Every little helps? YouTube, sousveillance and the ‘anti-Tesco’ riot in Stokes Croft

On 21 April 2011, violence flared in the Stokes Croft area of Bristol following a police raid on a squat. Media coverage suggested that this riot was a manifestation of the campaign against the opening of a Tesco supermarket in the area. Footage later emerged on YouTube, which appeared to support claims by local residents that the violence was caused by heavy-handed police tactics rather than the anti-Tesco campaign. This study uses a critical thematic analysis to explore the comments left by those who viewed these acts of ‘inverse surveillance’, or sousveillance. Results indicate that YouTube provided a space in which alternative views on the Stokes Croft riot were seen and heard but the views of many commentators still appeared to be influenced by the news media.


Introduction
Social media played a key role in transmitting information during the disturbances seen in English cities such as London, Birmingham and Manchester in August 2011. Messages posted on sites such as Facebook and Blackberry Messenger (BBM) helped mobilise many of the rioters by creating an impression that the police had lost control of these cities (Morrell et al., 2011;Riots, Communities and Victims Panel, 2012). Eyewitnesses and law enforcement agencies used Twitter to refute a number of unsubstantiated claims in relation to the scale of the disorder that had circulated on the microblogging site (Procter et al., 2011). Many of these messages would later be used as evidence to prosecute those who participated in the disturbances or incited others to do so. In one such case, two men were sentenced to four years in prison for posting messages on Facebook that were said to have incited violence in their respective home towns. 1 This paper adds to the debate over how social media may be used to transmit citizen perspectives on civil unrest by examining audience responses to YouTube footage of the disturbances in the Stokes Croft area of Bristol a few months prior to the English riots. Both local and national press claimed that the violence on 21 April 2011 was linked to the ongoing campaign against the opening of a Tesco Express store in the area. 2 Local activists used online spaces, such as Bristol Indymedia, to refute these claims and to highlight the heavy-handed tactics of the police during an operation to evict the residents of a squat 3 known as the Telepathic Heights. 4 This study focuses on the comments left in response to videos uploaded to YouTube by those who witnessed the violence that flared in the area. The paper will examine the extent to which the commentators : felt that the police tactics were heavy-handed and whether they contextualised these events by referring to the anti-Tesco campaign. It does so by reviewing the relevant literature on the use of social media for sousveillance, and presenting the findings of a critical thematic analysis of 1018 comments posted in response to the four most commented upon videos. In this way, it will explore whether the use of social media for sousveillance has the potential to elicit support for groups whose narratives do not always feature in mainstream media coverage of civil disturbances.

Social media and the 'Battle of Stokes Croft'
A police raid on the Telepathic Heights squat on Cheltenham Road, a main thoroughfare in the Stokes Croft area, was said to have been the catalyst for the clashes between the police and rioters. 5 The Assistant Chief Constable of the Avon and Somerset Constabulary, Rod Hansen, defended this raid as 'positive action' by the police to protect the public from the petrol bombs that were allegedly being made inside the squat for use during a planned attack on the controversial Tesco store (Hall, 2011). Local and national media coverage suggested that the 'Battle of Stokes Croft' had been 'prompted by protests against the opening of a Tesco Store' (Dutta and Duff, 2011) rather than the police operation to evict the squatters. Local residents, who referred to the police tactics as 'unfathomable', refuted these claims and questioned Hansen's assertion that petrol bombs were being made inside the squat. The news media were criticised not only for the inaccurate report that the Tesco store had been firebombed by protesters but also for 'copying and pasting' the police news release into coverage (Gallagher, 2011). One eyewitness went as far as to suggest that the media framing of the violence had benefited Tesco as it had not only 'criminalised the squat' but also turned the anti-Tesco protesters into 'evil petrol bombers' (Bristol Indymedia, 2011).
Claire Milne, spokesperson for the No Tesco in Stokes Croft group, condemned the violence but acknowledged that it was perhaps inevitable given that 'local voices were not being listened to ' (Bristol Evening Post, 2011). The No Tesco in Stokes Croft campaign argued that the development would not only threaten the future of a number of local independent traders but would also destroy the 'unique character' of Stokes Croft (People's Republic of Stokes Croft, 2011). It was ultimately perceived as unnecessary because of the presence of several supermarkets in neighbouring areas. The group had conducted a survey in March 2010, which found that 93 per cent of local residents opposed the plans for a new Tesco store on Cheltenham Road (No Tesco in Stokes Croft, 2010). The group used blogs such as 'Boycott Tesco' (http://.boycotttesco.wordpress. com) and its eponymous Facebook page to publicise its campaign, with the proceeds of the sale of merchandise bearing the 'Think Local, Boycott Tesco' logo used to fund the application for a judicial review of the decision to grant planning permission for the store. 6 The campaign had also been publicised by the People's Republic of Stokes Croft (PRSC), the local community association set up in February 2009 to remodel the area as a 'Cultural Quarter'. The PRSC asserted that the 'ethos of sustainability' would inform its plans to develop the area and criticised the plans to open the Tesco store on Cheltenham Road against the wishes of local residents (People's Republic of Stokes Croft, 2011). Therefore, it was perhaps no surprise that these groups used their websites to condemn the police raid on Telepathic Heights and to refute reports in the media that suggested that the ensuing violence was linked to the anti-Tesco campaign.
Although there was little evidence to suggest that the violence had been organised on sites such as Facebook, real-time information from eyewitnesses could be accessed under the #stokescroft hashtag on Twitter. A selection of these tweets, collated and published by a local blogger, suggested that the violence had started after a 'police charge' to disperse local residents who had gathered on the streets in protest against the arrest of the four alleged 'petrol bombers' in the squat (City Interhacktives, 2011). Clearly, it is difficult to establish with any degree of certainty whether the information conveyed in these tweets was accurate, as demonstrated by the recent study of how Twitter was used to spread rumours during the English riots in August 2011 (Procter et al., 2011). However, several eyewitnesses corroborated these allegations in first-hand accounts of the riot that were published on the Bristol Indymedia site a few days later. One local resident criticised the police for 'randomly charging around the place' and felt that the use of 'horses, vans and the police helicopter' had antagonised the Stokes Croft community (Bristol Indymedia, 2011). It was claimed that the 'bungled' police operation had resulted in many unprovoked attacks by riot police officers upon local residents who had gathered to see what was happening in their area (Conner, 2011). This study will explore the extent to which YouTube footage of the riot, uploaded by eyewitnesses, elicited sympathy amongst commentators for these claims that the police operation was heavy-handed.

Social media, sousveillance and the public sphere
Eyewitness perspectives can now be shared through the act of recording footage on a mobile phone and sharing it on sites such as YouTube. This is one of the emerging social practices which Benkler (2007) suggests have fundamentally challenged the idea of the 'mass-mediated' public sphere, the concept that highlights the role of the traditional media in providing spaces, free from direct interference from the state, for citizens to discuss issues of mutual interest (Habermas, 1996). This could be viewed as a shift in informational power that has the potential to redefine journalism, allowing previously marginalised voices to be heard in a public sphere that is co-created by both citizen and professional journalists (Deuze, 2009). Social media has enabled new forms of citizen journalism, the process whereby citizens play an active role in the 'process of collecting, reporting, analyzing and disseminating news and information' (Bowman and Willis, 2003: 9). The mainstream media has often used content posted online by eyewitnesses to illustrate the human suffering caused by natural disasters such as the South Asian tsunami in December 2004 (Allan et al., 2007). Yet, a critique of citizen journalism has emerged in relation to the uneven nature of its coverage and the problems associated with the verification of content posted online by eyewitnesses. Critics argue that these are 'accidental journalists' who happen to be 'somewhere that something was happening' but who do not adhere to professional journalistic norms, such as fact-checking (Allan et al., 2007: 378). Mainstream news organisations might also be vulnerable to online hoaxes if they are unable to verify the authenticity of this 'amateur' footage, leading some journalists to question the credibility of user-generated content (Singer and Ashman, 2009). Moreover, a sceptical interpretation of the networked public sphere suggests that it promotes homophily in online communities (Sunstein, 2007), leaving new voices increasingly isolated (Lax, 2009). Hence, scholars such as Hindman (2009) have argued that the participatory cultures of Web 2.0 are more likely to help facilitate new strands of investigative journalism than to increase public participation in policy debates.
This study adopts the theoretical framework of sousveillance, defined here as a form of inverse surveillance that empowers citizens through their use of technology to 'access and collect data about their surveillance' (Mann et al., 2003: 333), to analyse the Stokes Croft riot footage on YouTube. This framework emerged from the critique of surveillance practices that suggests that the pervasiveness of organisational surveillance threatens the autonomy of the individual (Stanley and Steinhardt, 2003). Clearly, viewer perception of user-generated content may determine whether it is defined as inverse surveillance. Many viewers might acknowledge how technology empowered those who witnessed an event but still define this content as 'accidental' or 'citizen journalism'. Arguably, eyewitness perspectives can be defined as sousveillance if they focus specifically on the actions of the police. Some may be clear examples of hierarchical sousveillance, when the witness purposively records the actions of authority figures, such as police officers, and often has a clear political agenda for sharing this material (Mann, 2004). The video footage of the police assault on Rodney King in January 1991 is probably the most well known example of this. Los Angeles resident, George Holliday, covertly filmed this assault and passed the video footage to local television station, KTLA. The tape was subsequently picked up by global news networks such as CNN and later used in the trial of the four police officers charged with the assault (Mann et al., 2003). However, it is reasonable to presume that not all those who witnessed the Stokes Croft riot set out to collect video evidence in this way. Rather, the use of mobile phones by local residents to record personal experiences may, in many cases, have been transformed into a form of inverse surveillance through its dissemination on YouTube, thus raising questions about the policing of the 'Battle of Stokes Croft' that was captured on camera. This personal sousveillance, as conceptualised by Mann et al. (2003), would appear similar to the accidental strand of journalism insofar as it is not a purposive activity designed to capture data about authority figures such as the police.
Both personal and hierarchical sousveillance impulses can be detected in the work of Salam Pax, also known as the 'Baghdad Blogger' (Bakir, 2010). His 'Dear Raed' blog provided an eyewitness account of everyday life in Baghdad before and after the 2003 invasion of Iraq. The Guardian newspaper in the United Kingdom started to print extracts from the blog in March 2003, turning Pax into a minor celebrity (Pax, 2003). The blog that Pax claimed that he started to provide information to his friend based in Jordan became a focal point for public debate about the impact of the war upon the Iraqi civilian population. This would appear to suggest that sousveillance requires analysis and dissemination in order for it to be transformed into citizen journalism (Häyhtiö and Rinne, 2009).
The individuals who choose to document their experiences by posting video footage online are under no obligation to contextualise the events depicted therein. This may allow for multiple interpretations of this content, leaving the witness unable to predict the outcome of the release of this material. For example, the footage shot by George Holliday was later used to illustrate Rodney King's aggressive behaviour during the incident, leading to the acquittal of the four police officers that was said to have sparked the 1991 LA riots. 7 There may also be unexpected consequences for those witnesses who do not set out to hold authority figures to account, as demonstrated by the release of pictures in April 2004 showing the abuse of prisoners in Abu Ghraib prison. Many of these 'trophy shots' had been taken by the prison guards to record their own experiences in Iraq, presumably with the intention that they might be shown to close family and friends but would not be broadcast in the mainstream media (Beier, 2007). The publication of these images by the global news media not only led to the trial of the nine US army reservists responsible for the abuse but also made it increasingly difficult for the Bush administration to deny the existence of the covert policy of torture in Iraq (Bakir, 2010). Hence, eyewitness perspectives may be perceived as a form of hierarchical sousveillance even if this was not the original intention of the witness.

Approaches towards the study of YouTube comments
Sousveillance will be used in this study as a conceptual framework for the analysis of videos that show the actions of the riot police during the Stokes Croft riot. The analysis of the comments will provide an insight into how online audiences respond to this use of technology by citizens to mirror state surveillant practices (Mann et al., 2003), and provide first-hand accounts that contradict the dominant media narratives on events such as 'Battle of Stokes Croft'. There are several theoretical approaches that can be used to explore how users respond to user-generated content uploaded to YouTube. Activation theory suggests that content will not stimulate interest amongst viewers if it is either too powerful or too weak (Harrington et al., 2006: 142). The videos that activate the optimal level of stimulation for the viewer are most likely to hold their attention. The social identity model of deindividuation effects (SIDE) proposes that users are more likely to be influenced by content if it is both created and validated by individuals who appear to belong to their social groups (Postmes and Baym, 2005). The salience of social identities and group norms amongst anonymous online peers can be illustrated by the homophily in online discussion forums and social networks (Cho and Lee, 2008). However, some scholars suggest that the SIDE model is best applied to relatively small groups and remain sceptical about whether ingroup-outgroup dynamics are influential in interactive groups that form spontaneously (Hogg and Tindale, 2005). Walther et al. (2010) argue that traditional media and peer networks are just two of the many sources of influence in online spaces; neither can fully explain how individuals respond to social cues in online communities. Clearly, it may be difficult for the eyewitness to predict how YouTube viewers will respond to their footage, particularly given that lurkers may view it but choose not to comment (Papacharissi, 2010). Previous research has suggested that there may be only one comment for every 204 views of a video on the site and that the majority of commentators are male and in their mid-to-late twenties (Thelwall et al., 2012;Walther et al., 2010).
Therefore, the comments left below YouTube videos only show how a small section of the audience are 'making meaning' of this content, thereby claiming their right to a voice within a 'placeless public debate' (Van Zoonen et al., 2011: 259). Large-scale quantitative studies of YouTube have found that most comments tend to be moderately positive in sentiment with the notable exceptions of entertainment and activist videos that generate mixed responses (Thelwall et al., 2012). Comments on videos that focused on Geert Wilder's anti-Islam film, Fitna, illustrate the extent to which activist content may polarise opinion on the site. Commentators frequently traded insults and tried to silence those who held different views on the controversial film (Van Zoonen et al., 2011). The videos did not appear to facilitate dialogue between commentators about the claims made about Islam in the film. Similar results emerge from analysis of comments generated by the 'Bus Uncle' incident in Hong Kong in April 2006. Footage of the argument between a middleaged man and a young male passenger on a bus generated sarcastic responses rather than rational debate over the use of mobile telephones on public transport that had led to this row (Chu, 2009). Nevertheless, the use of derogatory language by commentators might be conceptualised as a mode of expression that shows some level of engagement with the subject matter. For example, a number of themes emerged from analysis of the provocative and often inflammatory reactions of commentators to YouTube footage of the shooting of Oscar Grant by US Transit police in January 2009, ranging from criticism of bystanders for their passivity to support for the citizen journalists who had recorded the events on their mobile telephones (Antony and Thomas, 2010).

Research questions
Specifically, two research questions emerge from the literature in relation to the use of social media for sousveillance during the Stokes Croft riot: 1) To what extent were the comments congruent with the assumed expectations of those who posted the videos? 2) To what extent did commentators appear to perceive this footage as a form of hierarchical sousveillance?
A qualitative thematic analysis of comments on the YouTube footage of the Stokes Croft riot was conducted in June 2011. Data was collected on the 19 May, allowing four weeks after the disturbances for those eyewitnesses who intended to share their footage to do so. Three steps were required to create this corpus. First, several terms relating to the police action during the Stokes Croft riot were entered into Webometric Analyst (http://lexiurl.wlv.ac.uk). 8 This link analysis software allows users to submit key word searches to YouTube via its application programming interface (API). The use of these terms generated a corpus of 72 videos. The next step was to remove content, such as news media reports, that did not provide eyewitness perspectives on the policing of the riot. Footage of the violence that occurred in the area a week later was also excluded, leaving a corpus of 52 videos (N = 52).
Webometric Analyst was used to identify the four most commented upon videos in this corpus, all of which had been uploaded to YouTube within 24 hours of the disturbances. Although these were decontextualised events caught on camera, they appeared to provide some evidence to support the claims made by local residents that the police tactics were 'unfathomable'. All four videos feature the scuffles that broke out between bystanders and riot police who had sealed off Cheltenham Road during the disturbances. One video (Tesco Gets Ramraided) shows a man dressed in a tracksuit being pushed away by a police officer seemingly without any warning, prompting a wave of insults from the crowd. Another video (Stokes Croft Riot) captures the moment when riot police jump out of a stationary police vehicle and run after a crowd of bystanders who appear to be recording the unfolding events on their mobile phones. The third (Bristol Stokes Croft Riot) depicts footage of the police helicopter that was the subject of many complaints from local residents on sites such as Bristol Indymedia. In the final video (Stokes Croft, Bristol Riots), a convoy of police vehicles is shown moving through the area as one unidentified eyewitness is heard saying, 'Are they going to that squat again? What are they doing?' A recurring theme in all of the videos is the confusion amongst the crowd caused by the police tactics. The police are shown marching up and down Cheltenham Road from multiple perspectives in one of these videos before positioning themselves in front of the vandalised Tesco store (Bristol Stokes Croft Riot).
The witnesses also managed to record footage of the anti-social behaviour of the rioters. A group of hooded individuals is shown smashing the windows of an abandoned police vehicle in one video and then seen carrying police riot shields as they use burning litter bins to block the road in another. People are heard laughing as the witness records the attack on the Tesco store by masked individuals (Stokes Croft, Bristol Riots). However, some eyewitnesses are vocal in their condemnation of this anti-social behaviour. In Bristol Stokes Croft Riot, a bystander says 'that's not fair' after watching a young man in a red baseball cap push over a car trailer that had been parked in the street. Another member of the crowd screams, 'pacifism -stop being violent!' as the sound of broken glass can be heard in the background and bins are set on fire nearby. The peaceful protest of local residents against the Telepathic Heights squat eviction only features in Stokes Croft Riot, not surprisingly given that the study focused on footage of the police actions during the disturbances.
Only a small subset of the total online audience that had accessed this content left comments, as demonstrated by the comments per view ratio, (see Table 1). This was a finding congruent with previous research that suggested that the majority of individuals who access online spaces rarely make their presence felt in online spaces through public contributions (Papacharissi, 2010). It should also be noted that 652 of the earliest comments for Bristol Stokes Croft Riot were not included as Google only allows access to the last 1000 comments left under each video. 9 Comments were exported from YouTube into an Excel spreadsheet for coding. Two coders read each comment in order to ensure that they conformed to the requirements of the study; namely, that they should demonstrate some form of engagement with the sousveillance videos under review. Angry back-andforth exchanges between commentators, almost invariably involving the use of offensive language about the residents of Stokes Croft, were excluded as per previous studies (Antony and Thomas, 2010), leaving a sample of 1018 text-based comments. The number of individual users responsible for these comments varied across the videos. The Bristol Stokes Croft Riot video generated much discussion amongst a handful of viewers, with 30 comments linked to a conversation between two users about the policing of the scenes captured on camera. However, there were few such exchanges in the corpus, probably on account of the sampling strategy, which focused on comments that engaged directly with the footage rather than the dynamics of these conversations. Orbe and Kinefuchi's (2008) criteria of repetition, recurrence and forcefulness were then used to identify themes emerging from the comments. Codes were inductively coconstructed through manual data analysis in a similar vein to previous studies of YouTube comments (Antony and Thomas, 2010; Lee, 2012). There were three phases of data analysis required in order to identify key themes. First, an analysis of the entire dataset was conducted in order to identify the words and phrases most commonly used by commentators. The next step was to examine how these phrases related to the commentators' interpretation of the police operation and the link between the violence and the anti-Tesco campaign. For example, some comments characterised the campaigners as 'anarchists' who had instigated the violence and expressed sympathy for the police officers who were 'only doing their jobs' (Tesco Gets Ramraided). However, there were also comments that refuted these claims and appeared critical of the 'heavy-handed policing' that was said to have contributed to the violence (Stokes Croft Riot). The final step was to identify elements of these comments that corresponded to the criterion of forcefulness. Drawing on previous research (Antony and Thomas, 2010), the use of excessive punctuation, upper case letters and derogatory language in phrases such as 'BRISTOL BORN AND BRED!!' (Tesco Gets Ramraided) meant that they were categorised as forceful. This was considered necessary in order to identify emphasised reactions to both the events captured on camera and the broader debate over who was responsible for the violence. Although YouTube's terms of service state that they cannot 'guarantee any confidentiality with respect to any content' submitted by a user (YouTube, 2012), ethical guidelines for online research emphasise the obligation to protect online participants from any additional harm that might arise from the use of their data in academic publications (Ess and AoIR Ethics Working Committee, 2002;The British Psychological Society, 2007). Recent scholarship has focused on the need to develop new strategies for the anonymisation of social media datasets, particularly when it is not possible to obtain informed consent from users prior to the use of the data (Zimmer, 2010). These issues were addressed by adopting an ethical stance which acknowledged the rights of these 'data producers' as human participants and the responsibilities of researchers to adopt working practices that protect their privacy when appropriate to do so (Neuhaus and Webmoor, 2012;Whiteman, 2012). Previous research conducted on YouTube has tended to focus on less sensitive issues and commentators have been identified through the publication of their username (Antony and Thomas, 2010;Lee, 2012). This was not considered appropriate for the analysis of comments generated by the Stokes Croft riot footage. Clearly commentators who expressed support for the rioters might feel that their identification could make them more vulnerable to prosecution, as discussed earlier. Therefore, personally identifiable information (PII), such as pseudonym, age and gender, was removed from the results presented below. Extracts from these comments are also paraphrased, because the use of direct quotation might enable the original post to be located using a search engine such as Google. The granularity provided by the inclusion of direct quotes was not considered necessary in order to convey the key themes emerging from the data analysis. Comments are identified only with reference to their respective host video. While acknowledging the inherent limitations of this stance and the problematic nature of interpretation and representation of the intentions of these users in the presentation of the findings, this study was influenced by the work of Markham (2012) which suggests that all forms of research, both on and offline, involve a 'process of abstraction from the reality of lived experience' (p. 344).

Results
The majority of comments focused on the rioters rather than the much maligned police operation. Nonetheless, there was evidence to suggest that the police tactics had polarized opinion amongst commentators, with some calling the police operation heavyhanded and others claiming that it was not robust enough.

Commentators were critical of the police operation
A few comments referred to the police tactics as 'heavy-handed ' and 'brutal' (Tesco Gets Ramraided). The use of horses to disperse the peaceful protest outside Telepathic Heights was condemned as 'dangerous and despicable' by one commentator (Bristol Stokes Croft Riot). One commentator claimed that he had been hit by the riot police despite not being part of the peaceful protest shown in the video (Bristol Stokes Croft Riot). Another commentator referred to the police dogs that had started attacking one another before being used to disperse the crowd of bystanders. These commentators appeared to be familiar with the Stokes Croft area and described the squatters as 'well liked' by the local community.
However, it was the scale and timing of the operation that received the most criticism from commentators. Several referred to the 'army' of 150 riot police officers sent to arrest four people suspected of making petrol bombs in Telepathic Heights, with one commentator suggesting that a 'smaller anti-terrorist unit' might have been more appropriate. The decision to raid the squat at 10pm on a bank holiday was referred to as 'stupid' by several commentators with one commentator mocking the claims made by Avon and Somerset Police that it was necessary in order to 'protect the public' (Stokes Croft Riot). These commentators believed that the heavy police presence had antagonised the crowd, which had gathered for the peaceful protest against the raid on Telepathic Heights. However, not all commentators criticised the police tactics and some praised the Avon and Somerset Constabulary for the restraint they had shown during the riot. One commentator went as far as to suggest that US riot squads could learn a lot from the footage (Bristol Stokes Croft Riot).

Some commentators felt that the police response was not robust enough
The restraint shown by the police during the Stokes Croft riot was heavily criticised by a few commentators. There were two sub-categories within this grouping of comments. Some commentators characterised the police as 'cowardly' and questioned why they had not done more to protect private property in the area. One commentator appeared to mock the police cordon around the vandalised supermarket by claiming that they had arrived just in time to 'do nothing' (Tesco Gets Ramraided). A recurring theme in many of the comments was that the police officers were 'just doing their jobs' and that their personal safety had been jeopardised by these tactics (Stokes Croft Riot). The cordon placed across Cheltenham Road was said to have left the police officers increasingly vulnerable to the missiles that were being thrown in their direction. One commentator claimed to be a member of the South Wales Police that had entered the area to support the Avon and Somerset Constabulary in the raid on the Telepathic Heights squat. He stated that police officers had 'risked their lives to restore order' to Stokes Croft and rejected criticism that the operation was heavy-handed (Bristol Stokes Croft Riot). This view received much support from other commentators, who appeared to feel as much sympathy for the police as they did for local residents caught up in the violence. There were also many complaints that taxpayers would have to bear the cost for the repair of the vandalised police vehicles shown in the footage (Stokes Croft Riot).
A few commentators argued that more aggressive tactics were needed in order to police disturbances such as the Stokes Croft riot. It was frequently suggested that crowd control methods, such as baton rounds and water canon, would have proven more effective than the strategies deployed by the riot police to disperse the crowd that had gathered outside Telepathic Heights. One commentator expressed support for the use of 'real bullets against the idiots' that were throwing bottles at the riot police (Bristol Stokes Croft Riot). Further analysis revealed that many of these commentators believed that the Avon and Somerset Constabulary had been lenient in comparison to police forces outside the United Kingdom. Commentators argued that baton rounds should be considered as an option for riot control situations as they were frequently used for these purposes in countries such the United States and Zimbabwe. While it was difficult to verify the location of these commentators, many appeared to be referring to their own countries of origin. What is clear from the analysis is that there was no consensus amongst the online audience in relation to the allegations that the police operation was heavy-handed. Commentators were critical of the police raid on Telepathic Heights, with the exception of the riot police officer that defended this operation. However, there was tacit support for this officer's interpretation of events amongst commentators who argued in favour of the use of more aggressive crowd control methods. Thus, YouTube appeared to have provided a space in which commentators projected their own views on riot control onto the disturbances seen in Stokes Croft.

The 'No Tesco' campaign polarised online audiences
The media framing of the riot as manifestation of the anti-Tesco protest appeared to have resonated with the majority of commentators. There were two sub-categories that addressed the link between the 'No Tesco' campaign and the violence. The majority believed that the violence was an expression of local opposition to the controversial Tesco store. Some referred to the British Broadcasting Corporation (BBC) news coverage of the events that had suggested it was linked to the anti-Tesco campaign. Most of these commentators did not mention the context in which the riot had occurred, with few referring to the raid on Telepathic Heights that was said to have sparked the violence. Many of the reactions to the sousveillance footage consisted of two words, 'Why Tesco?' (Bristol Stokes Croft Riot). Some commentators expressed doubts about whether the vandalism of the Tesco store would force the supermarket chain to leave the area. One commentator urged local residents to express their hate for Tesco by 'not going there', neglecting to mention the boycott that had already been organised by local campaigners (Tesco Gets Ramraided). This was just one example of the lack of local knowledge demonstrated by some users who commented on this footage. Several commentators claimed that they had never heard of the Tesco supermarket chain and could not understand why local residents had opposed the opening of the store. Therefore, it was perhaps no surprise that these commentators conflated the anti-Tesco protestors with the rioters given that this frame emerged from the news media coverage of the riot.
A small but vocal minority rebuked the media for framing the events depicted in the videos as a 'violent protest' and others emphasised that the police raid on Telepathic Heights had caused the violence. One commentator cited the Channel 4 news report about a petrol bomb being thrown at the Tesco store as evidence of their 'lack of impartiality' (Stokes Croft, Bristol Riots). There was also an effort by some commentators to contextualise the events for those commentators who had queried why there was such vehement opposition against the opening of the supermarket in the area. One commentator referred to the 'No Tesco in Stokes Croft' poll (2010) as evidence that Tesco had 'run roughshod' over the local community (Bristol Stokes Croft Riot). Stokes Croft was characterised as an independent local community that supported 'small businesses run by local people' (Stokes Croft Riot). Commentators voiced their opposition to the proliferation of Tesco stores across Bristol, expressing sympathy for the local residents who had been unable to prevent the opening of the controversial store on Cheltenham Road. However, it was recognised that violence was neither inevitable nor justifiable in this context. Rather, the consensus amongst these commentators was that the actions of a 'violent minority' had greatly damaged the campaign.
Nevertheless, many commentators appeared to take umbrage with the anti-Tesco campaign rather than to criticise the violent mob that clashed with police during the riot, claiming that the 'No Tesco' campaign did not represent the 'genuine will' of the local community and criticising the group for telling residents where they 'should or should not shop' (Stokes Croft Riot). One commentator referred to the campaigners as middle class people who had moved to a run-down area but were 'oddly opposed its gentrification' (Stokes Croft, Bristol Riots). It was also claimed that the 'anarchist squatters' in Telepathic Heights had made the lives of local residents 'a misery', contradicting other posts that suggested that the squat had never caused any trouble for local residents (Stokes Croft Riot). These comments appeared to have been left by Bristolians who lived in the vicinity of the Stokes Croft district, with one commentator expressing their anger at the damage that had been done to their car during the riot (Bristol Stokes Croft Riot). Although it is difficult to verify these claims, the information provided in these posts strongly suggested that the commentators were at least familiar with the area.
What is clear from the analysis is that the 'No Tesco' campaign also polarised opinion amongst the section of the online audience that commented on this footage. There were many forceful comments made about the rioters and the police but little rational debate about either the policing operation or the reasons why the community opposed the Tesco store. It was also evident that only a very small number of users perceived this footage as hierarchical sousveillance. Hence, a handful of comments called for the videos to be removed as they might be used as evidence to prosecute rioters.

Discussion
If the intention of sharing this footage on YouTube was to hold the police to account for their 'heavy-handed' operation, then there was little evidence here to suggest that this strategy had been successful. Some commentators did believe that the police tactics were 'unfathomable', particularly in relation to the timing of the raid on Telepathic Heights. These comments appeared to corroborate the allegations of police brutality made by eyewitnesses on sites such as Bristol Indymedia. However, the majority of the comments did not describe this footage as a form of hierarchical sousveillance and the police were criticised for not adopting more aggressive crowd control methods to disperse the rioters. These commentators expressed little sympathy for the residents who claimed the operation was heavy-handed. The media framing of the disturbances, which suggested that the violence was a manifestation of the anti-Tesco campaign, was both challenged and supported by different sections of the online audience. Local residents argued that the provocative and heavy-handed actions of the police had sparked the riot, while commentators who were unfamiliar with the area tended to believe that the violence was an expression of local anger against the controversial Tesco store. The anti-Tesco campaign polarised opinion, particularly amongst those commentators that claimed to be from Bristol. Some Bristolians attempted to provide some context for those commentators who could not understand why local residents so vehemently opposed the Tesco store. There were other commentators that questioned the legitimacy of the campaign in light of claims that some local residents were actually in favour of the Tesco development. The study found that many of these commentators made no distinction between the anti-Tesco protesters and the rioters. These users often dismissed the broader debate sparked by groups, such as UK Uncut, over the impact of corporations such as Tesco upon local retailers. 10 The anti-social behaviour of the crowd shown in this footage had generated sympathy amongst sections of the online audience for the riot police officers that had been the focus of this footage.
Clearly it was not possible to verify the claims made by some commentators in relation to their proximity to the Stokes Croft area and other factors, such as political views, that might have influenced reactions to this content remain unknown. Thus, this study illustrates the potential limitations of using YouTube comments, consisting of only 500 characters, to analyse audience responses to sousveillance. Further research is needed to analyse the influences of traditional media coverage and peer networks upon those who commented on this footage. The role of Twitter in the information flows that surrounded the Stokes Croft riot should also be explored, as well as its role in directing online traffic towards the content analysed in this study. This would allow for further exploration of the themes identified in this study and provide further evidence of the extent to which YouTube provided a platform for rational debate about the meaning of the riot, hence furthering understanding of the role of sousveillance in promoting debate around social issues in online spaces more generally. While this paper did not set out to compare discourses generated by mainstream media outlets and those generated by citizen journalism supplemented by this sousveillance footage, such research would certainly add to this debate.
The findings from the study do, however, appear to support the notion that 'multiple publics' can be identified through their reactions to video footage of a single event (Antony and Thomas, 2010: 1291). The supporters and opponents of the anti-Tesco campaign traded insults alongside the competing narratives that emerged in the comments section below each video. There was no consensus amongst commentators in relation to the broader issues of how the police should respond to civil disturbances and the legitimacy of local campaigns to protect small businesses from large corporations such as Tesco. Nor did the sousveillance footage prompt many of these commentators to engage with the micro-level factors that contributed to the Stokes Croft riot, ranging from the timing of the police raid on Telepathic Heights to the decision by Bristol City Council to grant planning permission for the Tesco store against the wishes of many local residents. Analysis of the comments arguably provided the reader with a more nuanced understanding of the riot than the media coverage that portrayed the violence as a manifestation of the anti-Tesco campaign. The study suggested that YouTube played an important role in the broadcasting of opinions that were not heard in the mainstream media. The audience reactions to the Stokes Croft footage were similar to those generated by the videos of the Oscar Grant shooting insofar as many often used derogatory language to describe the participants but a vocal minority showed some level of engagement with the reasons why the incident had occurred (Antony and Thomas, 2010). However, no conclusions at all may be drawn about the responses of those who watched this sousveillance footage but chose not to leave a comment.
This study suggests that eyewitnesses may not be able to predict how commentators respond to footage presumably shared for the purposes of hierarchical sousveillance. The use of social media to share first-hand perspectives on the policing of civil disturbances may raise more questions about the behaviour of members of the public rather than that of authority figures. It may also lead to angry back-and-forth exchanges between commentators that involve the use of racist and offensive language and show little or no engagement with the events captured on camera. YouTube may provide a public space in which alternative perspectives may be both seen and heard, but these findings suggest that there is little rational debate about the meaning of events, with the views of many commentators still strongly influenced by the news media.

Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.