To what extent does politicization drive agenda-setting? Insights from the Dutch Safety Board’s response to four crises

(1)

Leiden University - Facultyof Governance and Global Affairs

Master Thesis Crisis and Security Management

To what extent does politicization drive agenda-setting?

Insights from the Dutch Safety Board’s response to four crises.

Student Anne Kodde Student number s1627392

Date of admission January 9th, 2018

Word count 21,185 words (excluding references and annexes)

Thesis Supervisor Drs. Wout Broekema, Assistant Professor at Leiden University Second Reader Dr. Ellen Verolme, Dutch Safety Board

(2)

Abstract

The present study explores the relation between politicization and agenda-setting within crisis investigation organizations, specifically the Dutch Safety Board. A comparative, in-depth multiple case study was performed, researching the context and events of four Dutch types of crisis cases. Cases were selected on the dependent variable, agenda-setting, which was considered present when a crisis was investigated by the Dutch Safety Board. Politicization, the independent variable, was validated when a crisis had become subject to increased political conflicting attention and was measured by analyzing parliamentary documents and newspaper articles. It was expected that politicization would promote issues on the agenda of the Dutch Safety Board. Results were somewhat mixed. The hypothesis was confirmed in two out of four cases. This points to the substantial role of politicization in the process of agenda-setting within crisis investigation organizations. The present study thereby provides a framework to serve as a foundation for further research on the topic.

(3)

Acknowledgements

Foremost, I would like to express my gratitude to my supervisor Drs. Wout Broekema, for his knowledge, time and guidance. You definitely provided me with the tools that I needed to choose the right direction and – hopefully - successfully complete my thesis.

Besides my supervisor, I would like to thank Dr. Sanneke Kuipers for her support, insightful comments and for leading the capstone group project.

My sincere thanks also goes to Dr. Ellen Verolme for making it possible to do this research in cooperation with the Dutch Safety Board and for her encouragement.

Finally, I would like to express by gratitude to my fellow students of the capstone, for supporting each other and discussing our questions, problems and findings.

(4)

List of figures and tables

Figures

1. Kingdon’s (1984) Multiple Streams Theory……… 13 2. Multiple Streams Theory for agenda-setting within crisis investigation organizations...19

Tables

1. An overview of the concepts, their definitions, indicators and examples……... 27 2. Agenda-setting and politicization (MP attention and media attention) per case... 31

(7)

1. Introduction

On February 23rd

2016, a passenger train at high speed collided with a lifting ramp on a railway crossing in Dalfsen, the Netherlands. The train was derailed and the engine-driver of the train lost his life in the accident (ProRail, 2016a). Less than a year later, on November 18th

2016, a very similar incident occurred in Winsum. A train crashed into a truck on an unprotected railway crossing and 18 people were injured (Prorail, 2016b). The incident in Dalfsen was investigated by the Dutch Safety Board (DSB), a crisis investigation organization in the Netherlands, to draw lessons from the event and to prevent its occurrence in the future (DSB, 2016). In contrast, the accident in Winsum later that year, was not investigated by the DSB. This is remarkable given the similarity of the geographical scope, severity and events of both cases. So, what made the DSB decide to investigate the Dalfsen case, but not the Winsum case? Why do some crises get addressed by crisis investigation organizations, while others do not? And what factors determine this choice? The present study will delve deeper into these questions and attempt to provide answers to them.

1.1 Deciding what crises to investigate

The objective of independent post-crisis investigation organizations is to examine how it was possible for a crisis to occur with the aim of learning lessons for the future and to prevent these incidents from happening again. This is of importance to be able to respond adequately to future crises and to ensure such events will not occur again (Broekema, 2016: 382). Ultimately, the aim of crisis investigation is to improve (inter)national safety. The reason for the investigation of a crisis can be formally driven by an official legal mandate, obliging these investigation organizations to examine certain events. Besides this legal obligation, these inquiry organizations also have the alternative of independently selecting the crises they consider worth an investigation. Given the variation of crises that can threaten the safety of individuals and the purpose of learning for safety improvement in the future, deciding which cases to investigate can be a complicated task. It requires the prioritization of the crises to address, and central to this practice is how investigative organizations allocate attention to these problems (Baumgartner & Jones, 2005: 12, 205). This process can be described as agenda-setting.

There is a large literature on agenda-setting, as it is a key process in decision-making (Kingdon, 1984; Baumgarnter & Jones, 2005; Howlett, Ramesh & Perl, 2009). Only a limited number of these studies, however, have delved in a systematic way into the process of agenda-setting in the context of crisis investigation (some important exceptions are Olson, 2000; Elliott

(8)

& McGuinness, 2002; Ansell, Boin & Kuipers, 2016). So far, the process of agenda-setting within crisis investigation organizations remains not well understood – especially regarding what factors drive the process.

Various scholars mention that political developments often play a large role in determining what problems and solutions are to be considered on the agenda (Cobb & Elder, 1972; Kingdon, 1984; Keeler, 1993; Baumgartner & Jones, 2005; Resodihardjo, 2009). The aim of this explorative study is to gain insight into how politicization – when an issue becomes subject to increased politically conflicting attention - affects the process of agenda-setting within crisis investigation organizations and to provide a framework for further research on the subject.

1.2 Academic and practical relevance

The political aspects of crises have been given a considerable amount of attention in literature (Brändström & Kuipers, 2003; Boin, McConnell & ‘t Hart, 2008; Broekema, 2016; Broekema, Van Kleef & Steen, 2017). Whenever a crisis occurs, political systems must respond to the event, making the crisis political in itself (Olson, 2000: 266). A broad range of actors becomes involved in this process who discuss ideas about why the crisis happened and how to address the problem (Birkland, 2006: 23). These actors compete in producing a dominant understanding of the implications of the crisis (Boin et al., 2008: 9).

However, the exact role this politicization plays in the agenda-setting process in the context of post-crisis investigations remains as yet unclear and ambiguous. The present study therefore takes a first step in exploring this relation. Clarification of underlying processes and observational substantiation of theoretical presumptions are still lacking. Empirical findings should help streamline current theory and future research on this process.

Practical insights of this study will contribute to improving the processes and logistics within crisis investigation organizations. The aim of these organizations is to improve safety through the investigation of (a series of) crises by drawing lessons for the future from the results of these evaluations by considering the causes explaining the occurrence of an incident. It is important to learn from crises to prevent or competently react to future ones, as the consequences of these events are often grave (Broekema et al., 2017: 1) Therefore, it is essential to know which crises are attended to and what factors drive this process of agenda-setting. The current research will provide guidelines for crisis investigation organizations to augment the methods through which they address and decide to investigate certain crises.

(9)

1.3 Research question

The following research question is posed: To what extent does politicization drive the decision

agenda of the Dutch Safety Board? Using a comparative case study design, the effect of

politicization on the agenda-setting process in crisis investigation organizations is explored, specifically within the DSB. Four Dutch cases are selected and analyzed. The DSB is a relevant object of study because it is a strongly independent organization which holds great responsibilities in the Netherlands regarding crisis investigation. The DSB possesses a strong mandate. On the one hand, this legal mandate obliges the organization to investigate certain (specified) disasters. On the other hand, it gives the Board the freedom to take up any (unspecified) investigation topic.

The present study will analyze four similar series of crises and explore the relationship between the politicization of these crises and agenda-setting within the DSB. It will do so by analyzing the degree of politicization of two cases in which agenda-setting was present - which the Board itself decided to investigate - and by analyzing the degree of politicization of two cases in which agenda-setting was absent – crises that were not investigated. These four cases will subsequently be compared to answer the research question. From the cases, an attempt will be made to induce a generic pattern of politicization of crises in relation to agenda-setting within crisis investigation organizations. Data are derived from internal and external documents, including investigation reports, protocols, legislature, media articles and parliamentary records.

1.4 Thesis outline

This explorative research will focus on politicization as a determinant of the decision agenda, i.e. the choice of investigation topics, of crisis investigation organizations. It will do so by building on theories from public administration, particularly in the stream of crisis management. This will provide useful insights into the process of agenda-setting and the role of politicization. These theories will subsequently be linked and applied to crisis investigation organizations. Thereafter, a hypothesis will be formulated regarding the expectations of the present study. Following this theoretical framework, (background) information on the DSB will be presented. Subsequently, in the methodology section, the research design will be described including the means of case selection and data collection and analysis. Finally, the findings will be presented and the results and their implications discussed. This last chapter will also provide an answer to the research question.

(10)

2. Theoretical framework  

In the following section, the main theoretical concepts and key mechanisms of agenda-setting and politicization are described. Attention will be given to the objectives of crisis investigation to sketch a complete image of the push and pull factors that influence this process. Building on theory, a hypothesis will be formulated concerning the influence of politicization on the agenda of crisis investigation organizations.

2.1 Crisis investigation

Crises are “situations which subject a community of people - such as an organization, state or municipality - to a serious threat to its basic structures or fundamental values and norms, which, under time pressure and uncertainties, necessitates making crucial decisions” (Boin & ‘t Hart, 2006: 42). Crises can be perceived as severe threats to the core values of a social system and change its existing sociopolitical order (Resodihardjo, 2009: 14). They are by definition extraordinary in their kind and/or scope, test the resilience of a society and expose the vulnerabilities of its institutions (Boin et al., 2008: 3). Following a disaster, there are questions on causality that always arise (Parker & Dekker, 2008: 261). What went wrong? How could this have happened? How can we prevent it from happening again? Crisis investigations provide an answer to these questions and give insights into the crisis outcome(s) (Gilpin & Murphy, 2008: 147).

After any crisis, governmental officials and/or organizations are confronted with the need to explain (the occurrence of) the event (Olson, 2000: 266). Organizations or policy sectors in which the crisis took place experience “a relatively strong decline in (followed by unusually low levels of) legitimacy” after the incident (Boin & ‘t Hart, 2000: 13). Post-crisis investigations, also referred to as inquiries or evaluations, are conducted to reveal the underlying causes of a crisis and the faults in an organization. These problems can then be addressed by implementing changes (Parker & Dekker, 2008: 261; Deverell, 2009: 180; Broekema et al., 2017: 5). “When experience systematically alters behavior of knowledge” within an organization, learning takes place (Schwab, 2007: 233 in Deverell, 2009: 180). The purpose is to “effect change is some tangible way” (Birkland, 2006: 9). The aims of investigations are to collect evidence and find facts to prevent or competently respond to future crises. This is of importance because the consequences of crises are grave, making the tolerance for mistakes low (Broekema et al., 2017: 1).

(11)

Besides this operational component of fact finding, there is also a symbolic dimension to crisis investigation, which serves a deeper purpose than merely formulating lessons (Bytzek, 2008: 107; Parker & Dekker, 2008: 271). Evaluations can evolve into interpretative authorities because they contribute to shaping public discourse and debate (Resodihardjo, 2009: 178). They can for instance reassure the public, help restore confidence in organizations and institutions or provide an outflow for the grief of victims (Elliott & McGuinness, 2002: 16; Parker & Dekker, 2008: 271).

The essence of crisis investigations is inquisitorial, rather than accusatorial (Elliott & McGuinness, 2002: 15). The purpose of an investigation is to systematically improve (future) safety through the objective analysis of prior events. Accountability questions in the criminal sphere are not part of the purpose of crisis investigation organizations. This is the responsibility of prosecution services.

Governments often appoint independent investigation organizations to examine disasters by collecting and analyzing information through a fair and comprehensive approach (Elliott & McGuinness, 2002: 16). This underlying idea of objectivity, in which hard data are brought together and examined, is assumed to provide an unbiased view on the crisis (outcome) (Gilpin & Murphy, 2008: 147). Crisis investigation organizations often enjoy strong authoritative structures, a supportive political environment, and well defined and respected goals, values and tasks (Ansell et al., 2016: 418).

Crisis investigations may be started due to the severity or magnitude of a crisis. But other determinants, such as for instance the degree of media coverage, the mobilization of interest groups and the need of public reassurance may also be crucial in determining whether to start an inquiry or not (Elliott & McGuinness, 2002: 14; Boin et al., 2008: 4). The factors underlying this choice, in which investigation organizations choose to review certain crises and not others, can theoretically be linked to the process of agenda-setting.

2.2 Agenda-setting

Why is it that some crises get addressed by crisis investigation organizations while others are ignored? Literature states the instruments and systems by which problems are recognized as requiring attention are complex. The concept of agenda-setting deals with the way issues emerge - or not - as candidate for attention. The ways in which these problems are recognized, if they are recognized at all, determine how and whether they will be addressed (Howlett et al., 2009: 92).

(12)

The word agenda is used in various meanings in daily conversation. The present study will focus on the public administration perspective of the agenda and apply this to crisis investigation organizations. Kingdon (1984) defines this agenda as followed:

The agenda is the list of subjects or problems which governmental officials, or others closely associated with these officials, are paying serious attention to. They attend to some rather than others. So, the agenda-setting process narrows this set of conceivable subjects to the set that actually becomes the focus of attention. (3)

The potential solutions to problems are thus winnowed down, wherein one or several are picked and readied for use. This requires the prioritization of information. Central to this issue prioritization is how decision-makers react to information from multiple sources and subsequently allocate attention to these problems (Baumgartner & Jones, 2005: 12, 205).

Attention is scarce to allocate: think of it as a bottleneck, in which problems at the wide bottom of the bottle must fittingly be poured through the neck. Because attention is limited, decision-makers can only handle a restricted number of issues (Olson, 2000: 274). When there is an abundance of information, decision-makers therefore need to find ways to dig through the large amount of signals, decide what data are relevant and what are not, and estimate the quality of information to ultimately address an issue. This attempt to solve problems based on interpreting information is policy-making (Baumgartner & Jones, 2005: 12).

In their work, Cobb and Elder (1972: 85) define two types of agendas. The systemic agenda consists of “all issues commonly perceived … as meriting public attention … within the legitimate jurisdiction of existing governmental authority”. The institutional agenda is “that set of items up for the active and serious consideration of authoritative decision-makers”. Kingdon (1984: 4) delves deeper into the latter and divides the institutional agenda into two sub forms. The first is the governmental agenda, which is “a list of subjects to which officials are paying some serious attention at any given time”. The second is the decision agenda, consisting of “a list of subjects that is moving into position for an authoritative decision”. The institutional, and specifically the decision agenda, is thus the agenda for action. Decision-makers identify and formulate problems by paying attention to some aspects of their organization while ignoring others features (Gilpin & & Murphy, 2008: 129).

There is difficulty in identifying a single source of factors which drive agenda-setting (Howlett et al., 2009: 99). Several multivariate models have been developed, and the present

(13)

study will combine some of these designs and their identified variables into the theory of agenda-setting by drawing on Kingdon’s (1984) Multiple Streams Theory (MST). This conceptual model will thereafter be applied to agenda-setting within crisis investigation organizations, and thereafter specifically to the DSB.

2.3 The Multiple Streams Theory (MST)

Imagine all problems in a line awaiting their turn on the decision agenda. In some way, these items must be arranged in the line. The opening of a window establishes the priority of the issues in the sequence. This window of opportunity is a metaphor used by Kingdon (1984) to describe a period of time in which a problem appears on the decision agenda. Windows open when a problem is recognized, a solution is developed for this issue in a policy community and a political change makes the time right for policy change (165). As Kingdon (1984) argues in his MST of agenda-setting, there are three sets of variables that interact: the problem, policy and political stream (178). When these streams confluence, a window of opportunity opens and at that point an issue enters the decision agenda (Figure 1).

Figure 1. Kingdon’s (1984) Multiple Streams Theory (MST).

2.3.1 Problem stream

The problem stream is composed of the perceptions of problems as public issues requiring attention. It is acknowledged that problem recognition is a socially constructed process as it involves the creation of a mutually accepted definition of what is normal or desirable and what is not (Howlett et al., 2009: 93). Birkland (2006: 19) mentions that only few issues will gain most consideration. Kingdon (1984: 95) notes that policy problems are not self-evident: “They need a little push to get the attention of people in and around government. That push is sometimes provided by a focusing event like a crisis or disaster that comes along to call attention

(14)

to the problem”. Organizations often experience long periods of stability, which are then abruptly interrupted by troubling events, creating opportunities for major policy change (Broekema et al., 2017: 3). In literature, there is consensus concerning the fact that crises will lead to great amounts of attention to the issues revealed by these focusing events (Kingdon, 1984; Stern, 1997; Birkland, 2006; Howlett et al., 2009; Ansell et al., 2016). Birkland (2006), in his book, applied the idea of focusing events to disasters and accidents and found that these events indeed cause a big increase in the attention paid to a policy problem. He defines a potential focusing event as:

An event that is sudden, relatively rare, can be reasonably defined as harmful or revealing the possibility of greater potential future harms, inflicts harms or suggests potential harms that are or could be concentrated on a definable geographical or community of interest, and that is known to policymakers and the public virtually simultaneously (2).

2.3.2 Policy stream

The policy stream consists of the experts and analysts exploring problems and proposing solutions for them (Kingdon, 1984: 104). People need power and authority to participate in agenda-setting and in carrying out decisions (Rainey, 1991: 171). As mentioned above, problems that are part of agenda-setting are constructed in the social realm of public and private ideas (Howlett et al., 2009: 96). The actors who decide to act on a problem are those people who have sufficient knowledge of an issue area. This allows them to participate in the process of developing attainable courses of action to address the problems raised at the agenda-setting stage (Howlett et al., 2009: 12). These groups of analysts are formed by professionals from mixed backgrounds, who produce knowledge on complex and highly specialized issues and who are expected to reduce uncertainty around these problems. These professionals have often been brought together by governmental officials (Dunlop, 2016: 5) and must act within specific sets of laws and regulations governing their behavior (Howelett et al., 2009: 140).

Experts, or policy makers, are under pressure to come up with solutions during or after a crisis (Resodihardjo, 2009: 16). As Birkland (2006: 19) notes, these solutions are likely to vary due to the different interests and motivations of the policy participants. Another point that he makes is that people and the organization in which they make decisions are boundedly rational, meaning that they seek to make rational decisions within the limits of their information

(15)

gathering and analysis capacity. Furthermore, it is also possible for solutions or policy ideas to be floating around, without having a specific problem to hatch on to (Kingdon, 1984: 142).

2.3.3 Political stream

The political stream in agenda-setting refers to just about anything “related to the authoritative allocation of values” (Kingdon, 1984: 145). Apart from the experts and analysts mentioned above, there are also other actors who are involved in setting the decision agenda. These actors can engage in various political activities aimed at encouraging, discouraging and persuading authoritative decision-makers (Woll, 2007 in Howlett et al., 2009: 140). These policy

entrepreneurs are willing to invest their time and resources to push for their favorite (or pet)

solution or views and can be found inside and outside of government (Resodihardjo, 2009: 19). Examples of actors engaged in the political stream are interest groups, the public and the media. All actors judge whether or not a problem needs action (Kingdon, 1984: 145; Keeler, 1993: 440).

2.4 Applying the Multiple Streams Theory to crisis investigation

The stream model as described above can be applied to agenda-setting within crisis investigation organizations. For a crisis to be investigated, it must appear on the agenda – a list of those issues moving into position for an authoritative decision – and, more precisely, it must end up high on the decision agenda – the agenda for action – of a crisis investigation organization (Kingdon, 1984: 3, 4; Howlett et al., 2009: 104). For an issue to rise on the decision agenda, a window of opportunity must open. This window of opportunity opens when the problem, policy and political stream of Kingon’s (1984) MST confluence. These three streams thus influence the decision agenda of crisis investigation organizations.

The present research specifically studies the decision agenda of the DSB. For a crisis to be investigated by the Board, i.e. appear on its decision agenda, the three streams must thus come together. When applying MST to the current study, crises can be identified as part of the problem stream, for they are public issues requiring attention. The policy stream includes (the composition of) the crisis investigation organization – the DSB – as it consists of the experts and researchers exploring crises and proposing solutions for these issues. The motivation of the experts of the DSB is entrenched in the knowledge they produce. However, this does not mean they are apolitical entities (Dunlop, 2016: 10). We must treat these professionals as political actors who take on a variety of roles in the agenda-setting process. Their belief system is not

(16)

static, but created and recreated through interaction with structures, interests and is tied to the wider social culture (Dunlop, 2016: 9). Thereby, the courses of action they develop to address crises are likely to vary due to the different interests, motivations and backgrounds of the individual experts (Birkland, 2006: 19). Also, as mentioned earlier, the members of crisis investigation organizations act within specific sets of regulations governing their behavior (Howlett et al., 2009: 140). Apart from the experts of the DSB, there are also other actors who influence the decision agenda. The political stream consists of those actors inside and outside of government engaged in the agenda-setting process of the Board. These actors press their solutions or views concerning a problem to the fore and include interest groups, the public and the media (Kingdon, 1984: 153). The present study will dive into the influence of this last stream - the political stream - on the decision agenda of the Dutch crisis investigation organization the DSB.

2.5 Politicization of crises

Cobb and Elder (1972: 12) mention that (pre-)political processes often play the largest role in determining what problems and solutions are to be considered on the agenda. Likewise, Baumgartner and Jones (2005: 211) and Kingdon (1984: 145) mention that the primary variable affecting agenda-setting is the leverage of outside influence by powerful individuals and groups who seek to impact policy. Developments in the political stream thus have a powerful effect on agendas. Crises, disrupting a society, invariably increase the number of political demands, as well as the complexity and quantity of these interests (Olson, 2000: 267).

In his work, Broekema (2016: 384) notes that an increase in conflicting political attention can be defined as politicization. This concept can occur inside and outside of formal political institutions. For a problem to become politicized, it must be raised from its status as subject of concern to that of a private or social problem and finally to that of a public issue by certain actors (Howlett et al., 2009: 93). Politicization stems from disagreement on interpretations of a crisis, such as the course of events, the underlying causes and outcomes, and the questions of responsibility and accountability (Broekema, 2016; Olson, 2000). Actors consider the current situation as alarming or worrying and believe measures must be taken or something needs to be done as the present circumstances of the issue cannot continue (Resodihardjo, 2009: 29).

How a subject of concern becomes to be interpreted as a public issue raises questions about how problems are socially constructed. Baumgartner and Jones (1991: 1047 in Howlett et al., 2009: 106) note that the image of a policy issue is significant. When a problem is portrayed as

(17)

a technical problem, experts will dominate the decision agenda. But when a problem is portrayed as social, ethical or political, a much broader range of participants will become involved. For a problem to subsequently become a public issue, participants in the policy arena must assign responsibility or urgency to a crisis so that someone will have to give attention to the problem (Resodihardjo, 2009: 16). Thus, for a crisis to become politicized, public values must be believed to be at stake if the crisis is not attended to by “government decision and resource allocations, or, more rarely, some other form of communal governance” (Buzan, Waever & de Wilde, 1998: 23).

Because the political stakes are high after a crisis, different kinds of stakeholders will try to push through their interests (Broekema et al., 2017: 4). Powerful interest representation might shape the crisis response strategy concerning the investigation of a disaster in favor of those interests (McConnell, 2008: 191). Crisis can thus be presented as “an opportunity to be exploited” (Damgaard, Gerlich & Richardson, 1989: 186). Crises allow proponents of certain policy formulations to slide their ideas forward, usually at the expense of other ideas (Birkland, 2006: 166). Crisis investigations can become venues in which political actors can pursue their own agendas in the post-crisis political process (Parker & Dekker, 2008: 265).

The intensity of politicization after a crisis varies widely, depending on various factors such as policy domain, actor mobilization, timing and the severity and cause of an incident. (Brändström & Kuipers, 2003). Political claims can be made by different political actors in multiple political arenas at different operating levels (Broekema, 2016: 384). Several (f)actors that underlie politicization can thus be distinguished. In the current study these are broadly defined under two categories: attention by Members of Parliament (MP attention) and media

attention.

2.5.1 MP attention

MP attention is put forward in literature as a factor contributing to the politicization of an issue (Kingdon, 1984; Keeler, 1993; Stern; 1997; Brändström and Kuipers, 2003; McConnell, 2008; Broekema, 2016). Birkland (2006: 20) mentions group mobilization and advocacy coalitions as political factors pressing for action after a disaster. Baumgartner and Jones (2005: 282) stress the importance of political movements. These categories can be grouped under MP attention. When a crisis occurs, it is expected that a discussion will arise amongst MPs concerning the issues and concerns it has revealed (Birkland, 2006: 23). This MP attention will be negative in

(18)

tone, because the organization or environment the crisis has occurred in will experience a decline in legitimacy (Resodihardjo, 2009: 22).

2.5.2 Media attention

In literature, there is unanimity that media play a role in politicization (Brändström & Kuipers, 2003; Baumgartner and Jones, 2005; Boin et al., 2008; Resodihardjo, 2009; Broekema, 2016). When a crisis occurs, the event will be scrutinized. Media attention makes a crisis prominent and provides swift feedback to those who seek to construct problems in a particular way (Elliot & McGuinness, 2002: 16; Birkland, 2006: 173; Ansell et al., 2016: 420). Crises which do not get much media attention are often small, have done little damage or require no action of anyone (Birkland, 2006: 22). Mass media are closely tied to the opinion of the public, a factor indirectly also determining whether a crisis becomes politicized or not (Kingdon 1984; Stern, 1997; Baumgartner and Jones, 2005). Mass media can be seen as an indicator of public opinion and vice versa. They portray the national mood or climate. At the same time, the direction media take concerning an issue will influence and determine the opinion of the public (Resodihardjo, 2009: 28). Media can thus be used as a measure of public interest. Media frame selection or direction will affect public debate (Boin et al., 2008: 21). A crisis symbolizes the performance problems of an organization or sector and will therefore gain negative media attention (Ansell et al., 2016: 420). Policy making is not directly influenced by media attention. Mass media exert pressure to ‘do something’ and accompany a crisis, which leads to efforts to adapt existing ideas of the problem (Birkland, 2006: 175). Therefore, mass media attention can be perceived as a factor that underlies politicization.

2.6 The effect of politicization on the crisis investigation decision agenda

As stated earlier, the probability of an issue rising on the decision agenda increases excessively if all three streams – problem, policy and political – of the Multiple Streams Theory are joined (Kingdon, 1984: 178). It has been substantiated that crises are part of the problem stream and that the policy stream includes the experts of the DSB. Politicization of an issue, when Members of Parliament and media attribute attention which is negative in tone to crises, can be identified as part of the political stream, as it related to the authoritative allocation of values by actors engaged in activities aiming to influence decision-makers. The present study explores the effect of this political stream on agenda-setting within crisis investigation organizations, specifically for the DSB, as it is one of the three important drivers of the decision agenda (Figure 2).

(19)

Figure 2. Multiple Streams Theory for agenda-setting within crisis investigation organizations.

When reflecting on the literature, one can conclude that awareness and apprehension increase after a crisis. Crises lead to a great amount of attention to the issues the event reveals. A broad range of actors – inside and outside of government – becomes involved in exchanging opinions, beliefs and theories on the causes and solutions of the crises (Olson, 2000; Birkland, 2006). Politicization stems from disagreement on these interpretations (Olson, 2000; Broekema, 2016). MP attention and media attention are put forward in literature as (f)actors contributing to the politicization of crises (Kingdon, 1984; Keeler, 1993; Stern; 1997; Brändström and Kuipers, 2003; Baumgartner & Jones, 2005; Boin et al., 2008; McConnell, 2008; Resodihardjo, 2009; Broekema, 2016). Following a crisis, these actors will question and scrutinize the functioning of an organization or policy sector in which the event occurred. The attention allocated to the crisis is thus negative in tone. MPs and the media will assign responsibility and/or urgency to the event so that someone will attend to the problem (Buzan et al., 1998; Birkland, 2006; Resodihardjo, 2009; Broekema et al., 2017). This has a powerful effect on the decision agenda of those who can – be expected to – address the crisis (Olson, 2000). In the case of the present study, this is the DSB.

Therefore – the other two stream being equal – it is assumed that the pronounced concern of MPs and the media (i.e. politicization) regarding a crisis will promote issues on the decision agenda of the DSB and lead to an investigation of the problem. Accordingly, the following hypothesis can be formulated:

H1: Politicization promotes issues on the decision agenda of the Dutch Safety Board (DSB).

(20)

3. The Dutch Safety Board as a crisis investigation organization

The next section will provide some information on the DSB, including its history, competences, (legal) responsibilities, composition and working method. After this background sketch, the methodology of the present study will be elaborated upon.

3.1 Foundation and competences

The DSB is a Dutch crisis investigation organization founded in 2005, in which the Safety

Investigation Board Act1

came into force and the Board was officially installed by the Minister

of Interior and Kingdom Relations. The establishment of the Board stems from an increase in calls for independent investigations following crises, in which the entire evaluation would be carried out by one autonomous investigatory organization.

The DSB’s objective is to improve safety in the Netherlands by investigating situations in which civilians are – largely - dependent on government and/or other (private) organizations for their safety. The Board can launch investigations into a broad field of crises and aims to draw lessons for the future from the results of these evaluations by considering the direct and underlying causes for the occurrence of an incident. Often so, structural safety deficiencies can be found in these root causes. Knowledge about underlying explanations may therefore enhance safety in the future. The DSB deals with several crises and investigations every year, and has published about 90 reports (DSB, 2017c).

The DSB is competent to investigate incidents in any conceivable sector: in the various transports sectors - including aviation, rail traffic, road traffic and maritime - and in the fields of defense, industry and trade, construction and services, nature and environment, crisis management and relief and health care (DSB, 2015a). The Safety Board is objective, independent and impartial in its judgement. Thereby, it is transparent in its choices why it conducts an investigation or not and will account for the investigation, its methods and recommendations.

3.2 Investigation topics

The DSB does not launch research into incidents that other organizations, such as inspection services, prosecution services or private organizations, conduct research on and it does not focus on the question of blame or liability. Information obtained by the Board may not be used in legal proceedings. The Safety Board can start an investigation after one or a series of

(21)

incidents. It is often the case that when the Board incorporates multiple incidents into one investigation, these incidents are all a certain type of crises. The limited capacity of the DSB - it employs about 70 people - results in that not all crises that are worth investigating can actually be explored. Therefore, the Board must decide which incidents to investigate and which not. These choices are – initially – based on a substantive assessment framework, incorporating various criteria and elements, to prioritize investigation topics. For an incident to be investigated, at least one of the six following questions must be answered positively concerning the (possible) occurrence of an incident (DSB, 2017c2_):

1. Is there a high degree of (potential) social commotion? 2. Are civilians exposed to a risk caused by others?

3. Are there signs that the system in which the incident occurred (in which civilians should be secured) is failing or knows ambiguities?

4. Can the DSB help to solve or end a complicated problem?

5. Are the specific competences of the DSB of clear added value to an investigation? 6. Are knowledge and expertise available within the DSB to investigate the crisis?

Apart from this free choice of investigation topics, the Board is also obliged to conduct research on certain crises. Topic selection can be formally driven by its official legal mandate (Legislative Act DSB3

, 2004). A great part of the legal framework of the DSB’s mandate stems from international treaties and European jurisdiction.

3.3 Composition and working method

The DSB consists of three permanent board members, including a chairman, which each have profound knowledge of a broad range of security issues. The underlying idea of the small number of board members is that it facilitates and speeds up the process of agenda-setting. Aside the board members, the DSB employs about 70 investigators, who are knowledgeable about (certain) security sectors (DSB, 2015a). The investigators of the Board have specific competences which are legislatively regulated, enabling them to obtain relevant information for the investigations they conduct. They are for instance authorized to enter incident sites and to gather information such as statements by witnesses and tape recordings. The DSB’s mandate obliges every organization to assist its investigators in the exercise of their powers.

2_{Extracted from the latest version of the DSB’s Research Protocol. However, the questions regarding the} assessment framework have remained constant over time.

(22)

Following a crisis, the DSB starts a preliminary research and decides whether there is question of a structural safety failing in need of further investigation. If this is not the case, the investigation stops after this exploratory research. If there is an underlying safety deficiency, the investigators draw a plan of action to continue the investigation. Eventually, the DSB publishes its investigation reports, that include recommendations for the involved organizations to attend to. These organizations are obliged to follow these recommendations. Examples of published reports include the investigation of the 2012 Westerpark train accident in Amsterdam, the MH17 plane crash in 2014 and the Mali mortar accident in 2016 (DSB 2012a; 2015c; 2017b).

(23)

4. Methodology

The present chapter elaborates on the methods used for the current study. The research design, case selection and operationalization of the dependent and independent variable are discussed. Also, the measures for data collection and analysis are explained. Finally, the challenges concerning the research methods are mentioned.

4.1 Research design

The present study is a qualitative research and the central research question is explanatory. A comparative multiple-case study design was used, taking the context and events of various cases into consideration and allowing for detailed and intensive analysis. The relation between agenda-setting and politicization was mainly studied at the issue level, in terms of responses to four thematic cases.

An in-depth case study design was chosen due to the complexity of the process of agenda-setting in relation to crisis and politicization, which is strongly embedded in the specific crisis contexts. A multiple-case study entails the intensive examination of numerous cases and is concerned with the complexity and specific characteristics of these cases (Bryman, 2012: 66). The comparative approach of the current research allows for comparison of four cases using identical methods. It ensures the researcher can understand social phenomena to a greater degree by comparing them in relation to multiple meaningfully contrasting cases and thereby better positioning himself or herself to establish whether a theory will hold or not. A feature of the comparative design is that is allows the different characteristics of the four cases to ‘act as a springboard’ for theoretical reflections about contrasting findings (Bryman, 2012: 72-75).

The DSB as a crisis investigation organization was selected as the object of study based on several criteria. First, the DSB holds important authorities and responsibilities regarding the investigation of crises in the Netherlands. Second, the Board provides an opportunity to analyze agenda-setting in relation to different crises within one organization. Finally, the DSB can be identified as exceptional, since it has the ‘freedom’ to pick up any investigation topic in any conceivable sector.

4.2 Case selection

For the empirical analysis, four cases were selected based on four criteria. These were maintained to minimize the chance of the influence of spurious relationships between politicization and the decision agenda of the DSB (Bryman, 2012: 345). The criteria were

(24)

formulated in accordance to the MST, as described in the theoretical framework. Since the effect of politicization (i.e. the political stream) was researched, the other two streams - the crises (i.e. problem stream) and (composition of) the investigation organization (i.e. policy stream) - were kept equally as far as possible.

The first criterion was that obligation of investigation by the legal mandate had to be absent because these types of crises would not appear on – or influence - the decision agenda of the DSB. Crises that were excluded from the list of cases were aviation accidents, grave railway accidents, very grave shipping accidents and grave toxic substance accidents4_{. (Possible)}

investigation thus had to be crisis driven opposed to mandate driven (Keeler, 1993: 442). The underlying idea of the mandate is that it designates the DSB to make ‘regulations’ that are required to ensure the successful implementation of the principles and aims of the enabling legislation (Howlett et al., 2009: 161). The recognition of crises that do not fall under this legislation are not objectively but socially constructed since it involves the creation of accepted definitions of normalcy and what constitutes an undesirable deviation from that status (McRobbie & Thornton, 1995: 572). This takes part in the agenda-setting process.

A second criterion was that at least one of the six questions of the substantive assessment framework of the DSB (2017c) had be answered positively concerning (the occurrence of) an incident, as it is a condition for the Board to investigate a topic. If all questions were answered negatively, a crisis would not appear on the decision agenda of the Board, and thus not be investigated.

A third criterion was that all crises had to be similar in severity and size, since these factors are indicators of its impact (Brändström & Kuipers, 2003; Baumgartner & Jones, 2005; McConnell, 2008: 189). The bigger the crisis in terms of fatalities, the geographical area covered and the values of damaged goods or property, the more attention it will get (Birkland, 2006: 5). Therefore, this variable had to be kept equal. The similarity of incidents links to constancy of the problem stream discussed earlier. It was decided to select crises which took place in the Netherlands and which consisted of a series of thematically similar incidents, in which numerous people suffered (serious) injuries and/or lost their life.

Fourth, and finally, only crises that occurred after August 2014, the year the latest of the three board members joined the DSB, were regarded (www.onderzoeksraad.nl). This condition was chosen because the board members of the DSB decide on which crises to investigate and

4_{This is in accordance with Articles 3-8 of the DSB’s Legislative Act of 2004, which contains an extended list of} crises (types) the Board is obliged to investigate.

(25)

thereby determine the decision agenda. Composition of this group of experts - part of the policy stream - was therefore constant.

After implementing the four criteria, cases were selected on the differences they exhibited on the dependent variable, namely agenda-setting. This resulted in four cases, of which two were investigated by the DSB and two were not. These were (1) New Year’s Eve (NYE) incidents, (2) drownings, (3) railway crossing incidents and (4) traffic incidents. All four cases comprise a series of similar incidents, or a type of crisis. For every case, obligation of investigation by the DSB by its legal mandate was absent. The cases are comparable in severity, size and the time they took place, as all cases lead to several deaths and injuries in the Netherlands yearly. In 2016, 473 people were hospitalized – of which 1 person eventually died – as a result of NYE incidents; 86 people drowned; 26 railway crossing accidents occurred, of which 4 were lethal; and 629 people died due to a traffic incidents (ProRail, 2016c; CBS, 2016b; 2017; Nijman & Valkenburg, 2017). As only two of the cases have been investigated by the DSB, and two have not, the present study enables comparing the relation between agenda-setting and politicization in terms of responses to these four similar thematic cases. It will assist in establishing whether the theory will hold or not and allow for theoretical reflections about the findings.

4.3 Operationalization of variables

4.3.1 Agenda-setting

The dependent variable for the present research was agenda-setting. The presence or absence of an investigation (report) by the DSB was operationalized as an indicator for agenda-setting. Investigations were chosen as indicator for agenda-setting for two reasons. First, the presence of an investigation (report) demonstrates that a crisis has been addressed on the decision agenda of the Board. Second, investigations hold a certain legitimacy, since they are carried out by the DSB, consisting of an independent team of experts, weighing different interpretations.

4.3.2 Politicization

The independent variable for the present research was politicization. Crises were regarded

politicized when they became subject to conflicting attention, which was negative in tone, in

both parliament and the media. The mere description of crises in parliament and the media is not considered to be negative in tone. Instead, this only occurs when the functioning of an organization or sector is questioned. Performance of an organization should not be in

(26)

accordance with the wishes of its environment. Therefore, only MP attention and media attention was acknowledged when it was negative in tone.

MP attention

To operationalize MP attention, existing Dutch parliamentary records were found online. Documents were selected by searching on three different key words for each crisis in the body of text. Examples include [fireworks], [drowning], [train collision] and [road traffic death]5

. As explained, attention regarding the crises had to be negative in tone. Only documents written from August 2014 – the moment the latest of the three board members joined the DSB – till November 2017 – the present moment of study – were included in the search. MP attention was measured through the analysis of (1) plenary and/or committee meetings of the Tweede

Kamer6

, (2) letters (by MPs) to the chair of the Tweede Kamer, (3) written questions and (4) submitted motions.

Media attention

To measure a - rise in negative tone of - attention in mass media, national newspaper articles from NRC Handelsblad (NRC), De Volkskrant, Algemeen Dagblad (AD) and De Telegraaf were selected. These four daily-newspapers have the highest circulation rate in the Netherlands and portray both the left and right of the political spectrum. Articles were retrieved via LexisNexis by searching on three keywords per case – corresponding to those used in the the parliamentary record search – in the body of the text.

The categories and indicators of the concepts of agenda-setting and politicization can be found in Table 1. Conceptions were formulated in accordance to the theory on agenda-setting and politicization in the theoretical framework. Also, some examples of the operationalization of the concepts are given, derived from the analyzed documents.

5_{Translated from Dutch: ‘vuurwerk’; ‘verdrinking’; ‘treinongeluk’; ‘verkeersdode’.} 6

(27)

4.4 Data collection and analysis

The present study was carried out through document analysis, a systematic method for reviewing or evaluating documents. This approach is suitable for qualitative case studies, since the interpretation of documents assists in eliciting meaning, gaining understanding and developing knowledge concerning the research question (Bowen, 2009: 27). It allows for intensive study of cases and produces rich and detailed description of these events. Documents provide background and context. Also, they offer a broad coverage of events, in which various sources over a long span of time can be consulted. Document analysis combines characteristics

(28)

of content analysis and thematic analysis. Content analysis entails “the process of organizing information into categories related to the central questions of the research” and thematic analysis comprises “a form of pattern recognition within the data, with emerging themes becoming the categories for analysis” (Bowen, 2009: 32).

The current research adopted features of content analysis, by organizing information into predefined categories, and employed characteristics of thematic analysis, as these predetermined categories were broad and more specific themes emerged during the process of analysis. Document analysis for the present study involved skimming, in which data were superficially examined, reading, in which data were thoroughly examined, and interpretation (Bowen, 2009: 32).

Data collection and analysis was structured in four steps. First, online statements and reports of the DSB were searched on their website and Twitter account, to establish whether crises had appeared on the decision agenda or not. Whenever the DSB confirmed an investigation, as indicated in Table 1, agenda-setting was validated.

Second, newspaper articles and parliamentary records were searched for every case. As mentioned above, articles were found by searching for keywords in the body of text. The documents were skimmed and those records that - for some reason - ended up in in the search, but in fact did not cover the crises, were excluded. Additionally, only documents which were

negative in tone were included in the analysis. For a document to be negative in tone, it had to

question (the functioning of) an organization or policy sector. The documents that remained7

were read and searched for statements which were of potential theoretical relevance. The unit of analysis was (parts of) sentences and the significant sentences were assigned to the concept of politicization and its indicators as described in Table 1. Taken together, parliamentary and media data were employed to demonstrate the intensity of politically conflicting interest graphically. In addition to this graphical display of politicization, quotes8

were used to show the tone of the debate.

The third step of the analysis was an exploration of the specific relationship between agenda-setting and politicization for every case individually.

Fourth, and finally, this relationship was compared between the four cases to obtain understanding of the extent politicization possibly drives the decision agenda of the DSB.

7_{A list of all analyzed documents can be found in the Annexes (p. 77).} 8_{These were translated from Dutch.}

(29)

4.5 Challenges

Taking on a comparative multiple-case study design and using document analysis has various challenges and limitations.

A main challenge of the present research regarding measurement validity stemmed from difficulties in measuring the dependent and independent variable. Both agenda-setting and politicization are concepts which are difficult to grasp due to the complexity of the processes. Therefore, they are subject to ontological, methodological and normative issues. A pitfall is that their operationalization does not actually, or indisputably, indicate agenda-setting or politicization. However, some simplification, through the concrete operationalization of the concepts, allowed for a systematic analysis of a quite large amount of data.

Another limitation is that – most likely – other factors besides politicization influence the decision agenda of the DSB, and that these factors are not accounted for in the current research. This is a threat to internal validity and relates to the issue of causality. Exploring the presupposed direct relation between politicization and the agenda-setting process risks oversimplification as it is possible that politicization is interrelated to – the direct or indirect role of – other factors.

External validity is the extent to which the results of the study can be generalized beyond the specific research context (Bryman, 2012: 47). It can represent a problem for qualitative research because it tends to employ case studies and small samples (Bryman, 2012: 390). For the present research, external validity was moderate. As the study compared four cases, the generalizability of the research was higher than with a single case study. However, more than four cases would substantiate theory and findings to a greater extent.

Reliability refers to the “consistency of measures” and whether the results of a research are repeatable (Bryman, 2012: 46, 168). A true replication of a qualitative research is almost impossible to perform, as there are no standard procedures to follow (Bryman, 2012: 405). By developing concrete indicators to measure the concepts of agenda-setting and politicization, the present study attempts to counter this issue. The analysis procedure is set out clearly, so that replications and follow-up studies become more feasible. Nevertheless, linking empirical findings to the indicators involves subjective judgement and document analysis invariably entails interpretation on the part of the researcher. Additionally, the perspectives concerning the measured concepts can differ between researchers.

Distinguishing whether attention given to the crises by MPs and the media is negative or neutral in tone, and should thus be included or not in the analysis, was not invariably

(30)

self-evident. This is a threat to reliability. An example is an article which mentions that the operator and driver of a train, involved in a collision on a railway crossing, are identified as suspects of the incident. It is thereafter stated that this is just a standard formal procedure and one is not to worry, because the individuals are not actually accused of anything (“Ontspoorde Trein van de Rails”, 2016). It could be argued that this article calls the functioning of the prosecution system for railway incidents into question. However, since the article is not explicitly negative in tone, it was eventually excluded from the analysis, along with all documents which implicitly question the function of an organization or sector.

A final limitation of the present study is biased selectivity, if collection of documents is incomplete (Bowen, 2009: 32). Only three keywords were used per crisis case to search the body of text of parliamentary documents and newspaper articles. This possibly excluded documents and articles which in fact covered the crises, but did not come to the fore using these specific keywords. Nevertheless, quite a lot of documents were retrieved, and these can be expected to be a reasonable representation of the attention payed to the crises.

(31)

5. Findings

The current chapter presents the findings for the four cases. Table 2 displays the findings for all cases graphically. For each individual case, agenda-setting, politicization and the relation between these two concepts were explored in-depth. These results, in which quotes were used to exhibit the attention given to every crisis (type), can be found per case below. Finally, the relationship between agenda-setting and politicization was compared between the four cases to gain an understanding of the extent politicization drives the decision agenda of the DSB.

Table 2. Agenda-setting and politicization (MP attention and media attention) per case.

5.1 Case 1: New Year’s Eve incidents

Across the world, New Year’s Eve (NYE) is a festive night where people celebrate the start of a new year. What characterizes this night in the Netherlands specifically, is that citizens can set off fireworks with friends and family on the 31st

of December. Every NYE, numerous people get – severely - injured due to this tradition. To illustrate: during NYE 2016-2017, 3,525 firework related incidents took place and 473 people were treated in hospital due to firework injuries (Nijman & Valkenburg, 2017).

Apart from these firework accidents, several other incidents take place every NYE. During NYE 2016-2017, 10,016 incidents were reported to the police and 340 cases, in which people were arrested for a crime related to the turn of the year, were submitted to the Public Prosecution Service. That same night, 83 incidents were counted concerning violent behavior against policemen and emergency service providers. (DLIO9

, 2017).

The occurrence of these various incidents lead to a spark in public debate every year, in which various actors – inside and outside of government - express their opinions on the topic and come up with potential solutions or measures.

9_{The DLIO (in Dutch: ‘Dienst Landelijke Informatie Organisatie’) is the operational center of the Dutch police} and manages all operations and ensures overview and coherence.

(32)

5.1.1 Agenda-setting

As the DSB states on their website, the Board is currently conducting a research on the problems and incidents caused by fireworks, disruption of public order and violence against policemen and emergency service providers, which take place during NYE (DSB, 2017e). The Board aims to find out what safety risks there are and what measures have already been taken to prevent or limit incidents and problems during this night. The research particularly focuses on the question whether preventative measures can be taken to ensure safer NYEs in the future. Agenda-setting is thus present in this first case.

5.1.2 Politicization MP attention

A substantial amount of attention was given to NYE in parliament: two meetings of the joint committee of the Ministry of Justice and Security (MoJS) and Ministry of Infrastructure and Environment (MoIE), thirteen letters, five questions and six motions were found and analyzed.

In the committee meetings, MPs repeatedly mention that government should take measures to prevent the incidents, continually taking place every NYE, from occurring. The meeting of 2015 focuses on the damage and injuries fireworks cause and the criminal behavior certain citizens display during this night (Committee of MoJS and MoIE, 2015). Urgency to ‘do something’ is assigned to the event, so government will give attention to the problem. This implies politicization (Resodihardjo, 2009: 16). One MP expresses that these incidents “are unacceptable and should be given priority” (Recourt, 2015). Other committee members mention that it is the duty of government to address and find solutions for these incidents. An MP stresses that “mayors, the police, the fire department and citizens are calling on government to take measures” and asks why “government is not prepared to listen to this request?” (Ouwehand, 2015b). Another MP states that “raising awareness for the use of fireworks is a task not only for government, but also for the fireworks industry” (Tellegen, 2015). Different interpretations and ideas concerning accountability and responsibility can thus be found, pointing to politicization (Broekema, 2016).

In the questions submitted to the Minister of Justice and Security, MPs state that “we still must work hard to ensure New Year’s Eve can take place without any noteworthy incidents” (Bontes & Klaveren, 2015) and ask if the Minister agrees that “safety of emergency service providers is at stake when citizens set off and throw (heavy) fireworks?” (Van Dam, 2017). In

(33)

the committee meeting of 2016, an MP mentions the number of incidents related to fireworks (4,000) and firework victims (438) and stresses the only way to prevent these events from happening is by “banning consumer fireworks” (Wassenberg, 2016). A party member submits two motions concerning the banishment of consumer fireworks (Ouwehand 2015a, 2016). The other committee members loudly disagree with a firework ban (Committee of MoJS and MoIE, 2016). There is thus disagreement regarding the course of events and proposed solutions, indicating politicization as mentioned by Broekema (2016) and Olson (2000).

In a letter to the chair of the Tweede Kamer in January 2015, the third-last Minister of Justice and Security sketches a national picture of incidents concerning that previous NYE, wherein he mentions that incidents concerning violence against police and emergency service providers have declined, but that firework related incidents have increased (Opstelten, 2015). In a letter in March that year, the (temporarily acting) second-last Minister of Justice and Security points out the measures that have been taken to decrease firework incidents, such as the introduction of limited times in which citizens can light fireworks, the establishment of a permit system for firework importers and the renewed approach to detect illegal fireworks (Blok, 2015). An MP sends a letter in April 2015 in which she underlies that “it is priority to not only tackle the import of illegal fireworks, but also to reduce the number of incidents resulting from legal fireworks” (Mansveld, 2015b).

In another letter in January 2016, the former Minister announces that the total number of incidents that previous NYE has decreased and in a letter in March 2016 he states that “this is positive, but does not mean that we can be satisfied with the current state of affairs” (Van der Steur, 2016a, 2016b). The events of the night of December 31st

2016, in which a policeman dies because of a serious incident, give an even more negative twist to the tone of debate. The Minister responds that “this incident taints New Year’s Eve”, “these violent incidents are unacceptable” and that “additional measures must be taken to prevent this from happening”, indicating politicization (Resodihardjo, 2009: 29). According to the Minister, solutions to NYE related incidents should be discussed between the Minister of Infrastructure and Environment, mayors, municipalities, prosecution services, the police, ambulance care, the fire department, the firework industry, researchers, experts and himself and “that this approach requires a joint effort by all parties”. The use of fireworks - against the police and emergency service workers included - should be given priority, and the present Minister mentions that “the municipalities of Den Haag, Amsterdam, Rotterdam and Utrecht wish for the DSB to research which concrete measures can be taken to tackle violent behavior towards emergency service providers, the

To what extent does politicization drive agenda-setting? Insights from the Dutch Safety Board’s response to four crises