An empirical analysis on how citizens evaluate different attributes of violent incidents.
Master thesis submitted to Leiden University in partial fulfilment of the requirement for the degree of
Master of Science
In Crisis and Security Management Faculty of Governance and Global Affairs
By
Safak Topuz
Student number: S2467348 To be defended August 2020 Graduation committee: Supervisor Dr. H. Mazepus- i -
PREFACE
To be honest, this section was reserved to talk about my academic career and to emphasize some highlights of this journey. Although I would like to talk about my accomplishments, the truth is that there is beauty in the struggle and at the moment it feels better to talk about the issues I faced.
Last year might have been one of the most difficult years in my life, after fainting and falling on my head in the hospital it took almost four months to conclude that I had no health conditions. Meanwhile my dad had his heart surgery and was recovering from it. When in September I heard the conclusions of my cardiologist I felt relieved and was happy to be
healthy. Unfortunately, the happiness did not last long, on October the 27th I was shocked when
I saw my father and the condition he was in. Apparently, he had cancer and it was metastasized to his liver and he decided to battle this disease on his own. When I saw him, it was in his last
phase and unfortunately, he passed away on the 13th of November. The relationship with my
father was never what you could call ‘good’, we never had a decent discussion or a chat and only spoke a couple times a year. Maybe this is why his passing hit me so hard. If I could go back in time, I would have invested more energy in our relationship. I would rather be mad and angry at him after a talk, than not speaking like we did now. After his funeral I was in a sunken dark place where I had nightmares every single night for approximately three months. I have been battling anxiety, cried randomly at random moments, and felt like nothing made sense. Although I’m still struggling with the loss of a parent, my head is calmer now. I cannot pinpoint an event or moment that made me feel better, I guess you just learn to live with it.
In loving memory……
However, that being said I would still like to thank a number of people. First and foremost, my mother, who is my rock and my world. I am proud of her and she is the motivation to be a better version of myself. Secondly, my brother thank you for all your support in my life. Without this superman I would not be the person I am today, I am not sure where or what I would be
- ii -
without this role model. To my girlfriend, thank you for listening, advising, talking and being here with me. I hope you enjoy this journey. I would also like to thank my friends (you know who you are) for all the support and love. Finally, I would like to express my gratitude towards my supervisors. Dr. Mazepus thank you for the constructive feedback and guidance during this process. I would also like to thank Prof. Dr. Koops for taking the time to read and evaluate this thesis. Lastly, I would like to thank Dr. Akdogan for going through the statistical analysis. Hope you enjoy reading this thesis.
Beauty in the struggle, ugliness in the success. Feeling blessed. God is good.
- iii -
ABSTRACT
Some violent incidents are classified as an act of terror, while others are not. Even if these incidents have similar characteristics, they are evaluated and categorized in different ways. This thesis aims to examine this categorization process and is addressing the following research question: ‘Under what conditions do Dutch citizens categorize violent incidents as an act of
terror?’. According to the literature, it can be stated that individuals are programmed to think
in a certain way based on the following three drivers. Firstly, the media focuses on particular characteristics and hereby have an impact on the perception of the viewers. Secondly, the membership to a certain social group leads to in-group favouritism and out-group discrimination. Lastly, the securitization of terrorism legitimizes controversial actions and ideas. For this study, an experimental model has been created in order to generate 18 violent incidents. These incidents differ from one another based on six independent variables, which are: ‘tactic’, ‘target’, ‘casualties’, ‘actor-characteristic, ‘actor-type’ and ‘motivation’. The rated-based conjoint analysis has been distributed among the public (N=90). This study shows that the predictors ‘target’ and ‘actor-characteristic’ are not significant. Meaning that the background and target of the perpetrator did not matter in the evaluation of violent incidents. This is not in line with multiple studies, where the indication is given that a Muslim perpetrator is more associated with terrorism. Furthermore, motivation had the biggest impact on the evaluation of a violent incident. In case of a violent incident with political motives, the target group would find it more terroristic compared to an incident with socioeconomical motives. Moreover, the study shows that the used tactic, number of casualties and the composition of a group has an influence on the categorization process.
Keywords: violent incident, terrorism, conjoint analysis, multiple regression analysis, public perceptions.
- iv -
LIST OF ABBREVIATIONS
CBS Central Bureau for Statistics
CS Choice set DV Dependent Variable DS Dataset EC Effect Coding GA General Assembly IV Independent Variable IS Islamic State LR Linear Regression LWT Left-Wing Terrorism
MLR Multiple Linear Regression
PvL Planbureau voor de Leefomgeving
RG Research group
RP Reveal Preference
RWT Right-Wing Terrorism
SP Stated Preference
SPSS Statistical Package for the Social Sciences
TG Target group
UN United Nations
- v -
TABLE OF CONTENT
PREFACE ... I ABSTRACT ... III LIST OF ABBREVIATIONS ... IV TABLE OF CONTENT ... V ... 1 INTRODUCTION... 1 1.1 BACKGROUND ... 2 1.2 PROBLEM DEMARCATION ... 3 1.3 KNOWLEDGE GAPS... 4 1.4 RESEARCH QUESTION ... 5 1.5 THESIS OUTLINE ... 6 ... 7 LITERATURE REVIEW ... 7 2.1 TERRORIST THREATS... 8 2.2 AMBIGUITY OF TERRORISM ... 92.2.1 Role of the media ... 10
2.2.2 Social identity theory ... 10
2.2.3 Securitization of terrorism ... 12
2.3 CHARACTERISTICS OF TERRORISM ... 12
2.3.1 Objective component ... 13
2.3.2 Subjective component... 13
2.4 CONCEPTUALISATION OF CORE CONCEPTS ... 14
2.4.1 Tactic... 14
2.4.2 Target ... 15
2.4.3 Casualties ... 15
2.4.4 Actor ... 16
- vi -
2.5 HYPOTHESES AND VARIABLES ... 16
... 19
METHODOLOGY ... 19
3.1 ANALYSIS METHOD ... 20
3.1.1 Research Method ... 20
3.1.2 Experimental design ... 21
3.1.2.1 Attributes and their levels ... 22
3.1.2.2 Experimental design ... 23
3.1.2.3 Dependent variable ... 26
3.2 DATA GATHERING... 27
3.2.1 Targeting the citizens and distribution strategy ... 27
3.2.2 Consequences of strategy ... 27
3.3 PROCEDURE ... 28
3.3.1 Consent and introduction ... 28
3.3.2 Survey process ... 28
3.3.2.1 Assessment of violent incidents ... 28
3.3.2.2 General questions ... 29
3.3.2.3 Sociodemographic and descriptive data ... 29
3.3.3 Debrief ... 29
3.3.4 Terminology ... 29
3.4 LIMITATIONS ... 30
... 31
DATA PROCESSING AND SAMPLE CHARACTERISTICS ... 31
4.1 DATA PREPARATION ... 32
4.1.1 Missing values ... 32
4.1.2 Coding ... 33
4.2 VERIFICATION AND VALIDATION ... 34
4.2.1 Outliers... 34 4.2.2 Normality test ... 34 4.2.3 Representativeness ... 34 4.2.3.1 Distribution of age ... 35 4.2.3.2 Distribution of gender ... 36 4.2.3.3 Distribution of education-level ... 37
4.2.3.4 Distribution of living environment ... 37
4.2.3.5 Distribution of income ... 38
4.2.3.6 Conclusion representativeness ... 38
... 39
RESULTS ... 39
- vii -
5.1.1 Transforming data ... 40
5.1.2 Model selection ... 42
5.1.2.1 Full model... 43
5.1.2.2 Finding the right model ... 44
5.1.3 Model introduction ... 45
5.2 MODEL INTERPRETATION ... 49
5.2.1 Insignificant variables ... 49
5.2.2 Does the tactic matter? ... 50
5.2.3 Is the number of casualties relevant? ... 51
5.2.4 What influence does motivation have? ... 52
5.2.5 How does Actor-type affect the perception? ... 53
5.3 ANSWERING THE MAIN RESEARCH QUESTION ... 54
5.4 SEGMENTATION ANALYSIS ... 55
5.4.1 Gender ... 56
5.4.2 Age ... 56
5.4.3 Education ... 56
5.5 RELIABILITY AND VALIDATIONS ... 57
... 58
CONCLUSION ... 58
6.1 RESEARCH GOAL ... 59
6.2 RESULTS AND IMPLICATIONS ... 59
6.2.1 Feedback to the literature ... 60
6.2.2 Practical application ... 61 6.3 REFLECTION ... 61 6.4 RECOMMENDATIONS ... 61 REFERENCES ... 63 APPENDIX A ... 72 EXPERIMENTAL DESIGN ... 72 APPENDIX B ... 76 OUTLIERS ANALYSIS ... 76 APPENDIX C ... 80 NORMALITY TEST ... 80 APPENDIX D ... 82 REPRESENTATIVENESS ... 82 APPENDIX E ... 86
- viii -
APPENDIX F ... 88
TESTING RESULTS ... 88
APPENDIX G ... 122
- 1 -
- 2 -
1.1 Background
Following the Utrecht tram shooting incident on the 18th of March 2019, a debate on social media occurred whether this incident should be classified as terrorism. While Prime Minister Rutte spoke about a potential terrorist attack he indicated that the motives were still unclear (RTL Nieuws, 2019a). On different social media platforms the public, led by widespread of actors, organisations and media outlets, had a more severe reaction to this incident calling it a terroristic act immediately. According to Bradshaw and Howard (2019), political parties and governments use social media to spread political propaganda, the pollution of this digital platform with disinformation leads to polarization, distrust and shaping of the public opinion. In case of the Utrecht tram shooting incident it can be argued that different actors shaped the public opinion through social media.
In order to analyse this phenomenon further a brief summary of several characteristics of terrorism will be given. ‘’Act of terror can be seen as an act of violence (1), where citizens
have fear (2) or felt threatened (3). Aside from this, these cases show that it is a planned/systematic attack (4) with a certain strategy/tactic (5)’’ (Schmid and Jongman, 1988,
as cited in Weinberg, Pedahzur & Hirsch-Hoefler (2004, p.780). Bear in mind that there are dozens of other characteristics that can define terrorism. In case of the Utrecht incident the perpetrator, is convicted to a shooting with terrorist intent (Rechtspraak, 2020). The question that rises now is: how come other violent crimes that fit these criteria are not evaluated the
same?
To give some examples in 2019 there were 72 shootings in Rotterdam (NOS, 2019) and there was a significant rise of the use of grenades that were found in public space in Amsterdam (AD, 2019). Yet, none of these cases were being referred to as an act of terror by either the traditional media, social media, the public or the court. It can be argued that these types of crimes fit some of the criteria that are mentioned. For example, if the focus is set on the use of grenades in public space, it can be said that criminals use this method to extort and intimidate businesses (RTL Nieuws, 2019b). Therefore, the placement of a grenade is a form of violence that leads to fear for the surrounding. Subsequently, it is a planned attack, with a certain
strategy. This makes one wonder, does something become more terroristic when certain
- 3 -
1.2 Problem demarcation
‘Terrorism’ is a term that can be used in different ways and/or for different purposes. Schmid (2004) introduced five conceptual lenses in which terrorism can be viewed from. Depending on the lens, some aspects of terror get emphasized and included in the definition, while others are being neglected. These lenses are as follows;
1) Terrorism as/and crime 2) Terrorism as/and politics 3) Terrorism as/and warfare
4) Terrorism as/and communication
5) Terrorism as/and religious fundamentalism
These frames can be used as a tool to comprehend some aspects of terrorism. It is important to note that one frame does not exclude another. It is very reasonable that a violent act can consists of criminal and political components, which could be defined as a political crime. These lenses illustrate how diverse and broad the concept of terrorism can be, which again shows how contested this concept is. Aside from these lenses, the legal dimension complicates the issue even further.
While a lot of national and regional definitions of the concepts are proposed, no universal legal definition for terrorism which is approved by the General Assembly of the United Nations exists (Schmid, 2004). According to Bruce (2013) this can be seen as problematic for the following reason: counter-terrorism measures will not be effective, due to the wide international character of terrorism. The command and control of terrorist groups can be based on different states; therefore, all nations need to cooperate and agree to the characteristics of terrorist groups and their activities. A common universal definition regarding the concept would be a step towards universal cooperation. After the 1972 Munich Olympic massacre the United Nations unsuccessfully tried to get a universal agreement on the term. It became clear that different states were unwilling to put a ‘terrorist’ label to groups if they sympathised with their aims. An example of this is the support to the Nicaraguan Contras by former United States president Reagan, while in hindsight their actions could be seen as terroristic. This indicates that the question whether something is terroristic depends on the definer, which makes an objective definition difficult (Ganor, 2002). The notion is set to see whether an objective definition that is based on general observations is possible to find.
- 4 -
1.3 Knowledge gaps
Terrorism is not a new phenomenon. However, its intensity varies across times and its impact is different across regions. Therefore, each time a new wave of terrorism starts, it is regarded as a new phenomenon (Laqueur, 2017). As outlined by Rapoport (2001) in ‘The four waves of Modern Terrorism’, terrorism was well established long before the Islamic wave came up, in 1970. Throughout this period a lot of research has been done on this subject, resulting in various levels of terrorism that has been examined. It can be found that most of the literature has an explorative nature, where emphasis is placed on understanding the developments and dynamics that are at place. For example, the fact that democracy, education and finances play an important role in reversing acts of terrorism (Lawrence, Hamilton & Hamilton, 1983). In addition, lots of studies are looking into characteristics or elements that define terrorism and the phycological side where the group or individual is examined. These studies are trying to understand the motivations and strategies of perpetrators, McCauley (2004, p.35) outlined a wide variety of motivations that play an important role for individuals to become terrorists.
As indicated earlier the perception of an act of terror can be subjective, this means that states, organizations and citizens might have divergent views on whether something is terroristic. In this study the focus will be put on the citizens and their perception.
When it comes to the perception of ordinary citizens and their views on terrorism, different studies can be found. Lemyre, Turner, Lee and Krewski (2007) researched the public perception of terrorist threats in Canada. In this study the respondents were asked whether a terrorist attack could occur and if so what kind of types and effect it would have. Overall, it became clear that Canadians do not consider terrorism to be a threat. Although this study falls under perception of terrorism, it does not give insights on how characteristics of terrorism, as introduced in the first section, relate to each other. When investigating terrorism, as it is perceived within the public domain the research of Huff and Kertzer (2018) stands out. The researchers conducted a study to examine which various elements of incidents are perceived as terrorism by the public. This study used the combination of a conjoint experiment and machine learning to describe what the public exactly classifies as terrorism. Their study indicates that the perception of the citizens is dependent on the extremity of the incidents and the organization carrying out the attack. It is found that if the perpetrator is Muslim it is more likely associated with terrorism. Subsequently, the media coverage has a significant effect in the shaping of opinions of violent incidents. Moreover, this study also found out that the citizens do not necessarily think that violent incidents needs to target something specific in order to be seen as a terrorist act. To
- 5 -
clarify, Audi (2009) argues that an attack on a military base can be seen as more terroristic due to the political and governmental factors that are involved. Huff and Kertzer (2018) proved in their experiment that according to their respondents it did not matter if the target was either a civilian or military personal. One important drawback of the Huff and Kertzer (2018) study is that it is carried out in the United States only. This is problematic for broader conceptualisation of the research because it is case specific. To illustrate this problem, in the United States there is a lot of gang and non-gang affiliated gun violence due to the availability of firearms (Ludwig and Cook, 2004). In European countries, such as the Netherlands, with strict gun control laws it is very likely that the citizens would view gun violence as more severe due to the rarity than in the United States. The Huff and Kertzer (2018) study focus on how Americans think about terrorism, it would be interesting to see whether people groups in different regions of the world have similar judgements, this is also encouraged by Huff and Kertzer. The objective of this thesis is to see if a similar experiment in the Netherlands would result in similar observations. This research aims to contribute to a better understanding of relationships between elements of an act of terror and public perception.
1.4 Research question
Some of the violent incidents are classified as an act of terror within the public domain, yet it is not entirely known how Dutch citizens categorize violent incidents as an act of terror. It is unclear how different characteristics of an incident interrelate and shape the opinion of an individual. This thesis aims to contribute to a broader conceptualisation on the relationships of this public perception. The main interest lies on how citizens differentiate characteristics of violent incidents and how these characteristics weigh in the evaluation of whether something is an act of terror. Therefore, this thesis aims to answer the following research question:
Under what conditions do Dutch citizens categorize violent incidents as an act of terror?
With conducting the proposed study, the limited amount of existing literature concerning the important characteristics of violent incidents according to Dutch citizens will be further extended. The objective of this study is to contribute to a better understanding of the important relationships between characteristics of a violent incident and public perception.
- 6 -
1.5 Thesis outline
This thesis is structured as follows. In the next chapter a theoretical outline will be presented. This chapter will try to find the core concepts with the use of a literature review. These core concepts will be operationalised in order to promote measurability. Subsequently, in chapter three the methodology will be discussed. The emphasis here lies on the set-up of the experimental design, data-gathering method, the analysis method and the limitations of the research approach. In chapter four sampling and processing of the data will be highlighted. This chapter will elaborate more on how the data is prepared for analysis and will go through the verification and validation procedure. In chapter five the results of the analysis will be presented, first the estimation model will be explained, and the hypotheses will be testes. Based on the results of the analysis, the main research question will be answered. Chapter five will also perform a segmentation analysis where different sub-groups of the sample are being examined. Finally, in chapter six a conclusion will be formulated, followed up by recommendations for further research and a discussion regarding the methodology used in this thesis.
- 7 -
LITERATURE REVIEW
In this chapter a short introduction of terroristic threats will be given. Subsequently, three drivers that influence the categorization process of individuals will be explained. Furthermore, the search will continue to the key characteristics that determine whether a violent incident is an act of terror. After identifying these elements the conceptualisation of core concepts will be made, and the focus of this study will be set. Finally, the variables and hypotheses for this thesis will be drawn.
- 8 -
2.1 Terrorist threats
The 20th century faced different challenges, in which the military of a nation was the only form of self-defence. Until the very end of the Cold War, Neorealism, was the dominant factor in the studies of International Relations. Neorealism advocates that the core value of a state that needs to be secured is the national survival (Waltz, 1979; Morgenthau, Thompson and Clinton, 1985). After the Cold War there was more room to talk about other security threats that did not only threatened the state, but also the well-being of individuals (Krause, 2019).
The World Bank Group (2016) indicated that the amount of inter-state rivalry and/or wars are greatly reduced in the 21st century. Important to note that modern day conflicts are not only restricted to these wars or states. The increase in organized crime networks, such as terrorist groups, has caused a situation where non-state actors pose a threat to the national security. After 9/11, terrorists and insurgency groups were the most recognizable threats to the security of nation (Srikanth, 2014). According to Hoffman (2013) terrorism can be defined as ‘’deliberate creation and exploitation of fear through violence or threat of violence in pursuit
of political change’’ (as cited in Srikanth, 2014, p.62). This suggests that these groups always
have a political goal. In order to seek attention for their political agenda terrorist groups target the civilian population. An important security threat with non-state actors is that these groups do not recognize state boundaries and the jurisdictional laws that are in order. For governments it is incredible difficult to wipe out these terrorist groups due to their tactics. It can be noted that these groups use technology and social media, have sleeper and/or phantom cell networks and the element of surprise (Srikanth, 2014).
Islamic State (IS) will be given as example to illustrate the complexity of terrorist organisations. IS can be seen as a group that applied different tactics to reach their objectives (Valasik and Philips, 2017). They used multivariant activities to target different vulnerabilities of Westerns states (Jasper and Moreland, 2016). If the focus is put on the threats the following operations and tactics become clear:
• The organization used the so-called blended tactics. This can be seen in the traditional ‘military’ unit, incorporated with autonomous cells. The military unit was active in the fight in Syria, while the autonomous cells were committing attacks on European countries.
• The organization used information warfare. The internet functioned as a place where jihadist propaganda could be placed in order to recruit people, raise monetary funds and also spread disinformation to influence individuals.
- 9 -
• IS can be seen as an organization of criminals. They had the opportunity to generate revenues for their operations and campaigns.
What can be noticed from the operation of IS, is that terrorist groups use different tactics, both regular and irregular and that they deploy non-conventional methods to reach their objectives. It is difficult to exactly pinpoint which element of IS makes the group more or less terroristic. Scholars have tried to look for an answer and a wide variety of elements are identified. However, has stated before, this thesis will try to understand how the Dutch citizens categorizes these different elements. The upcoming paragraph will investigate the main drivers that influences the categorization process of the citizens.
2.2 Ambiguity of terrorism
In the United States a widespread debate followed whether the Las Vegas gunman, Stephen Paddock, should be labelled as terrorist (BBC, 2017). Together with cases of Dylann Roof (Charleston attack) and James Fields (Charlottesville attack) the notion was set that events with mass violence were not necessarily terroristic (Bump, 2015; Touchberry, 2018). However, if the focus is put on Omar Mateen (Orlando attack) and Sayfullo Habibullaevic (New York City Attack) the terrorism label is quickly applied to these cases (Ellis, Fantz, Karimi & McLaughin, 2016; Yan & Andone, 2017).
One comment that is frequently being used in social media platforms to address this difference is the following: if the ‘white’ attacker would happen to be Muslim, people would faster label the attacks as terrorism (Vedantam, 2017). This ‘gut feeling’ has been confirmed by different scholars. In Western countries it can be stated there is a correlation between the perception of terrorism and people from Middle Eastern decent and Islam (Park, Felix & Lee, 2007; Saleem & Anderson, 2013). West and Lloyd (2017) showed in their research that incidents committed by Muslims, were more often evaluated as an act of terror. Kearns, Betus and Lemieux (2018) even claim that if all the factors of a violent incident are exactly the same, and the only difference is whether the perpetrator is Muslim or not, the Muslim committing the crime would more likely to be associated with terrorism. In addition, Muslims especially foreign born, are a national security threat (Allouche & Lind, 2010). A very important limitation in these studies is that they do not take the differences among the public into consideration (Dolliver & Kearns, 2019). Thus, it can be stated, that the categorization process of individuals is still unclear.
- 10 -
This thesis aims to investigate this categorization process from the publics point of view. It is not made to decide when an incident should or should not be classified as an act of terror, but it is believed that understanding the public will help to understand the concept of terrorism. In this study three drivers are identified that can explain why some citizens consider some incidents as terroristic, while others don’t. These drivers are: the role of the media, social identity theory and securitization of terrorism.
2.2.1 Role of the media
McCombs and Reynolds (2019) state that the media plays a very important role in forming perceptions, preferences and ideas about humans, locations, events and other things which individuals have not personally experienced. The media could be seen as a tool that amplifies various incidents and ideas by spreading perspectives towards more people (Bekkers, Beunders, Edwards & Moody, 2011). This can lead to a situation where the media amplifies one incident, while ignoring a different incident. It can also be stated that the media is determining what perspectives, positions and preferences the viewers (citizens) are exposed to (Bekkers et al, 2011). In a study that was conducted among California residents, Callanan and Rosenberger (2010) showed that media coverage had a major influence on how the residents viewed security issues. It can also be seen that the coverage of terrorism in the media has a positive correlation to the perceived risk and fear of viewers (Nellis & Savage, 2012).
The spread of information can be noticed when a focussing event occurs. Focusing event can be seen as an attention-grabbing event where public awareness towards a certain topic is created (Kingdon, 1995). During a focusing event the discussion and debates regarding some topics receive more attention than others. The media does not really determine how an individual feel or thinks about the topic; however, it does set the tone for the issues to discuss (McCombs, 2003). Whenever a focussing event occurs, where the perpetrator is Muslim, the following stands out. First, there is more media coverage about the event (Arya, Idris & Pervez, 2017). Second, the event is more likely to be associated with terrorism (Kearns et al, 2017). It can be stated that the media focuses on particular characteristics of violent incidents and hereby have an impact on the perception of people, these individuals are than classifying the situation based on this information.
2.2.2 Social identity theory
Tajfel and Turner (1979) introduced the social identity theory to explain the psychological and behavioural relations in groups. Their studies demonstrated that a membership to a group
- 11 -
could be so powerful that it leads to in-group favouritism and out-group discrimination. In other words, people that are in the same social group are more positive against each other, while on the other hand people in other groups are viewed more negatively and sometimes even discriminated. Social identity theory claims that individuals go through three processes to make in-group/out-group classifications. First, social categorization process, Tajfel and Turner (1979) argue that individuals tend to organize people into different social groups in order to mitigate a situation. This results that people tend to categorize individuals based on their social group, instead of their individual characteristics. It is possible to put an individual in different categories, but it will only play a role if the categories are relevant to the social group. The second process is the social identification, which focuses on the identification of group members. This leads to individuals that are acting on the basis of how they believe people in the social group should act (Brewer, 1999). Last process is the social comparison, this refers to the fact that individuals compare their social groups with other social groups (Tajfel, 2010). Social identity theory is a widely studied phenomenon, where the in-group and out-group behaviour of various social groups, based on ethnicity, political orientation, religion, studies, organizations, professions, hobbies and many more, are studied.
When this theory is applied in the context of terrorism and counterterrorism, it can be noted that people associate members of the out-group as more violent (D’Orazio & Salehyvan, 2018). For the in-group it becomes easy to generalise and even stereotype the out-group behaviour to a certain narrative (Haslam, Oakes, Reynolds & Turner, 1999). This means that when someone from the out-group is committing an act of terror, the individual characteristics are not as important as the social group this person is in. When someone from the same social group commits a similar crime, it becomes difficult for the members of this social group to identify themselves with this individual. This results in looking for alternatives to describe this person. For example, due to the social identity theory individuals can label groups as ‘terrorist’, ‘rebels’, ‘activists’, ‘guerrillas’ or ‘dissidents’ (Norris, Kern & Just, 2003, p.3). It is important to note, as described in the previous section, that the role of the media in this categorization process is very important. In practice, the narrative in news stories is that terrorism is framed as an Islam problem (Sultan, 2016), while if someone of the in-group is the perpetrator the person is framed as either a lone-wolf or ill-perpetrator (Powell, 2011). This framing is not only done by the media but also by politicians.
- 12 -
2.2.3 Securitization of terrorism
Securitization theory refers to transforming subjects into matters of security. It can be seen as boxing the subject as a ‘threat’ and by doing so exclude other potential meanings or interpretations (Hansen, 2000). According to Buzan, Wæver, Wæver and de Wilde (1998) the following three characteristics are essential in securitization: identifying the existential threats, emergency actions and effects on relations.
The 9/11 attacks were major events for Americans, in which the comparison with Pearl Harbor was quickly made (Edmunds, 2012). The Western world was united in their reactions towards this act and the attack was seen as a new security threat. Terrorism threatened the well-being of the lives of individuals. Although the concept of terrorism was not new, the 9/11 attacks changed the politic on a domestic, regional and global level (Afzal, Iqbal and Inayay, 2012).
These attacks can be seen as a pivotal moment, where different politicians in various Western countries were focused on securitizing subjects in order to legitimize their actions (Abrahamsen, 2005). An example that can be given is that in the United States, President Trump has repeatedly indicated that the only form of terrorism in the United States comes from Islam (Beauchamp, 2017). This way he can justify some actions, which are controversial, such as closing the borders for different minorities. It is important to note that social identity theory can still play a role under these circumstances. For example, conservatives could evaluate a violent incident as an act of terror, only if the perpetrator is Muslim. The reason can be because they are in the same social group as President Trump or due to way he securitizes the subject and influences their perception.
2.3 Characteristics of terrorism
It can be argued that citizens are programmed to think a certain way when it comes to categorizing violent incidents. The impact of the media, the social group someone belongs to and the securitization of subjects influences their perceptions, preferences and ideas. This study aims to investigate what characteristics individuals find important when evaluating violent incidents.
When individuals classify an act of terror two components are found important (Huff and Kertzer, 2018). First component is the objective facts that are known, which contain information about the type and severity of the incident in terms of violence, target and location. This information is mostly known immediately and given to the public. The second component has
- 13 -
a more subjective character concerning the identity of the actors involved and the motivation they have.
2.3.1 Objective component
This component investigates the fundamental characteristics of the incident. According to basic definitions of terrorism two factors are always being discussed during/after terrorist attacks. The first factor is violence (Weinberg, Pedahzur and Hirsch-Hoefler, 2004) and the second one is the information about whom the attack was against.
For violence this works in various ways, according to Kalyvas and Balcells (2010) the different technologies characterized the different tactics that could be employed in an incident. It can be argued that hostage taking is a whole different type of tactic than planting a bomb. There is a correlation that any bombing incident will be classified as terrorist attack by the citizens (Abrahms, 2013; Huff and Kruszweska, 2016). According to Bloom (2005) bombing is such an act of violent crime it is automatically associated with terrorism.
Subsequently, the target and the location of an incident matters to the public. Firstly, it is especially important if the target is either an innocent bystander/civilian or a non-combatant (Coady, 2004). It has been found that political targets receive more attention in the media (Zhang, Shoemaker & Wang, 2013). According to Authors (2017) a governmental target will result in higher chances of defining an attack as an act of terror by the public.
Furthermore, the location of the target is important. Following the terrorist attacks in Paris and Lebanon in 2015, a wide variety of individuals criticized Western media outlet for the selective media coverage regarding terroristic attacks occurring in different states, which were outside European borders. Pizam and Smith (2000) found out that in 79% of the cases the location of an attack influences the perception of the citizens.
Finally, the severity of an attack is important to the people’s perception. If the casualty rate is higher, the shock value of the public increases (Zhang et al, 2013). Chermak and Gruenewald (2006) found out that even just one casualty would influence the perception of an attack. The media attention to the topic also increases when there is more violence (Miller & Albert, 2015).
2.3.2 Subjective component
After an incident the public, media and governments look for those who are responsible for the attack. Usually people tend to think that terroristic activities are organized by a collective, where the objective could be seen as a part of a political or social agenda (Spaaij,
- 14 -
2010). An individual perpetrator that is responsible for an attack creates doubt in the eyes of the public, in where the question rises if there is a broader agenda to the cause (Brewer, 1999).
The psychological factors for a perpetrator to commit an act of terror is not applicable for all circumstances (Crenshaw, 1981). The motivations to commit an act of terror varies immensely. In general, it can be stated that the grievances play an important role to commit an attack. For example, the lack of political representation leads to the rise of separatist movements (Crenshaw, 1981). In most of the attacks the motives become clear in order to promote the political agenda, in other cases the motives are not defined (Las Vegas attack).
It is important to note that the three drivers that were identified are active in the subjective component of the categorization process. These drivers affect the emotions of citizens and influence how they evaluate the violent incident (Dumont, Yzerbyt, Wigboldus & Gordijn, 2003). The media will inform and emphasize on the background and motivations of the perpetrator. The citizen that receives this information will categorize the act based on the in-group and out-in-group dynamics that are at place.
2.4 Conceptualisation of core concepts
The previous sections tried to identify the drivers that have an impact on the minds of individuals. In addition, it is explained how the objective and subjective components of terrorist acts influences the perception of the public. This section will elaborate how the understanding of different notions will be used in this thesis. In order to eliminate the ambiguity in the concept of terrorism the core concepts that will be investigated in this study will be explained. In this thesis the following simplistic definition of terrorism will be used: ‘the systematic use of
coercive intimidation against civilians for political goals’ (Norris, Kern and Just, 2003, p.3).
The key elements of terrorism that are identified with this definition are: techniques, targets and objectives. By combining these elements with the objectives and subjective components that were identified the following core concepts become relevant for the citizens in this study, these are: tactic, target, casualties, actor and motivation. It is believed that these core concepts influence the perception of the citizen and determine whether a violent act is an act of terror. The core concepts will be delineated in order to use them in the proposed study.
2.4.1 Tactic
An act of terror can be seen as a tactic that involves intimidation, threats or the use of violence in order to destroy a property or physically harm other individuals (Norris et al, 2003). From this perspective three different tactics are identified that terrorists can use in order to
- 15 -
destroy or harm people. The first tactic is a hostage incident. This tactic was used in almost 15% of all terroristic events between 1968 and 2005 (Mickolus, Sandler, Murdock & Flemming, 2006). A hostage incident receives media attention, and due to the publicity, that is generated, it results in recruitment and money (Brandt & Sandler, 2010). Between the 60s and 70s Palestinian terrorists actively used this tactic in commercial airlines to get media attention for their political views and inspire others to do the same (Nacos, 2014). The second tactic is the use of conventional firearms to intimidate or threaten individuals. Firearms are one of the most common weapons used by lone-wolf terrorists (Spaaij, 2010). In the examples that were given (Las Vegas, Orlando, Utrecht) but also others like (Christchurch, Utoya) firearms were used to commit terrorism. The third tactic that terrorists predominantly use are bombs, these account for almost 75% of all international attacks (Williams, 2004).
All of these tactics are used as a mechanism to control and are related to the other core concepts. The chosen tactic is based on the operation and is designed to threaten the target population.
2.4.2 Target
The procedure to select a target to attack, begins after the terrorist makes the decision to launch an attack. The target selection is a complicated process which is affected by different factors (Hemmingby & Bjorgo, 2018). A very important factor in choosing a target is the achievement of high media attention (Torres-Soriano, 2019). An important development that can be seen is that transnational terrorists have shifted their focus from military and governmental targets to business and private targets (Brandt & Sandler, 2010). This shift can be explained due to the simple fact that the governments are invested in stopping these groups, so the terrorists are focusing on targets that are harder to defend (Brandt & Sandler, 2010). The targets that will be used in this study are the logical extremes. Firstly, a governmental building is selected. As described, the theory indicates that initially an attack on the government was seen as terroristic. The second and third targets that are defined are places that are hard to defend, namely a school and public area.
2.4.3 Casualties
The number of casualties an incident has is related to the tactic and target. Multiple studies indicated that even one casualty will increase the perception of the public. On the other hand, if the aim of a terrorist is to gain media attention, they would prefer more numbers. For this
- 16 -
study it has been chosen to limit these values, to see how the citizens react on smaller differences. The number of casualties that will be used is 0, 5 and 10.
2.4.4 Actor
The actor in this study is divided in ‘actor-characteristic’ and ‘actor-type’. Actor-characteristic refers to the background of the perpetrator. According to the Dutch government, the main threats of terrorism in the Netherlands comes from Jihadism. However, due to recent developments the government is also focusing on threats of left-wing (LWT) and right-wing (RWT) terrorism (AIVD, 2020). According to Hewitt (2003) there is a correlation between right-wing terrorism and lone-wolf terrorism. It can even be argued that Islamic extremists that are seeking the Jihad are stimulating more right-wing extremists to react in a counter-jihad (Shaffer, 2016). For this study the background of the perpetrator will only be described as left-wing orientated, right-left-wing orientated or Muslim.
The actor-type refers to the composition of the group, a clear distinction can be made between lone-wolf terrorism and terroristic groups. The difference lies in the single-minded actors versus the decision-making units and how this influences the tactics, targets and motivations of the perpetrator (Horgan, 2005).
2.4.5 Motivation
According to Zalman (2018, p.1) couple motivations can be defined when investigation motives of terrorists, which are political and socioeconomical. Political terrorism is violence from a non-state group or individual to gain political dominance over a different group. In general, these terrorists perceive some social and political elements as wrong an try to correct it. For example, guerrilla warfare or abortion clinic bombings. Socioeconomic terrorism can be explained as divergent set of factors (poverty, education) that drive people to participate in terrorism, their situation makes them more vulnerable for recruitment by organizations. In this study these two types of motivations are defined as either changing governmental policies (political motives) or personal vendetta (sociodemographic motives).
2.5 Hypotheses and variables
In order to investigate the perception of citizens an experiment will be conducted. In this experiment 18 hypothetical scenarios of violent incidents will be presented to the respondents. The respondents are asked to evaluate these incidents on a scale from 1 to 10, The 18 violent incidents will differ from one another based on the core concepts that are defined. Chapter 3
- 17 -
will elaborate more on the experimental design and the survey set-up. Table 3.1 in chapter 3 gives a summary of the independent variables that are used to determine whether a violent incident is an act of terror (dependent variable). This study will focus on the effects of the following independent variables: Tactic, Target, Casualties, Actor-characteristics, Actor-type and Motivations. From this point on these independent variables can also be called attributes.
All of the attributes consist of three layers, the so-called levels. These attribute-levels are based on the literature found in academic work and in grey literature. In chapter 3 a combination of these attributes and their attribute-levels will be used to generate hypothetical incidents that will be used in the experiment. Figure 2.1 shows a conceptualisation of what the proposed study will investigate. The objective of this study is to find out which independent variables influence the perception of the citizens and how big this effect can be. The three drivers that are identified that could influence the perception of the individuals are minimized in this study. The focus is primarily on how the characteristics that are chosen influences the respondents, however the three drivers can still influence the categorization process.
The following hypotheses are formulated:
Hypothesis 1: If the target of an attack is a governmental body, it is more likely to be
evaluated as a terrorist attack.
Hypothesis 2: The background of the perpetrator matters to the public, incidents in where
the perpetrator is Muslim are expected to be seen as more terroristic.
Hypothesis 3: In comparison with hostage-taking and a shooting, a bombing should be
perceived as more terroristic.
Hypothesis 4: The number of casualties in violent incidents has an effect on the perception
of the citizens, more casualties in an incident will more likely be associated with terrorism.
Hypothesis 5: If the motivation has a political element, such as changing governmental
policies, the incident is more likely to be associated with terrorism.
Hypothesis 6: If the attack is carried out by an organization, it would more likely to be
- 18 -
Figure 2.1
Conceptualisation of the study
- 19 -
METHODOLOGY
In this chapter a clear research approach will be developed based on the included variables and the demarcation of the scope. Furthermore, the research method will be explained and data-gathering will be discussed. This chapter ends with the possible limitations of the chosen method.
- 20 -
3.1 Analysis Method
In order to answer the main research question a quantitative analysis will be used. The aim of a quantitative research is to achieve reliable, generalizable, and valid results (Srnka & Koeszegi, 2007). To quantitatively analyse the influences of attributes, data associated with these attributes are needed. To represent the target group, which is the Dutch population, it is important that the method is non-exclusive, and it should cover as much as sociodemographic as feasible, in order to sample participants that are a reflection of Dutch society. Due to the wide variety of the population the expectation is that multiple predictors influence and determine the characteristics of violent crimes. A quantitative analysis will be used to look for some of these predictors.
The methodology for this choice of analysis is the use of a questionnaire. A fixed survey will allow the respondents only to answer categories that are predetermined. This survey aims to strive for the maximum explanatory power that is available. However, it is important to note that not all factors can be researched with this method.
3.1.1 Research Method
To investigate or model the perception of Dutch citizens it is important to know the context in which some choices are made. For this type of study, the decision process of the target group can be searched with either revealed or states preferences (Hanley, Mourato & Wright, 2001). These two paradigms are primarily used to quantitatively analyse the importance of different determinants. Reveal Preference (RP) refers to the choices that are actually made in real markets, while Stated Preference (SP) refers to choices that are made in hypothetical situations (Liao, 2014). In SP the researcher is able to formulate scenarios and attribute levels, which are not available at the moment, making it more suited for hypothetical scenarios. As a part of the choice modelling family, conjoint analysis and SP will be used to answer the research question of this study.
A conjoint analysis has two purposes. Firstly, a conjoint analysis is based on a survey and uses statistical techniques to determine what value people tend to give to different attributes (characteristics, features, elements) that make up for a product/service (Orne, 2006). In other words, the preferences of consumers become clear. Secondly, compared to traditional survey experiments, a conjoint analysis makes it possible for the researcher to vary multiple factors in one experiment and study the multidimensional objects (Leeper, Hobolt and Tilley, 2020). In this case it becomes possible to study and compare causal effects (Knudsen & Johannesson,
- 21 -
2019). This leads to a method where the trade-off behaviour of the respondents can be investigated, without showing the test objective to the respondents. It is also possible to develop an experimental design, that is flexible for the needs of the researcher (meaning that inclusion and exclusion of factors can be determined).
As indicated earlier, the important characteristics of violent incidents are subject to the investigation of this study. The choice-tasks will be given to the respondents in a SP survey, the choice tasks are defined by several attributes and attribute-levels. To develop violent incidents an experimental design will be used to specify the attributes and attribute-levels. The chosen experimental design will also determine the number of choice-sets that needs to be given to the respondents.
In order to perform a proper analysis on the different attributes of this study a multiple regression analysis and conjoint analysis will be used. The conjoint analysis will be rated based experiments in where the respondents are presented experimentally designed hypothetical situations (Oppewal & Timmermans, 1993). The rating that a respondent gives to the hypothetical incidents will be used to determine the partial valuation of attributes.
A conjoint analysis is used in economic market studies, where the fundamental idea is that respondents need to make a consideration between characteristics of a product. The relationship of characteristics can be corroborated by the sample and the influences of the characteristics. Aside from market studies, an increasement in the use of conjoint experiments can be observed within social science (Mutz, 2011). Due to the nature of the conjoint analysis different multidimensional studies have been conducted, such as the attitudes towards immigrants by American citizens (Hainmuller & Hopkins, 2014), the labour market reform policies (Gallego & Marx, 2017), examination of health care systems (McClain and Rao, 1974), and many more. This study aims to make it clearer how the preferences of citizens are related to the evaluation of terroristic elements of violent incidents. A conjoint analysis is the perfect method to compare multiple attributes that are relevant.
3.1.2 Experimental design
In order to perform a conjoint analysis a couple steps need to be taken. This section will elaborate further on the design process of the experiment. First the attributes and attribute-levels are defined, subsequently the type of design will be highlighted. Finally, the choice-sets of the experiment is generated.
- 22 -
3.1.2.1 Attributes and their levels
Ideally a conjoint analysis would be performed to see what attributes (characteristics) the target group would find important. For example, a new high-end electric vehicle might have various attributes such as: price, range, charging time, seats, and etc. For all of these attributes different attribute-levels can be developed. The attribute-levels for the attribute ‘range’ could therefore be: 100, 500, 1000 and 2000 miles. For all of the other attributes different attribute-levels can be developed. The difficulty for the proposed study is that there are no consumers and therefore no demand for a certain product or service. In fact, it can be assumed that no-one is waiting for a violent act to happen. In addition, the purchase of a vehicle is easier to imagine and it is easier to decide which elements are important. For the proposed study, it is more difficult to define these important attributes.
Yet during the literature review an attempt was made to elaborate on the included and excluded attributes of this study. Currim, Weinberg & Wittink (1980) stated that attributes that are defined on three levels are more important than two level attributes. Furthermore, in order to investigate non-linear effects, and attribute must have more than two levels. Therefore, in this study it has been chosen to include three levels in each attribute. Table 3.1 shows all the attributes with their attribute-levels.
Table 3.1: Chosen attributes and their corresponding attribute-levels
Attributes Level Attribute-levels
Tactic 0 1 2 Hostage Shooting Bombing Target 0 1 2 Governmental body School Public area Casualties 0 1 2 0 5 10 Actor-characteristic 0 1 2 Left-Wing Right-Wing Muslim
- 23 - Actor-type 0 1 2 Organization Lone-wolf Unknown Motivation 0 1 2 No known motivation Changing governmental policies Personal vendetta
3.1.2.2 Experimental design
There are different experimental designs that can be applied. First of all, a full factorial design, is an experimental design that includes all possible choices to the respondents. This design can be used to estimate all effects to the dependent variable (main and interaction effects). The downside is that in order to include a full factorial design for this study an amount of 36 (three levels and six attributes) 729 scenarios need to be created, it is unrealistic to include these many scenarios for practical reasons.
A fractional factorial design is a design that consists of a subset from the full factorial design. In order to preserve the correlation between attributes-levels an orthogonal design is chosen as well. This means that in the orthogonal design the parameter estimates are unbiased, but it can occur that some choice-sets are more dominant than others.
To construct an orthogonal fractional factorial design a so called ‘basic plan’ will be used. A basic plan is a blueprint of an experimental design made by mathematicians (Molin, 2017). A basic plan is used to assure that the design is orthogonal, which again means that attributes of the experiment are not correlated. There are multiple basic plans to choose from, but the decisive factor it the number of variables that are included in the study. Due to the six attributes that are chosen and the three layers that they consist, basic plan 4 is the most suitable to use for this study. Figure 3.2 below shows this basic plan, if it is necessary another attribute (7th
variable) can be added to this study with this basic plan. The numbers 0,1 and 2 in each column represent the attribute-levels that can be used. The number of rows are the profiles that will be generated with this basic plan. From basic plan 4 it can be noted that there are 18 rows, which will design the 18 profiles. In order to keep the orthogonal design, it is not possible to exclude rows and generate less profiles, however it is possible to include or exclude columns and use less or more variables.
By giving all attributes and their attribute-levels a code, a profile can be generated. For example, for the attribute ‘tactic’ the 0 is assigned to a hostage situation, 1 to shooting incident
- 24 -
and 2 for a bombing scenario, as shown in table 3.1. If then column 1 is given to tactic it becomes possible to categorize this attribute through all the rows. Table 3.3 shows the incorporation of all attributes in the basic plan. From this the profiles (scenarios) can be generated, the first row shows the following sequence: 0, 0, 0, 0, 0, 0. This means that the first profile, violent incident 1, will consist of: a hostage situation, on a governmental body, with zero casualties, by left-wing orientated, unknown actor and with unknown motivations. If we look at the 4th row the following sequence: 0, 1, 1, 1, 2, 0 this leads to a: a hostage situation, in a public area, with 5 casualties, by a right-wing orientated, unknown actor, without known motivations. Table 3.4 shows the first couple profiles of the experiments. Appendix A elaborates further on the experimental design and other profiles can be found here as well.
Figure 3.2
- 25 -
Table 3.3: Attribute-levels of the independent variables that are incorporated in basic plan 4
Profile number Attribute level (Tactic) Attribute level (Target) Attribute level (Casualties) Attribute level (Actor-c) Attribute level (Actor-t) Attribute level (Motivation) 1 0 0 0 0 0 0 2 1 1 2 1 1 1 3 2 2 1 2 2 2 4 0 1 1 1 2 0 5 1 2 0 2 0 1 6 2 0 2 0 1 2 7 0 2 2 1 0 2 8 1 0 1 2 1 0 9 2 1 0 0 2 1 10 0 2 1 0 1 1 11 1 0 0 1 2 2 12 2 1 2 2 0 0 13 0 0 2 2 2 1 14 1 1 1 0 0 2 15 2 2 0 1 1 0 16 0 1 0 2 1 2 17 1 2 2 0 2 0 18 2 0 1 1 0 1
Note. Each column represents an attribute. All attribute-levels are assigned to a number (0, 1 or 2). This means that in each column every level comes back six times (balanced design). For example, hostage-situation in the attribute tactic is assigned to 0. It can be seen that profile 1, 4, 7, 10, 13 and 16 consists of a hostage-situation. In the same fashion all other attribute-levels are assigned to a profile.
- 26 -
Table 3.4: First four violent incidents of the experiment
Profile 1 2 3 4
Tactic Hostage Shooting Bombing Hostage
Target Governmental
body
Public area School Public area
Casualties 0 10 5 5 Actor-characteristic Left-wing orientated Right-wing orientated Muslim Right-wing orientated
Actor-type Unknown Lone-wolf Organization Organization
Motivation No known motivation Changing governmental policies Personal vendetta No known motivation
Note. There are a total of 18 profiles constructed in the same fashion, all of these profiles will be used in the survey. This table should make it clearer how the diversity and range of the attributes and attribute-levels will be used to generate these profiles (violent incidents). For more elaboration and more profiles see Appendix A.
3.1.2.3 Dependent variable
Each profile consists of the six independent variables (the attributes), the dependent variable is ‘evaluation of the violent incident’. The respondent is asked to evaluate to what extent the violent incident described can be seen as an act of terror, this is done on a 10-point Likert scale. Each respondent needs to evaluate all 18 profiles, otherwise the data is incomplete, and orthogonality is not preserved. By evaluating all profiles, it becomes possible to analyse how different predictors influence the dependent variable. The objective is to understand the variation in the dependent variable, based on the independent variables. By investigating this variation, the importance of the attributes and attribute-levels become clear. To give a brief example, if ‘an act of terror’ for someone is primarily focussed on the number of casualties, by definition he/she should evaluate all violent incidents with 10 casualties, with a higher grade than other violent incidents. In practice the complexity of the design, a large number of variables (6 attributes with 3 levels -> 18), will make the trade-off even more difficult.
- 27 -
3.2 Data gathering
To carry out the proposed conjoint analysis, data of the preferences of the public is needed. With the use of a survey the necessary data will be collected to do the analysis. There is a variance in recommendations regarding the sample size that is needed in order to perform the suggested analysis. Green (1991) addressed this question considering the purposes of a multiple regression analysis and statistical power. Statistical power refers to the probability that the analysis rejects the null hypothesis, when an alternative hypothesis is true. When the statistical power increases, the chances of making a mistake to wrongly reject a null hypothesis decreases. According to Green (1991) if the researcher is interested in the beta weights the following formula should determine the sample size: N≥104+k, where k is the number of independent variables. However, one of the most commonly accepted rule of thumbs that is being used in SP experiments comes from Orme (1998). Orme (1998) proposed the following equation to the determine the sample sizes in SP experiments, where the focus is on the main effects only:
N≥500×𝐿𝑚𝑎𝑥
𝐽∗𝑆 ,where Lmax is the largest number of attribute-level of all attributes, J is the number
of alternatives and S=1. For this study the Lmax is 3 and J is 18 in order to estimate the main
effects, while remaining its statistical power a minimum of 83.3 respondents are needed.
3.2.1 Targeting the citizens and distribution strategy
Given the scope and research aim of this study, this explorative survey is interested in the perception of an individual that is based in the Netherlands. The survey was distributed in a digital format, in order to stimulate the respondents to fill it in. ‘Qualtrics’ was used to programme the survey and gather data. The expectation was that the easy interface and smart-phone compatibility would make it more accessible for the respondents to participate in the survey. The survey was spread out in multiple ways and it can be found in Appendix G. For the distribution of the survey the following strategy was at order. Firstly, with the use of social media various people were targeted. Secondly, different cities (Arnhem, Delft, Rotterdam and the Hague) in the Netherlands were visited to ask people to fill in the survey. Lastly, various friends were contacted to distribute the survey amongst their co-workers.
3.2.2 Consequences of strategy
The distribution strategy that is described above can be categorized as convenience sampling, where the respondents are from a group that is approachable. Notably, there are some issues regarding this approach. The researcher of this study is not active on all social media
- 28 -
platforms, aside to this not every citizen is active on social media either. This would result that only a specific group of people will be represented in the survey. Aside to this, the visited cities will only give an indication of the perception of people in urban areas. There is no proof that the opinion of this group is in line with more rural areas or other cities in the Netherlands. In addition, friends and their co-workers are currently active in governmental bodies and consultancy firms. All of this could mean that the age, education, interests and therefore the preferences are not representative to the population.
3.3 Procedure
This section will elaborate the process that a respondent is going through, the complete survey can be found in Appendix G
3.3.1 Consent and introduction
When a respondent enters the digital environment, a brief introduction is presented to the reader. The purpose of the research and the expected time that is required to successfully complete the survey is also shown. Furthermore, it is highlighted that aside from the researcher the direct thesis supervisors had access to the anonymized data and that the final results will be published in the university repository. The respondent was asked to consent to the following conditions:
1) The data is completely anonymized before use.
2) The data will not be used or sold other than graduation purposes. 3) The respondent cannot go to previous questions to alter their choices.
3.3.2 Survey process
This section will briefly go through the different blocks of the survey and highlight what is sought to analysis with these questions.
3.3.2.1 Assessment of violent incidents
The survey starts with presenting a set of 18 violent incidents that were created with the use of basic plan 4, the respondents are asked to evaluate these fictitious violent crimes one by one on a scale from 1 to 10. The scale-variable (dependent variable) is the perception of whether something is terroristic. To make sure that the attributes are known, a short summary of each attribute is displayed below the incident that is described.
- 29 -
3.3.2.2 General questions
The next part of the survey is regarding the background of the respondent. In this part the respondent is asked to define their own definition of terrorism, followed by general questions regarding the topic.
3.3.2.3 Sociodemographic and descriptive data
In the last section of the questionnaire the interest lies on the sociodemographic information of the respondent. The questions in this section were focused on age, income, education level, gender and living environment. These questions will be used to look at the representativeness of the survey and for segmentation analysis.
3.3.3 Debrief
After completing the survey, the respondents are asked if they wish to be kept informed about the outcome of the study. In case the respondents are interested, they can contact the researcher.
A total of 165 respondents went into the survey environment and 98 of them completed the survey. During this data-gathering phase five people reached out and were interested in the outcome of the study and asked to be kept informed. Subsequently, three people made suggestions to the survey, stating that the political orientation of the respondents would be interesting to include in the survey. Unfortunately, a couple respondents were not amused by the survey. One person called it unnecessary and was upset that the research was limited to its scope. This person thought that the survey is focussing on the wrong elements and had a ‘self-fulfilling prophecy’. ‘If the focus is set on Muslims, you will mostly find out that this will be
displayed in the results.’ Finally, an important draw-back that came back as feedback was the
difficulty of the survey. Respondents had a hard time in understanding the concepts and answering the questions. This can be seen in the amount of respondents that did not complete the survey (67), which shows that the drop-out rate of 40%.
3.3.4 Terminology
In order to prevent confusion or form a barrier for respondents, the terminology in the survey is kept as simple as possible. In this order no misunderstanding could occur, and the respondent will not be influences by certain terminology. For the sake of an impartial research it is important to measure the preferences and perceptions of the respondents. Their views are essential; therefore, no definitions of certain words are given. For example, the concept of terrorism is not explained, and it is not controlled what this term means.
