Testing a Threat Model of Terrorism: A Multi-method Study About Socio-Economic and Psychological Influences on Terrorism Involvement in the Netherlands

In this chapter, we presents the results from a multi-method study in the Netherlands into the role of socio-economic and psychological factors underlying terrorism involvement. Building on theories and findings of previous researchers in the field, we present a descriptive model of terrorism that categorizes distal and proximal ‘threat triggers’. In the quantitative part of the study, we analysed a combined data set on suspects of terrorist offenses, a control sample of the general population and a sample of general offenders. Terrorism suspects were more often lower educated, unemployed, and previously involved in crime compared to persons from the general population with the same gender and age. Relatively often, they had lost their job or became imprisoned for another crime a year before they were charged with a terrorist offense. In the qualitative part of the study, we conducted interviews with four detainees from terrorist units, eight detainees charged with traditional crimes (as reference group), and 18 professional informants that had personal experience with current and former detainees on terrorism and other offenses. The results of these interviews suggest that among terrorist offenders, early family experiences, attachment problems, and mental health issues increase feelings of perceived threat, which further justify violent narratives of belonging and significance.


Introduction
In this chapter, we presents the results from a multi-method study in the Netherlands into the role of socio-economic and psychological factors underlying terrorism involvement.Building on theories and findings of previous researchers in the field, we present a descriptive model of terrorism that categorizes distal and proximal 'threat triggers'.In the quantitative part of the study, we analysed a combined data set on suspects of terrorist offenses, a control sample of the general population and a sample of general offenders.Terrorism suspects were more often lower educated, unemployed, and previously involved in crime compared to persons from the general population with the same gender and age.Relatively often, they had lost their job or became imprisoned for another crime a year before they were charged with a terrorist offense.In the qualitative part of the study, we conducted interviews with four detainees from terrorist units, eight detainees charged with traditional crimes (as reference group), and 18 professional informants that had personal experience with current and former detainees on terrorism and other offenses.The results of these interviews suggest that among terrorist offenders, early family experiences, attachment problems, and mental health issues increase feelings of perceived threat, which further justify violent narratives of belonging and significance.
In this chapter, we present the results from a multi-method study (part of the PROTON collaborative project) that was aimed to investigate the role of socioeconomic and psychological factors and processes leading to involvement in terrorism.With this contribution, we attempt to serve three main goals.First, we formulate a 'threat model of terrorism' that incorporates major socio-economic, cultural and psychological determinants of terrorism that have been proposed in the literature.Second, we aim to present and summarize the most important findings of our two studies and provide quantitative as well as qualitative new insights on terrorism offenders in the Netherlands.Third, we aim to use the quantitative and qualitative findings to evaluate the basic features of our threat model of terrorism, and to comparing terrorist suspects with the general population as well as suspects of regular offenses.
Previous research in the Netherlands already highlighted the potential role of individual and group level adversities and strains, relative deprivation, personal characteristics and network dynamics (see for overviews e.g., Weggemans et al. 2014;Feddes et al. 2015;de Bie 2016).However, most of the existing research in the Netherlands has been based on open sources, government data and/or qualitative material retrieved from small and specific samples of respondents.There are still important gaps in contemporary knowledge and insights on who are involved in terrorist offenses in the Netherlands and what has led them to become involved or recruited.Only recently, studies began to employ quantitative data of larger samples (Weenink 2015;de Bie 2016;Bergema and van San 2019).In these studies, the majority of jihadists was found to be from the lower social strata, often failed to acquire adequate vocational education, and was unemployed or working in unskilled jobs (de Bie et al. 2015;Bergema and van San 2019).Previous studies also show, however, that there is not one common characteristic for all radicalized individuals.Researchers agree that the answer to why and how people become involved in terrorism should be sought in complex pathways towards terrorist engagement (see also Borum 2011).Both attitudinal and behavioral aspects of terrorism are important, and influenced by a variety of proximal and distal factors, shaped by different interpersonal and socio-cultural socializing agents, including family, peers, and religious leaders (Sageman 2004;Decker and Pyrooz 2011;Victoroff and Kruglanski 2009).
In our contribution, we focus not only on the characteristics of those who become involved in a terrorist ideology or network in comparison to the general population, but also investigate differences and similarities to those who commit regular crimes.
V. Ljujic et al.In this way, we might better tease out what processes are specific for becoming involved in terrorism, as opposed to the processes underlying more general involvement in developing criminal behavior.
Theoretically we will argue that the transition from radical thoughts to radical actions occurs in conjunction with a psychological need to (restore) clarity and positive self-image (significance) in face of perceived threat to one's material and immaterial wellbeing (building on Kruglanski et al. 2014).Building on past research and literature, we will propose a descriptive threat model of terrorism that categorizes distal and proximal 'threat triggers' of terrorism -experiences of (perceived) threat that sets in motion cognitive and emotional responses, and as a result of that, increases involvement in extremism and radical action.This model combines psychological and criminological notions of perceived threat (Stephan and Stephan 2013;Schmid and Muldoon 2015); strain (Agnew 2014) and significance loss (Kruglanski et al. 2014) to make an inventory of the factors associated to violencepermissive attitudes and violent action.The model also stresses the importance of proximal triggers and personal experiences in terrorist engagement.
We will report the results of two complementary studies, one with a quantitative approach and one with a qualitative approach.In the quantitative study, we combined and analyzed various existing data sets that were available for the complete Dutch population of terrorism suspects since the introduction of the Crimes of Terrorism Act in 2004.In the qualitative study (Study 2), we complemented the quantitative analyses with an in-depth account of the motives, experiences and life histories of persons involved in terrorist activities in the Netherlands.To achieve this, we conducted interviews in prisons with suspects and convicts of terrorist offenses.We conducted interviews with four detainees from terrorist units, as well as eight detainees charged with traditional crimes (to provide a reference group), and complemented this with interviews with 18 informants that had personal experience with current and former detainees that were incarcerated on the basis of the Terrorist Crime Act. 1n a concluding section, we will summarize these findings and compare them to our theoretical model to evaluate its validity.

The Socio-Political Context of the Study
According to Statistics Netherlands (2016), the Dutch ethnic majority covers 79.3% of the total population.Most non-Western immigrants come from Turkey (2.4%), Morocco (2.2%) and Suriname (2.1%).In terms of religious affiliation, 4.9% of Dutch population is Muslim. 2 For decades, constitutional freedom to religion and multicultural policy has facilitated integration of immigrants into the Dutch society.However, the pace of integration of immigrants in the socio-economic sphere has lagged behind the juridical sphere (Shadid 1991).In particular, there is a considerable amount of housing and school segregation (that is, concentration of minorities in certain neighborhoods and schools).Combined with increased political polarization, this facilitated a social distance between the host and Muslim population (Veldhuis and Bakker 2009).Ethnic associations, growing presence of radical imams from abroad, and experiences of discrimination and exclusion are further obstacles for intergroup contact.Until 9/11 however Muslim extremism did not pose an immanent threat to Dutch society.
After the assassination of Theo van Gogh in 2004, jihadi extremism was waning for 7 years, to be revoked again in 2011 with the start of the Arab Spring (de Bie 2016). 3Political turbulences in the Middle East and, especially, the emergence of Islamic State rekindled the emergence of jihadists' networks in the Netherlands.Since 2012, the risk of terrorism has been labelled as "substantial" by the Dutch Security Service (Bergema and van San 2019).In recent years, around 250 people from the Netherlands have travelled to Syria and Iraq to join Islamic State (ibid.).These foreign fighters are regarded as a major security concern in the Netherlands (Bakker and de Bont 2016;Wittendorp et al. 2017).After the military defeat of Islamic State, some of the combat-trained jihadi travellers returned to the Netherlands, increasing the perceived risk of terrorist attacks in the country.
In order to counter the security threats, the Netherlands amended several legislative measures to supplement the Crimes of Terrorism Act.In our study, we adopt the Dutch legal definition of terrorism, which explicitly refers to criminal offences 'committed with a terrorist intent'.According to Article 83a, terrorism entails a wide range of activities aimed at serious intimidation of the population (or part of it); the attempts to force the government or any organization to act or refrain from acting, and the disruption of the political, economic or social structure of a country.Several laws, such as The Code of Criminal Procedure and the Penal Code have been amended to investigate and prosecute crimes with terrorist intent.A list of these crimes has been widened to include various offences, such as recruitment for jihad; membership in a terrorist organization; the preparation of (and conspiring to) commit terrorist offence; spreading terrorist propaganda, and providing material and logistic support to terrorist groups.In an attempt to minimize prison radicalization and recruitment, the Dutch Ministry of Justice established special prison units for terrorism detainees.
The salience of terrorist threat (e.g. in media and political statements) adds its share to intergroup anxiety and tensions along ethnic and religious lines (Andersen and Mayerl 2018;Choma et al. 2015).PEW Research Centre (2016) found that 58% of surveyed Dutch believe that Muslims support extremist groups like ISIS and 35% reported overly negative attitudes towards Muslims. 4A survey from 20175 shows that 67% of Dutch population views ISIS as a major threat to the society.According to study by Statistics Netherlands (2017), 70% of Dutch citizens are worried about a terrorist attack, and 40% believe there is a "good chance" it may happen in the country. 6In addition to perceived threats to physical safety, recent polls indicate a perceived threat to national culture and the existence of xenophobic feelings.It was reported that 83% of the Dutch are concerned about Dutch norms and values, and 81% is concerned about immigration.Fifty percent of the respondents see non-Western migrants as a threat to their way of life, and almost one third of the Dutch population reported to feel threatened by the large influx of refugees. 7ight-wing extremists exploit the threat of terrorism and negative stereotypes against Muslims to pursue a discriminatory and -in some cases -violent Islamophobia agenda.Since 2004, two thirds of Dutch mosques have received threats or were vandalized by hate graffiti, smashed windows, dead pigs on their doorsteps, or even arson.These incidents indicate that Muslims as a religious group have been held accountable for terrorist attacks performed by militant extremists (Vellenga 2018).Right-wing extremists suggest through social media and militant public marches that it is time to defend "our way of life" violently against the rise of Islam.At the same time, an increasing number of non-Western (Muslim) minorities feel more insecure and less at home in the Netherlands (cf.Visser-Vogel et al. 2018).According to the Social and Cultural Planning Bureau, more than 50% Moroccan-Dutch and 40% Turkish-Dutch have accounted discrimination in different areas of life, particularly in the job market (Andriessen et al. 2014).
Additionally, there are also left-wing extremist networks active in the Netherlands. 8They sometimes confront right-wing extremists in public and violent anti-demonstrations.Left-wing extremist groups are less organized, and less known for attacks on citizens but may also pose a threat to society, as manifested in the murder of right-wing politician Pim Fortuyn by a left-wing extremist in 2002.However, left-wing extremists have not been terrorism suspects since the Crimes of Terrorism Act in 2004.Therefore, they are not part of our research population.
While only a small portion of Dutch Muslim gets involved in activities with terrorist intent, many (deprived) individuals from the general population may be susceptible to radical beliefs (Balen and van den Bos 2017;Verkuyten 2018).According to the Dutch security agency, there are a few thousand sympathizers of jihadi ideology in the country, which is attributed to ongoing polarization among majority Dutch citizens and their Muslim counterparts (AIVD 2016).9

Previous Insights on Socio-Economic Influences and Psychological Processes Related to Terrorism Involvement
Literature on terrorism proposes a variety of historical, political, religious, economic, socio-economic and psychological drivers of terrorism (see e.g., Schmid 2004Schmid , 2011)).Most studies postulate that terrorist involvement occurs in conjunction with social disadvantage and certain psychological responses to deal with collective and individual strains (cf.Agnew 2016;Kruglanski et al. 2014;Doosje et al. 2013).
Research in the Netherlands shows that marginalization and socio-economic deprivation are relatively prevalent among Dutch jihadi extremists and foreign fighters.Some disadvantageous neighborhoods have become a breeding ground for suburban terrorists (cf.Ljujic et al. 2017).Several authors (e.g. de Bie 2016;de Poot et al. 2011) argue that the interplay of unfavorable life circumstances and sociocultural factors determines both recruitment into, and bonding within, extremist groups.The majority of Jihadi travellers (65.5%) in the Dutch sample were friends with fellow fighters, often from the same neighborhood, before travelling to Syria (Bergema and van San 2019).Unemployment rates are relatively high among jihadi fighters; moreover, many of them are from poor backgrounds, have relatively low levels of education, and have dropped out of secondary or tertiary school (Feddes et al. 2015;de Bie 2016;Bergema and van San 2019;Weenink 2015).
Psychological research suggests that disruptive life events may trigger extremist behavior (violence) due to "significance loss," promoting a perception of life as meaningless (Kruglanski et al. 2014).In other words, the ultimate motivation for individual involvement in terrorism lays in a personal motivation of "significance quest" (the search for esteem, achievement, meaning, competence, and control) against the background of disturbing and threatening life events (cf.Kruglanski et al. 2017).Empirical terrorism studies distinguished several 'significance loss' triggers for radicalization, such as the death of a loved one, failure at work or in education, and contacts with the justice system (Feddes, Nickolson and Doosje 2015).These disruptive events may have provided a 'cognitive opening' for alternative life goals in the form of radical beliefs and extremist violence.
The social milieu from which terrorists are recruited is constantly changing in line with different needs and goals of terrorist groups (Gallagher 2014).Basra and Neumann (2016) have shown that ISIS-related radicals are typically recruited among suburban criminal networks.Likewise, several Dutch terrorism suspects and foreign fighters have been involved in various violent and property crimes prior to radicalization (de Bie 2016; Ljujic et al. 2017;Basra and Neumann 2016).This confirms the so-called crime-terror nexus described in previous literature (Hutchinson and O'malley 2007;Makarenko 2012, see also Ljujic et al. 2017).Several processes are suggested to explain this nexus (Basra and Neumann 2016).Terrorist networks may turn to criminal organizations for instrumental reasons.But offenders can also be actively recruited to become terrorists, as a way to be redeemed by God from previous criminal acts and sins (Basra and Neumann 2016).It is also possible that both criminal and terrorist involvement are ways to cope with the adversities of socio-economic deprivation and adversity; as a 'suburban response to strains' (McCauley and Moskalenko 2008).
Next to socio-economic and personal deprivation and strains, perceived threats to one's culture, status and physical wellbeing is mentioned as one of the most important influences on radicalization and terrorism (cf. Huddy et al. 2007;Koomen and van der Pligt 2009).Perceived threat does not always reflect real harm, but rather the anticipation thereof (Stephan and Stephan 2005).For example, people who have a lower socio-economic status or belong to a devalued minority group typically associate higher status with hostility, repression and domination (Glick 2002), whereas people who occupy a higher status may associate lower and devalued status with (undesirable distribution of) social benefits, immoral behavior and criminality (Fiske et al. 2002).Relationships between perceived threat and group violence are also well documented (e.g.Ranstorp 1996;Kakar 1996), including terrorism (cf.Koomen and van der Pligt 2009).Even a symbolic (nonmaterial) threat may determine how people think, feel and behave towards others (Feldman and Stenner 1997;Brewer 1999;Riek et al. 2006, for meta-analyses).In the past decade, ridiculing religious symbols and cultural values (Dutch film Fitna; Mohammed cartoons; Charlie Hebdo) have led to public outrage, death threats, even terrorist attack against satirical journalists.
Several theories provide explanations for the decisive role of perceived threat in modern society.According to evolutionary psychologists, being perceptive of threat is part of a human survival mechanism (cf.Buss 2015).Fears and phobias towards 'others' resemble instinctive reactions to threats to survival in an ancestral environment (e.g., snakes, predators, diseases) (Bracha 2004;Neuberg and Cottrell 2006;Schmitt and Pilcher 2004).This evolutionary 'threat management system' enables people to quickly recognize threatening stimuli in their environment, including pathogens (i.e., the disease avoidance system) and threats to one's physical integrity (Neuberg et al. 2011).Such physical threats include hostile coalitions or outgroups.Indeed, Bowles (2009) illuminated that intergroup conflict and warfare significantly shaped the process of natural selection among ancestral humans.This facilitated not only within-group cooperation but also psychological adaptations to compete for scarce resources (e.g.water, food, housing) and maintain group characteristics that promote bonding and group cohesion (e.g.rituals, language, culture).
Social psychologists argue that group categorization evokes the salience of group labels through discursive reinforcement of old-established fears and dislikes (Cuddy et al. 2011).Such stereotypical evaluations of 'outsiders' reflect different types of (perceived) threats and correspond to different levels of hostility towards different (religious, ethnic, national) groups.Historically, perceived threat has captured stereotypical dangers from 'others' that were predominantly defined in the line with perceived biological or physical differences (e.g.racism).According to the classical theory of intergroup conflict (Tajfel and Turner 1979), people exaggerate similarities within groups (in-group cohesion) and differences between groups (Li and Brewer 2004).Unsurprisingly, such polarization coincides with the emergence or invigoration of fear and aggression towards 'threatening' outgroups (Brewer and Campbell 1976).
Testing a Threat Model of Terrorism: A Multi-method Study About Socio-Economic…

A Threat Model of Terrorism
All in all, we propose that perceived group threat, associated with relative deprivation, socio-economic hardship and personal uncertainty, increase susceptibility to extremist viewpoints and ultimately, involvement in terrorist activities (Doosje et al. 2016;Kruglanski et al. 2014;Feddes et al. 2015;Agnew 2016;Brewer 1999).
Our model (presented in Fig. 1) postulates that perceived threat originates from different distal and proximal sources.Sometimes, the source of threat involves factors which are far beyond the personal zone of influence.For example, people perceive distal global affairs (perceived Western dominance in geopolitical distribution of power and wealth), and international events (military interventions against Iraq, Syrian war, Gaza conflicts, etc.) as unjust, threatening and harmful.These distal threats may lead to feelings of solidarity, empathy, but also anger, frustration and grievances along lines of affiliation, such as ethnic and religious backgrounds.
We also propose discrimination and segregation as a source of threat.Consistent to Sageman's (2008) emergent model of terrorist involvement, we argue that resentments associated to devalued group status resonate with extremism as a device to regain positive group identity.It involves uncritical acceptance of in-group's rules and values; and a willingness to defend the group 'in threatening situations' (Moghaddam 2005).Perceived threat also contributes to a process of dehumanization of outgroup members, and the rejection of mainstream norms and laws (van Prooijen et al. 2015;Zick et al. 2008).
We expect that perceived threat associated to one's personal zone (disruptive and traumatic experiences and individual strains and adversities) plays a crucial role in the step towards terrorist engagement.As previously mentioned, a person does not need to encounter an imminent threat to life to feel genuinely threatened.A threat to one's status may have the same profound effects, because it threatens a deep-rooted desire to matter, to be significant.According to Kruglanski et al. (2014Kruglanski et al. ( , 2017))  threats may lead to action when someone gets into contact with persons or means (i.e.violence) to combat the source of threat (see also Doosje et al. 2013;King and Taylor 2011;Klausen 2015).In extreme cases of suicide terrorism, perceived threats are so immanent that they require an ultimate sacrifice in order to achieve significance again (de Graaf 2017; cf.Versteegt et al. 2018).

Study 1: Socio-economic determinants of terrorism
In our quantitative study, we utilized two data sources.An anonymized database from the Dutch prosecution office was used to identify all suspects who were accused of 'crimes with terrorist intent' in the Netherlands (based on the Terrorist Act 2004) between 2004 and April 2017.This comprises the complete population of terrorism suspects since the introduction of the Crimes of Terrorism Act.This dataset could be combined with several anonymous datasets on general statistics from Statistics Netherlands (CBS) to identify quantifiable and registered indicators of factors related to terrorism involvement.The unique combination of these data enabled us to acquire insight into the relationship between various demographics, socio-economic indicators of adversity, previous crime involvement and being a suspect of terrorist activities.
We did not receive the data directly.Instead, the identifiable data about terrorism suspects (population database numbers) from the Dutch prosecution office were sent directly to Statistics Netherlands, where the population database numbers were removed so that the suspects were not identifiable.The data preparation and analyses were conducted via a protected server of Statistics Netherlands.
In total, the database from the Dutch prosecution office consisted of 353 suspects of terrorist offenses.However, a substantial portion of these persons could not be linked to the databases of Statistic Netherlands, because they were not registered in the Personal Records Database (BRP). 10Individuals who are not registered are most likely illegal immigrants or asylum-seekers whose applications have been rejected. 11They constituted 21.0% of the sample.For 279 suspects, we had sufficient information about demographics, socio-economic position and their criminal past.This 'terrorist' sample (n = 279) is the basis of our quantitative analysis.
We also constructed two control groups.For both the descriptive statistics as well as logistic regression analyses, we created a control offender sample (n = 279) that consisted of non-terrorism suspects of other offenses who are similar to terrorism 10 The database contains the personal data of people who (used to) live in Netherlands.Municipalities record the personal data of all residents in the BRP.For information, see: https://www.government.nl/topics/personal-data/personal-records-database-brp 11 Research on Dutch jihadi networks demonstrates the involvement of vulnerable immigrants in the Netherlands (De Bie et al. 2014).For additional information on who is excluded from the BRP, see the CBS-website: https://www.cbs.nl/nl-nl/achtergrond/2017/38/basisregistratie-personen(this information is only available in Dutch).
Testing a Threat Model of Terrorism: A Multi-method Study About Socio-Economic… suspects along the demographics of age and gender.12Also a general control sample (n = 6000) of citizens from the general population was selected, again similar with regard to age and gender.
The datasets from Statistic Netherlands gave us access to the necessary data.However, most of the datasets, which we combined with the dataset from the Dutch prosecution office, did not offer information on the complete time span (2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016) that we focused on.Therefore, the tables below show different sample sizes for different variables.It also appeared that there was a lot of missing data on education in general.For this reason, we chose to exclude the education variable from the logistic regression analyses.
Finally, for the dynamic variables that may change over time (e.g., work situation, level of education, suspected of other crimes, et cetera) we used data from the year before the suspicion. 13This way we ensured that potential relations we find cannot be interpreted as effects of being a terrorism suspect instead of a potential cause.
In Table 1, we present some general demographic descriptives for the total group of 279 suspects of terrorist offenses for which we had sufficient information.Males dominate within the terrorism suspect population (87.5%).The mean age is 30 (median: 27) during the time of suspicion and the majority of the population is between 18 and 35 years of age (62.6%).
In Table 2, we describe the characteristics of our sample of terrorism suspects together with our two control samples of general suspects and the general population.We present information about origin, level of education, whether they were officially employed or not, and previous involvement in general offending.
The table shows that suspects of terrorist activities have a migrant background more often than general suspects and the general population (81.3% versus 48.4% and 33.2%).This may be the result of the recent wave of jihadi terrorist activities but also reflects the political and police attention that is devoted to this type of terrorism at the moment.Further, suspects of terrorist offenses seem slightly less often officially employed (44.0%versus 50.9% and 58.7%), and they more often lost their job in the year before the suspicion (11.5% versus 6.1% and 3.5%).Later in the analyses we will scrutinize whether these differences also point to significant and independent effects of employment variables.Education levels of terrorist and general suspects are quite similar; in both groups, lower education makes up (almost) two third of the group.Within the general population, less than half of the sample falls within the lower education category.A substantial portion of the terrorism suspects, almost 30%, was involved in general types of offending before they became suspected of terrorist offenses.This is similar to the general suspects, but highly different from the general population of which almost none were suspected of offenses (2.4%).Finally, only a minority has spent time in detention 1 year before becoming a terrorism suspect (9.9%).However, this is comparable to the general The employment variable exists of data on whether or not a person had a job in a specific year.Those without a job can be unemployed on a voluntary or involuntary basis.Therefore, besides being unemployed, those who fall within the category 'No' could also be students (and have no need for a (official) job), live from social benefits, or have unofficial jobs Testing a Threat Model of Terrorism: A Multi-method Study About Socio-Economic… suspects (7.2%), and much higher than in the general population, were we find that almost no one in the sample (0.5%) was detained before.
In the next analysis, we investigated to which extent the variables from Table 2 had an independent and significant relationship with being a suspect of terrorism. 14 We included 204 terrorism suspects in this analysis for whom we had sufficient information. 15 Table 3 gives an overview of the correlations between different factors and being suspected of a terrorist offense.In the comparison with the general population, we mostly find values between 0.1 (small) and 0.3 (medium) and all of the correlations are significant.We observe a modest positive relation between being a terrorism suspect and having a 1st and 2nd generation immigrant background, being suspected of criminal offenses, and having been in detention.For non-immigrant background we find a negative correlation and for employment (having a job) a very small negative one.The comparison with general offenders, on the other hand, demonstrates that solely origin is significantly correlated with being a terrorism suspect instead of suspects of general types of crime.Here, the other variables are very weakly and non-significant associated with being a terrorism suspect instead of a general suspect.
Table 4 present the results from two logistic regression models that give us insight in what characteristics/factors are independently associated with an increased likelihood of becoming a terrorism suspect.In regression Model I, terrorism suspects compared to the general population, most of the effects are signifi- 14 As explained before, the education data was excluded from the logistic regression analyses due to the structural missings. 15For some terrorism suspects, we had missing data on a range of variables.Therefore, we excluded them from the analyses.cant. 16People with an immigrant background have a higher probability of becoming a terrorism suspect than others in the population.Also losing a job in the previous year predicts an increased likelihood of becoming a terrorism suspect.Interestingly, the opposite situation of finding a job does not have an effect.Also previously being suspected of other criminal offenses is related to an increased likelihood of becoming a terrorism suspect, and the effect is relatively strong (estimated odds ratio of more than 11).Finally, being in detention in the year before is also positively related to an increased likelihood of becoming a terrorism suspect.
Model II, terrorism suspects compared to general suspects, shows that none of the effects related to work situation or criminality are significantly related to the distinction between being a terrorism suspect and a general suspect.Solely origin, having an immigrant background compared to not having an immigrant background, predicts an increased likelihood of becoming a terrorism suspect instead of a general suspect.In line with the descriptive data, this confirms that terrorism suspects and general suspects are similar on most of the characteristics discussed. 16Inclusion of the variable unemployment in Model I resulted in a significant Hosmer en Lemeshow test, which is an indication that the model is not a good fit.Therefore, we ran Model I without this variable.In sum, origin (1st and 2nd generation immigrant), job loss, being suspected of other offenses and detention can be seen as risk factors in getting involved in terrorist crimes.However, terrorism suspects do not significantly differ from general suspects on SES characteristics and criminal history.

Study 2: Qualitative Analysis of Perceived Threat and Individual Involvement in Terrorism
In our qualitative Study 2 (for an extensive report, see Versteegt et al. 2018), we interviewed 30 people to capture the life events, socio-economic backgrounds and involvement mechanisms of people who were detained as a suspect of terrorist offenses.We conducted our study in the two prisons in The Netherlands that have a so-called terrorist unit.
We were able to recruit four inmates in the terrorism units, which we interviewed extensively.To contextualize our findings about the terrorism suspects, we interviewed eight inmates of other units of the prison.Because the participation among the detainees of the terrorist unit was low, we also interviewed 18 persons that were working in or around the terrorist units and had in-depth knowledge about current and former detainees that were suspected of terrorist offenses.These 'informants' included prison wardens and mentors, spiritual caretakers and psychologists at the prison, which ensured a variety of perspectives.We also conducted a focus-group interview with informants inside the prison.
We used a semi-structured questionnaire to assess the background, socialization and previous development of the detainees, as well as their ideology and experiences with socio-economic and personal setbacks and adversities.Additionally, we employed a calendar instrument, adapted from a version that was developed in a previous study on the lives of incarcerated women (Joosen and Slotboom 2015).With this instrument, detainees could indicate the socio-economic situation and events in their personal life during the last year prior to their offense.We also asked the informants to complete a life event calendar for one particular detainee that they had sufficient information on.Through various open questions, we looked for signals of socio-economic and personal adversities, we investigated their psychosocial wellbeing and mental health, and asked about their religious and ideological attitudes.We also asked details about their family background, childhood experiences and socialization.
The use of informants may seem limited because they can have instrumental reasons to express particular views about the inmates of the terrorist units.However, we have taken several measures to increase and evaluate the reliability of the informant reports.In order to prevent a biased response from only prison workers, we interviewed lawyers as well, and people working for the probation office -both professionals with their own point of view, and not employed by the prison.Informants from these different groups were generally in agreement about the ter-rorism suspects.Further, it turned out that two of the calendars taken from informants referred to an individual who were also in our sample as a respondent.This enabled us to compare the information in the life histories from two sources, and these appeared to be remarkably consistent.Therefore, we do not believe that the responses from the informants are systematically biased.Workers in the prison did express criticism of the prison system and specific legal procedures associated with the terrorist unit though, but we collected these statements anonymously in a separate document for the prison director and left this out of the analysis.
To guarantee anonymity, we are very careful in how we present our data.As the terrorist units house less than 30 prisoners in total, and the professionals working with them are limited in number as well, we present the quotes without specific reference to age, gender or profession.In our presentation of the interview quotes, the inmates are referred to as "detainees".The various people with professional personal contact with the terrorist suspects are referred to as "informants".
We provided all the respondents with a distinguishing code for their category, and a number for each individual.These are the codes that we used to present the quotes: A: terrorist suspects B: suspects of regular criminal offences (the control group) C: professionals working inside the prison D: professionals working outside the prison Our findings partly confirm the quantitative findings, but also nuance these and add some other elements that may be important to understand why and how people become involved in terrorism.
With regard to education, there appeared to be a difference between detainees from the terrorist units that were regarded as 'leaders' and detainees that were seen as merely 'followers'.The followers had a relatively low level of education, while the average educational level of the leaders was higher than that of the followers and also the regular detainees.However, the informants reported that in many cases, their acquired status and their economic situation did not match their educational level and their capacities.This difference between expectations and real outcomes may have contributed to feelings of significance loss in this group, leaving them with cognitive dissonance about the contrast between their intellectual abilities and their experienced opportunities in society.These experiences may have made them susceptible to an extremist ideology.
Informant D10, working outside the prison, states: "One person had a very high education and he literally said: so they keep rejecting me because of my ethnic background, I probably don't belong.Let me find a place where I do belong and where they do respect me." Similarly, informant C13, working in the prison, notes: They (the whole group, IV) are rather well educated; they are not doing so badly.There are many people in a similar situation, who do not radicalize.(Later): In their educational situation, they are not the needy.Very often, they have a better than average intelligence and they have read many books, too.
Testing a Threat Model of Terrorism: A Multi-method Study About Socio-Economic… With regard to employment, the interviews provide additional information to the quantitative study that could not distinguish between jobless people, students, and under aged persons. .In our qualitative study, we found that previous employment through jobs and independent companies is not uncommon among the interviewed suspects at the terrorism unit, while long term unemployment was more common in the group of regular detainees that we interviewed (4 of 8 cases).The informants confirmed that unemployment was not the rule for terrorist suspects.
Interviewer: Did they have jobs?Informant C3: Definitely.I: More than at the regular prison?C3: Oh yes, much more.
The jobs of the terrorism suspects varied from simple technical jobs at a low wage, to management jobs.The general impression of the informants was that suspects at the terrorist units had better jobs than the regular detainees.However, in several cases, terrorist suspects had lost their job a few years prior to their crime.This was reported by three of the four respondents in the terrorism unit.Two of our respondents explained how they had become obsolete in a company that they had worked in for a longer period of time.This is what a detainee (A14) at the terrorist unit told us about employment: At the moment of the terrorist offense, "I was unemployed.I had been fired after a large reorganization in a company where I had worked over 10 years, and where I had enjoyed working.After losing his job, I unsuccessfully applied for jobs."The terrorist suspect added that the strained job market made it difficult to find employment.
We found no indication, however, that the income situation was extremely stressful for the terrorism unit inmates prior to their offence.Debts, receiving social benefits and criminality were not mentioned more by the respondents and informants (even slightly less) for the terrorist unit than by the respondents from the control group.
With regard to housing, detainees at the terrorist unit seemed to be less deprived than regular detainees.One of our informants describes the general picture of the terrorist unit group in terms of housing: Informant inside the prison, C4: "Most of them lived with their parents or with a partner." A similar picture of relatively stable housing conditions emerged from a focus group interview with informants working inside the prison: Focus group informant: "In all those years in the terrorist unit, we had only five persons who were homeless." Overall, there are no outspoken objective differences between the two research groups (terrorist unit detainees and regular detainees).We did not find striking differences in socio-economic markers like education, jobs, and housing between terrorist and regular detainees -and there are indications that the terrorist unit detainees may even be slightly "better off".
Next to these socio-economic variables, psychosocial and mental health factors may have contributed to the development of extremist thought (see Weenink 2015).However,our informants did not agree on this topic.Some of them mentioned ADHD, post-traumatic stress, autism and attachment issues for the detainees of terrorism units.According to some informants, PTST and attachment disorders may have made detainees vulnerable to social failure and deprivation, but they may also have made them more susceptible to group pressure.Some informants mentioned a proneness to violence, possibly related to psychopathology and narcissism, for the supposed leaders.One informant explained the difference in mental illnesses and disorders between leaders and followers as follows: Informant C2, working inside the prison: "There are leaders and followers.Besides the difference in IQ, these people are also different in their personalities.The leaders are natural leaders, they are highly educated and could also have been leaders in some other organization.They tend to have these personality disorders like narcissism, bipolar disorder, or anti-social disorders.The followers have a low education and are easily influenced.They are very vulnerable, can have depressions and a low self-image".
In terms of adverse experiences, terrorist unit and regular detainees also appear to share many characteristics.Both groups report various experiences of discrimination, trauma, and personal setbacks.However, the terrorist detainees seem to perceive these experiences as more severe and as signs of social injustice, while the regular detainees seemed to accept their problems as part of life and as something they should cope with personally.
Informant C1, working in the prison: "I always wonder if they have experienced more losses, I don't think so.But they are triggered by these events to question life."Respondent A22: "If I had to describe them (his fellow inmates at the terrorist unit): they are sensitive guys.They cannot deal with injustice.I can let go of things.They cannot." These findings left us wondering why adverse events created more significance issues, and stronger feelings of relative deprivation in the inmates from the terrorist units.We found three possible, interrelated characteristics of the terrorist group, which may explain these differences in perception.
First, we have found indications that the early family life of the extremist group was less stable -this notion is based more on the accounts of our informants than on the stories of our respondents, who seemed to provide mainly social desirable answers about their family life.The informants mentioned that divorce of parents and dysfunctional, broken families are very common among the detainees of the terrorist unit.According to some of our informants, such circumstances occurred in more than 80% of the cases (expressed in a focus group interview in prison, as well as estimated by one informant outside prison, D10).In many families, parents were (said to be) unable to provide stability and unconditional love.Other factors that were mentioned included culturally transferred patterns of strict or abusive parenting, or 'spoiling' and pedagogical-moral neglect.Such impoverished settings may create trauma in young children -especially when their family stands out negatively in a larger cultural community, and when criticism of parents is a cultural taboo.
An interview with an informant, working at the terrorist unit, illustrates this: C4: "(….)There was more instability in the situation at home, which is my impression."I: "Can we speak of attachment disorders?" Testing a Threat Model of Terrorism: A Multi-method Study About Socio-Economic… C4: "Absolutely.Whereas the image of mother, within the culture, remains intact.There is a big discrepancy there.Parents must be talked about with a lot of respect.Mother is almost a saint.'Paradise is at the feet of the mother', that is how it is written in the Qur'an.But this particular mother often neglected things, or refrained from doing things, or was unable to do them." Another illustration from a focus interview with professionals working at the terrorist unit: Second, in interviews with detainees from the terrorist units as well as with the informants we found indications that a search for meaning, identity and belonging was strongly present in the lives of the terrorist unit detainees.They seemed, more so than the control group, sensitive to a narrative that provides them with an identity, a sense of belonging, and moral guidelines from authority figures they can love as their surrogate family.

Informant inside the prison, C4
: "There is a difference.These people really have brothers.
Normal detainees have friends that buy drugs from them.These people have a brotherhood, which really feels very intense to them.
Informant inside the prison, C1: "I can imagine that this brotherhood really means a lot to them.When you feel you have always been treated like an outcast actually…and suddenly there is someone who…uses these 'lover boy'-like methods…With all those hugs, too…" I: "Hugs?"R: "They keep embracing each other on every occasion".
This search for belonging may be interpreted as a result of to the reported early life instabilities.The (mental or physical) unavailability of a primary caretaker may have generated attachment problems in children, and as young adults, this may have left them with an anxiety-driven need to find social bonds.The narrative of 'brotherhood' in extremist organizations literally provides individuals with a stable and reliable family.Persons with attachment problems also often seek identification figures to admire.We heard stories among our informants about celebrity worship syndromes of detainees in the terrorist units (towards Mohammed B., the killer of Theo van Gogh, for example), and religious obsessions.Also the tendency of the terrorist unit inmates to hug each other, and form strong social pacts, may be seen as indications that the need for social bonding in these individuals has pathological and obsessive elements, which could stem from childhood trauma.A third factor that set apart the detainees from the terrorism unit from the regular detainees, were the coping strategies and narratives they used to deal with adversities.We found that regular detainees more often relied on religious coping strategies to deal with experiences of personal or social injustice or fell back on reasoning that enabled them to keep their faith in social justice and humanity ('Just World narratives').
Respondent from the control group, B2: "I found out, in life, you just have to make the best of it.It can be over any time" (…)."You have to be able to deal with difficult situations.You have to see through the darkness, towards the light.
Such examples of 'coping' narratives were not found in the stories from the terrorism unit respondents.Regular detainees were also less distrusting of the prison staff and held compassionate views of their fellow human beings and their mistakes.Terrorist unit detainees on the other side tended to judge all events as further contributions to their victimhood, according to multiple accounts of our informants.Events that had a negative outcome for them were interpreted as deliberate mistreatment and discrimination, even if they probably were not.
A discussion with a focus group of informants inside the prison illustrates this: I: "Do they talk about discrimination?"Focus group informant 1: Oh yes.That is the main storyline.I: The main storyline?Focus group informant 1 and 2: Yes, they will say they are victims, and the government hates them.I: Do you feel that they are more than the other group using stories of discrimination and deprivation?Several informants: "Oh yes.All the time" Focus group informant 1: "Anything.They return from the court, and things did not work out positively, then everybody is plotting against them.Here, at prison, everyone is plotting against them.
Adding to this, the inmates at the terrorist unit seem to worry a lot about social injustice, morality, world problems, and their own destiny, according to informants.They were described as psychologically vulnerable and less able to cope with setbacks in comparison to the regular detainees, who expressed more healthy coping strategies.
Informant C4, working inside the prison: "They never felt able to cope with injustice.But due to the IS ideology, they suddenly feel stronger (…)." Lastly, we found that the role of religion (Islam) in the development of involvement in jihadi terrorism is different than often thought.Most of the terrorist detainees seemed to be not very well informed about the Islam and did not have a thorough knowledge of the Quran.Instead, they became attracted to the powerful representation of the Salafi -Jihadi discourse on the internet of the Islam as "one universal truth" in which the hypocrites must be discarded from the true believers.For them, this resonates with the vague notions of Islam their parents have provided them with, notably that of 'being a good Muslim', and it makes them believe that this representation of Islam is correct.
Detainee A13 from the terrorist unit explains: "I had quit using alcohol, I was making my life better.I wanted to be a good Muslim.I was looking for something new, actually I was… looking for my identity.Then I talked with a couple of men, friends of mine, about Syria and Similarly, in the accounts of one detainee in our sample and several informants, it appears that right-wing extremist detainees were politically not very well informed.Instead, they were sensitive to the general narrative of 'the foreign threat to the Dutch identity and society' together with a desire to help protect society from these dangers.

Conclusion and Discussion
In this chapter, we first proposed a threat model of terrorism, as a theoretical framework for organizing and interpreting the major socio-economic and psychological determinants of terrorism involvement.Drawing upon intergroup notions of strain, perceived threat and significance loss, our model lays out a clear distinction between proximal and distal threat triggers, and identifies multiple emotional and cognitive processes linking those triggers to the behavioural outcome: unlawful activity with terrorist intent.We utilized recent quantitative and qualitative data to test the model.In the quantitative study we combined and analysed existing data sets that were available for the complete Dutch population of suspects of terrorist offenses since the introduction of the Crimes of Terrorism Act in 2004.The findings show that terrorism suspects belonged more often to ethnic minority groups, were lower educated, were unemployed, and were more likely previously involved in crime, than persons from the general population with the same gender and age.They also more often lost their job and were detained previous to becoming involved in terrorist offenses.Among the terrorism suspects, 30% were registered for other criminal offense(s) in the year before becoming known to police for terrorist offense.In the multivariate analyses we found that economic adversity and criminality substantially and independently increase the likelihood of becoming a terrorism suspect for the general population.These findings offer initial support the proposed threat model as well as its theoretical antecedents (Doosje et al. 2016; see also Ljujic et al. 2017;Pare and Felson 2014).However, terrorism suspects did not differ significantly from general suspects in terms of individual strains and adversities.This means that additional factors and processes are needed to fully understand why people become involved in terrorist activities.
In the qualitative study, we complemented the quantitative analyses with an indepth account of the motives, experiences and life histories of persons involved in terrorist activities in the Netherlands.Consistent with the first study, the findings suggest that terrorist offenders and other, regular, detainees have similar life courses to a certain extent, and both groups are not very different in socio-economic position and experiences of deprivation and discrimination.Moreover, the qualitative study shows that both groups also often encountered personal adversities in other life spheres and were reported to have various types of mental health problems.Different from general offenders, however, terrorist offenders perceive these setbacks and experiences of deprivation as more severe and signs of social injustice.We found that this was related to three other characteristics that were more typical for the terrorism offender group: an instable and impoverished early family life, an intense search for meaning and identity and need for acceptance, and coping strategies that were not adaptive but instead constantly reaffirmed their perceived victimhood.In the end, these suspects have encountered 'a bunch of guys' that introduced them to a network with an extremist narrative or were attracted by the strong but simplified narratives they found on the internet.This susceptibility to ideological (peer) indoctrination went hand in hand with their strong need for significance and belonging, and may be explained by several psychological conditions, such as attachment issues and mental disorders that impair social reasoning and impulse control.The childhood problems of the terrorism suspects may have created a susceptibility to find belonging in deviant social networks.This susceptibility was further aggravated by deprivation, and experiences of discrimination, social exclusion and loss.The narrative of the extremist networks specifically addresses these feelings of rejection and puts it in a global perspective of urgency, belonging and heroism, as well as personal gain.These processes were most outspoken for suspects with a jihadi background, but also seemed to be applicable to right-wing extremists.
In sum, the findings support the threat model but also extend it.We found that distal threat triggers such as conflicts abroad and socio-economic differences within the population can contribute to emotional and cognitive responses among extremists.Muslim extremism frames group threat as indications of a devalued and discriminated minority, whereas right wing extremism exploits popular resentments about immigrants, which threaten the ethno-national identity.Besides shaping one's cognitive and emotional responses towards 'threatening' others (Ljujic et al. 2011), threats also provide a 'convenient' rationale for intergroup violence (Feldman and Stenner 1997), including terrorism (Staub 2004).However, these extremist narratives may appeal to many, but only a few become actually involved in terrorism.The results of our research show that proximal personal factors (such as economic hardship and disruptive life experiences), may be crucial to our understanding of involvement in terrorist offenses.Adding to this, our results suggest that early family experiences, attachment problems, and mental health issues may have contributed significantly to feelings of perceived threat and a susceptibility to appealing narratives that promise feelings of belonging and significance.
To a certain extent, these life histories that are associated with involvement in terrorism and extremism may be a matter of chance and a series of bad luck, as well as meeting the wrong persons or internet sites at the wrong time.However, our study also provides evidence that one's path to terrorism is driven by tangible socioeconomic and psychological factors that might be addressed by professionals, policy makers and the criminal justice system.In this respect, Dutch society is facing a twofold challenge: to prevent engagement in terrorism, but also, to stimulate deengagement and re-socialization of convicted terrorism offenders.To these ends, a better understanding of the role of perceived threat in radicalization trajectories (cf.It is important to note that our study has various limitations.A strong feature of our quantitative study was that we were able to get access to data about the complete population of terrorism suspects since terrorism was distinguished as a separate offense in the Netherlands.However, a limitation was that we could only retrieve general information about socio-economic characteristics and previous involvement in crime, and could not go into detail about life histories and events.The findings were also limited in the sense that for a part of our sample, we lacked sufficient information on characteristics such as education and employment situation.A strong feature of our qualitative study that we gained access to respondents and informants in specialized prison terrorist units, that enabled us to get access to indepth information about the life histories and personal backgrounds of current and former terrorism suspects, as well as regular detainees.However, the findings were limited in the relatively low number of respondents that we could speak, and also had to rely on information we got from informants.Although the information from the informants was often consistent and uniform, and were quite similar for those working inside and outside prison, we cannot rule out the possibility that their observations are incomplete or that experiences with some detainees in the past were generalized too much to the complete population.
All in all, our findings offer interesting and plausible insights about the backgrounds of terrorism offenders in the Netherlands and the processes that may lead to their involvement in terrorist crimes, but our results need to be scrutinized in future studies to get more certainty about their validity and generalizability.We therefore hope that the findings presented in this chapter will act as a stimulus for further research in the field.

Fig. 1
Fig. 1 Threat model of terrorism Testing a Threat Model of Terrorism: A Multi-method Study About Socio-Economic…McCauley and Moskalenko 2008) is necessary to counter extremism and terrorism and to prevent recidivism.

Table 1
Basic demographics of terrorism suspects

Table 2
Origin, SES and criminality of terrorism suspects, general suspects and population

Table 3
Associations between being a terrorism suspect and origin, employment and previous offending variables

Table 4
Summary of logistic regression analyses for variables predicting being a terrorism suspect * p < .05. * * p < .01. * * * p < .001Testing a Threat Model of Terrorism: A Multi-method Study About Socio-Economic… Testing a Threat Model of Terrorism: A Multi-method Study About Socio-Economic… they said, that is where you should go.If you want to be significant as a Muslim.There is no war at all.There are just Muslims who need your help.So I wanted to go there, see what I could do, make a new start.