RADIMED
Explaining and understanding the role of exposure to new social media on violent extremism.
An integrative quantitative and qualitative approach

Summary

en kwantitatieve benadering/ Commissioners
Promoter/coordinator
Co-promoters

Researchers

Belspo
FOD Binnenlandse zaken – Directe algemene veiligheid
Lieven PAUWELS - UGent
Fabienne BRION - UCL
Brice DE RUYVER – Ugent
Marleen EASTON - Ugent
Nele SCHILS – Ugent
Julianne LAFFINEUR - UCL

1

1.

Introduction

Research problem
Worldwide, both policy and research pay a lot of attention to extremism, and radicalization as the process leading to extremism (Van de Linde & Rademaker, 2010). Scholars and policymakers increasingly focus on unravelling the processes of radicalisation, hoping to prevent the violent radicalisation of their own youth and eventually political violence (Van de Line & Rademaker, 2010). The Internet and its constant technological developments in particular are causes of concern (Conway, 2012).
Recently, we have seen a boom in new social media (NSM) and other web 2.0 applications, bearing a large potential for communication and networking (Conway, 2012). These developments have transformed the world in an online village, with every offline actor being represented online. It is therefore no surprise that criminals, radicals, violent extremists and terrorists also use this medium to their advantage (Benschop, 2006; Stevens &
Neuhmann, 2009; Weimann, 2004). By means of the Internet and NSM, violent extremist organisations and individuals are able to easily reach each other and address a broad, global audience, using an extensive and dynamic set of narratives.
This has caused a growing fear that recruitment and violent radicalisation will increase under influence of the
Internet (Thompson, 2011). According to the AIVD1 (2006) the Internet may even be seen as the turbo of contemporary violent Jihad and one of the principal instigators of bottom-up processes of radicalisation and
‘Jihadisation’. This has caused policymakers to increasingly worry about what governments are able to do to prevent (online) violent radicalisation and recruitment. In this regard, it is problematic that there is hardly any empirical evidence confirming the relationship between exposure to extremist messages and attitudes that are supportive of violent extremism (Conway, 2012, Silke, 2008). The few empirical studies available that do tackle the topic of Internet exposure and violent extremism are limited to content analysis of extremist forums and websites, focus on the practical usage of the Internet and NSM by extremist groups or analyse virtual communication. Still, it is important not only to understand how the Internet and web 2.0 applications are used by extremist groups and individuals, but also to increase our understanding of the relationship between NSM and violent radicalisation. This is necessary to come to a better understanding of contemporary violent extremism and subsequently successfully tackle processes of racialisation and recruitment, online and offline (Zhou, Reid, Qin,
Chen, & Lai, 2005).
The present study2 assesses the impact of exposure to extremist content through NSM on the processes of radicalisation and recruitment among Belgian adolescents. The central research question focuses on the role of
NSM in the process of violent radicalisation. The following key research questions will be addressed.
1.

Is exposure to extremist content through NSM related to offline violent extremist activities and how strong is this relationship?

2.

What is the relationship between NSM and offline (risk)factors of violent extremism?

3.

What is the relationship between exposure to extremist NSM and individual extremist propensity?

In order to answer these questions, quantitative research based on a web survey in the general population has been conducted, followed by in-depth interviews with radicalised individuals.

1

The AIVD is the Dutch ‘general Intelligence and Information Office and falls under the Dutch ministry of Internal Affairs.
This project has been conducted on demand of the Federal Sience Policy (Belspo) and the Federal Ministry of Interior (FOD
BIZA). The project is embedded in the research line ‘radicalization’ within the Institute for Urban Security and Policy
Studies, University of Ghent. The institute has built some expertise on radicalization during the research project ‘Polarisering en radicalisering: een integrale preventieve aanpak’ by order of the Ministry of the Interior. During this research, as the first empirical research on radicalization in Belgium, a first model of radicalization was developed.
2

2

Concepts
Radicalisation is a highly contested and politicised container concept (Schmid, 2013). So far, there is no universally accepted definition of what exactly constitutes radicalisation. This may partially be explained by the fact that the perception of radicalisation is context-dependent and as such based on social, political and economic factors. Violent radicalisation primarily refers to the process of the development of violent extremist convictions and ideologies that challenge the status quo and denounce compromise (Borum, 2011, Schmid, 2013). The transfer to engagement in violent extremism or terrorism is not inevitable and may follow different pathways, radicalisation only being one of them.
Radicalism is a relative construct, and what is considered to be radical depends on who we are and when and where we come from. In the context of Western democratic societies, radicalism may best be defined in relation to mainstream political activities and Western (liberal, democratic and secular) core values (Segwick in Schmid,
2013). With this in mind, Schmidt describes radicalism based on attitudes and behaviour. Radicalism strives for a sweeping political change based on the conviction that the status quo is unacceptable, while, at the same time, proposing a radically different alternative. The means used to obtain this goals may be violent or non-violent, democratic or non-democratic in nature.
Violent extremism, on the other hand, is by definition violent and intolerant. Compared to radicalism, violent extremism takes it one step further and fully denounces pluralism, strongly emphasises (dogmatic) ideologies and uses violent and oppressive methods to achieve their political goals (Schmid, 2013). Defined in this way, violent extremism leaves no place for diversity or compromise. Violence is always accepted as a legitimate means to obtain and hold on to political power, which manifests itself either in violent attitudes or violent actions or both. In this research we focus primarily in the behavioural component of violent extremism: political violence. 2.

Theory
a.

Risk factors

Radicalisation may best be described as a process (Kundnani, 2012), but it is hard to say with certainty what phases this process comprises and what makes an individual progress from one phase to another. Most researchers do agree that the process of radicalisation is preceded by what may be described as a pre-phase
(Bjørgo, 2002; Koomen & Van der Pligt, 2009; Van der Pligt & Koomen, 2009; Van der Valk & Wagenaar,
2010). In this pre-phase, the seeds for further radicalisation are planted. Whether or not these seeds develop into a fully grown three depends on interactions of contributing risk factors. The literature provides a long and diverse list of such possible risk factors for radicalisation into violent extremism.
1.

Contextual factors
a.

Broad (global) contextual factors or broad long-term societal processes, like segregation, overpopulation, etc. Mostly, it constitutes structural political, social and economic processes beyond the scope of individuals or even individual states.

b.
2.

Local societal circumstances like unemployment, (political) inequality, etc.

Push factors at the individual level
a.

Personality traits like a need for kicks, impulsivity and sensationalism may make individuals more sensitive for certain experiences.

3

b.

Social psychological mechanisms determine the extent to which individuals are sensitive for certain societal and personal circumstances. The most important are perceived injustice, perceived group threat and perceived insecurity.

c.

Social mechanisms determine the social situation of the individual in relation to others in the same group. Individuals searching for belonging and social acceptance on the one hand and meaning and identity on the other hand are particularly susceptible for violent radicalisation.

d.

Emotions like frustration, hate, anger, revulsion and fear may have an influence on behaviour and action readiness.

3.

Pull factors
a.

Extremist groups try to fulfil certain fundamental, social and psychological needs of young people. They offer things that are wanted by adolescents, but that some among them have difficulties finding in regular society, like friendship, identity and security.

b.

Ideology is often given as a justification for certain actions, but is only rarely a central factor leading to violent extremism. In fact, initially, most individuals joining an extremist group are not ideologically or religiously schooled. On the other hand, the choice for a specific group does depend on ideological recognition.

4.

Catalysts
a.

Trigger events may have recruiting effects.

b.

Violence may be part of the what makes extremist groups attractive, but violence by third parties may also be a reason to join.

c.

Biographical availability.

d.

Significant others like friends and family often form a first link the extremist ideology and extremist group.

Individuals in this pre-phase usually experience feelings of frustration and discontent with certain aspects of their lives, society in general and/or the political policy. What radicalising individuals have in common is that they seem to be at a crossing point in their lives and are ‘searching’. This means especially socio-psychological mechanisms like insecurity, perceived injustice and group threat are of importance, in addition to the search for social inclusion, meaning and identity. Typically, these individuals meet (intentionally or unintentionally) other like-minded individuals, and go through the further process of radicalisation together that may eventually lead to violent extremism. Ideological arguments are often given as a post hoc justification for memberships of an extremist group, but the real explanatory factors are social in nature most of the time (Bjørgo, 2012). General violent extremist movements offer these individuals three things (Fermin, 2009):
(1) an answer to existential life questions and a clear identity,
(2) a political activist answer to injustice and
(3) a warm home and sense of belonging.
This corresponds to the three most important grounds for violent radicalisation. If these things are missing in the lives of individuals and they cannot be found in regular society, extremist groups become very appealing and attractive. (1) a need of meaning and significance,
(2) a reaction to (experienced) injustice and
(3) a need for social inclusion (Buijs et al, 2006).
Based on this, two preconditions for violent radicalisation may be distinguished: (1) Violent extremist viewpoints providing an answer to experienced grievances have to be present and accessible, (2) Social bonds,

4

networks and/or extremist groups that may pull individuals further into violent extremism have to be present and accessible. b.

NSM and violence

The general consensus, shared by political actors and policymakers, is that NSM facilitates the search for information and subsequently the immersion into online violent extremist milieus. This creates the dangerous possibility of becoming involved in violent extremist groups or movements, online and/or offline (Conway,
2012). Most concerns focus on the amount of online extremist impressions that young people are subjected to as being so extensive that they are not counterbalanced enough by real offline experiences (Klein, 2009a).
Academics, on the other hand, do not agree on the Internet playing a significant role in the process of violent radicalisation. Messages of hate, the recruitment of others in name of this message and the use of media may be found throughout history (Klein, 2009a). What is continuously changing are the context and receivers of this message as are the methods to reach them. In other words, it is not so strange that the Internet and NSM are also used by violent extremist groups.
Unfortunately, little to no research has been conducted into the relation between NSM and (violent) radicalisation. On the contrary, the relationship between violent media exposure and violent extremist behaviour has been thoroughly researched, providing first indications for the relationship between NSM and extremist behaviour. Concerns about violent and sexually explicit images as the root causes of all sorts of deviant behaviour, especially violence, among young people long predate the emergence of the Internet. Similar concerns have been expressed about the radio, cartoons, comic books and even newspapers. The idea is that exposure to violent media content leads to undermining social norms and values and the learning of deviant norms through cultural transmission, resulting in crime and violent behaviour (Newburn, 2007).
The debate about the effects of exposure to violent images on behaviour is highly polarised with, on the one hand, researchers convinced of a moderate (negative) effect of violent media content on behaviour (Anderson et al, 2010; Bushman, Rothstein, & Anderson, 2010), and on the other hand, more sceptical scholars (Ferguson,
2007; Ferguson & Kilburn, 2010).
The first group of researchers seems to be able to draw upon a large amount of scientific literature and empirical studies to reinforce its statements, pointing out both short-term and long-term effects (Anderson et al, 2010;
Anderson et al, 2003; Huesmann, 2007). However, when looked at more closely, most of these studies only find small and marginal effects, of which the direction is unclear and which, most of the time, cannot be reproduced
(Savage & Yancey, 2008; Sherry, 2001). It seems to be a question of a specific combination of specific personal factors and specific environmental factors that may lead to specific behaviour. If one factors changes, the outcome changes as well. Also concerning the formation of attitudes, exposure to violent media content appears to be only one of several environmental factors that, in combination with individual characteristics, may lead to aggressive and/or extremist attitudes (see Wood, Wong, & Chachere, 1991).
However, it should be kept in mind that NSM differs fundamentally from traditional media in one aspect. The content of traditional mass media is determined top-down and offered to the public in its final and fixed form
(ISD, 2012b). On the contrary, NSM is based on interaction and revolves around constantly changing usergenerated content (Thompson, 2011). In fact, criminological theories, or at least some key etiological variables in contemporary criminology, may be applied to the online world, as NSM is more interactive than any other medium, making real-life interaction possible in cyberspace. This importance of virtual communities has been recognised earlier (Pauwels, Weerman, Bernasco, & Volker, 2012; Soudijn & Monsma, 2012; Weerman, 1998).
This makes NSM, compared to traditional media, especially effective in providing the ‘preconditions for

5

radicalisation’, as we have discussed earlier. Through the Internet 1) radical narratives encouraging political violence are easily accessible at any time and place and 2) also the necessary social bonds and networks are easily accessible. So, the question is whether or not NSM forms a risk factor in the explanation of radicalisation into violent extremism.
3.

Risk factor approach: criticism

However, some serious shortcomings are linked to such a risk factor approach. Wikström (2004) states that it essentially is a non-theoretical approach that does nothing more than gathering and inventorying characteristics that are associated with delinquent behaviour in a non-random pattern. A factor is seen as a risk factor when it increases the chance of committing an act of crime. In reality, these are all correlates ‘predicting’ crime. This listing of correlates has become a goal in itself, causing the problem of causality in etiological criminology to be neglected. There are too much correlates, making it impossible to see the wood for the threes. A number of those risk factors may correctly be interpreted as causal but the majority cannot.
Wikström and Bouhana (2008; 2011) state that, if we really want to explain violent extremism, we have to evolve from a risk factor approach to a more explanatory method by conducting research into violent extremism by looking for explanatory mechanisms that link background characteristics to the real causal factors (see also
Wikström, 2004). A distinction has to be made between the direct causes or mechanisms that have a direct influence on violent extremism and indirect causes. In other words, we have to use our knowledge on how background characteristics (both genetics, biological psychological and social) influence extremist behaviour to assess their impact on the real direct causes (what individuals move to action). This allows us to look for those correlates that qualify as causes (or part of a cause) and to discover causal mechanisms.
What was and is needed, is a comprehensive theoretical framework capable of differentiating causal factors from the rest (Wikström, 2010). Although a lot of research has already been conducted into radicalism and violent extremism, the domain is still lacking integrated theoretical frameworks. The existing frameworks are fragmented and not integrated. According to Wikström, this may be solved by applying the situational action theory (SAT) on the explanation of violent extremism (Bouhana & Wikström, 2008). The SAT has been developed by Wikström (2014) as an answer to similar problems in the explanation of general offending in criminology. 4.

Towards an integrated model for the explanation of violent extremism
a.

Situational action theory

The Situational Action Theory (SAT) is based on insights and research from the social sciences in general and criminology more specific. The SAT offers a comprehensive and integrated approach for the study of crime as moral action and its causes. Violent extremism and political violence are, as are other forms of crime, equal to the intentionally breaking of these rules. By placing violent extremism under the same umbrella, it is possible to use the SAT as a framework to explain political violence. The basic arguments of the SAT are the following (Wikström, 2004,
2010, 2014; Wikström et al., 2012).
1) Acts of crime are moral actions, guided by what is right or wrong to do, or is simply not done, in certain situations, and should be explained as such. More specifically, crime may be seen as the breaking of moral rules as stated in the law.
2) Actions, including political violence, are ultimately the outcome of a perception choice process. It is necessary to 1) see political violence as a valid action alternative and 2) choose to carry out this action alternative over others.

6

3) This perception choice process is initiated and guided by relevant aspects of the person-environment interaction. In other words, the probability of political violence depends on the individual propensity for violent extremism and the interaction with exposure to extremist settings.
4) What kinds of people and what kinds of environments are present in a jurisdiction is the result of historical processes of personal and social emergence. Processes of social and self-selection place kinds of people in kinds of settings (creating particular kinds of interactions).
Wikström defines crime as ‘acts that breach moral rules as defined by law’. This is what all forms of crime, from theft to murder, have in common. Using this definition, the focus rests on breaking the moral rule (as defined in law) and not on the moral rule itself (Wikström, 2010). Defining crime in this way makes it possible to cradle different kinds of criminal behaviour, including violent extremism (political violence), in the same explanatory process. According to Bouhana and Wikström (2008), violent extremism and political violence may be defined in exact the same way, i.e. as an offence against the moral rules as defined by law. Although expressions of violent extremism differ, all forms of violent extremism are violent in nature, and violence refers to violations of the (penal) law. This has as the advantage that all expressions of violent extremism, be it the throwing of rocks during a manifestation, the taking of hostages or violent attacks, are of the same category. By doing so, conceptual discussions are circumvented and the problem of violent acts being perceived as violent extremism in some jurisdiction and/or time periods and not in others is solved. The focus lies on the explanation of the breaching of the moral rule. The validity of the moral and/or whether or not the breaking of a rule may be justified is not addressed.
In order to explain violent extremism, two basic principles of human behaviour must be kept in mind. First, humans have agency, meaning that they are able to react deliberately to their environment. Second, humans essentially are rule-guided actors. They base their actions on moral and social rules about what is the right or wrong thing to do in certain situations. Based on this, individuals perceive a number of action alternatives.
Which action alternatives are perceived depends on the personal moral rules of the individuals and the moral rules of the social setting where the action takes place. Out of the different perceived alternatives, one is chosen to carry out. This means that identified risk factors, both individual and environment factors, may only be accepted as causes of political violence if they directly influence 1) the perception of action alternatives in a particular situation and 2) the process of choice.
This perception choice process arises from the person-setting interaction (situation). Crucially, the SAT states that the decision choice process is shaped by a combination of characteristics of the individual (crime/extremist propensity) and the environment (exposure to criminogenic/extremist settings).
According to the SAT, propensity and exposure are causally relevant, meaning that changes in propensity and/or exposure will lead to changes in action, through their impact on the perception choice process (Wikström, 2014;
Wikström & Treiber, 2007).
Individuals vary in the likelihood of perceiving political violence as an action alternative (propensity) and settings vary in the extent to which their characteristics promote extremist behaviour (exposure). Propensity is determined by 1) the morality of a person (moral values and emotions) and 2) the capability of self-control.
Exposure is determined by 1) the moral rules of the setting and 2) the level of enforcement of these moral rules.
Propensity and exposure are considered to be the direct causes of violent extremism. Factors influencing propensity and exposure are seen as the causes of the causes (emergence) (Wikström, 2007). They are mostly factors of personal or social development, like education and social context. They explain 1) why individuals differ in extremist propensity, 2) why settings differ in extremist character and 3) how certain individuals end up in certain settings. This is shown in figure 1.

7

Figure 1: Model Situational Action Theory

b.

The integrated conceptual model for the explanation of violent extremism

Based on the SAT, we have developed an integrated model for the explanation of violent extremism (political violence) that makes a distinction between direct causes and causes of the causes, and pays attention to both the individual and the environmental level. Elements from key theories in the explanation of offending are integrated in the model as causes of the causes: general strain theory (Agnew, 2006), social control theory (Hirschi, 1969), social learning theory (Sutherland, 1947) and procedural justice theory (Tyler, 2006). The integrated approach argues that perceived strains leads to weakened societal bonds, which in turn affect personal beliefs and exposure to extremist settings. The study’s explicit starting point is that poor social integration, perceived procedural justice and perceived discrimination may positively affect moral support for violent extremism and therefore be of importance in the explanation of individual differences in the study of violent extremism and political violence. This model may be found in figure 2.
Since the main focus of this project lies on the influence of NSM media, our attention will mainly be directed towards exposure and its influence on political violence. More specifically, the focus lies on exposure to extremist content through NSM. Other forms of exposure may be equally important, like peers, but are not explicitly incorporated in this model.

Figure 2: Integrated model for the explanation of violent extremism

8

5.

Quantitative research

This study aims at gaining insight into the relationship between exposure to extremist content through NSM
(ENSM) and violent extremism among Belgian adolescents and young adults. In our definition of violent extremism, it was made clear that the concept encompasses both an attitudinal and behavioural component.
Following the logic of the SAT, we focus only on explaining this behavioural aspect of violent extremism: the use of political violence. The cognitive aspect of extremism (extremist attitudes) is incorporated in the model as a dependent variable (part of propensity). Data were collected (1) through a traditional paper and pencil survey of pupils in the third cycle (ages 16-18) of secondary education in Antwerp and Liege and (2) through a web survey of young adults – both students and young adults who have left school. The web survey could be accessed through the Facebook page of the project.
a.

Analysis

Both dependent variables are highly skewed. This means that most respondents group around the low values of these scales. After all, the majority of respondents have never committed an act of political violence. From an analytical point of view this poses problems for analysing the data, since regression analysis demands a normal distributed dependent variable. Therefore, these variables are dichotomised and binary logistic regression is used to gain insight into the independent effects of the available set of independent variables (Hosmer & Lemeshow,
2000). The emphasis is on the effects of different measures of exposure to violent extremism through new social media (ENSM) on self-reported political violence. A distinction is made between more active forms of exposure
(extremist contact through NSM, online communication with extremists via NSM) and more passive forms of exposure,(exposure to violent extremism through NSM and exposure to violent extremism through TM). Active exposure refers to the deliberate search for certain information and content. Passive exposure refers to accidently encountering certain content while doing other things online. Distinguishing between these types of exposure allows for studying the differential effect of ENSM. According to social learning theory/differential association theory, active forms of communications would have a stronger impact than passive exposure.
b.

Results
i. The effects of active and passive exposure through NSM controlled for other risk factors. First, we estimate the effect of a series of independent variables on the likelihood of having committed political violence versus not having committed political violence. Various logistic regression models are presented, showing the net effects of indicators of NSM exposure, independent of socio-demographic background variables and low social integration (model 1), variables indicating strain (model 2), personality traits (model 3) and measures of peer deviance (model 4). The results for model 5, both for political violence towards property and political violence towards persons are shown in table 1.

9

Table 1: Binary logistic regression analysis of political violence on ENSM controlling for other risk factors
Table 1: Binary logistic regression analysis on political violence towards property/persons
Political violence towards property
Model 5 Exp (B)

Political violence towards persons
Model 5 Exp (B)

Active extremist contact NSM (ref: no)
Yes
1.947***
1.306
Online communication extremism NSM (ref: no)
Yes
3.061***
1.911**
Exposure extremism NSM (ref: low)
Medium
1.109
1.455*
High
1.431***
2.102***
Exposure extremism TM (ref: low)
Medium
1.219
0.961
High
1.335
0.759
Native background (ref: other)
Belgian
0.965
0.611**
Gender (ref: women) men 1.476**
2.746***
Age (ref: -18)
19-22
1.348*
0.899
22+
1.930**
0.751
Attending religious service (ref: never)
1x year
1.030
1.123
2x year
1.119
1.067
1x month
1.243
1.139
1x a week
0.820
0.720
1.098
1.224*
Importance of religion
1.091
1.236**
Low social integration
0.892
0.918
Police procedural justice
0.915
1.121
Perceived personal discrimination
1.304**
1.031
Perceived group discrimination
0.953
0.846*
Overall police legitimacy
1.108
1.080
Support for violent extremism
0.980
1.073
Religious authoritarianism
1.264**
0.978
Thrill seeking behaviour
1.128
1.847***
Impulsivity
0.968
1.193**
Peer racism
1.477***
1.325***
Peer delinquency
PCP
91.2%
92.2%
-2ll
2054.676
1742.396
Nagelkerke R²
0.217
0.304
*** significant at 0.00 level ** significant at 0.01 level *significant at 0.05 level
Reference category = Low/0

In sum, the following should be remembered from these results. First of all, we clearly see that different measures of exposure to violent extremism through NSM are related to political violence. The relationship between these measures and political violence is strong and continuous to exist, even when controlling for measures derived from rival theoretical explanations. This is true for both active measures of exposure as well as passive measures of exposure. However, it appears that active measures of NSM exposure (especially online extremist communication) are more consistently related to political violence than the more passive modality of exposure. In other words, passive ENSM cannot be disregarded when explaining political violence, but the strongest and most influential effects may be expected from active ENSM, especially when actively and deliberately engaging in direct communication with other extremists using NSM. Second, offline associations with peers with racist attitudes and delinquent peers are also strongly related to political violence, meaning that we must not forget that harmful and influential exposure comes not only from NSM and the Internet. Other offline sources, like peers, are of extreme importance as well. Third, the offline world and individual experiences in it stay important in providing relevant strains, facilitating radicalisation into violent extremism. In other words, it is likely that the causes of the causes are still located in the real world. Finally, the individual and the presence of certain characteristics must be taken into account when assessing the influence of ENSM. ii. The effects of propensity and exposure through NSM controlled for causes of the causes Next, we grouped the relevant independent variables into a series of risk scales reflecting our integrated model.
(See figure 2). By performing a logistic regression for each dependent variable using these risk scales we were

10

able to test the integrated model for the explanation of political violence. Given our research focus, our main attention lies with the role and function of exposure and more specifically ENSM.
First, a logistic analysis was conducted with active exposure (total active NSM exposure) and and passive exposure (passive NSM exposure) as two separate independent variables. The results may be found in table 2.
What should be remembered is that the influence of total active NSM exposure is far more important in the explanation of political violence than the effect of passive NSM exposure. The odds ratio 3 for the highest measure of total active exposure is twice as substantial as the OR for the highest measure of passive exposure for political violence towards persons and even three times as substantial for political violence towards property. For example, individuals experiencing high levels of total active exposure have 6.29 times more chance of committing political violence towards property than individuals only experiencing low levels of total active exposure. For high levels of passive exposure, this is only 1.98 times more likely than individuals experiencing only low levels of passive exposure.
Table 2: Binary logistic regression analysis of political violence on active/passive ENSM and propensity controlling for causes of the causes
Tabel 2 : Risk scales, active versus passive exposure
Political violence towards persons Political violence towards property
Exp (B)
Exp (B)
1.361***
1.204**
Poor social integration
1.136*
1.108 NS
Perceived discrimination
1.334***
1.193
Subjective alienation
1.497***
1.343***
Propensity
Ref
Ref
Passive exposure low
1.462*
1.230 NS
Passive exposure medium
1.976***
1.832***
Passive exposure high
Ref
Ref
Total active exposure low
2.714***
Total active exposure medium 1.365*
4.146***
6.293***
Total active exposure high
*** significant at 0.00 level ** significant at 0.01 level *significant at 0.05 level
Reference category = Low/0

Second, the effect of overall NSM exposure (combining total active NSM exposure and passive NSM exposure) on political violence was assessed, again using a logistic analysis. The results may be found in table 3.What is noticeable here is that the effects of overall exposure remain more or less constant over different levels of overall exposure but rise exponentially for the highest measurement of overall exposure. Where the odds ratio for individuals experiencing high levels of overall exposure is only 2.6 for political violence towards persons, this becomes 7.52 for individuals experiencing very high levels of overall exposure. For political violence towards property, the odds ratios are even further apart with, 4.6 for individuals experiencing high levels of overall exposure and 13.16 for individuals experiencing very high levels of overall exposure. Furthermore, there are only very low rates for the lower and medium levels of overall exposure, the OR for low level of overall exposure not even being significant.

3

Odds ratios (Exp (B)) mirror the chance (or odss) that an individual scores positive on the dependent variable. For categorical independent variables they indicate the odds of political violence between a certain comparison category
(coded 1) of the independent variable and a reference category (coded 0). An individual in the comparison category has x times more likelihood of committing political violence then an individual in the reference category. For metric variables the odss ratio indicates the x rise in odds of political violence when the independent variable rises with 1 unit. An odds ratio >1 indicates positive effects (a rise in odds), an odds ratio 30 y.o.

Left-wing extremism

4

2

4

-

2

6

-

Right-wing extremism

6

-

4

1

1

4

2

Religious extremism

2

-

-

1

1

-

2

TOTAL

12

2

8

2

4

10

4

Left-wing extremism is understood as: Individuals who seek to abolish all forms of hierarchy, particularly the unequal distribution of wealth and power, and support the use of violent measures in order to pursue their goal.
They strive for a society in which everyone is provided equal economic and social opportunities, and no-one has excessive wealth or power over others (Woshinsky, 2008).
Right-wing extremism is understood as: Individuals with political positions or activities that accept or support social hierarchy and/or social inequality and support the use of violent measures in order to pursue their goal.

16

Those affiliated with the right-wing extremism consider social hierarchy and social inequality as inevitable, natural, normal, or desirable, typically justifying this position on the basis of natural law or tradition (Carlisle,
2005). Religious extremism is understood as individuals that promote, impede, or direct social, political or economic change within a set of religion-orientated revendications and support the use of violent measures in order to pursue their goals. However, it must be noted that, in this study, also individuals trying to advance religion-oriented ideas within the democratic system have been interviewed. So, two types of activism have been observed : 1) activism within the system, using it in order to enhance Muslim right recognition and equality of rights between Muslim and non-Muslim people (not extremist) and 2) activism outside of the system by struggling against it in order to impose a certain view of Islam (extremist).
“Leader” is understood as a person in charge of decision-making processes within the group. A “subaltern” is a person in charge of great responsibilities within the group, but not a part of the decision-making process.
“Executives” do not have many significant responsibilities, but take part in the group activities. Respondents who have claimed not being involved in any organised group have been qualified as “isolated”. iii. General perception
Given our main variable of diversification, we have made a synthesis of the particularities of each ideological trend according to the different categories, based on the analysis results (see further). An overview may be found in table 5.
Left-Wing Extremism (LWE)


Level of violence acceptance : exclusion and instrumentalisation,



indirect negative experiences towards a wide reference group (human beings) : feeling of inequality and unfairness, 

opponent: diffuse and difficult to identify: the "Great Capital",



family is usually engaged,



conventional confrontation with authority (police),



perceived social injustice based on information and speeches rather than personal experience,



indignation about the failure of the authorities regarding their tasks : to restore equality and reduce inequities between human beings,



strong social diversity.

Right-Wing Extremism (RWE)


Level of violence acceptance : instrumentalisation and resignation,



indirect and direct negative experiences towards a specific reference group : perceived discrimination from authorities towards non-natives,



opponent identified and identifiable: 1) non-natives, ‘who abuse social advantages, and 2) the authorities that ‘support’ this injustice,



family is less engaged,



conflictual confrontation with authority figures (police, justice, traditional parties, media),



perceived procedural injustice (compared to non-natives),



perceived discrimination towards the individual and the group (natives),



indignation about the failure of the authorities regarding their tasks: Protecting the rights and priveliges of natives or superior groups/individuals in society



little social diversity.

17

Religious Activism (focus on Muslim respondents) (RA)


Level of violence acceptation : instrumentalisation (thrill-seeking) and resignation,



direct and indirect negative experiences towards a specific reference group : perceived discrimination by authorities towards Muslims benefitting non-Muslims,



conflictual confrontation with authorities (police, justice, traditional parties),



strong basis of definitions based on social setting and community of belonging (Muslims),



perceived procedural injustice (compared to non-Muslims),



perceived discrimination towards the own community (Muslims),



opponent identified and identifiable: 1) Islamophobic persons ‘who discriminate Muslim people"’, 2) the authorities who ‘allow’ this discrimination and 3) opponents to a specific ‘Muslim cause’,



indignation about the failure of the authorities regarding their tasks: to restore equality and reduce inequities, 

either strong or weak social diversity.

Table 5: Features of ideological trends regarding the main categories of data
LWE

RA

Exclusion &

Instrumentalisation &

Instrumentalisation &

instrumentalisation

Violence acceptance

RWE

resignation

resignation

Indirect ; wide group

In/direct ; small group

In/direct ; small group

Diffuse

Identified & Identifiable

Identified & Identifiable

Conventional

Conflictual

Conflictual

Social

Procedural

Procedural

Indirect

Direct

Direct

Strong/neutral

Neutral/weak

Neutral/weak

Social diversity

Strong

Weak

Strong or weak

Reference group

Community > individual

Individual > community

Individual = community

Information / Action

Information / Action

Information / Action

Experiences
Adversary
Confrontation
(Perceived) injustice
(Perceived)
discrimination
Legitimacy of authorities Internet use

b.

Results

Following the SAT, factors which may have an influence on propensity towards violent extremism and exposure to extremist content are here understood as « the causes of the causes ». More specifically, we formerly identified them as 1) poor social integration, 2) perceived injustice 4 and 3) perceived discrimination.
i. Level of extremist acceptance
We observed a difference in the degree of violence acceptance. Three scenarios were presented in the interviews.
They differ in the presence of moral barriers and the perception of moral discomfort (see below). See table 6.

4

More correctly, it’s about subjective alienation of wich perceived injustice is an element.

18

-

Exclusion: Violence is not accepted under any circumstances. Violence is not considered as an action alternative because it contradicts the personal norms and values of the individual. The end does not justify the means. This especially happens when moral discomfort is little or not present.

-

Instrumentalisation: Violence is a way "as good as any" to achieve an end. There is no moral imperative that prohibits the use of violence. Several reasons may explain this. o Thrill-seeking: The need to show of force, the importance of the social impact and the excitement provided by the reprehensible actions take over.

o

The penalisation of actions is not a problem for the individual. Punishment is not a deterrent because the consequences are not seen as outweighing the benefits of not being penalised.

o

The individual has not benefited from learning moral rules according to which "the use of violence is not acceptable".

-

Resignation: Violence is the only alternative when all other non-violent means to achieve an end are exhausted. At first, violence is not a morally envisaged action alternative, but may be a last resort when non-violent alternatives have been depleted. These non-violent alternatives may be: a petition, demonstration, organised political action, involvement in a political party to defend ideas, etc. When these alternatives have been depleted, the individual may resign themselves to use violence, but this requires the weakening of moral barriers. Resignation is particularly interesting, because it reflects a trend that may be explained by the particular experiences of the individual. These experiences may influence the moral barriers of the individual and/or the experienced moral discomfort and have an influence on the degree of political violence acceptance.
Table 6: Moral barriers and moral discomfort in the three levels of violent extremist acceptance
Moral
barriers

Moral discomfort Exclusion





Instrumentalisation





Resignation





ii. Moral barriers
Moral barriers prohibiting the use of violence are obtained through an individual learning process. If such learning does not take place, or alternative violent, moral values are learnt, the moral imperative will not be able, by definition, to hinder the use of violence (instrumentalisation). If such a learning process does take place, the individual will not resort to violence (exclusion) or violence will only be perceived as a resource under certain conditions (resignation).
In case of resignation, we have seen that the moral imperative can be circumvented. Indeed, if the individual finds that the conventional means used to achieve a primary goal are not enough to achieve this goal, he finds himself in an uncomfortable tension between the impossibility of achieving his goal and needs. This is amplified by negative experiences perceived as discriminatory: that is to say, when the individual is the subject of a perceived unfair or illegitimate discrimination. This uncomfortable tension is understood as "moral discomfort". iii. Moral discomfort
Moral discomfort may be defined as the tension between the status quo or life situation as it is and fundamental needs or the life situation that the individual wishes to have. The goal of the individual will therefore be to

19

change the balance of power so that it is more favourable to him and his group. These basic needs differ from one individual to another and from one community to another, but usually refer to needs of recognition and appreciation. When this moral discomfort is too high, change of the status quo becomes a necessity for the individual or the community. It is precisely this need for change that may weaken the moral barriers and push individuals to resign for the use of political violence in order to achieve their goals.Moral discomfort is primarily shaped by the experiences of individuals, particularly perceived negative experiences.
It is the outline of these two dimensions (learning moral barriers against the use of violence and the imperious need to change the balance of power to be more favourable) that may lead to violent radicalisation. The more the balance of power will be perceived as negative by an individual or a group, the more the moral barriers of an individual or a group against the use of political violence will be undermined. iv. Direct and indirect experiences
In case of direct experiences (events experienced first-hand by the individual), the way these are perceived and defined by the individual will be decisive whether are not the event will contribute to a feeling of moral discomfort. For example, in case of a violent assault perpetrated by someone of foreign descent, whether or not the victim will make an association between ‘foreign descent’ and ‘violent aggression’ will have an impact on the perception of the event, the individual beliefs and subsequently on the definition of the status quo. If such an association is not made, beliefs will be affected differently.
Indirect experiences may also have, different, but well established influences on the development of beliefs, especially if the victim in question is a significant other of the interviewee (for example, parents or friends) or belongs to a wider reference group to which the individual relates to (for example the Islamic community).
Negative experiences by a broad, general, reference group (e.g. ‘males’, ‘young people’, ‘immigrants’) will be less likely to cause tension in the moral discomfort of an individual, because the perceived injustices or discrimination will be more diffuse.
It is important to keep in mind that the perception of the event is crucial. Whatever the true nature of the facts, it is the perception of the individual that matters and determines the impact of an experience. In other words, it is useless to try to influence the factual nature of an experience while it is the perception of an individual, with all subjectivity it entails, which will come into play.
The impact of those experiences on an individual’s beliefs contributes to a disillusioned and disenchanted worldview. The individual perceives his norms and values to be wrong all along. The tension between the status quo and basic needs (recognition and appreciation) may increase as these experiences destabilise the individual.
Subjective alienation
Perceived procedural justice
Several RWE and RE respondents referred to experiences with different forms of authority (police, justice, traditional political parties, media), during which they felt they had been treated unfairly. According to them, it is this lack of consideration that would have caused a shift to a greater level of violent extremism in their speech and / or their actions. Especially direct negative experiences with the police, justice and the traditional political parties where mentioned, like unfair treatment of them and/or their community and the use of double standards.
This perceived injustice is indeed a driving force of radicalisation.
LWE respondents mainly mentioned experiences with the police during demonstrations (see below Conventional confrontation). Such experiences, when negative, do have consequences. They question the police’s legitimacy

20

as their actions do not correspond to the expectations of individuals. However, those experiences have no direct impact on the degree of violent extremism and on the involvement of LWE respondents.
Perceived legitimacy of authorities
The legitimacy of the authorities is questioned throughout our interviews, mainly when a breach of "duty" is observed. Respondents had forged an opinion about the expected roles and tasks of the authorities before any encounters, and they question the legitimacy of these authorities if reality does not support these prior definitions. The respondents’ first reaction to this ‘disappointment’ is not to question their expectations but to question the legitimacy of the authorities. If, for example, an individual believes that the police "don’t do their jobs", he will first question the legitimacy of the police before questioning what he thinks police’s job is.
Expectations that individuals have of the authorities are of various nature: people expect the police to protect them, justice to protect their interests in a fair way, political parties to allow the exercise of democracy, etc.
When these, as legitimate perceived, expectations are not met, individuals face disillusionment and a loss of confidence. But more importantly, if the expectation was required to achieve a primary goal (to mitigate moral discomfort), the fact that this expectation is not met will force the individual to turn to other options to achieve its goal.

Perceived discrimination
Of all the experiences, those reflecting (perceived) discrimination, both of the group and personal, have the greatest impact on the level of violent extremism, by increasing feelings of moral discomfort. In addition to the perceived discrimination, respondents also have a need for recognition and consideration of their beliefs
(especially RWE and RA). Disrespect for extremist beliefs defended by individuals contributes to the marginalisation of this group of individuals and may lead to the further violent radicalisation of the group.
v. Exposure
Exposure to extremist content takes place on multiple moments: during the formation of beliefs, during the engagement (making contact), and throughout the further process of violent radicalisation..
a)

During the formation of beliefs, searching for information is one way to enhance, refine by comparison and validate beliefs in a dynamic process.

b) During the period of involvement, information is practical and may help an individual to get in touch with an organised movement (e.g., to learn about the beliefs defended by an organised movement and / or look for the coordinates of an organised movement).
c)

During the period of involvement in a movement, the individual may continue to learn, based on the framework provided by the extremist movement and then producing an enhancing effect nourished back and forth between beliefs and information.

We have observed in the interviews that the search for extremist information is initiated only if the individual is already in a process of forming (radical) beliefs.
The Internet is currently the most popular alternative for mainstream media to search for alternative information.
This may be explained, according to the interviewees, by a need for critical and correct information, and mistrust developed against traditional media and mainstream information. Our data show that even when faced with uncertainty about the reliability of information on the Internet, it seems nevertheless preferable for interviewees to use this information to compare and criticise the "official" information.

21

Active exposure
Online communication
The importance of active exposure to radical content is supported by our data. We have seen that respondents deliberately sought out extremist online, and less frequently offline, content. This active exposure is facilitated by the quick, easy and anonymous access to online content on radical websites, forums, Facebook pages, etc.
Two types of online communication related to political activity may be distinguished. Both cases were observed among respondents in the three ideological trends.


The individual seeks information but does not want to chat with other people.



The individual seeks information and wants to exchange ideas with other people.
Searching for extremist contacts

Regarding the search for extremis contacts, some respondents have actively searched to communicate online with others who share the same ideas. When the goal is to join an organised group, the Internet and NSM are useful to find an organised group, regardless whether the person has access to a strong offline like-minded environment or not. The use of the Internet and NSM for this purpose is quick and easy, but the Internet and
NSM are not the only possibilities. Other respondents have said they are in contact with violent extremists by means other than the Internet (meetings, events, by word of mouth, etc.). In the first stage of making contact with a group, individuals have no preference between the use of online or offline access. However, when it comes to joining and actively being involved in a group, this requires an offline meeting. The respondents said that the offline meeting is a guarantee of trust and good faith that the Internet does not provide.
Self-selection
In the search for information (online or offline), individuals select information of interest for themselves. This does not mean that the selected information will defend their point of views. Some respondents claim to have looked for information on ideological trends other than their own. That being said, a self-selection (or targeting) from the individual will inevitably narrow down the fields of information accessed. This concerns both LWE,
RWE as RE respondents.
We did not observe a single case in which the respondent made a dramatic volte-face of his opinions after any
"counter-speech" information. This, together with the process of self-selection as a whole, leads us to think that the emergence of radical beliefs is a prerequisite for a more advanced information search and commitment in a group (which does not subsequently prevent that beliefs evolve within this group). This self-selection creates a dynamic of considerable restriction both within and outside of an organised group.
Passive exposure
Some respondents have said that they had been accidently in contact with people defending extremist ideas at events or meetings. These moments are occasions of passive exposure to violent extremist speech, possibly more extremist in nature than what the individual had come for. The individual will have opportunities to compare different settings and possibly choose a setting that better satisfies the need to mitigate moral discomfort. None of the respondents has testified that he was passively subjected to radical speeches without any voluntary initiative prior to this situation. We cannot support the hypothesis that a non-predisposed person exposed to online extremist content would be influenced by this content. For the extremist content to have an impact on an individual, this content must be supported by the propensity of the individual.

22

We have not distinguished important differences between the three ideological trends regarding the use of
Internet and NSM differences. These are used for the purposes explained above, regardless of ideological orientation. In addition, given the interviews, we cannot support the hypothesis that the Internet and NSM would be a triggering element of violent extremism (at most a facilitator, as a privileged means of communication and exchange of information). vi. Social settings
Social integration
The data tell us that social anchorage does not influence the likelihood of an individual to engage in an extremist movement. However, social isolation of an individual may be an incentive to join a movement. This obviously requires that the convictions of the one and the discourse of the other have a common base (ideological appeal).
We have observed people with both strong and weak social ties in each of our group’s diversification. We cannot support the hypothesis that low social integration leads to radicalisation or that a strong social integration prevents radicalisation.

Social diversity
That being said, social integration should not be understood here too restrictively. During the interviews it became clear that also well integrated individuals can become involved in violent extremism but it should be noted that this theoretical variables should be nuanced by another dimension of "social diversity". What is of importance, is the homogeneity of the setting in which the individual develops strong social ties. In fact, although an individual may claim to be socially integrated, this does not specify whether he is integrated in a highly homogeneous community or not. And a strong homogeneity of moral discourse

within a social

environment may affect the learning of moral barriers and have very homogeneous beliefs as a consequence.
c.

Conclusion

In summary, we have seen that experienced moral discomfort is the central element on which the process of violent radicalisation is based. This moral discomfort may collide with the moral barriers of the individual, creating an excess of unfavourable definitions towards respect of moral rules in relation to extremist definitions, which could lead to political violence.
Moral discomfort occurs when a person experiences too great a difference between what he considers to be a
"satisfactory balance", between his basic needs and the life he actually lives and / or the perception of it. Each individual defines for himself what his basic needs are. They may be of personal order (need for recognition, promotion, respect) and / or of group order (need for recognition, promotion, respect). Fundamental needs might for example be , social and professional integration, recognition of himself and / or community, respect for individual freedoms, fair and equal treatment, etc. Non-compliance with these basic needs crystallises in experiences of perceived discrimination and perceived injustice that amplify the moral discomfort perceived by the individual and undermine the perceived legitimacy of the autorities, especially in the case of RWE and RE respondents. Discrimination or injustices perpetrated against an individual by an authority will be experienced more strongly by that individual when he has a high esteem of these forms of authority (cf legitimacy of authorities). The main difference we are able to observe between the three ideological trends is this moral discomfort that does not seem very present among the LWE respondents, unlike for the RWE and RE ones. That being said, we see individuals with low moral discomfort and others with a high moral discomfort within the

23

three trends. This means that the elements that could lead to a process of radicalisation are not specific to an ideological trend but rely on broader structural causes (causes of causes).
Although the Internet and NSM facilitate the exchange of information and making contact, the Internet and social media are not the starting point of the violent radicalisation process. They form a facilitator which will be involved when the radicalisation process has already begun. In addition, its use, although easy and fast, has its limits in terms of confidentiality and reliability of exchanges.
The perception of authorities as legitimate is crucial. Discriminant experiences (direct or indirect) perpetrated by the authorities towards an individual and / or a community have a very important impact on the perception of individuals. It is not the direct opponent that will have an impact on the degree of violent extremism, but the feeling that the authorities maintain a situation of injustice between the individual and his adversary. In other words, "facing an adversary is not a problematic situation, what is problematic and frustrating is that the authorities do not allow a fair and equitable fight against this adversary". Violent extremism is then a response to the failure of the authorities: "Our group provides the means to restore a more equitable balance of force into this struggle." Finally, we have seen that the social environment of radicalised individuals may help define the moral rules of an individual. They will leave this environment if their fundamental needs require the adoption of another set of moral rules. More than social integration, it seems more effective to focus on meeting those basic needs, rather than trying to influence the social environment of an individual. In this context, a counter narrative approach would be relevant only if they aim to provide an alternative to meet those needs and not just an alternative deconstructing discourse.
7.

General conclusion and recommendations

General conclusion
Early on in our research it was clear that violent extremists are not mentally insane or delusional when committing acts of political violence. On the contrary, during the interviews it became clear that political violence was perceived and used as an efficient (and often last) resource to achieve a, in their eyes, reasonable and necessary goal. By acting and reacting to a situation that is perceived as highly problematic, extremists are able to resolve a tension between the current (problematic) political and social situation and what they perceive is needed to fulfil some essential and fundamental needs (moral discomfort). The more urgent and prominent this tension is, the more likely these individuals will resort to political violence, as an immediate and effective means to an end. This implies that violent extremism and political violence may be better understood as a part of a deliberate strategy to reach a goal that has strong moral significane for the individual (Braeckman, 2006). This idea is backed up by Wikström and his Situational Action Theory (Bouhana and Wikström, 2008), which clearly states that actions, including political violence, are ultimately the outcome of a perception-choice process, in which an individual perceives certain action alternatives (based on the setting he or she is in and his/her own propensity) and then chooses, deliberately or out of habit, one of those alternatives, taking both moral rules and social controls into account.
It also became clear that there are numerous motives and reasons that supposedly tell us why a certain act of political violence is committed and that, at the same time, are used to justify and legitimise these acts for those that commit them. These motives are mostly ideological or religious in nature. For example, political violence may be committed to install the Sharia in Europe, to establish a white nationalist state, to overthrow the capitalist elite, to demand equal rights, to liberate test animals, to save unborn children, to get rid of immigrants, to safeguard the own culture, etc. Such justifications are most of the time only provided after the facts and as such

24

can not explain the initial reasons for resorting to violence. Still, motives like these are often central in media reports on acts of violent extremism, especially in the case of religious (Muslim) extremism, contributing to an image of irrational and mentally ill perpetrators. Even some scholars and terrorism and violent extremism experts highlight ideology and religion in their assessment of political violence and explanation of why some individuals radicalise into violent extremism. This is problematic because motives are not and cannot be causes of political violence. Motives explain why it is important to reach a certain goal, but they do not explain why a specific action out of several possibilities is chosen to achieve this goal. They are necessary but insufficient conditions to explain action (Davidson, 1963). What is needed for extremist motives to become causes is the contributing role of the interaction between situational exposure and extremist propensity. The literature review has already indicated that ideology is only fully learnt and incorporated after joining an extremist group and therefore is not a primary factor leading to radicalisation into violent extremism.
Instead of focusing on motivations and ideology, attention should go to structural issues, group processes and perceived individual strains as the breeding ground of violent radicalisation. The real (compelling) reasons
(Davidson, 1963) to join extremist groups are often social in nature and based upon feelings of indignation and being lost (Bjørgo, 2002). Individuals searching for social inclusion, searching for identity/meaning and experiencing injustice are particularly susceptible for violent extremism. Extremist groups often enhance, cultivate and even install these grievances by using a polarising discourse and providing in simple, logic and hands-on answers. More specifically, extremist groups answer to the needs of individuals by offering them 1) a strong sense of identity, 2) a political activist answer to injustice and 3) a warm home and sense of belonging
(Fermin, 2009). If these things may not be found in the rest of society (or society fails to offer them), extremist groups may become very attractive to the individual. The fact that this, initially, has nothing to do with ideology is also indicated by research that shows that in areas where individuals have easy access to organised crime
(gangs), offering them the same things, violent extremism is absent (Roy, 2008).
A huge problem in the research on violent extremism and political violence is the excess of risk factors, and confusion about which of these factors can be seen as causal and which ones are mere correlates. To address this problem, an integrated, theory-based model for the explanation of violent extremism (political violence) has been developed based on Wikström’s (2014) situational action theory, by incorporating key elements from theories explaining (youth) offending in general as causes of the causes. This was possible because, like general acts of crime, political violence can be defined as ‘the breaching of moral rules as defined by law (Bouhana &
Wikström, 2008). This definition circumvents conceptual discussions and makes it possible to reach a general explanation of political violence by applying the logic of the SAT to the explanation of political violence. Our analysis has confirmed the applicability of this model on the explanation of violent extremism (political violence). This means that
1) a difference has to be made between the direct causes of political violence and the causes of the causes
(emergence) and
2) individual extremist propensity and exposure to extremist settings form the direct causes of political violence in interaction with each other.
Given the focus of our study, we mainly studied exposure to extremist content through NSM. It proved to be fundamental to make a distinction between active exposure or deliberately searching for certain information and communication, and passive exposure, or encountering certain content by accident while doing other things online. The results of the qualitative interviews have clearly shown that extremists do not simply refer to NSM as causes. NSM is mainly used as a useful resource or instrumental tool to online pursue offline developed interests by looking for information and communication with like-minded individuals (active exposure). NSM is further

25

used to keep up to date with the movement and to organise networks. It is very unlikely that an individual will radicalise only by using NSM (passive exposure). Furthermore, our results have clearly shown that the ‘danger of exposure to extremism’ can not only be found online, but also that offline exposure to violent extremism should be taken into account. Exposure to delinquent and or extremist peers can particularly influence the radicalisation process.
On the other hand, our analysis did show that the active use of NSM for extremist goals does have a major impact on the likelihood of committing political violence. Individuals deliberately seeking out other extremist groups and individuals for communication are indeed at higher risk of violent extremism and, in the end, using political violence. This effect is amplified by propensity. It was very clear that those adolescents having a high propensity towards violent extremism had most of all been influenced by ENSM. The effects, and subsequently, danger of exposure, even active exposure, for individuals with only low propensities towards violent extremism are negligible. This means that the danger lies not with NSM themselves, but in adolescents wanting to visit certain sites and wanting to communicate with extremists. In other words, in the prevention of violent radicalisation, it is crucial to prevent adolescents to become attracted to violent extremism. In order to do this, attention should be directed towards the structural causes of violent extremism or, in other words, those factors determining exposure and propensity, the causes of the causes.
According to the SAT, some risk factors identified by research and literature are, in fact, (elements of) causes of the causes. In this regard, our analysis has identified several elements of perceived strain/injustice that contribute to the breeding ground for violent extremism. This indicates that more attention should be directed towards elements of perceived injustice and strain as structural root causes in the explanation of violent extremism and political violence and not to relative deprivation and poverty. In studies addressing general offending, this has already been recognised. So far, the latter has received a lot of attention without much result. Researchers trying to discover socio-demographic profiles of extremists have only found mixed and contradicting results. Some individuals come from a wealthy background, while others are poor. Some are highly educated, while others are illiterate, etc. Focusing on perceived injustice can put these results in perspective. Feelings of injustice can be present among all social groups and layers of society and hold strong explanatory power when addressing political violence. When developing prevention policies, this should be kept in mind and efforts should be made to address the (structural) problems that are at the base of these perceptions.
Elements of perceived injustice can be found in different aspect of individuals’ lives. Potential elements of perceived injustice can be found in their own personal situations (e.g. family situation, career opportunities, etc.), regarding general social, political and economic aspects of life (e.g. discrimination on the house or job market, being refused a loan at the bank, etc.) and when coming in contact with the authorities (e.g. feeling singled out during police controls, etc.). More specifically, our research has identified poor social integration, elements of perceived discrimination and elements of perceived procedural justice as elements of injustice that form the causes of the causes, influencing extremist propensity and exposure. Our results have clearly shown that 1) elements of poor social integration, 2) perceived group discrimination and perceived personal discrimination and
3) the perception of the authorities as being unjust and subsequently illegitimate greatly contribute to political violence. Zooming in on one of these elements, perceived procedural justice, the results have clearly shown the importance of trust in the authorities for the prevention of violent extremism and, subsequently, the need to restore this trust among the populations at risk5. All respondents have clearly indicated having negative
5

Previous research has already shown that minorities attach the same importance to procedural justice and use the same criteria for its assessment as majority groups do (Tyler, 1994, 2001). Both in the UK and the US (Huq, Tyler, & Schulhofer,

26

perceptions of the authorities and the police and not perceiving them as legitimate. Furthermore, measures of perceived procedural justice are clearly related to political violence, even controlling for other risk factors.
Perceived (procedural) injustice, as part of subjective alienation, has been found to be a ground cause of violent extremism, influencing individual violent extremist propensity and violent extremist exposure. This is problematic, not only because perceived injustice forms a ground for violent extremism, but also because cooperation of populations at risk is necessary to identify violent extremist groups and preventing radicalisation.
This means that strict enforcement of law without any attention for cultural and local sensitivities and/or harsh actions by the police may, in fact, increase the risk of violent extremism, because it creates cynicism regarding the law. The same is true for policies and specific measures focusing on a particular group (e.g. a ban on headscarves instead of a ban of all religious symbols). This undermines the willingness to cooperate with the police and to participate in society and civic actions. Consequently, policy should pay attention to this problem and work on enhancing general perceptions of trust and legitimacy of the authorities and more specifically address (perceptions of) unjust and biased police actions.

General remarks
Based on the results of both the quantitative and qualitative research and the conclusions linked to them, some recommendations can be made regarding the prevention of violent radicalisation among adolescents and more specifically the prevention of violent extremism. Before lining out our recommendations, it is important to stress four remarks.
(1) It has to be kept in mind what one wants to achieve. There is a difference between measures aiming at the prevention of violent extremism and political violence and strategies aiming at deradicalisation. They each address a different population and they each ask for a specific approach. Our study only allows us to make recommendations regarding the prevention of radicalisation. In order to do so for deradicalisation, further research is needed.
(2) The position of the government has to be taken into account. This is challenging because two different and conflicting signals have to be given. First, it must be clear that violent extremism will not be tolerated. Second, it is crucial not to push movements and individuals underground. With this in mind, policymakers have to be careful not to develop strategies that contribute to the problem. In this regard, it is problematic that, so far, only a few projects are scientifically evaluated. However, research into projects aimed at reducing youth violence in general show that only those projects starting with prevention from a very early stage are effective on the long term (see also Christmann, 2012).
(3) The main concern of practitioners and field workers is often how to recognise individuals at risk in order to focus prevention. We like to stress that general structural problems are at the base of violent radicalisation and violent extremism. This can be seen as ‘bad news’, because it makes it hard to aim prevention strategies specifically at violent extremism. On the other hand, it also means that we can learn from what is already known from the prevention of youth offending general resilience and youth programs. Although it is difficult and not always possible to solve structural problems, wherever possible, they have to be addressed in order to prevent violent extremism on a long term. Structural problems may influence the perceptions of individuals (exposure), possibly guiding them towards certain violent extremist settings (propensity) and in the end political violence.
2011; Tyler, Schulhofer, & Huq, 2010), empirical research has shown that the perceived fairness of the criminal justice system, and especially the police, police intervention and manners highly predict legitimacy and cooperation. However, research has also shown that minorities have the most negative image of the authorities as unjust and unfair. They feel singled out by authorities based on prejudice and stereotypes (Tyler & Wakslak, 2004). This has been confirmed by our study. 27

Whether or not these perceptions are one hundred percent objectively justified, they are at least partially based on problems of inequality and discrimination (causes of the causes). On the short term, young people can be made more resilient to violent extremism by training in how to deal with problems and difficult situations.
Resilience-training addresses first of all propensity as the direct cause of violent extremism by learning youth how to control their impulses and develop positive non-violent attitudes. Second, this will also reduce the need for (active) violent extremist exposure. These kinds of strategies can be usefully for a broad range of groups and do not need to be aimed specifically at young people at risk of radicalisation. By addressing general groups and offering positive skills, individuals at risk are also reached without stigmatising while benefitting the young in general. (4) In order to put our recommendations into practice, it should be clear who should take on which responsibility. However, although theoretically relatively easy to answer, in practice this is a difficult task, especially since we are dealing with federal, regional and local levels of government. Inevitably, there will be overlapping action domains and shared responsibilities. The main issue is coordination. When not accurately coordinated, this could cause certain issues to be overlooked or ignored, while other issues will be addressed by several different actors, leading to confusion. Based on these concerns and on the need for information present in the field, we repeat our suggestion for the formation of a knowledge and information centre on violent radicalisation and violent extremism, grouping relevant actors and knowledge and supported by research (see
Noppe et al, 2011). Different actors are concerned about radicalisation into violent extremism on both the short and the long term. Different actors have already developed some initiatives to tackle this radicalisation. What is missing, is a central and neutral partner that is able to bring existing knowledge together, advise in the development of an integrated prevention policy, coordinate responsibilities, give concrete information on how to recognise and tackle vioent radicalisation and support and evaluate existing initiatives. All of our proposed recommendations could be organised by or would benefit from such a centre.
Specific recommendations
Following the recommendations made by Noppe et al (2011), we maintain that it is important to develop a prevention policy based on two axes. First, existing social policy and initiatives should be reinforced, especially in the domains of education, employment and integration. Our results have shown that poor social integration and perceived discrimination and perceived (procedural) injustice are important elements of the breeding ground for violent extremism. It is important that structural grounds for these strains are tackled. Second, regional and local work packages have to be developed. Violent extremism is often tied to personal experiences and local circumstances and can therefore best be recognised and addressed on this local level. Our recommendations refer to these local initiatives.
First, for the prevention of online radicalisation into violent extremism, the virtual lifestyle of adolescents must be taken into account. The kinds of websites adolescents visit differ in frequency, nature and persuasiveness of violent extremist messages. As already explained, there are different virtual settings giving access to violent extremist content, but not all of them pose the same danger for violent radicalisation. Although violent extremists might be present on passive mainstream virtual settings 6, these settings are not violent extremist in nature causing possible violent extremist comments and narratives to be countered by other users.
This can result in discussions on different social and political issues and how to interpret them. In other words,

6

Mainstream passive virtual settings refer to websites and fora that are not extremist in nature but that do aim at a target audience that also can be attractive for extremist groups. So extremist will be present on these websites, trying to recruit new members. 28

passive virtual settings can give rise to a bottom-up and pluralistic discourse and offer alternatives to the violent extremist narratives. Young people in the pre-phase of radicalisation might visit these websites in search for answers, before ending up in more active violent extremist virtual settings. The alternatives they encounter here might put the violent extremist discourse in another perspective and keep them from further radicalising.
Moreover, our results indicate that for some adolescents being able to ventilate on these websites fulfils an important social function, keeping them from further radicalising into violent extremism. If this possibility would fall away, it is likely that these adolescents will end up in active violent extremist virtual settings being at risk of further radicalisation into violent extremism. Finally, the interactions on these websites and forums reflect interactions in real life. Therefore, censuring and controlling them would be a waste of time and resources. As has already been noted, peers and offline social contacts remain crucial in the process of radicalisation.
For those reasons we recommend supporting these kinds of websites and forums instead of censuring them and closing them down. This can be done by helping these websites to ensure their continuation (financially or logistically) and/or make them more effective in 1) recognising and countering violent extremist propaganda and
2) recognising and responding to individuals at risk of violent extremism. Most of these websites do not welcome violent extremist content but do not know how to react. So a possibility would be to offer support and training to existing moderators of these websites. Furthermore, this does not exclude, and even provides further opportunities for the monitoring of these websites. The presence of violent extremists and their activities and networks can provide valuable information for police and security services.
Concerning active violent extremist settings 7 or genuine violent extremist websites and forums, the only option might be to close them down. However, it is questionable whether the closing down of these websites would be very effective, since they most likely will reappear under another name or format. Still, consistently closing them down could form a clears signal that such content is not acceptable and will not be tolerated. However, it is unlikely that young people will repeatedly end up on these websites by accident, implying a more dangerous active exposure. Furthermore, adolescents visiting these websites are expected to already have a certain propensity towards violent extremism. Therefore, prevention strategies targeting these websites are probably not effective, since their audiences have already bypassed the pre-phase of radicalisation and have entered the process of violent radicalisation. This means deradicalisation would be needed, rather than prevention. Previous studies have shown that deradicalisation strategies are only effective when they are personal and face-to-face, so it is unlikely that the necessary bonds of trust can be established through these websites. As already been mentioned, further research on deradicalisation is needed, both in terms of understanding the processes of deradicalisation in general and in order to determine the role of the Internet and NSM in this process. NSM can provide unique opportunities to reach violent extremists and facilitate deradicalisation, but, on the other hand, elements of online influence and online social ties may prevent deradicalisation. Further research is needed to address this question.
Second, there has been a lot of talk about the use of counternarratives for the prevention of online radicalisation (see also Leuprecht, Hataley, Moskalenko, & McCauley, 2010a, 2010b). We think counternarratives can only be useful in a passive virtual setting. In these settings, individuals with a low or medium propensity to violent extremism can be found that are still looking for ‘the right answer’ on the questions they are struggling with. It could be useful to offer clear and positive messages that counter the narrative of the violent extremist groups. However, one has to be careful not to elapse into an ideological battle for the one and only right answer. The most effective would be to freely offer some strong and positive arguments by neutral actors (Weilnböck, 2005). Often, this has already been done by the other non-violent
7

Active extremist virtual settings are websites and forums that are extremist in nature and intention.

29

extremist members of these websites and forums. For active virtual settings, it is unlikely that counternarratives would have any effect. Individuals active in these settings have already found their (violent extremist) answer and will not be open for counterarguments. It is more likely that counternarrative attempts will be interpreted as confirming the violent extremist narrative of a deceiving and untrustworthy government. As has already been said, what is needed here is not prevention, but deradicalisation.
Third, actions should be taken to restore the trust in the authorities. This is not an easy task. What is expected from the authorities differs for different groups and these visions can contradict each other. For example, what the political right-orientated population expects form the government on the subject of immigration is probably not the same as what people of foreign descent expect. Furthermore, the objective situation and treatment by the authorities might not respond to the perception among the population or the individuals involved. We believe that the most effective way to rebuilt trust is by sensibilising public servants and especially the police.
Therefore, we recommend that they are made aware of the existence of perception of distrust and injustice, how their actions contribute to these perceptions and how best to react in certain situations. Research has already shown that it is not so much the outcome of interactions with the police that determines whether are not they are trusted and seen as legitimate. On the contrary, it has repeatedly been proven that the perception of being treated fairly and respectfully by neutral police during encounters more strongly contributes to the perception of the police as being legitimate, even in case of a negative personal outcome (Tyler, 1994, 1997, 2000, 2001; Tyler &
Huo, 2002; Tyler & Lind, 1992). With this in mind, it is also important that police are trained in how to interact with different populations and learn to recognise possible sensitive situations and how best to handle these. It is important for police to stay neutral in their daily encounters with populations at risk and not to let their actions be guided by political sentiments (Deflem, 2004a). We expect this to be equally true for other public servants.
We do repeat that perceptions of injustice are not only a consequence of encounters with the authorities, but can also be a consequence of the personal situation (poor social integration) and more general social, economic and political situations (perceived discrimination). Our focus on strengthening trust in the authorities does not mean that other elements leading to feelings of injustice should be disregarded.
Fourth, it is crucial to prevent young people from joining violent extremist groups. Therefore they have to be made more resilient to violent extremist thoughts and hate discourses and made more adept in coping with difficult situations. As part of this, it is necessary to educate the young on what exactly constitutes a hate discourse, how it can be recognised and what alternatives there are. Minors in particular may not always have the necessary social knowledge on politics and social issues to accurately judge the content of certain propaganda. In other words, more attention should go to critical reflection (Braeckman, 2006). Second, young people have to be educated into confident and intellectual autonomous individuals with a strong, democratic consciousness. There should be more focus on the acquisition of values related to respectful and peaceful coexistence. This can be started from a young age by showing them how to work together, solve problems together, communicate, reach compromises and handle differences in the daily life. Third, knowledge and understanding of our political system and how society works should be increased. Often, young people have no idea of how the political decision-making process works, what the historical roots are of certain trends or laws in our society (e.g. migration, division state and religion, etc.) or how to understand different cultural political and social attitudes in society. This can also help the young to recognise injustice and teach them how to react and change their situation. Fourth, there should be more focus on Internet literacy. It is often difficult to distinguish legitimate from false information on the Internet, especially in case of well-designed websites. This may be an important task for schools, youth workers and other youth organisations. It is important for young people to feel

30

included and valued in society. Once again, it has to be kept in mind that violent extremist exposure does not only occur online, but may also, and sometimes mainly, occur offline through peers and other offline social ties.
Following the two previous remarks, it is important that to note that international research (PIDOP project) has shown that the place where the young feel treated unfairly the most is the school environment. (This can be by teachers, but also by other personnel, other students, perceived opportunities, etc.) This confirms the importance of the two previous recommendations and the important role schools play. After all, young people spent 35 hours a week in the school environment. In case of increasing resilience towards violent extremism, it is also important to start early on because 1) violent extremist groups target individuals as young as 13 or 14 years old and 2) individuals at risk are more likely to drop out of school after a certain age. Furthermore, teachers, but also other professionals in the field, like youth workers, social workers, etc. often do not know how to recognise violent radicalisation, when it becomes a problem and how to handle it. However, because of their close contact with young people they are extremely valuable for prevention. Therefore, specific training improving their skills regarding radicalisation and violent extremism could be very useful.
Finally, when developing a prevention strategy, it should be kept in mind that friends and family have to be made aware that they are essential for prevention and deradicalisation. Social support and contact are crucial in keeping young people from radicalising into violent extremism. Radicalising persons are in need of extra support from their environment, especially when they want to leave. Lack of social ties and having no place to go often inhibits youth to return on their steps (Bjorgo, 2002). In this regard, families have to be given the right information and tools to handle situations of radicalisation and to facilitate prevention and deradicalisation. In that regard, it is also important to keep communicating and talking with both individuals at risk and violent extremists. Open communication can be an important tool in preventing violence. It has to be stressed that this does not mean that family and friends can be to blame for individuals becoming violent extremists. Recruits are often deliberately alienated from their families and friends by violent extremist groups. When this happens, the influence of friends and family becomes limited, but they remain important as a safety network when youngsters do want to abandon violent extremism.

31

8.

References

Agnew, R. (2006). Pressured into crime: An overview of general strain theory: Oxford University Press, USA.
AIVD. (2006). De gewelddadige Jihad in Nederland. Den Haag: Algemene Inlichtingen-en Veiligheidsdienst.
Anderson, A., Shibuya, A., Ihori, N., Swing, E., L., Bushman, B., J., Sakamoto, A., Saleem, M. (2010). Violent video game effects on aggression, empathy, and prosocial behavior in Eastern and Western countries: A meta-analytic review. Psychological Bulletin, 136(2), 151-173.
Anderson, C. A., Berkowitz, L., Donnerstein, E., Huesmann, L. R., Johnson, J. D., Linz, D., Wartella, E. (2003).
The influence of media violence on youth. Psychological science in the public interest, 4(3), 81-110.
Benschop, A. (2006). Virtuele Jihad en de cultuur van de grote bekken. In S. Harchaoui (Ed.), Hedendaags radicalisme. Verklaringen en aanpak. (pp. 140-179). Apeldoorn: Het Spinhuis.
Bjørgo, T. (2002). Exit Neo-Nazism. Reducing recruitment and promoting disengagement from racist groups.
NUPI working paper 627. Oslo: Norwegian Institute of International Affairs.
Borum, R. (2011). Radicalization into violent extremism I: A review of social science theories. Journal of
Strategic Security, 4(4), 7-36.
Bouhana, N., & Wikström, P. H. (2008). Theorizing terrorism: Terrorism as moral action. A scoping study.
London: University College London.
Braeckman, J. (2006). Nucleair terrorisme: enkele sociaal-psychologische bedenkingen. In G. Cornelis & G.
Eggermont (Eds.), Nucleaire terreur: Reflecteren over voorzorg en ethiek (pp. 107-115). Gent:
Academia Press.
Bushman, B. J., Rothstein, H. R., & Anderson, C. A. (2010). Much ado about something: Violent video game effects and a school of red herring: Reply to Ferguson and Kilburn (2010). Psychological Bulletin,
136(2), 182-187.
Christmann, K. (2012). Preventing religious radicalization and violent extremism. A systematic review of the resarch evidence: Youth Justice Board for England and Wales.
Conway, M. (2012). From al-Zarqawi to al-Awlaki: The emergence of the internet as a new form of violent radical milieu. Combating Terrorism Exchange, 2(4), 12-22.
Davidson, D. (1963). Actions, reasons, and causes. The Journal of Philosophy, 685-700.
Deflem, M. (2004). Social control and the policing of terrorism: Foundations for a sociology of counterterrorism.
The American Sociologist, 35(2), 75-92.
Fermin, A. (2009). Islamitisch en extreem-rechtse radicalisering in Nederland. Een vergelijkend literatuuronderzoek. Rotterdam: Risbo Research Training Consultancy.
Glaser, B.J. & Strauss, A. (1967). The discovery of grounded theory. Chicago: Aldine.
Hirschi, T. (1969). Causes of delinquency. Berkeley: University of California Press.
Hosmer, D. W., & Lemeshow, S. (2000). Model‐building strategies and methods for logistic regression. Applied
Logistic Regression, Second Edition, 91-142.
Huesmann, L. R. (2007). The impact of electronic media violence: Scientific theory and research. Journal of
Adolescent Health, 41(6), S6-S13.
Huq, A. Z., Tyler, T. R., & Schulhofer, S. J. (2011). Mechanisms for eliciting cooperation in counter-terrorism policing. Evidence from the United Kingdom. Journal of Empirical Legal Studies, 8(4), 728-783.
Klein, A. (2009). A space for hate: The white power movement's adaption into cyberspace. Washington: Howard
University.
Koomen, W., & Van der Pligt, J. (2009). Achtergronden en determinanten van radicalisering en terrorisme.
Tijdschrift voor Criminologie, 51(4), 345-359.
Kundnani, A. (2012). Radicalization. The journey of a concept. Race & Class, 54(2), 3-25.

32

Laperriere, A. (1997). La théorisation ancrée: démarche analytique et comparaison avec d’autres approches apparentées in J. Poupart et al., La recherche qualitative. Enjeux épistémologiques et méthodologiques.
Montréal: Gaëtan Morin Editeur.
Leuprecht, C., Hataley, T., Moskalenko, S., & McCauley, C. (2010a). Containing the narrative: Strategy and tactics in countering the storyline of global jihad. Journal of Policing, Intelligence and Counter
Terrorism, 5(1), 42-57.
Leuprecht, C., Hataley, T., Moskalenko, S., & McCauley, C. (2010b). Winning the battle but losing the war?
Narrative and counter-narratives strategy. Perspectives on Terrorism, 3(2).
Neumann, P. R., & Rogers, B. (2007). Recruitment and mobilisation for the Islamist militant movement in europe. London: The International Centre for the Study of Radicalization and Political Violence.
Newburn, T. (2007). Criminology. Devon: Willan Publishing.
Noppe, J., Ponsaers, P., Verhage, A., De Ruyver, B., & Easton, M. (2011). Preventie en radicalisering in België.
Antwerpen: Maklu.
Pauwels, L., Weerman, F., Bernasco, W., & Volker, B. (2012). Ruimtelijke criminologie. Tijdschrift voor criminologie, 54(4).
Roy, O. (2008). Al Qaeda in the West as a youth movement: the power of a narrative. CEPS Policy Briefs(1-12),
1.
Savage, J., & Yancey, C. (2008). The effects of media violence exposure on criminal aggression a meta-analysis.
Criminal Justice and Behavior, 35(6), 772-791.
Schmid, A. P. (2013). Radicalization, De-Radicalization, Counter-Radicalization: A Conceptual Discussion and
Literature Review ICCC research paper. The Hague: International Centre for Counter Terrorism.
Sherry, J. L. (2001). The effects of violent video games on aggression. Human communication research, 27(3),
409-431.
Silke, A. (2008). Holy wars: Exploring the psychological processes of Jihadi radicalization. European Journal of
Criminology, 5(1), 99-123.
Soudijn, M., & Monsma, E. (2012). Virtuele ontmoetingsruimtes voor cybercriminelen. Tijdschrift voor
Criminologie(4).
Stevens, T., & Neumann, P. R. (2009). Countering online radicalization. A strategy for action. London: The
International Centre for the Study of Radicalization and Political Violence.
Sutherland, E. H. (1947). The principles of criminology. Philadelphia: J.B. Lippincot Company.
Thompson, R. (2011). Radicalization and the use of social media. Journal of Strategic Security, 4(4), 167-190.
Tyler, T. R. (1994). Governing amid diversity: The effect of fair decisionmaking procedures on the legitimacy of government. Law and Society Review, 28(4), 809-831.
Tyler, T. R. (1997). The psychology of legitimacy: A relationaal perspective on voluntary deference to authorities. Personality and Social Psychology Review, 1(4), 323-345.
Tyler, T. R. (2000). Social Justice: Outcome and procedure. International Journal of Psychology, 35(2), 117125.
Tyler, T. R. (2001). Public trust and conficende in legal authorities: What do majority and minority group members want from the law and legal institutions? Behavioral Science and the Law, 19(2), 245-235.
Tyler, T. R. (2006). Why people obey the law. Princeton: Princeton University Press.
Tyler, T. R., & Huo, Y. J. (2002). Trust in the law. Encouraging public cooperation with the police and courts.
New York: The Russel Sage Foundation.
Tyler, T. R., & Lind, E. A. (1992). A relational model of authority in groups. In M. Zanna (Ed.), Advances in experimental social psychology 25 (pp. 151-191). New York: Academy Press. Retrieved from http://books.google.nl/books?hl=nl&lr=&id=z9vHEy0osBAC&oi=fnd&pg=PA115&ots=Pfqsk4nfB&sig=CNJW2lbu7-yFDaaRmY-KVfjpZG8#v=onepage&q&f=false 33

Tyler, T. R., & Wakslak, C. J. (2004). Profiling and police legitimacy: Procedural justice, attributions of motive, and acceptance of police autority. Criminology, 42(2), 253-281.
Tyler, T. R., Schulhofer, S., & Huq, A. Z. (2010). Legitimacy and deterrence effects in counterterrorism policing: A study of Muslim Americans. Law and Society Review, 44(2), 365-402.
Van de Linde, E., & Rademaker, P. (2010). Een toekomstverkenning van de invloed van brede maatschappelijke trends op radicaliseringsprocessen. Den Haag: Wetenschappelijk Onderzoek- en
Documentatiecentrum.
Van der Pligt, J., & Koomen, W. (2009). Achtergronden en determinanten van radicaliseirng en terrorisme. Den
Haag: Wetenschappelijk Onderzoek- en Documentatie Centrum.
Van der Valk, I., & Wagenaar, W. (2010). In en uit extreemrechts. Monitor racisme en extremisme. Amsterdam:
Anne Frank Stichting.
Weerman, F. M. (1998). Het belang van bindingen. De bindingstheorie als verklaring van verschillen en veranderingen in deliquent gedrag. Groningen: Rijksuniversiteit.
Weilnböck, H. (2005). Psychotrauma, narration in the media, and the literary public—and the difficulties of becoming interdisciplinary1. Narratology Beyond Literary Criticism: Mediality, Disciplinarity, 6, 239.
Weimann, G. (2004). Www.terror.net: How modern terrorism uses the internet. Washington: United States
Unstitute of Peace.
Wikström, P. H. (2004). Crime as alternative: Towardz a cross-level situational action theory of crime causation.
In J. Mc Coard (Ed.), beyond empiricism: Institutions and intenetions in the study of crime. (pp. 1-37).
New Burnswick: Transaction.
Wikström, P. H. (2007). In search of the causes and explanations of crime. In R. D. W. King, E. (Ed.), Doing research on crime and justice. Oxford: Oxford University Press.
Wikström, P. H. (2010). Explaining crime as moral actions Handbook of the Sociology of Morality (pp. 211239): Springer.
Wikström, P. H. (2014). Why crime happens: A situational action theory. In G. Manzo (Ed.), Analytical sociology: actions and networks: John Wiley & Sons.
Wikström, P. H., Oberwittler, D., Treiber, K., & Hardie, B. (2012). Breaking rules: The social and situational dynamics of young people's urban crime. Oxford: Oxford University Press.
Woshinsky, O., H. (2008), Explaining Politics: Culture, Institutions, and Political Behavior. New York, New
York, USA: Routledge.
Zhou, Y., Reid, E., Qin, J., Chen, H., & Lai, G. (2005). US domestic extremist groups on the web: Link and content analysis. IEEE Intelligent Systems, 20(4), 44-51.

34