Introduction

During the terrorist attacks in Oslo and at Utøya on 22 July 2011, in which 77 people were killed, the head of communication at the Norwegian Police Security Services (PST) was monitoring Twitter communication related to the attacks. He found much relevant information circulating on the micro-blogging service. In an interview with the head of communications at The Norwegian PST, we were surprised by a statement regarding 22 July 2011. He stated that he followed Twitter and was thinking that he hoped that Buskerud police district was watching this since they had the operative responsibility. As a contrast to this way of engaging with social media content during the terrorist incident, soon after the attack on the government building in the centre of Oslo, the head of the blood bank at Oslo University Hospital used Twitter to urge citizens to give blood to help out in the treatment of casualties. The response was overwhelming, and the blood bank continued to use Twitter throughout the evening to coordinate blood donations.

These two ways of using Twitter during the terrorist attack represent two different social media strategies employed by authorities during the crisis: the silent bystander (PST), and the active participant (the blood bank). In this chapter, we will take a closer look at the Norwegian authorities’ use of Twitter during the attack. Based on an analysis of the presence of authorities and public bodies on the Norwegian Twitter-sphere during the hours of the attack, the chapter seeks answers to the question: How and to what degree did the Norwegian authorities use Twitter to communicate with the public on 22 July 2011?

The chapter starts with a literature review of social media use in crisis situations, before the specifics of the attack are presented. The methodological details on how the Twitter data were obtained and analysed are then presented, before the findings reveal the number of tweets, retweets and mentions of different authorities and their different communicative strategies (or their absence) during the attack. The different communication strategies used by PST and the blood bank are then analysed in detail, before possible implications of our findings are discussed.

Situation Awareness, Social Media and Crisis Communication

Technology in general and social media in particular are increasingly seen as important factors in dealing with crises in a modern society (Bruns & Burgess, 2014; International Federation of Red Cross and Red Crescent Societies, 2013; Kaufmann, 2015). Historically, the Internet is a relatively recent phenomenon in media management of crisis, but today it is unthinkable to develop media strategies to cope with disasters without implementing what is often referred to as Web 2.0 technologies (Kaufmann, 2015). In 2011, the OECD report Future Global Shocks tried to identify new communication technologies that could be useful during disasters and concluded that social media ‘have the potential to prevent communication breakdown through reliance on just one platform and thereby to reinforce the diffusion of warning messages but also present policy makers with new challenges’ (Wendling, Radisch, & Jacobzone, 2013, p. 6).

One obstacle to implementing social media in the media management of crises is that many communicators have reservations about the credibility of social media content (Wright & Hinson, 2009). However, findings from recent research seem to indicate that there is a growing feeling in most organisations that decision makers ‘no longer have a choice about whether to integrate social media into crisis management; the only choice is how to do so’ (Jin, Liu, & Austin, 2014, p. 76).

Social media can play a vital role in the creation of ‘situation awareness’ (SA) during a crisis. Endsley (1995, p. 36) defines situation awareness as ‘the perception of the elements in the environment within a volume of time and space, the comprehension of their meaning, and the projection of their status in the near future’. SA is, in other words, being aware of what is going on at a given time and in a given space, how to understand it, and how to act on that knowledge immediately and properly. Creating SA is vital during an emergency for all parties – the authorities, the police, rescue workers, those affected and the public – to get control of the situation and minimise harm.

In recent years much attention has been given to how SA is not only an individual process, but something which is created in collaboration (Salmon et al., 2008). For the authorities, it is becoming established knowledge that the public can serve as a resource in risk and crisis management and that ‘crisis communication best practices would emphasize a dialogic approach’ (Seeger, 2006, p. 238). Social media are potentially an ‘awareness system’ (Maireder & Ausserhofer, 2014) and tools for making sense of an emergency event (Heverin & Zach, 2012). They thereby represent arenas for collaboratively created SA – arenas authorities can take advantage of in order: (1) to monitor how a crisis situation like a terrorist attack is understood and acted upon; (2) to spread information, correct misinformation and interact with the public in order to secure that a best possible SA is created for the general public and (3) to get information in order to secure a best possible SA for themselves.

However, how such collaborative SA is ‘configured’ via dialogue and distributed through social media during an emergency is not properly addressed by research, according to Perng et al. (2012). Recent research has shown how various authorities have managed to take advantage of social media to configure SA during crisis situations. During the flood in Queensland, Australia in 2011, the Queensland Police played a leading role on Twitter in ‘disseminating timely and relevant information to the public, and in coordinating and guiding the wider discussion’ (Bruns, Burgess, Crawford, & Shaw, 2012, p. 8). The authorities in Kenya used Twitter actively during the Westgate mall terrorist attack in 2013 to communicate with the public and among themselves (Simon, Goldberg, Aharonson-Daniel, Leykin, & Adini, 2014).

There are, however, several challenges for authorities using social media to configure SA during an emergency. One challenge is related to coordination and synchronisation of information. During the Westgate mall attack in Kenya, Simon et al. (2014, p. 9) found that 15 Twitter accounts belonging to various authorities were in use, which caused ‘information overload and synchronization problems’. Another challenge is related to how responding authorities choose to communicate. Simon et al. (2014) found that emergency dispatch operators were more negative in their Twitter activity than managers and responding authorities during the Westgate mall terrorist attack. Seeger (2006, p. 241) argues that the best practice of official crisis communication implies honesty, candour and openness, and that the public responds more positively to spokespersons who ‘demonstrate appropriate levels of compassion, concern, and empathy’.

Background: The Events of 22 July

On 22 July 2011, the lone terrorist Anders Behring Breivik planted a bomb in the ‘Regjeringskvartalet’, the executive governmental quarter in Oslo. The bomb exploded at 15.25, killing eight people. The second attack was a series of shootings by the perpetrator at a youth camp organised by the youth organisation (AUF) of the Norwegian Labour Party (AP) on the island of Utøya in Tyrifjorden, 40 km from Oslo. At approximately 16.55, Behring Breivik arrived at the shore opposite Utøya. From here, he was transported by the ferry MS Thorbjørn to Utøya. He immediately started shooting and killed 69 people, mostly youngsters. More than an hour later, at 18.34, he was apprehended by the police.

Previous research has shown that social media played a vital role during and in the aftermath of the attack. Social media were pivotal in the ‘meet-hate-with-love’ discourse that dominated the Norwegian public sphere in the aftermath of the attack (Kverndokk, 2013), and such media represented arenas for ‘backchannel discourse negotiations’ (Eriksson, 2016, p. 1) especially related to differing understandings of the event. Activity in the Norwegian Twitter-sphere doubled during the crisis situation (see Chapter 1 for a thorough analysis of dominant actors and themes in the Norwegian Twitter-sphere). Kaufmann (2015, p. 14) found that social media during 22 July created new forms of “selfinitiated and mediated resilience governance,” which emerged spontaneously and have since been implemented in programmes for emergency management. Social media also made it possible for the individuals at Utøya to ‘watch out’ for each other and created a sense of sharing and caring (Kaufmann, 2015, pp. 14–15).

The 22 July Commission appointed by the Norwegian government to investigate what exactly happened on 22 July 2011 concluded that the events offer a unique possibility to analyse the effects of digital communication, since a successful operation by the authorities demanded coordination and knowledge-based leaderships (NOU, 2012). The police were heavily criticised in the report, among other things for their lack of knowledge of, and routines in the use of social media. The police were side-tracked and lacked efficient means and routines to deal with the dramatic situation, according to the report.

In contrast to the critique of the police, the commission reports favourably on the way the health authorities in general, and especially Oslo University Hospital, dealt with the challenges, even though the report does not go into specific details on the use of social media by these authorities.

Methodology

To answer the research question, the chapter is based on a triangulation of methods involving: (1) content analysis of the Norwegian Twitter-sphere on 22 July 2011; (2) qualitative text analysis of the Twitter communication involving the Norwegian Security Police and Oslo University Hospital on 22 July 2011 and (3) a semi-structured, qualitative interview with the head of communication at PST.

The tweets in (1) and (2) were drawn from a dataset of Twitter messages acquired from Gnip, Twitter’s enterprise API platform. This dataset consists of 2.2 million tweets posted between 20 July and 28 August 2011. The dataset represents almost the entire Norwegian Twitter-sphere during that period, limited only by the weaknesses of the Twitter API search mechanisms. These weaknesses derive from the limitations of Twitter’s language detection and geotagging capabilities in 2011. The dataset was, therefore, obtained based on a Boolean search involving information derived from the bio and other metadata of Twitter users (see Chapter 1 for further details on the dataset). ¹

We then constructed a database containing all 2.2 million tweets, and for the purpose of this chapter, we extracted all tweets (58,538) posted between 15.25 (the time of the bomb blast) and 21.00 (when the terrorist had been apprehended and the situation was under control) on 22 July from the database. All user-IDs, retweets and mentions in this set of tweets were then manually analysed to identify activity by any kind of authority or public body. The purpose here was not only to find out which authorities were active on Twitter, but also which authorities the public turned to, and how information from the authorities was diffused on Twitter.

Furthermore, all the 58,538 tweets were read manually to get a sense of dominant themes and communicators. Based on this manual reading and the authorities identified, the communication from the Norwegian Security Police (PST) and Oslo University Hospital was selected for further analysis, because of these authorities’ different approaches to Twitter as a communicative tool during the crises. These two case studies extend beyond the critical hours to also include tweets posted past 21.00 on 22 July.

Lastly, we conducted a semi-structured in-depth interview with the head of communication at PST. ² The research group in the RESCUE project has conducted several additional interviews with representatives of other authorities on social media strategies related to emergencies, and the findings from these interviews, which are presented in Hornmoen Backholm, Frey, Ottosen, Reimerth, and Steensen (2018), will also be drawn upon in this chapter.

Findings

In the following, we will present the findings of the empirical investigation. First, we will present the general picture of how Norwegian authorities were represented on Twitter during the critical hours on 22 July 2011. Then we will present the social media strategy of the two cases further analysed; PST and Oslo University Hospital. Finally, we will present the analysis of the activities of these two authorities on Twitter during and in the immediate aftermath of the attacks.

Discussion

The blood bank case serves as an example of what Ling and Yttri (2002) label ‘micro-coordination’ of an emergency response, in which social media helped to obtain situation awareness (SA) related to the potential sub-crisis of acquiring enough blood to save casualties. In line with the findings of Perng et al. (2012), our analysis demonstrates how SA was acquired through collaboration between the hospital and the public. The hospital expressed their need for blood and urged the public to give blood, and the public helped with spreading the information on Twitter and giving instructions and clarifications as the situation changed. Members of the public acted as ‘remote operators’ in awareness configuration (Heath, Svensson, Hindmarsh, Luff, & Lehn, 2002) and through this collaboration, the University Hospital and the public quickly and efficiently managed to establish a common understanding of what the potential sub-crisis of blood shortage was, and how to act upon it. Within a few hours, the potential sub-crisis had therefore been resolved. However, the lack of precision in the first tweet on registered blood donors almost caused the situation to get out of control.

The blood bank case is a rare exception to how the authorities used social media during the 22 July crisis. Neither the police nor other emergency institutions used Twitter to communicate with the public on what was going on, and the PST case shows how such an absence of activity could lead to speculation and misinformation about, for instance, how many attacks were going on and who the perpetrator was. Such speculation, when not corrected or in other ways addressed by the PST, obstructed the creation of a proper SA and probably caused more confusion and fear about the crisis than necessary among the public.

The number of retweets and mentions about the authorities during the crisis found in our material suggests that the public’s need to get in contact with – and get information from – authorities by far exceeded the authorities’ activity on Twitter. However, absence of activity was not the only problem with how the Norwegian authorities used social media during the 22 July crisis. As Table 1 suggests, the authorities that actually did post something during the critical hours are not representative of those one would expect to take the lead in official communication about the event. There was not a single, official institution that took on the role as coordinator of crisis communication in social media during the critical hours, and therefore people were unlikely to know who to turn to. This might be an explanation as to why so many tweeters turned to the personal accounts of the prime minister of Norway (@jensstoltenberg) and the mayor of Oslo (@fabianstang) instead of the official accounts of the authorities, as shown in Table 2. This need to get in contact with individuals instead of institutions echoes Seeger’s (2006) finding that the best practice in official crisis communication implies honesty, candour and openness, and that the public respond more positively to spokespersons who are compassionate, concerned and empathetic, because these are subjective traits associated with individuals, not institutions.

Conclusion

The research question for this chapter was: how and to what degree did the Norwegian authorities use Twitter to communicate with the public on 22 July 2011? Our findings show that the Norwegian authorities were almost completely absent on Twitter during the critical hours of the terrorist attacks, and that there was no coordination and synchronisation of communication from the authorities. Neither the police, other official emergency institutions, nor governmental bodies took on the role of crisis communication coordinator, despite a clearly articulated public need to get information from, and get into contact with the authorities.

This official silence gave space for the diffusion of speculation and misinformation, which were not corrected or addressed, as the analysed PST case shows. This clearly demonstrates the danger highlighted by the OECD about the effects of leaving social media out of crisis communication preparedness (Wendling et al., 2013). It seems safe to assume that the official silence related to social media crisis communication during the critical hours of the attacks obstructed an efficient configuration of situation awareness, both for the public and for the authorities, especially when we consider the degree to which Twitter was used as a channel of communication in Norway during the attack, with almost 60,000 Norwegian tweets posted during the critical hours.

The blood bank case illustrates the potential of social media in a crisis situation like a terrorist attack with a high number of casualties. Oslo University Hospital had a potential sub-crisis of their own related to not having enough blood to help save the injured, and they dealt with this sub-crisis by configuring situation awareness in collaboration with the public on Twitter. Based on this case, the other findings of this study and previous research, we conclude by offering the following advice to authorities on how to use social media to configure situation awareness during an acute and unforeseen crisis:

• Social media are important arenas for configuring situation awareness during a crisis, and authorities need to coordinate and synchronise their crisis communication in social media. However, messages distributed on social media should be well constructed so as to avoid unintended consequences like misunderstandings and information overload.

• Members of the public can become remote operators in configuring situation awareness during a crisis, and authorities need to acknowledge that situation awareness is best and most efficiently configured not as directives but as a result of collaboration with the public.

• Previously established relations are important to define who the public will choose to turn to and follow during a crisis, and relations with individuals are stronger than institutional relations. The authorities, therefore, need to establish good, personal relations with the public based on compassion, empathy and openness in their everyday communication, so that the public knows who to turn to in times of crisis, and know that they will be listened to and met with understanding and honesty.