these issues did receive a little bit more attention in the most frequently retweeted URLs; that is in the linked blogs, newspaper articles, and activist websites. In total, 8 percent of the weighted share of referenced sources devoted attention to the original reasons for the protests. Of course, 8 percent is still not a high figure, especially when considering that no less than 62 percent of the referenced pages were focused on police presence, arrests, and violence.

What makes Twitter a fascinating object for content analysis is that its real-time nature makes it possible to reconstruct how an account unfolded from day to day. The analysis of the top retweets shows how particular stories or themes took central stage in the Twitter account and subsequently withered to the background. In this sense, Hermida is right to argue that ‘Twitter can be considered as an awareness system’ (2010: 301).

Examining the top URLs and retweets, it becomes clear that the protest issues were not completely invisible for the entire period under analysis. As visualized in Figures 5 and 6, respectively showing the volume of the retweets and retweeted outlinks from a particular account in terms of the times these have been retweeted, it becomes apparent that in the days preceding the summit, the issue-focused retweets did make it to the most relevant results. For example, on 23 and 24 June respectively one-quarter and one-third of the weighted share of the top retweeted URLs referred to webpages discussing among others the ‘native land rights’, ‘queer resistance to the G20’, and ‘women’s response to the G20 health agenda’.

OPEN IN VIEWER

Figure 5.

Proportion of top retweets, per account type

OPEN IN VIEWER

Figure 6.

Proportion of top outlinks from Twitter, per account type

Yet, rather quickly, police activity became the main focus of the #g20report retweets. First, the top retweets and referenced webpages highlighted the police presence itself, describing Toronto as a ‘police state’ and discussing the ‘patrolling cops’, and ‘undercover police’. As the summit was approaching, the focus shifted slightly to ‘illegal searches’ and ‘journalists are being detained and arrested’. During the subsequent summit days, these types of retweets were accompanied by reports of police violence: ‘police beat people with batons’, and ‘cops fired on crowd’. At this stage, none of the top retweets and linked pages any longer reported about indigenous and queer rights, or peace and environmental issues.

This even continued to be the case after the summit ended, when it became clear that over a thousand people had been arrested. Most of the post-summit top results were coded under the label ‘condemning the police violence and the mass arrests’. Another intriguing category that emerged contained the retweets and referenced pages claiming that the black bloc rampage in downtown Toronto was effectively part of a police conspiracy. Many reports observed that the police had done very little to stop black bloc demonstrators (not actually a particular activist group but a protest tactic) from destroying shop windows and police cruisers during the first day of the summit. Subsequently, the conclusion was drawn that the black bloc rampage was used to legitimize the arrests and police violence that followed. No fewer than 13 percent of the top retweets were concerned with this ‘agent provocateur’ theory. One of the top retweets during three consecutive days (26–28 June) read, for example, ‘confirmed: @torontopolice left junk cars in streets to be torched by agents provocateurs’. These types of claims were also visible on YouTube. The weighed share of videos that could be coded in the black bloc category was 8 percent. In absolute numbers, the ‘black bloc’ videos were even more evident as 8 out of 46 videos fall in this category.

Taken together, it is clear that violence, arrests, and the overwhelming police presence, as well as the anger about this, completely dominated the #g20report account on all three social media platforms. Eventually these issues became the protestors’ main focus, overshadowing the original reasons for the G20 protests.

Discussion

How can we understand the #g20report social media accounts within the larger history of alternative media? On the one hand, the attempt to appropriate the major social media platforms for protest reporting can be considered as at least an interesting attempt to crowd-source alternative journalism. One of the main problems of activist reporting was, even on open publishing platforms such as Indymedia, that ultimately the resulting account was based on the observations and experiences of a small group of insiders. This obviously runs counter to the alternative journalistic ideal to include a wide diversity of voices in news reporting. This research project showed, however, that the same problem can, and does, reappear on social media platforms. Especially on Flickr, and to a lesser extent on YouTube, the #g20report account was based on the observations of only a small number of users. Clearly, the most promising platform for crowd-sourcing alternative reporting was Twitter. Not only did this platform have the most contributors, retweeting also involved users, albeit in a light weight fashion, in allowing specific accounts to become dominant.

On the other hand, when examining the #g20report accounts themselves, the use of activist social media appears much less of a success from the perspective of alternative journalism. Examining these accounts, it becomes clear that a highly problematic reversal appears to be taking place. In 1970, Halloran et al. criticized the mainstream media for adopting an event-oriented approach to protest reporting. As a result, the mainstream press is, according to these authors, heavily focused on the spectacle, especially the violence, which accompanies many protests. Although the conceptual approach to mainstream protest reporting has been considerably refined over recent decades, the observation that the press pays scant attention to the reasons for protest has been repeated over and over again.

Today, it appears that precisely the same observation can be made about the activist social media protest accounts. Like mainstream reporting, these accounts were dominated by the violence that accompanied the protests. The particular dynamic between the overwhelming police presence and violence, and the protestors’ efforts to document the police actions, produced accounts that were squarely focused on the violence and spectacle that accompanied and eventually overtook the protests. Of course, a major difference from traditional mainstream protest reporting is not that the protesters in the social media account are depicted as deviants, or outlaws, but that attention is paid to excessive police violence. Nevertheless, the result of activist social media reporting is just as tragic: the attention is drawn away from the original issues at stake in the protests.

From a democratic perspective, it is vital that these issues, ranging among others from minority rights to environmental degradation, are opened up to public contestation. Particularly in the absence of transnational democratic structures and procedures, the international political and financial summits are the most important sites to contest these issues on the street (Bennett, 2004; Castells, 2009: 299–415; Dean, 2003; Sassen, 2006: 338). Yet, to do so successfully it is crucial that these protests, and more importantly the issues at stake, are actually covered by the news. Since the mainstream press has historically done a rather poor job, social activists have rightfully taken protest reporting in their own hands.

In the light of this research project, it is, however, very much the question whether social media, such as Twitter and YouTube, should be the main platforms to facilitate this type of reporting. Currently, this is a particularly pertinent question, as activists are increasingly turning to these media as their preferred platforms of communication.