From grief to grievance: An analysis of the dominant disaster narratives after the 2001 café fire in Volendam and the 2000 fireworks disaster in Enschede

(1)

From grief to grievance

An analysis of the dominant disaster narratives after the 2001 café fire in

Volendam and the 2000 fireworks disaster in Enschede

Master Thesis

Crisis and Security Management

Leiden University – Faculty of Governance and Global Affairs

August 16, 2020

Jelle Sangers

S1689665

Supervisor: Dr. P.G.M. Aarten

Second reader: Dr. A.V. van der Vlies

(2)

2

Abstract

This study, that is part of the capstone project ‘Community resilience after disasters’ aims to identify dominant disaster narratives in media reports during the aftermath of the fireworks disaster in Enschede (May 13, 2000) and the café fire in Volendam (January 1, 2001). Moreover, it tries to explain differences in the appearance of certain disaster narratives between both disasters. First, it will be argued that media framing is relevant for the academic field of crisis and security management because it can affect the resilience of communities after a disaster. Hereafter, several expectations about dominant disaster narratives will be formulated in line with the articles of Albrecht (2017), Hughes and White (2006), Bohensky and Leitch (2014), McKay (1999), Houston et al. (2012) and Vasterman and IJzermans (2002). According to these scholars, dominant disaster narratives were believed to be predominantly focused on short-term consequences, on scapegoating and on worst case scenarios and alarming scenarios. These theories were tested by performing a content analysis supplemented with a critical discourse analysis on a total of 460 Dutch newspaper articles that were published in the first 5 years after the café fire and the fireworks disaster. The academic database of Nexis Uni was used to select relevant articles, in line with previously determined inclusion and exclusion criteria. It turned out that in both disasters, a critical attitude towards the government, an increased focus on short-term consequences, a focus on scapegoating and a focus on alarming scenarios were the most dominant disaster narratives. Moreover, differences between the disaster narratives of both disasters could be explained by insecurity about the cause and the culprit (in Enschede), the distribution of responsibility (in Enschede) and the reputation of the involved municipality (in Volendam). Finally, a major finding of this study is the discovery of a new group of scenarios in media coverage after disasters. These positive scenarios were neglected in academic literature about disaster scenarios in traditional media outlets so far. Therefore, further research is much needed to elaborate the characteristics of these types of scenarios, and to generalize the results of this thesis to a greater extent.

(3)

3

Foreword

Before you lies my thesis, titled ‘From grief to grievance: an analysis of the dominant disaster narratives after the café fire in Volendam and the fireworks disaster in Enschede’. This document has been written in order to finish the Master study in Crisis and Security Management at Leiden University. From January until August, I used all of my academic fascination and inspiration in this project. Combined with a good portion of time and effort, I believe the results will speak to mind and enrich academic knowledge in the field of Crisis and Security Management. I would like to thank my supervisor, Dr. Pauline Aarten for the possibility to engage in this capstone project and for her support and confidence during the thesis trajectory. The feedback moments were very useful and really helped me in finding the best way to conduct this research. I’m also thankful for the support and feedback I got from my fellow students while conducting this thesis. Finally, I would like to thank my parents for their overwhelming care, love and support during my entire academic career, but especially during this thesis trajectory.

Jelle Sangers

(4)

4

List of Tables and Figures

• Table 1: Overview table of articles showing the importance of media p. 13 discourses for crisis and security management

• Table 2: Overview table of expectations about the dominant disaster narratives p. 20 • Table 3.1: Overview table of assigned codes in the overall coding process p. 33 • Table 3.2: Fragment of overview table about expectation 1 p. 38 • Table 3.3: Fragment of overview table about expectation 2 p. 40 • Table 3.4: Fragment of overview table about expectation 3 p. 46 • Table 3.5: Fragment of overview table about expectation 4 p. 49 • Table 3.6: Fragment of overview table about expectation 5 p. 51 • Table 4: Overview table of assigned codes, specified per disaster p. 57

• Figure 1.1: Pie chart of news articles about the fireworks disaster per period of 5 years p. 28 • Figure 1.2: Pie chart of news articles about the café fire per period of 5 years p. 28 • Figure 2: Flow chart showing the selection process of suitable newspaper articles p. 29 • Figure 3: Network view of the relation between codes and quotations within p. 33

the article ‘Gezondheidsonderzoek na vuurwerkramp’

• Figure 4: Timeline showing distribution of newspaper articles about the disasters p. 44 over the years, including the distribution of articles used for analysis

(6)

6

Introduction

Disasters and catastrophes are of all times, but the response of societies and media has changed over time. Throughout history, the majority of natural and technological disasters were often classified as ‘acts of God’, since it was not appropriately understood what caused these kinds of disasters (Park, 2016). But in modern times, this explanation lost its credibility and more focus was given to the human attribution of responsibility during disasters instead (Arceneaux & Stein, 2006). Since citizens are always seeking for an explanation whenever a disaster happens, much of the responsibility is often attributed to the government, that plays an important role in preparing for and responding to (natural) disasters (Arceneaux & Stein, 2006). The most common policy response from governments after disasters is therefore a revision on safety and security measures.

However, attention for the long-term effects of crises on victims and other involved actors is often limited (McFarlane & Williams, 2012; Quast & Feng, 2019). In the direct aftermath of a disaster, society provides care and concern for the victims. Hereafter, most people tend to recover and forget (Weisaeth, 1996). This leads to negative effects for the small proportion of survivors that are in most need of mental health services on the longer term (McFarlane & Williams, 2012). Moreover, the long-term impact of a disaster on health care staff is often neglected or underestimated as well (McFarlane & Williams, 2012). Another example of a long-term effect is given by Quast and Feng (2019), who argue that a reduction of health care utilization takes place on the long-term. More specifically, they found that screenings related to diabetes were less implemented in the 3 years after Hurricane Katrina compared to the 3 years prior to the disaster (Quast and Feng, 2019). This leads to negative consequences such as a lower quality of life and higher health care costs.

Moreover, most attention in the disaster response usually focuses on individuals, whereas the impact on a community as a whole can be severe as well (Lindell & Prater, 2003; Zhang, Lindell & Prater, 2009). For example, the loss of dwellings can have a significant sociodemographic impact on a community. It can take a long time before a stricken community can recover from the loss of their assets (Lindell & Prater, 2003). These problems are especially apparent in lower income households, who are often disproportionally affected by a disaster (Lindell & Prater, 2003). Other impacts that are especially applicable to communities are political impacts. During the recovery period of a disaster, a lot of victim grievances come to light, which in turn can lead to opportunities for political disruption (Lindell & Prater, 2003). An example of a potential problem within a community in the aftermath of a disaster is what the goal of the recovery process should be. A clash between people that want to rebuild the community as it was before versus people that want to renew and improve the community can lead to a cultural conflict (Rubin, 1991; Dash, Peacock & Morrow, 1997).

(7)

7 A stimulating and complicating factor in these attribution and coping processes is the role of the media in the aftermath of disasters. In modern society, the impact of the media has increased dramatically and has influenced the way society thinks about definitions of risks, health perception and one’s own personal well-being (Vasterman, Yzermans & Dirkzwager, 2005). Subsequently, the way how disasters are framed can have far-reaching consequences for the survivors and victims of disasters as well (Cottle, 2014). It’s impossible to provide solid advice to the affected about this media exposure without theory that helps us to understand which mechanisms underlie the media discourse after disasters. A failure in understanding these mechanisms can lead to negative outcomes such as health issues in the aftermath of a disaster (Vasterman et al., 2005). Unfortunately, most research about these media narratives focuses on the role of the media in the aftermath of disasters in big countries such as the USA (Miles & Morse, 2007; Fleetwood, 2006; Crow et al., 2016). A focus on smaller (European) countries such as the Netherlands is often limited. Especially in the case of the Netherlands this limited attention seems undesirable, since the Netherlands is one of the countries that has already had its fair share of disasters in the previous century (Ellemers, 2001). Moreover, in line with research from Vasterman and IJzermans (2002), Dutch media outlets are believed to affect community resilience in the aftermath of disasters. This makes that a further investigation on the effects of the media discourse in the Netherlands during the aftermath of its disasters seems much needed.

Therefore, this thesis focuses on two disasters in the Netherlands that gained massive media attention from traditional media (newspapers, TV and radio). More specifically, it will focus on the fireworks disaster in Enschede on May 13, 2000 and the café fire in Volendam on New Year’s Day, 2001. The reason why two disasters instead of only one disaster will be investigated is because this thesis not only aims to identify dominant media narratives after disasters, but also aims to explain possible differences in the narratives between disasters. Therefore, it was necessary to select Dutch disasters that were most similar in a lot of aspects, but differed in the dominant media narratives during the aftermath of the disasters. The three criteria beneath are believed to be sufficient to guarantee such a fair comparison between media narratives. First of all, I sought for disasters where the long-term effects are still clearly visible (Tubantia, November 19, 2019; NH Nieuws, June 22, 2020). Second, both disasters gained overwhelming media attention from the traditional media (Ellemers, 2001). Third, both disasters happened before the social media era (Lindsay, 2011). This last criterium is important because it enables a clear comparison between the same types of ‘traditional’ communication means (such as newspapers, tv broadcasts and radio reports) in order to identify a complete view of the dominant disaster narratives. In modern day disasters, the media discourse is spread out over both traditional communication means and ‘modern’ communication means such as Twitter, Facebook (groups) and Instagram. This multitude of sources makes it much more complicated to grasp the main disaster narratives after each disaster. Besides this practical argument, traditional media outlets still play an important role in creating and adjusting disaster narratives in this day and age. For example, in crisis situations involved actors tend to

(8)

8 rely on traditional media above social media to frame their messages (Liu & Kim, 2011). Moreover, Liu and Kim (2011) found that organizations are way more inclined to incorporate emotions in crisis responses in traditional media compared to social media messages. Finally, these scholars found that actors are using a larger variety of frames via traditional media compared to social media (Liu & Kim, 2011). In other words, most actors use traditional media for a more comprehensive crisis response. Because of these reasons, traditional media outlets are still very relevant with regard to research about disasters and media narratives. In contrast to social media outlets, traditional media outlets are effectively able to offer more comprehensive crisis response messages from involved actors. Moreover, traditional media seems more suitable to frame certain messages and to incorporate emotions (Liu & Kim, 2011). Although social media functions keep improving and can reach a bigger audience, traditional media still plays an important role in shaping and substantiating the dominant disaster narratives.

In Volendam, a café caught fire because a sparkler hit Christmas decorations on the ceiling (Commissie onderzoek cafébrand nieuwjaarsnacht 2001, 2001). At that moment, this café was packed with approximately three hundred youngsters (Commissie onderzoek cafébrand nieuwjaarsnacht 2001, 2001). Fourteen youngsters deceased as a result of the fire, and a lot of other victims got scars of their burns (de Vries, van Loey & Janssen, 2004). In Enschede, a warehouse with fireworks caught fire and exploded. Twenty-three people died and approximately nine hundred and fifty people got injured (NOS, October 28, 2019). Moreover, two hundred houses in the surrounding area got damaged (NOS, October 28, 2019). Even though it is twenty years later, the communities of Volendam and Enschede are still dealing with the aftermath of the fire. Examples of this are remembrances that are still held by citizens of Volendam (Hart van Nederland, October 6, 2019) or ongoing discussions about the convictions of individuals in Enschede (EenVandaag, June 12, 2019). Several years after the disasters, negative effects such as mental health problems and a lower quality of social life were measured as well (IJzermans, Dirkzwager, Kerssens, Cohen-Bendahan & ten Veen, 2006; Dorn, ten Veen & IJzermans, 2007). Although no further research addressing the long-term effects of these disasters has been conducted ever since, these problems could still be relevant today when speaking the residents about the disasters. This is in line with other longitudinal studies about long-term effects after disasters, that focus on effects such as traumatic bereavements and posttraumatic stress disorders (Arnberg, Eriksson, Hultman & Lundin, 2011; Boe, Holgersen & Holen, 2011).

Gap of knowledge and academic/societal relevance

Multiple scholars agree that the media is important in contributing to community resilience factors such as identifying potential threats, advocating for changes and creating a narrative for the community (Houston, Pfefferbaum & Rosenholtz, 2012; Hughes & White, 2006). However, some scholars argue that community resilience can be undermined by mass media as well (Bohensky & Leitch, 2014). This

(9)

9 contrast of opposing views indicates a gap of knowledge with regard to the effects of media framing on involved parties after a disaster. If more knowledge about dominant disaster narratives and how they affect a community is gathered, the positive and negative effects of the role of the media after disasters will become clearer. Right now, there seems to be no consensus about the positive or negative effects of this framing (Houston, et al., 2012; Vasterman et al., 2005; Albrecht, 2017). Therefore, this contradiction of theories seems very worthwhile to investigate. Moreover, the reason why some media frames or narratives after disasters are more dominant than others is not paid much attention to (Entman, 1993; Vasterman et al., 2005). This thesis wants to make an effort to expand knowledge about the most dominant disaster narratives and wants to explain why in some cases a certain narrative is more often used than in other cases. It will enrich academic theories about media narratives and media framing to a further extent. Because of these motivations, this thesis will be academically relevant for the scientific field of crisis and security management.

In addition, this thesis will bring societal relevance since it can improve the understanding of society in the complexities that come with media attention after big disasters. An important factor in the bereavement of communities is massive media attention after the disasters. For example, a study by Nivel (2005) showed that media exposure after the fireworks disaster in Enschede led to more doctor visits and physical health problems of the victims. When more understanding about the mechanisms of this media exposure is created, healthcare professionals and other professionals involved in the aftermath of disasters can better prepare for the negative effects of this media exposure when a disaster takes place. This can help improve responses to a disaster in a more effective, just and compassionate way (Wald, 2008). Moreover, the way how traditional media outlets portray certain events can have a big impact on the narrative of society (Curran, 2002). When a certain narrative gains credibility, it can promote or mitigate the stigmatization of certain groups, individuals, lifestyles and behaviours (Wald, 2008). In media reports after the café fire in Volendam for example the media named the informal way of ‘gettings things done’ as a cause of the disaster, thereby contributing to negative stigma’s about Volendam (Thissen, 2010). It is therefore important to understand the appeal and persistence of a narrative in the media. When it is better understood how a narrative shapes attitudes and stigmas towards certain groups, the negative effects of these framing processes can be mitigated be making people more aware of these media dynamics. Finally, this thesis contributes to society since it can help in the aftercare of the victims of the Volendam fire and the fireworks disaster in Enschede. After all, more knowledge about disaster narratives and media frames help to put the newspaper reports about the café fire and the fireworks disaster into a broader perspective. This can help victims of future disasters as well.

(10)

10

Research question

The central research question that will be answered in this thesis will be as follows:

What were the dominant disaster narratives after the café fire in Volendam in 2001 and the fireworks disaster in Enschede in 2000 and how can their differences be explained?

This question primarily focuses on the most dominant media narratives in the aftermath of the respective disasters. So this part of the question is purely describing, and can be answered by simply identifying the most dominant narratives in the media coverage. However, an explanatory question will be posed as well, namely how the dominant narratives of the café fire and the fireworks disaster relate to each other. There are probably some differences between the dominant narratives of both disasters, which are interesting to investigate a bit further. So when the dominant disaster narratives of both disasters have been identified (descriptive part of the question), they will be compared with each other and their differences will be explained (explanatory part of the question).

Reading guide

This thesis will continue with the vast body of knowledge about media discourses and framing. It will be explained how these mechanisms are relevant for the study of crisis and security management. Moreover, some widely accepted theories about dominant disaster narratives will be discussed. In line with these theories, a couple of expectations will be formed that will be tested by performing a content analysis supplemented with a critical discourse analysis. The thesis will then be continued with a research design chapter, in which the methodological approach that will be used to answer the research question is explained into detail. This chapter also consists of a paragraph in which the case selection will be justified. Furthermore, the expectations that were formulated in the body of knowledge will be visualized in the form of conceptual models. Important concepts will be operationalized and paragraphs about data collection and data analysis will conclude this methodological chapter. Hereafter, the results of the content and critical discourse analysis will be presented and analysed. This new chapter starts off with some interesting findings that were found during the coding process. It will then continue with discussing and analysing the findings that were relevant for the expectations and the research question. In the conclusion, it will be judged whether the formulated expectations need revision or seem to be accurate for describing the dominant disaster narratives in the case of the café fire and the fireworks disaster. Moreover, the dominant disaster narratives in both disasters will be discussed and differences in the narratives between those disasters will be explained. This will be continued by a discussion paragraph, in which the findings of the thesis will be evaluated on the basis of previous studies. Hereafter, the limitations and strengths of this study will be discussed. After the main implications of this study for the academic field of crisis and security management have been named, the thesis is then concluded by some final remarks.

(11)

11

Body of knowledge

Bohensky and Leitch (2014) claim that media discourses and dominant media frames affect the academic field of crisis and security management, because the media ‘reflects and shapes public opinion by defining and limiting public discourse around key events (…) and contributes to the sense making and informal learning of the community’ (Bohensky & Leitch, 2014, p. 476). Therefore, it is an important factor to take into account in the aftermath of a disaster. Furthermore, the media can influence risk perceptions and plays a role in the coping processes of the actors involved in a crisis and broader society (Bohensky & Leitch, 2014). In his article, Cottle (2014) explains how certain conditions can increase the intensity and expansion of media discourses. One of these conditions is the increased availability of interactive and bottom-up communication means alongside top-down communication via traditional communication structures. These new ways of communicating increase the surveillance capacity of actors such as civil society groups, contemporary media and the incumbents (Cottle, 2014). As a consequence, governments have to put much more effort in attempts to keep disasters out of sight. Hurricane Katrina is an extensively researched case for measuring the impact of the media discourse on dominant disaster narratives. For example, Cottle (2014) argued that media framing was used to construct the hurricane as a so-called ‘focusing event’ that was able to concentrate the political discourse. In addition, actions of the government were often legitimized by the dominant media frame. Therefore, Cottle (2014) concludes that crises such as hurricane Katrina are seen as opportunities to legitimize political authority. Moreover, they can be seen as moments of critical reflexivity that are used by a multitude of political actors in order to mobilize themselves or to gain support for their cause (Cottle, 2014). Other scholars that were intrigued by hurricane Katrina are Martinko, Breaux, Martinez, Summers and Harvey (2008). These scholars were specifically interested in the attribution of responsibility during this disaster. In their study, these scholars argued that observers such as news outlets tend to blame the outcome of a disaster on characteristics of the actors that were involved (Martinko et al., 2008). With this strategy, they indirectly downplay situational factors during the disaster such as the location of the crisis and the available information the actors had (Martinko et al., 2008). As a consequence, observers tend to give the involved actors more credit or blame than they actually deserve. Therefore, it often occurs that public leaders are unfairly blamed when a disaster goes wrong or unfairly praised when a disaster turns out to be manageable (Martinko et al., 2008).

Another interesting study in the academic field of crisis and security management was conducted by Vasterman and IJzermans (2002). In their study, they found that media hypes affect the amount of people that are willing to register themselves for a health inquiry after a disaster (Vasterman & IJzermans, 2002). Therefore, they claim that peaks of media attention can be aligned with new waves of patients who claim to be ill from a disaster. This leads them to the conclusion that an increasement in media attention for a disaster leads to an increasement of patients as well (Vasterman & IJzermans, 2002).

(12)

12 When asking these patients about their motives to contact a doctor, it became clear that these people began to doubt about the impact the disaster had on them because of the media attention. Even people who never thought their health issues had anything to do with a disaster began to doubt (Vasterman & IJzermans, 2002). Because of this, the scholars concluded that these new waves of patients probably became ill because of the media attention that came with the disaster, instead of the actual disaster itself (Vasterman & IJzermans, 2002).

Other scholars, such as Houston, Pfefferbaum and Rosenholz (2012) are more interested in the different roles the media can play during crises. For example, these scholars expected that the media had an important role in contributing to individual and community preparedness, resilience and recovery of crises. In their study, this role turned out to be smaller than expected (Houston et al., 2012). On the other hand, roles such as communicating with the audience and keeping society updated were adequately fulfilled. Moreover, these scholars argued that ‘media coverage can help communities to identify potential threats, advocate for needed changes in the built environment, and inform personal and family disaster readiness’ (Houston et al., 2012, p. 619). In addition, media provides a forum for community planning and information in the recovery process of a disaster (Houston et al., 2012). When analysing resilience functions of the media, the scholars found that the media is able to create a common narrative in a community (Houston et al., 2012). But this narrative can be negatively affected by media attention as well, when the media doesn’t pay enough attention to the cause of the disaster, the rationale behind the disaster responses or long-term effects of the disaster (Houston et al., 2012). Finally, when media coverage doesn’t pay attention to the past and the future of a disaster, this can have a negative effect on a community as well since it will be less able to critically reflect the disaster and to develop a realistic vision for the long-term (Houston et al., 2012).

The studies above can be summarized in the following table:

Subject Author(s) Country Method(s) Main finding(s)

Media analysis of themes of resilience in the 2011 Brisbane flood Bohensky & Leitch (2014)

Australia Systematic newspaper analysis

Media coverage of the floods can both reinforce and weaken community resilience.

Disasters and their interrelationship with media Cottle (2014) No specific country

Literature study • Disasters are more and more dependent on media how they are understood and responded to.

(13)

13 • Disasters give

opportunities for the legitimation of political power and for actors that seek support for their goals. Hurricane Katrina and attributions of responsibility Martinko et al. (2008)

USA Analysing statements of actors and observers

Differences in perceptions of blame and responsibility in the aftermath of Katrina can be explained by the actor-observer bias.

Influence of news about the Bijlmer disaster on disease symptoms of the affected Vasterman & IJzermans (2002)

Netherlands Newspaper analysis and data about disease symptoms

Peaks of media attention align with new waves of people who claim to be ill because of a disaster

Framing and frame changing in coverage of major U.S. natural disasters

Houston et al. (2012)

USA Content analysis of newspapers and tv news coverage

Media coverage tends to focus on:

• The current impact of disasters

• The current state-of-events

Table 1: Overview table of articles showing the importance of media discourses for crisis and security

management

These findings indicate that media discourses and frames can have a huge impact during the aftermath of a disaster. Therefore, it is relevant for the academic field of crisis and security management. When knowledge about the exact mechanisms and processes underlying dominant disaster narratives is expanded, this can help crisis and security management professionals. This is because the knowledge could be used to predict which narratives are most likely to thrive after a future disaster will occur. For example: when the results of this thesis indicate that a big majority of media reports focus on short-term consequences of a disaster, it can be expected that this focus will be the same in ‘new’ disasters. With this knowledge in mind, crisis and security management professionals can account for this in their analysis of a disaster or in their advice towards involved actors. In the remainder of this body of

(14)

14 knowledge, I will first discuss general theories about media discourse and framing. Hereafter, I will specify how these processes affect disaster narratives.

Media discourse and framing

Media discourse is an important concept in this thesis. It consist of multiple dimensions, such as ‘what precisely gets represented in media texts, who gets to do the talking, what knowledge people need to share in order to understand the media and how power relations are reinforced or challenged’ (Matheson, 2005, p.1). Certain labels are often part of a media discourse as well. According to Matheson (2005), labelling an organization, group or person is a way for society to understand and judge actions of this actor. Since labels are superficial and opportunistic, a label can lead to generalization of these actors (Matheson, 2005). Because of this, labels are powerful means to portray actors which in turn can lead to prejudices. Journalists are in possession of a large range of labels in order to construct the media discourse (Pan & Kosicki, 1993). The label that is ultimately used by these journalists can be explained by news routines, organizational processes and professional considerations (Pan & Kosicki, 1993). According to Pan and Kosicki (1993), the ideology of the journalist is a factor as well.

Other definitions of a media discourse exist as well. For example, it can also be defined as ‘a set of interpretive packages that give meaning to an issue’ (Gamson & Modigliani, 1989, p. 3). According to these scholars, a set of packages can reflect and contribute to the public opinion. The public opinion in this regard is seen as a representation of the struggle between various institutions, social groups and ideologies (Gamson & Modigliani, 1989). In their article, these scholars came up with three determinants that were believed to contribute to the appeal of a certain media discourse or package. One of these determinants is cultural resonances (Gamson & Modigliani, 1989). Cultural resonances mean that some media discourses have a perk because their language and ideas refer to broader cultural themes (Gamson & Modigliani, 1989). According to the scholars, this determinant can increase the appeal of a certain discourse, because its ideas are familiar and are believed to be natural. Mankowski and Rappaport (2000) investigated these cultural resonances as well but named them differently, namely dominant cultural narratives. These scholars argue that cultural narratives have the power to affect values, ideologies and beliefs of most people within a culture (Mankowski & Rappaport, 2000). According to them, these dominant cultural narratives can be identified when they relate to certain folk tales, stories and myths that are part of one’s cultural heritage (Mankowski & Rappaport, 2000). The second determinant that presumably increases the appeal of a certain media discourse is sponsor activities. Gamson and Modigliani (1989) describe these sponsors as agents that promote a collective agenda. Moreover, sponsors are able to provide media outlets with symbols or catchphrases that can be used to make a certain media discourse more appealing (Gamson & Modigliani, 1989). Finally, media practices are named as the last determinant that is able to increase the attractiveness of a media package. These media practices can be understood as the unconscious preferences of journalists to value news

(15)

15 packages of official institutions such as governments as more important than news packages of alternative sponsors (Gamson & Modigliani, 1989). In addition, journalists often have routine relationships with certain sponsors, which in turn affects their contribution on the media discourse (Gamson & Modigliani, 1989). To summarize, certain media packages or discourses are able to become dominant by combining cultural resonances, sponsor activities and a successful fit with media norms and practices. The public opinion has an indirect effect of the dominant media discourse by influencing journalists’ beliefs about what their audiences are thinking (Gamson & Modigliani, 1989).

Framing can be defined as selecting certain aspects of a perceived reality and making them more apparent in a text (Entman, 1993). The goal of this process is to promote a certain interpretation, evaluation or problem definition for the matter of attention (Entman, 1993). A frame can have multiple functions, such as tracing causes, making moral judgments, defining problems or suggesting remedies (Entman, 1993). It is possible that some sentences within a text display multiple framing functions, while other sentences display no frames at all. Moreover, a frame does not necessarily have to consist of all of the functions described above (Entman, 1993). According to Entman (1993), frames can be found on multiple locations within the communication process. In its article, he identifies four of these locations: in the text itself, at the communicator, at the receiver and in the overarching culture (Entman, 1993). Communicators can be interpreted as actors such as journalists that have their own considerations about what they are communicating to their audience. In this thought process, they are influenced by frames that foster their beliefs (Entman, 1993). Another location where frames can be identified are within texts. These frames are recognizable when assumptions in the text are backed up by a specific source of knowledge, or when certain keywords and stereotypes are used (Entman, 1993). The third location in this process is the receiver, which is the reader of the text. On his turn, the receiver judges if the framing intention of the communicator makes sense (Entman, 1993). Finally, Entman (1993) defines the culture as a collection of commonly presented frames, which in turn demonstrate the dominant narratives of most people in society.

Other scholars such as Pan and Kosicki (1993) apply a different interpretation of frames. According to them, frames ‘function as both ‘internal structures of the mind’ and ‘devices embedded in political discourse” (Pan & Kosicki, 1993, p. 57). This leads them to the conclusion that one the one hand, frames are used as a strategy to construct the media discourse. On the other hand, these frames are part of the media discourse itself as well (Pan & Kosicki, 1993). In their article, the scholars identify three types of actors that are part of the news discourse, which are sources, audience members and journalists. According to them, these actors all have certain roles that are defined by society, and can be linked to each other by the same news discourse that they construct, design and act on (Pan & Kosicki, 1993). The dominant media discourse then operates in a domain of shared beliefs by these actors. Boundaries of this dominant discourse consist of these same beliefs (Pan & Kosicki, 1993). In addition, Pan and Kosicki (1993) argue that every media story has a certain theme, which is interpreted as an overarching

(16)

16 idea that is able to connect different aspects of a story in a whole story (Pan & Kosicki, 1993). The idea of such a story is to direct attention or restrict certain perspectives for the audience.

To summarize, media discourses consist of multiple dimensions and can be seen as ‘a set of interpretive packages that give meaning to an issue.’ (Gamson & Modigliani, 1989, p. 3). Cultural resonances, media practices and sponsor activities help to make a certain package more attractive. Moreover, media discourses operate in a domain of shared beliefs between sources, journalists and audience members (Pan & Kosicki, 1993). Labels are seen as constructions that allow actors to judge, understand and generalize other actors (Matheson, 2005). Finally, the process of framing is regarded as a way to make some aspects of a perceived reality more apparent than others (Entman, 1993). These frames can be found on different locations in the communication process.

Disaster narratives and expectations

In her article, Albrecht (2017) argues that media reports can explain political attitudes in the aftermath of natural disasters. More specifically, she argues that political claims and media frames are explanations for possible differences in the political consequences for governments after equivalent disasters (Albrecht, 2017). For example, when citizens perceive that the government has failed to manage the disaster adequately, this can lead to negative effects for the political system and the officials that were involved (Albrecht, 2017). On the other hand, disasters could lead to positive consequences for incumbents as well when the overall perception is that the government did great in disaster management (Albrecht, 2017). The tricky part in this regard is the perception of failure and success. According to the scholar, this can be perceived differently depending on how citizens frame and percept the crisis. This perception is in turn influenced by dominant actors in the political system. Moreover, media narratives can have a huge impact on satisfaction with the government and political trust as well (Albrecht, 2017). Therefore, the scholar considers the media discourse as a key feature to explain changes in political attitudes in the aftermath of a disaster. This is because the media can be seen as an arena in which various frames and critics about the government can be contested (Albrecht, 2017). Examples of mechanisms that come into place in media discourses are attempts of responsible actors to prevent blame when disaster management has been framed as a failure (Albrecht, 2017). However, actors can also exploit the media discourse when their crisis management is perceived to be successful. The scholar notes that when a crisis breaks out, several actors try to exploit this crisis for their own benefit (Albrecht, 2017). The perception of citizens can be influenced in this process dependent on the way how involved actors interpret blame and how government responds to it. As a result, actors make a judgment in which blame is accepted, challenged or avoided (Albrecht, 2017).

Another finding of Albrecht (2017) was that certain journalistic styles and strategies such as interpretive journalism and game framing can have an effect on the probability of crisis exploitation in the media. For example, interpretive journalism is used by journalists when their stories are meant to explain or

(17)

17 analyse (the management of) a disaster (Albrecht, 2017). This style is often characterized by the use of public statements and value-laden terms or phrases (Albrecht, 2017). The opposite of interpretive journalism is descriptive journalism. This style is more objective and focuses on neutral language and facts (Albrecht, 2017). Moreover, it can have a huge effect ‘whether journalists focus on the actual events, including the actions that take place in relation to the reported situation (issue framing), or whether they discuss political strategies or power plays that are observed or suspected to underlie the occurring events (game framing)’ (Albrecht, 2017, p. 33). According to Albrecht (2017), publications that are rich of issue framing and descriptive journalism are predominantly positive about the government. Media reports with a high degree of game framing and interpretive journalism on the other hand indicate a critical attitude towards the governments’ crisis management (Albrecht, 2017). These findings indicate that media framing is an important factor for the perception of failure and success in crisis management. Finally, Albrecht (2017) notes that a disaster can be utilized by political opponents to increase criticism towards the government in order to create a political crisis. This is especially effective in situations where the government is already under fire (Albrecht, 2017).

In accordance with these findings, two expectations can be formed. It seems that reports that are characterized by interpretive journalism and game framing are predominantly critical towards the government. Moreover, reports with a high degree of issue framing and descriptive journalism are believed to be predominantly positive towards the government. Therefore, the first two expectation will be as follows:

Expectation 1: Media reports about the Volendam and Enschede disasters with a high degree of interpretive journalism and game framing are predominantly critical towards the government.

Expectation 2: Media reports about the Volendam and Enschede disasters with a high degree of descriptive journalism and issue framing are predominantly positive towards the government.

Hughes and White (2006) discussed how the media can help to develop resilience in response to bushfires. Their study showed that most people rely on the media when they try to understand the specifics of the bushfire (Hughes & White, 2006). Moreover, the media can be used by emergency services as a tool to warn people in the affected area. In addition, the scholars note that the media can serve to inform people about the measures they should take to prevent the risks of bushfires as much as possible (Hughes & White, 2006). The scholars identified that the media was used to give updated information about the status of the bushfires as well. In the aftermath of the fire, the media played a role in discussions about the causes and consequences of the fires, and in discussions which policies should be implemented in order to prevent these bushfires from happening in the future (Hughes & White, 2006). Overall, it seems that the media played an important role in multiple stages of the crisis. Therefore, the scholars believe it is important to focus on some negative aspects of this media attention as well. One of these problematic aspects is the creation of inaccurate stories and myths about the

(18)

18 bushfires (Hughes & White, 2006). Moreover, the scholars point to a lack of a long-term perspective in media reports. This last finding is supported by Bohensky and Leitch (2014) as well. They argue that media sources tend to neglect long-term effects such as learning processes by focusing on short-term constructions of resilience that align with political consequences (Bohensky & Leitch, 2014). They conclude that when the media speaks about learning, it is often in a limited and narrow way.

With relation to this finding, another expectation can be created. According to these scholars, media reports tend to focus on short-term effects of disasters, thereby neglecting possible effects a disaster can have on the long-term. Therefore, the follow expectation is formed:

Expectation 3: Media reports about the Volendam and Enschede disasters will mostly focus on short-term consequences of the disasters, thereby overlooking possible long-short-term effects of these disasters. Other scholars focus on scapegoating in media reports. In his article, McKay (1999) gives an example of scapegoating by newspapers, when he indicates that newspapers tend to undermine information campaigns by being to fatalistic. According to him, this can be seen as a discouragement towards individuals to adopt self-protecting measures against disasters (McKay, 1999). Hughes and White (2006) agree with this finding, stressing that media reports have a negative effect on community resilience because they present the bushfires as uncontrollable for citizens. According to them, this leads to passive attitudes of citizens instead of a feeling of self-reliance (Hughes and White, 2006). The scholars note that this tendency can be identified in other disaster narratives as well. They stress that media is only interested in looking for a scapegoat to blame (for example in the case of a bushfire: arsonists or fire commanders), thereby overlooking deeper underlying causes which allowed the disaster to happen (Hughes and White, 2006). Houston, Pfefferbaum and Rosenholz (2012) agree with this finding. They found that media outlets prefer official information, that news about a disaster doesn’t sustain for a long time, and that media likes to focus on a dramatic description of the disaster instead of possible causes of the disaster (Houston et al., 2012).

According to these findings, media tend to be more interested in blaming a certain actor for a disaster (scapegoating) than finding out what the underlying causes of a disaster really are. In line with these mechanisms, the following expectation can be formed:

Expectation 4: Media reports about the Volendam and Enschede disasters will be more interested in scapegoating, than finding out the underlying causes that enabled the disasters to happen.

Vasterman and IJzermans (2002) were interested in analysing media hypes after the Bijlmer disaster in Amsterdam. They believed the media played a huge role in framing and dramatizing this disaster. In their article, the scholars stated that the impact of the media on a disaster narrative was dependent on factors such as the extent of media coverage, the language that was being used to describe the disaster and the point of view from the journalists (Vasterman & IJzermans, 2002). Moreover, the scholars found

(19)

19 that extensive media reports of a disaster lead to more unrest among society. An example of this mechanism is when people become more alert to signals that confirm their own risk perception (Vasterman & IJzermans, 2002). This mechanism however does not only impact individuals, but media sources as well. This results in media outlets that are focused on news that confirms their point of view. Ultimately, seemingly insignificant facts can turn into big news this way (Vasterman & IJzermans, 2002). In other words, a self-reinforcing process is created that is characterized by media hypes that lead to the reinforcement of the same frame. According to the scholars, a problematic characteristic of such a hype is that filters and critical voices are ignored (Vasterman & IJzermans, 2002). Therefore, only the dominant frame is more and more implemented. When a certain media hype is omnipresent, more and more new facts pop up that confirm the existing situation (Vasterman & IJzermans, 2002). This affirms the so-called Thomas-theorem: ‘if men define situations as real, they are real in their consequences’ (Thomas & Thomas, 1928, p. 572). Another finding of the scholars is that most disaster news was characterized by fear, guilt, outrage and anger. More nuanced news or news about scientific risk assessments barely got attention (Vasterman & IJzermans, 2002). This led them to the conclusion that alarming sources get more attention than sources that want to ease the audience. Moreover, it seemed that media reports tended to give more attention to worst case scenarios compared to more realistic scenarios (Vasterman & IJzermans, 2002). These developments bring a huge pressure on governments to react to disasters under all circumstances, even if the risk is not as high as displayed in the media (Vasterman & IJzermans, 2002). Ultimately, this leads to a dilemma: not acting on behalf of the government will lead to a trust crisis, whereas drastic measures on behalf of the government will strengthen the current media hype even more (Vasterman & IJzermans, 2002).

With relation to this finding, another expectation can be formed. The scholars argue that disaster narratives tend to focus more on worst case or alarming scenarios compared to more realistic scenarios about the disasters. Therefore, the following expectation can be formed:

Expectation 5: Media reports about the Volendam and Enschede disasters will be more interested in worst case or alarming scenarios, compared to more realistic scenarios about the consequences of the disasters.

In conclusion, a couple of triggering observations have been made that stimulate expectations about possible characteristics of dominant disaster narratives. These expectations can be summarized in the table underneath:

(20)

20

Subject Author(s) Derived expectation(s)

The effect of journalistic styles and strategies on the perception of success and failure in crisis management

Albrecht (2017) Expectation 1: Media reports about the Volendam and Enschede disasters with a high degree of interpretive journalism and game framing are predominantly critical towards the government.

Expectation 2: Media reports about the Volendam and Enschede disasters with a high degree of descriptive journalism and issue framing are predominantly positive towards the government.

Effects of media influence during crises on long-term learning processes

Hughes & White (2006), Bohensky & Leitch (2014)

Expectation 3: Media reports about the Volendam and Enschede disasters will mostly focus on short-term consequences of the disasters, thereby overlooking possible long-term effects of these disasters.

Scapegoating versus finding underlying causes of a disaster

McKay (1999), Hughes & White (2006), Houston et al. (2012)

Expectation 4: Media reports about the Volendam and Enschede disasters will be more interested in scapegoating, than finding out the underlying causes that enabled the disasters to happen.

The role of media in dramatizing and framing a disaster

Vasterman & IJzermans (2002)

Expectation 5: Media reports about the Volendam and Enschede disasters will be more interested in worst case or alarming scenarios, than in more realistic scenarios about the consequences of the disasters.

(21)

21

Research design

Method

The research method that has been used was mostly deductive. Such a deductive method entailed that existing theories regarding media frames and disaster narratives were analysed in the cases of the Volendam fire and the fireworks disaster in Enschede. A deductive method was chosen because this method best suited the overarching goal of this thesis: namely to bring more clarity about which disaster narratives were most dominant and how differences in the dominant disaster narratives between the disasters could be explained. However, this research was partly inductive as well, since new knowledge about disaster narratives and the effects of media framing has been gathered. This new knowledge could be found by analysing media reports about the Volendam and Enschede cases in order to identify frames that had not yet been explained by scientific theories. Moreover, new findings also popped up when it became clear that some disaster narratives were more apparent than was expected in line with previous academic theories. If this was the case, it indicated that current theories about dominant disaster narratives were not able to explain the results. Therefore, it was recommended that these theories were revisited. Moreover, this is a qualitative research. This type of research was most suited to identify patterns of dominant disaster narratives and to test the formulated expectations. In order to give a judgment about the feasibility of a formulated expectation, an in-depth understanding of the concepts was needed. This focus on mechanisms and processes underlying the disaster narratives called for a qualitative approach. Moreover, this thesis used a comparative case study design in order to get a valid answer to the research question. This study design was chosen because it was most able to identify under what circumstances newspaper articles used a certain narrative. For example: under what circumstances were newspaper articles using alarming scenarios? What explains for this? When multiple cases are compared (as is the case in this study), a comparative case study design is most easy to identify these dominant narratives and the conditions under which these narratives thrive.

Case selection

The thesis took the form of a between-case comparison. This is because the thesis attempted to explain differences in the dominant disaster narratives between the two cases. So the thesis focused on the characteristics of the disaster narratives in the two cases, and analysed why the narratives were characterized this way. For example: when the disaster narrative of the café fire could be characterized by a negative attitude towards the government, it was examined why this was the case. According to academic expectations, this sceptical tone could be explained by interpretive journalism and game framing (Albrecht, 2017). An investigation of the media discourse after the disasters could be used to deliver observations that supported this expectation.

(22)

22 A couple of criteria have determined what the relevant cases for this study would be. These criteria were selected in line with a most similar system design. In other words: I sought for cases that were as homogeneous or similar as possible. Therefore, disasters had to fulfil a couple of conditions in order to make the context of the disasters as similar as possible. This similarity in context was necessary in order to make a clear comparison between the cases. The reason why a most similar system design was applied is because this thesis focused on the outcome of cases. In this thesis, the outcome could be seen as the dominant disaster narrative in a case. After all, the dominant disaster narrative was the outcome of the continuous framing process by all involved parties. It was namely expected that the dominant disaster narratives in the aftermath of the café fire and the fireworks disaster differed from each other. On first sight, it seemed that the dominant narrative in Volendam mostly focused on alarming scenarios in which the municipality and the culture of Volendam were criticized. In the aftermath of the Enschede disaster on the other hand, the tone of the media articles seemed to be more reluctant and compassionate. Because both disasters met the essential conditions to make the context as similar as possible, this thesis was able to explain variations in the dominant disaster narratives between the disasters.

One of the selection criteria was the long-term impact of the concerning disaster on the community. Both the Enschede fireworks disaster as the Volendam café fire still had long-term impacts on the communities nearly 20 years later. This could be seen in the memorials and commemorations that were still held (Tubantia, November 19, 2019; de Volkskrant, January 1, 2011). This criterium had been created in order to make sure the disasters were still relevant, although they happened 20 years ago. A second criterium was the period in which the disasters happened. The disasters both took place in the year 2000, before social media and mass communication means were fully integrated (Our World in Data, September 18, 2019). This was important because it enabled a comparison between the same types of traditional communication means in order to identify the dominant disaster narratives via the same kind of sources (i.e. newspapers, tv broadcasts, documentaries). The last criterium that determined why the cases of Volendam and Enschede were investigated was the overwhelming media attention these two disasters got (Wijfjes, 2002; Kuttschreuter, Gutteling & de Hond, 2011). This criterium was important because more media attention led to more news reports and more questions about the disaster. As Vasterman and IJzermans (2002) made clear, extensive media coverage leads to a self-reinforcing process in which a certain media frame becomes more and more apparent. In line with this thought, a lot of media attention led to the reinforcement of a certain disaster narrative, making it easier to identify. This was crucial for the goal of this thesis: namely to identify the dominant disaster narratives. Therefore it was important there were a lot of media reports that could help to construct the dominant disaster narrative. When combining these three selection criteria, the Volendam fire and the Enschede fireworks disaster were the most suitable cases for a within-case comparison of the dominant disaster narratives.

(23)

23

Conceptual models

The expectations that were listed before could be translated into conceptual models. These can be shown as follows:

Expectation 1:

Expectation 2:

Expectation 3:

Media reports with a high degree of interpretive journalism and game framing

Negative attitude towards the government

Media reports with a high degree of descriptive

journalism and issue framing

Positive attitude towards the government

Media reports

More focus on the short-term consequences

Less focus on long-term effects

(24)

24

Expectation 4:

Expectation 5:

The variables on the left were the independent variables. In this thesis, media reports about the disasters were the independent variables. They were believed to contain some characteristics. On the right, the dependent variables are shown. For example, the first conceptual model shows that a high degree of interpretive journalism and game framing in media reports should lead to a negative attitude towards the government. The last conceptual model shows that it is believed it is more likely that media reports focus on worst case or alarming scenarios compared to more realistic scenarios.

Operationalization

One of the main concepts of the thesis is the media discourse. This abstract concept consists of multiple dimensions, such as ‘what precisely gets represented in media texts, who gets to do the talking, what knowledge people need to share in order to understand the media and how power relations are reinforced

Media reports

More focus on scapegoating

Less focus on finding underlying causes

Media reports

More focus on worst case/alarming scenarios

Less focus on realistic scenarios

(25)

25 or challenged’ (Matheson, 2005, p.1). In this research, the most interesting dimension of media discourse is about media framing, since this should have had an effect on the disaster narrative. The act of media framing could then be defined as ‘to select some aspects of a perceived reality and make them more salient in a communicating text, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation, and/or treatment recommendation for the item described.’ (Entman, 1993, p. 52). This dimension of media framing can be turned into concrete indicators. One of these indicators is when a media source gives a value judgement about the disaster. This could show the sentiment around the disaster and it could display possible stigmas revolving around the communities. Another indicator for identifying media framing was when a media source emphasized certain aspects of a disaster. This could shape the public debate in a different direction.

Other concepts that needed further explanation were ‘interpretive journalism’, ‘descriptive journalism’, ‘game framing’, ‘issue framing’, ‘short-term consequences’, ‘long-term effects’, ‘scapegoating’, ‘worst case scenarios’, ‘alarming scenarios’ and ‘realistic scenarios’. These concepts will now be discussed one after one.

Interpretive journalism

In line with the article of Albrecht (2017), interpretive journalism was defined as ‘a journalistic style that is about analysing, evaluating or explaining a situation’. An indicator to identify this journalistic style was when a journalist applied ‘value laden terms and overt commentary’ (Albrecht, 2017, p. 33) in his/her article.

Descriptive journalism

Descriptive journalism could be defined as ‘a journalistic style that is about describing a situation’. An indicator to identify this journalistic style was when a journalist used ‘neutral language and excludes evaluations and analyses of the occurrences’ (Albrecht, 2017, p. 33).

Game framing

The concept of game framing could be defined as follows: ‘a strategy in which journalists discuss political strategies or power plays that are observed or suspected to underlie the occurring events’ (Albrecht, 2017, p. 33). Political strategies in this case could be policies, priorities or certain administration cultures in which public officials consciously make certain choices. With regard to power plays, one could think about hierarchical boundaries or administrative pressures from (public) leaders. These indicators were able to identify processes of game framing.

Issue framing

Issue framing could be defined as ‘a strategy in which journalists discuss actual events, including the actions that take place in relation to the reported situation’ (Albrecht, 2017, p. 33). Therefore, logical

(26)

26 indicators to identify this strategy were when journalists discussed certain actions or events that had something to do with the disaster, thereby excluding thoughts about the motives behind these actions. Short-term consequences

In this thesis, short-term consequences of disasters one could think of were for example political consequences (such as the resignation of a mayor), financial compensations or updates about the amount of disaster victims. For this thesis, short-term consequences were coded as such when media reports focused on consequences which either (1) happened in the first 2 months after the disaster or (2) would happen within 2 months after the publication date of an article.

Long-term effects

In this thesis, long-term effects of disasters one could think of are for example learning processes, career prospects or mental health problems. For the coding of the articles, long-term effects were coded as such when media reports focused on effects which either (1) were visible at least 1 year after the disaster or (2) would happen at least 1 year after the publication date of a given article. The timeframe of 1 year was chosen in line with the article of Bohensky and Leitch (2014), who identified that few newspaper articles about the 2011 Brisbane flood focused on long-term effects that were visible after at least 1 year after the flood.

Scapegoating

In line with the article of Hughes and White (2006), scapegoating could be defined as ‘the blaming of actors that allegedly caused or were involved in a certain disaster, which precludes a deeper analysis of the underlying causes that enabled the disaster’. Furthermore, in this thesis the act of criticizing a certain actor/document was regarded as scapegoating as well, since this distracts the attention from finding the primary cause of the disaster. According to the scholars, an indicator to identify the process of scapegoating was when newspapers presented the causes of the disaster as ‘out of control of residents, thereby encouraging an attitude of fatalistic passivity’ (Hughes & White, 2006, p. 215).

Worst case scenarios

In line with Sunstein (2007), worst case scenarios were defined as follows: ‘scenarios with a low probability and/or a high impact’ (Sunstein, 2007, p.1). According to this scholar, people are susceptible to worst-case scenarios when they are either excessively overreacting the impact of a disaster, or utterly neglecting the small probability of the scenario to happen (Sunstein, 2007). Therefore, these two indicators (excessive exaggeration and utter neglection) could help to identify worst case scenarios. Alarming scenarios

In scientific articles there seemed to be no suitable definition for alarming scenarios in the context of this thesis. Therefore, a definition had to be composed during the coding process. While reading the

(27)

27 newspaper articles, I found some characteristics that were typical for alarming scenarios. In line with these characteristics, I decided to code sentences as ‘alarming scenarios’ when they either (1) described worrying developments/information, (2) focused on negative consequences of the disaster for a certain actor, or (3) contained speculations focused on danger. An indicator that was able to identify alarming scenarios, is when journalists had a fatalistic or desperate tone-of-voice in their articles. Another way to signal alarming scenarios was when journalists applied generalization to enlarge the problem of a single actor to multiple actors. This way, a negative development was made bigger than it actually was. Realistic scenarios

In his book, Sunstein (2007) argued that realistic scenarios needed to be likely and with a low degree of risk. The last part of this sentence (‘with a low degree of risk’) was removed from the definition, since it proved to be irrelevant during the coding process. Therefore, realistic scenarios were defined as follows: ‘scenarios that are likely to happen.’ Indicators to identify such scenarios were when journalist rationally argued for the probability for a certain scenario, or when they put things into perspective.

Data collection

In this thesis, data was gathered from newspaper articles about the disasters. These articles were gathered from Nexis Uni, which is an online database for national and international newspapers (LexisNexis, 2020). The reason why data from newspapers was collected instead of data from other traditional media sources is because newspapers are often regarded as more trustworthy than radio commentary for example (Steelman, McCaffrey, Velez & Briefel, 2015; Albrecht, 2017). Moreover, most newspapers have the possibility to analyse certain disasters in-depth, thereby increasing their impact on public opinion (Steelman et al., 2015). Therefore, newspapers were suitable materials for identifying media narratives after the disasters. Suitable data for the newspaper collection was found by using the search terms ‘Cafébrand’ and ‘Volendam’ for the café fire and ‘Vuurwerkramp’ and ‘Enschede’ with regard to the fireworks disaster on Nexis Uni. For this thesis, newspapers were collected from 2000 up until 2005. One reason why only newspaper articles from the first 5 years after the disasters were collected, was because most articles about the disasters were written in these years (see figure 1 and 2). The numbers in these figures underneath were determined by the amount of results that popped up after searching for ‘Cafébrand Volendam’ and ‘Vuurwerkramp Enschede’ in the Nexis Uni database (LexisNexis, 2020). After searching for these terms, a filter under the heading ‘timeline’ could be applied in order to show how many articles about the disaster were published each year or during a given time frame. Finally, double articles were removed by using the ‘group duplicates on’ option in the top right corner of the screen. Another reason why the timeframe from 2000-2005 was chosen is because it was expected that questions of blame, guilt and the causes of the disasters were more most posed in the first few years after the disasters. This assumption was based on an article of Kuttschreuter, Gutteling and de Hond (2011), who described how the framing and tone-of-voice in newspaper articles after the fireworks

(28)

28 disaster changed over the years. It was expected that the frames and tone-of-voice in articles about the Volendam disaster would develop in the same way.

Figure 1.1: Pie chart of news articles about the fireworks disaster per period of 5 years

Figure 1.2: Pie chart of news articles about the café fire per period of 5 years

The timeframe from 2000-2005 gave a combination of approximately 14.000 newspaper articles about the disasters. However, these media reports did not necessarily all focus on the disasters. Some of them discussed other news, in which the disasters were casually named once or a couple of times as an

10121

3891 3207

1492

Newspaper articles fireworks disaster

2000-2004 2005-2009 2010-2014 2015-2019

2413 754

336120

From grief to grievance: An analysis of the dominant disaster narratives after the 2001 café fire in Volendam and the 2000 fireworks disaster in Enschede