Urgent Invite - AFTER WOOLWICH: SOCIAL REACTIONS ON SOCIAL MEDIA

Lead Research Organisation: Cardiff University
Department Name: Sch of Social Sciences

Abstract

The research will analyse social reactions to the murder of Drummer Lee Rigby in Woolwich on 22/05/13 using social media (SM) data. Such data uniquely enable the tracking of the evolution of public perceptions and sentiments in real-time as key events occur. The work will produce new insights into the social dynamics of collective responses to high profile violent crimes, alongside methodological innovations developing text-mining methods for rigorous social scientific analyses of SM. Using a case study design applying qualitative, quantitative and geo-spatial data analysis techniques, the project will illuminate the signal event, conflict escalation and de-escalation, influence and resilience dynamics that arise in the aftermath of a major crime.

Specifically, the research will:

1. Conduct exploratory research into the dynamics of how publics interpret and react to high profile crime events, and how such views evolve and shift over time. This reflects how SM data afford opportunities to ask new and different questions than were previously possible with more established social science methodologies.

2. Develop new and prototype methods for social scientists to analyse SM in the future. This will involve the refinement of several social science based analytic frameworks and novel software 'apps' to operationalise and visualise these.

3. Reflect upon the learning derived from the case study analysis to inform future strategies and agendas in terms of how social science responds to the challenges of 'big data'.

The empirical data on the Rigby murder were collected as part of an initial test of the 'Sentinel' (Sentiment mining of Narratives using an Extremist Lexicon) tool - a research prototype text-mining application designed by an inter-disciplinary team of social and computer scientists as part of a European Commission funded project. When the Woolwich murder occurred, the team launched the tool to monitor SM sources and have continued to collect such material (over 10 million tweets and nearly 59000 other user-generated comments collected so far). However, analyses of these publicly available data were not part of the initial project's focus or funding. Accordingly, this application to the ESRC Urgent Priorities scheme has been prepared to enhance and make timely use of the data collected. The urgency relates to:

- Time-limited opportunities to improve the data collected by following and unpacking hyperlinks embedded in user-generated comments (e.g. tweets), providing a 'richer', more nuanced picture of public reaction. This data capture needs completing before too many links are broken.

- Funding would enable monitoring of the case through to the conclusion of the criminal justice process. Tracking the evolution of public opinion in detail and on an ongoing basis, from incident to post-trial, is not something accomplished before.

- A significant policy impact need. Rapid research would provide an evidence-base to inform responses to any future incidents and community tensions. ACPO has provisionally agreed to provide a quantitative time stamped dataset of all hate crime incidents in England and Wales by type, for the recent period - enabling comparison of public opinion with known events in a continuous chronological series.

The approach adopted involves some innovative emergent ideas about public reactions to crime, applying these to SM data and operationalising new methodologies. This will include: the Signal Crimes Perspective (Innes, 2004; forthcoming); Collins' (2012) model of the time-dynamics of conflict; and Sluka's (1990) work on the role of 'soft support' community opinion formation influencing the creation of operating 'spaces' for violent extremists. The key point being that timely and open-ended SM data, when harnessed by new informatics technologies and methodologies, enables social science to pose and answer new questions.

Planned Impact

The project from which this proposal originates is already attracting considerable interest from policy and practice users. The Metropolitan Police Service and Police Academy of the Netherlands are project partners in relation to the European Commission funded work.

In preparing this proposal we have received expressions of support and interest from the Metropolitan Police, the College of Policing, South Wales Police, Home Office and the Mayor's Office for Policing and Crime. This reflects the considerable and growing interest in the police about how to engage with social media, and how it might be used in the aftermath of critical incidents in particular. Specifically, we are planning for the following impacts to develop:

- Police involved in the 'consequence management' of terrorist attacks will be interested in garnering improved knowledge about the dynamics of public reactions in the aftermath of high profile attacks. Indeed, social media informatics opens up new possibilities for the conduct of community impact assessments.
- In mainstream policing, interest is focusing upon how social media might afford new opportunities for community engagement work and better understanding community needs and expectations. The work proposed will provide an important contribution to such efforts.
- One output of the research will be to provide an assessment of official communications by the authorities following an attack.
As such, the research work will generate new insights into what effects (if any) formal communications have in these kinds of context.

The impacts of the work will be, in part, facilitated by the involvement of the Metropolitan Police and College of Policing who have agreed to support and host two workshops to enable police practitioners to engage with the emerging findings of the study. These events will be important in shaping understandings within user communities about what social media analytics can and cannot do. In addition, the workshops will provide important user feedback to the team in terms of what forms of analysis would be most valued by potential users.

Publications

10 25 50
 
Description This research analysed social reactions to the murder of Fusilier Lee Rigby in Woolwich in 2013 using social media data. Based upon a dataset of 35 million data points, it has produced new insights into the social dynamics of collective responses to high profile terrorist incidents, focusing in particular upon understanding processes of social reaction and conflict dynamics. By understanding the social media data as providing 'digital traces' of social action, they have been utilised to provide a uniquely 'high resolution' view of what happens in the aftermath of a terrorist atrocity. Cast in this light, they enabled the illumination of aspects of social reaction that cannot be observed and are thus almost imperceptible when using more orthodox data collection and analysis methods.
All objectives for the work specified in the proposal have been achieved. They were:
1. To construct a unique dataset that uses social media data to track patterns of social reaction in the aftermath of a high profile crime. This involved cleaning and assessing data already collected by the research team, and engaging in additional data collection activities to provide a continuous chronological series on public opinions and attitudes towards the case beyond the conclusion of the criminal justice process.
2. To conduct analyses of these data to produce new knowledge about: How public perceptions and attitudes are shaped by high profile 'signal events'; How and why different segments of the public react very differently to the same stimulus; The influence such incidents have upon public opinion; How signal crimes cause a chain of secondary incidents and consequences. Comparative analysis was also undertaken with police data, to examine how public reactions as expressed via social media are associated (or not) with different types of event. The latter data were provided by the Metropolitan Police Service and provide a series of recorded hate crimes in the days and weeks following the murder of Lee Rigby. These have been analysed in conjunction with the social media dataset.
3. To develop opportunities for future research and impact through the design of several data analysis 'apps' and by engaging with policy and practice communities. This included collaborative work with units from the National Counter-Terrorism Functions Command and College of Policing exploring the potential for using the approaches developed by the research to understanding community perceptions of crime and disorder more generally. This has further seeded a broader reflection upon the opportunities presented by social media data for the social sciences.

The key findings from the work completed are thus predicated upon a set of interacting conceptual, empirical and methodological innovations. The analysis of the empirical data focused upon two aspects. The first involved using the fine grained 'high resolution' data that social media affords to develop an in-depth and nuanced understanding of the processes of social reaction in the aftermath of terrorist incidents. This has identified '10 Rs' of social reaction. Conceptually these constitute component parts of the overarching trajectory of reaction, each of which is an online behaviour freighted with 'offline' consequences and impacts. For example, immediately after the murder, people at the scene of the crime started 'reporting' what they had observed via social media. At around the same time, others in the vicinity were 'requesting' information about what was happening, as they had detected 'anomalous signals' in their environment, such as lots of police sirens and the air ambulance. As news of what had happened began to spread, many social media users communicated their emotional, cognitive and behavioural 'responses' to the news. Following on from which, others sought to employ social media to organise 'retaliatory' violence, to try and 'recruit' others to their ideologically motivated cause, and propagate 'rumours'. Later on in the reaction process, acts of 'remembering' became more important in the social media discourse, both in terms of the particular victim, but also allied incidents from the past. Some people used these collective memories to try and 'reheat' extant grievance narratives that connected the most recent crime, with previous atrocity events.
Importantly though, the analysis also identified that as well as being involved in the development of secondary problems and incidents, social media were also engaged in 'resiliencing' behaviours. These are instances where individuals, groups and communities engaged in counter-mobilisation collective actions in order to enact pro-social informal social control in response to anti-social risks and threats. In particular, two case studies of 'resiliencing' were developed through the analysis, showcasing how social media platforms are increasingly integral to this aspect of reacting to terrorism.
Importantly, there are suggestions within the data that there might be a degree of sequencing and patterning to these activities leading to the development of a sensitising concept of a 'reaction order'. Drawing an analogy with Erving Goffman's seminal concept of 'interaction order', this idea suggests that key elements of social reaction to major crimes may be organised around a set of common public sense-making rituals. Given that this was a rapid response mode grant and based upon a case study, further development and empirical testing of this concept is required. However, it is promising in terms of opening up new ways of studying processes of social reaction.
The second aspect to the findings focuses upon the study of conflict dynamics. Again, utilising the high resolution view that such data afford, they have been used to study how and why a number of secondary conflict events occurred following the killing of Fusilier Rigby. In the preliminary analysis of the dataset, it was identified that a number of conflict events involving far-right and left-wing / anti-fascist groups were being reported on. These were separated into discrete events and subject to more intensive study.
Adopting this approach has allowed us to empirically test some contemporary theories and hypotheses about the causes and consequences of violent conflict interactions. In particular, Randall Collins' (2012) work on the time dynamics of conflict has emerged as providing an especially relevant and interesting framework. Based upon limited data, Collins made a number of important claims and predictions about how violence evolves and unfolds at the micro-social level. We found that our data enabled an empirical test of some of these claims. The resulting findings suggest that whilst Collins' framework is broadly correct, there are important points of detail where the real social dynamics departed in important ways from what he hypothesised would happen.
Elements of these findings have been written up in two journal articles which are currently under review and a book chapter that has been accepted for publication. The two journal articles report the key empirical findings, and two further articles are in preparation developing the theoretical implications of these. The book chapter is scheduled for publication with Sage in their the Handbook of Online Research Methods, and as such, helps the project to deliver on its third aim of helping to shape and inform future strategies and agendas in terms of how social science responds to the challenges of 'big data'.
Exploitation Route The methodological and substantive focus of the study has generated considerable interest from policy and practice facing potential users. This is evidenced by the fact that the research team have delivered eight invited presentations to national and international policy and practice conferences, with two further such engagements planned. These presentations have included speeches at a conference organized by the Dept. of Homeland Security in Washington DC.
The level of user interest is further reflected in the recent award from the Police Knowledge Fund run by the College of Policing, Home Office and HEFCE to establish the Open Source Communications Analytics Research (OSCAR) Development Centre. OSCAR involves a strategic partnership between the Cardiff research team, and the National Counter-Terrorism Functions Command, South Wales Police, West Midlands Police, Surrey Police, Sussex Police, Cardiff Council, and the London Borough of Sutton. The funding for the OSCAR Centre and the engagement of the range of partners involved, especially the Counter-Terrorism Command, is directly attributable to the level of interest generated in the approach and findings emanating from the 'After Woolwich' work.
At a policy level, there is ongoing interest in the implications of the findings for current approaches to counter-terrorism. Notably, it has been suggested that the cross-governmental CONTEST strategy, rather neglects the consequence management dimensions of community impacts following high profile violent incidents. Aspects of the analysis conducted provide an evidence base about what the content of a post-event strategy should attend to and how it should be delivered.
Practitioner use of the findings has, in addition to the consequence management aspects outlined above, focused upon understanding the ways in which the kinds of analytic approaches applied during the study could develop more robust and rigorous methodologies for harnessing 'open source' information and intelligence.
The final group who will be able to take aspects of the work forward are the research community. Conceptually and methodologically the study has opened up new possibilities about how researchers can use social media data in new and interesting ways to develop unique insights into processes of social reaction. There has, after all, been a long-standing interest in how individuals and groups react to major crime events, as manifested in moral panic theory for example. What the analysis of social media affords is a far more fine-grained view of how different sections of the public react to such incidents and how their reactions evolve over an extended period of time. Thereby illuminating subtleties and nuances in these processes that were hitherto imperceptible.
Sectors Communities and Social Services/Policy,Digital/Communication/Information Technologies (including Software),Government, Democracy and Justice,Security and Diplomacy

URL http://www.tandfonline.com/doi/full/10.1080/09546553.2016.1180289
 
Description The findings have been taken forward through a collaborative project with five police forces the Open Source Communications Analytics Research (OSCAR) Development Centre (detailed under further funding). The original research funded by ESRC has obtained significant traction amongst practitioner communities and is fundamentally re-shaping how they conduct 'open source' analysis. As part of this influencing work, the findings are also having a baring upon the UK wide development of a national 'Digital Policing' strategy. The 'reach' of the findings has been extended through presentation of the findings to a number of policy and practice audiences, including: Dept. of Homeland Security Conference, Washington DC (August 2015); National Police Chief's Council Open Source Conference (2015); Home Office CONTEST Review Team; representatives from Defence, Science and Technology Laboratories; US Army Research Labs; Royal United Services Institute 'Digital Policing 2016 Conference. They are also being taken forward through membership of the National Police Chief's Council Open Source Working Group. In 2018 I was commissioned to lead an international research consortium on behalf of the governments of US, UK, Canada, Australia and New Zealand to develop policy and practice advice on using social media to manage the impacts of terror attacks. This work was directly framed by the ideas originally developed in the 'After Woolwich' study. The After Woolwich study also fed into work on disinformation that was cited by the Commons Select Committee report on Disinformation and Fake News.
First Year Of Impact 2017
Sector Digital/Communication/Information Technologies (including Software),Government, Democracy and Justice,Security and Diplomacy
Impact Types Societal,Policy & public services

 
Description Cited in DCMS Fake News Report
Geographic Reach Multiple continents/international 
Policy Influence Type Citation in other policy documents
Impact The research identified hostile state information influence operations following the 2017 terror attacks in the UK. It has been utilised by a number of UK and international government departments and cited in the report of a major UK Parliamentary enquiry into 'Disinformation and fake news'.
URL https://www.parliament.uk/business/committees/committees-a-z/commons-select/digital-culture-media-an...
 
Description Commissioned to write report on social media and aftermaths of terrorism by governments of US, UK, Canada, Australia and New Zealand
Geographic Reach Multiple continents/international 
Policy Influence Type Citation in other policy documents
Impact This work was commissioned collaboratively by five governments to aid policy development in relation to the management of terror events. It has been used to inform new policy and practice guidelines.
URL https://www.publicsafety.gc.ca/cnt/ntnl-scrt/cntr-trrrsm/r-nd-flght-182/knshk/ctlg/dtls-en.aspx?i=13...
 
Description Below the Radar Activity
Amount £10,000 (GBP)
Organisation Nesta 
Sector Charity/Non Profit
Country United Kingdom
Start 04/2014 
End 12/2014
 
Description Centre for Research Evidence on Security Threats
Amount £89,000 (GBP)
Organisation Economic and Social Research Council 
Sector Public
Country United Kingdom
Start 04/2017 
End 05/2018
 
Description Police Knowledge Fund
Amount £405,000 (GBP)
Organisation College of Policing 
Sector Private
Country United Kingdom
Start 10/2015 
End 03/2016
 
Description Article in Police Professional Magazine 
Form Of Engagement Activity A magazine, newsletter or online publication
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Professional Practitioners
Results and Impact The article has been influential in the police practitioner community

Has been picked up by policy-makers
Year(s) Of Engagement Activity 2014
 
Description Control Signals Presentation Policing Global Cities Conference Rotterdam 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Policymakers/politicians
Results and Impact The talk examined issues of perception management with an audience of senior police officers and Mayors from major world cities.

Following my talk I have been contacted by policy development teams from several international cities. There have been discussions of me attending events they are organising.
Year(s) Of Engagement Activity 2015
URL http://www.2015policingglobalcities.org/policing-global-cities/828/
 
Description Eight Rs of Social Reaction - Dept Homeland Security Conference (invite only) Washington DC 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Professional Practitioners
Results and Impact I was asked to share research information about UK Counter-Terrorism policy and practice with an international audience of professionals.

The feedback on my presentation was especially positive, I am currently discussing further presentations to UK agencies and departments.
Year(s) Of Engagement Activity 2015
 
Description Fear vs Terror: Signal Crimes, Counter Terrorism and the Charlie Hebdo Killings for OUP Blog 
Form Of Engagement Activity A magazine, newsletter or online publication
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Public/other audiences
Results and Impact The article contributed to public debate about the impact of terrorist attacks.

I was contacted by several additional media outlets to provide commentary at around this time.
Year(s) Of Engagement Activity 2015
URL http://upsi.org.uk/news/2015/1/20/fear-vs-terror-signal-crimes-counter-terrorism-and-the-charlie-heb...
 
Description From Community to Communicative Policing - Home Office Briefing at the RSA 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Policymakers/politicians
Results and Impact An in-depth discussion with Home Office officials about policy development issues relating to social media and other forms of big data.

A partnership event with officials from the Home Office and DCLG is being planned for later this year (2015)
Year(s) Of Engagement Activity 2015
 
Description Invited talk to National Community Tensions Team, London 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Professional Practitioners
Results and Impact The talk was to police officers engaged in monitoring and managing community tensions following terrorist incidents and other major crimes. A detailed discussion on the research findings was held afterwards.

We have had follow up contact with several police forces requesting further information.
Year(s) Of Engagement Activity 2015
 
Description Media article in Guardian on IS 
Form Of Engagement Activity A magazine, newsletter or online publication
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Media (as a channel to the public)
Results and Impact Media article in national newspaper on a topic of significant public interest

This activity resulted in several follow-up enquiries from journalists.
Year(s) Of Engagement Activity 2014
 
Description Police Futures: Social Control in the New Information Environment - Home Office Conference 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Policymakers/politicians
Results and Impact Talk was used as the basis for thinking about future policy development needs in respect of crime and security issues.

This talk has provided a platform for several subsequent engagements with Home Office officials.
Year(s) Of Engagement Activity 2014
 
Description Terrorism as a Teachable Moment - Hay Festival 2018 Talk 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Public/other audiences
Results and Impact Informed public debate around terrorism, attended by around 350 attendees at the Hay Festival
Year(s) Of Engagement Activity 2018