Scholarly Journal Article Review

1

Developmental Predictors of Violent Extremist Attitudes – A test of General Strain

Theory*

Amy Nivette

Griffith University

Manuel Eisner

University of Cambridge

Denis Ribeaud

ETH Zurich

2

ABSTRACT Objectives: This study examines the influence of collective strain on support for violent

extremism among an ethnically and religiously mixed sample of Swiss adolescents. This

study explores two claims derived from General Strain Theory: (1) exposure to collective

strain is associated with higher support for violent extremism and (2) the effect of

collective strain is conditional on perceptions of moral and legal constraints.

Methods: This study examines the effects of collective strain using data from two waves

of the Zurich Project on the Social Development of Children and Youth. This study uses

ordinary least squares procedures to regress violent extremist attitudes at age 17 on strain,

moral and legal constraints, and control variables measured at ages 15-17. Conditional

effects were examined using an interaction term for collective strain and moral

disengagement and legal cynicism, respectively.

Results: The results show that vicarious collective strain does not have a direct effect on

violent extremist attitudes once other variables are controlled. However, the degree to

which individuals neutralize moral and legal constraints amplifies the impact of collective

strain on violent extremist attitudes.

Conclusions: This study shows that those who already espouse justifications for violence

and rule-breaking are more vulnerable to extremist violent pathways, particularly when

exposed to conditions of collective social and economic strife, conflict, and repression.

3

Research on violent extremism has produced a wide array of risk factors in psychological,

social, and political domains (Bhui, Warfa, and Jones 2014; Borum 2011a, 2011b;

Dalgaard-Nielsen 2010; Gill, Horgan, and Deckert 2014; LaFree and Ackerman 2009;

McGilloway, Ghosh, and Bhui 2015). These include psychological characteristics (e.g. low

self-control), social context features (e.g. alienation) and political processes (e.g. exclusion

from politics). LaFree and Ackerman (2009) argue that part of the difficulty in

synthesizing information on extremist violence is due to the breadth of attitudinal,

behavioral, and group-based outcomes examined under one conceptual umbrella. In

addition, studies differ in their analytical approach, including for instance analyses of risk

factors using survey samples and individual interviews (Doosje, Loseman, and van den

Bos 2013; Goli and Rezaei 2010; Pauwels and De Waele 2014), or retrospective life

history analyses of known terrorists (Gill et al. 2014). As a result of this diversity in

theoretical domains, outcomes, and analytical approaches, empirical findings on the

causes and correlates of violent extremist beliefs and behaviors are understandably mixed.

In light of this, Freilich and LaFree (2015) call for a better integration of terrorism

and extremism research into broader criminological theory and analysis (see also Agnew

2010; Schils and Pauwels 2014). Following this call the present paper examines the

interplay between two potentially fruitful theoretical approaches to violent extremism,

namely strain theories and neutralization theories. Strain theories such as Agnew’s

General Strain Theory predict that support for violent extremism is more likely when

collective strain is experienced, such as perceived discrimination against a group one

identifies with, feelings of injustice, or vicarious or direct trauma from war and civil strife

(Agnew 2010; Bhui et al. 2014; Dalgaard-Nielsen 2010; Hagan, Merkens, and Boehnke

1995; LaFree and Ackerman 2009; Pauwels and De Waele 2014; Weine et al. 2009).

Neutralization theories predict that support for violent extremism is higher when actors

morally disengage from ethical standards that prohibit violence or when they legally

disengage from the obligation to comply with the law (Bandura 1986; Ribeaud and Eisner

2010; Nivette et al. 2015; Rattner and Yagil 2004). These theories are not mutually

exclusive. Rather, collective strain as a structural feature and neutralization as a

psychological process may mutually reinforce each other (Mazerolle and Maahs 2000).

This paper therefore examines a core prediction of strain theory, namely that support for

4

violent extremism should be particularly high when experiences of collective strain are

coupled with psychological mechanisms of moral and legal neutralization.

We investigate these hypotheses with data from the Zurich Project on the Social

Development of Children and Youth (z-proso). This is a cohort study of an ethnically and

religiously mixed sample of adolescents in Zurich, Switzerland, where support for violent

extremism was measured at age 17. A large proportion of study participants’ parents

immigrated from fragile and conflict-torn societies, making the sample particularly

relevant for examining the stipulated mechanisms. Also, it is one of very few studies

worldwide that can prospectively examine the developmental mechanisms associated with

the formation of violent extremist attitudes during late adolescence

Violent extremist attitudes are defined here as beliefs and attitudes that condone

the use of violence to achieve collective goals on behalf of a national, ethnic, political or

religious group. This is close the definition used, for example, by the International

Association of Chiefs of Police [IACP], which defines violent extremists as “those who

encourage, endorse, condone, justify, or support the commission of a violent criminal act

to achieve political, ideological, religious, social, or economic goals” (IACP 2014). We

note that the relationship between extremist beliefs and actual terrorist activities is poorly

understood. A number of conceptualizations of the extremist value-acquisition process

portray the pathways to violent extremist behaviors in a stepwise fashion (see Borum’s

[2011a] review). In these models, pro-extremist attitudes are typically acquired in the

“early” stages among a wider sample of the population, whereas engaging in extremist

acts occurs among a much smaller proportion of those with favorable attitudes at a

“later” stage (McCauley and Moskalenko 2008). However, the relationship is complex as

some violent extremists and terrorists have been found to have limited “radical beliefs”

(e.g. Simi, Sporer, and Bubolz 2016), and actors with high levels of support for violent

political strategies may never engage in violence themselves (Wikström and Bouhana in

press). Therefore it is likely that the development of beliefs and attitudes that justify

violent political action and involvement in terrorist activities are partly influenced by

different mechanisms. In this paper we focus exclusively on risk factors for individual

differences in extremist, violence-condoning attitudes.

5

THEORETICAL BACKGROUND

General Strain Theory

Generally, strain theories explain criminal attitudes and behaviors as manifestations of

negative coping in response to adverse events, conditions, or treatment (Agnew 1992,

2006; Merton 1938). Agnew’s (1992) revised General Strain Theory [GST] aimed to

improve upon earlier versions of strain theory by expanding the types of negative

relationships that produce strain, explicating the social-psychological mechanisms that

underlie the relationship between strain and crime, and examining the conditions under

which effects of strain may be buffered or amplified (Agnew et al. 2002).

Agnew (1992) outlined three types of strain resulting from negative relationships

with others. First, strain can result when individuals are prevented from achieving their

goals, which includes relationships or interactions that are perceived as unjust or

inequitable (Agnew 1992). The second type arises when positively valued stimuli are

removed, such as the loss of a parent, romantic partner, or employment. Third, strain can

result from noxious stimuli such as victimization, child abuse, and negative experiences

with parents, peers, police, and employers (Agnew 1992; Kalmakis and Chandler 2015).

Exposure to these strains can produce negative emotions like anger and frustration, which

demand corrective action (Agnew et al. 2002). According to GST, crime is a type of

corrective action that seeks to injure, damage, or seek revenge on the presumed sources

of the strain.

General strain theory offers a theoretical framework to conceptualize the effects

of strain on support for violent extremism. In particular, it outlines the types of strain that

are most relevant for extremist violence, and conditional influences likely to amplify or

buffer the effects of strain (Agnew 2010). Thus, Agnew (2010) criticizes the broad

conceptualization of strain used in much terrorism and extremism research. Such

approaches fail to account for the specific motivations for violent extremism as opposed

to ordinary crime or deviance. Specifically, he argues that extremist violence is typically

inflicted on behalf of a social, religious, or political group or ideology. In order to endorse

violence on behalf of a group or ideology, one must experience collective strain (Agnew

2010; Piazza 2012). Types of collective strain likely to facilitate the adoption of violent

6

extremist beliefs are high in magnitude, considered highly unjust, and caused by more

powerful political, social, or religious groups (Agnew 2010: 136).

Prior studies have highlighted a range of strains as potential sources of extremist

beliefs and behaviors, including adverse childhood experiences (Simi et al. 2016),

discrimination and feelings of injustice (Goli and Rezaei 2010; Pauwels and De Waele;

Pauwels and Schils 2016; Piazza 2012), vicarious or direct trauma from war (Bhui et al.

2014; Weine et al. 2009), and relative deprivation (Freilich et al. 2015). More specifically,

one key source of collective strain that is often high in magnitude, considered unjust, and

inflicted by powerful “others” is exposure to political violence, such as conflict, terrorism,

and war (Canetti et al. 2013; Gill et al. 2014; Hirsch-Hoefler et al. 2014; Muldoon 2013;

Pedersen 2002; Simi et al. 2016). Prolonged exposure to political violence can act as a

stressor that leads to anger, anxiety, and depression (Garbarino and Kostelny 1996).

Studies examining the effect of the Israeli-Palestinian conflict on support for extremism

find that both direct and indirect exposure to conflict increases negative emotions and

feelings that an individual or group is under threat from the “other” or out-group (Heath

et al. 2013; Hirsch-Hoefler et al. 2014; Hobfoll et al. 2009; Huesmann et al., in press).

Hirsch-Hoefler et al. (2014) found that Israelis and Palestinians exposed to political

violence were more likely to report psychological distress, perceive group threat, and less

likely to support peaceful means of political conflict resolution.

Exposure to collective strain need not be direct in order to induce negative

emotions and corrective action (Agnew 2002; Comer et al. 2007). Agnew (2002: 609)

argues that vicarious strains can cause distress, increasing the likelihood that individuals

will seek to “prevent further harm to those they care about, to seek revenge against those

they believe are responsible for the harm, and/or to alleviate their negative feelings.”

According to Agnew, vicarious collective strains are more likely to lead to negative coping

strategies when they are high in magnitude and considered unjust, when they affect

closely related others, when they are directly witnessed or experienced by the individual,

when they are unresolved, and seen to be likely to affect the individual. Vicarious

collective strains may be particularly salient for second generation immigrant adolescents,

who may feel “culturally homeless” during a key stage in identity discovery and formation

and consequently seek out groups that offer a clear identity and a sense of significance

7

(Lyons-Padilla et al. 2015: 2). In a review of research on violent radicalization among

Muslims in Europe, Dalgaard-Nielson (2010) finds that identity-seeking and lack of

societal trust increase susceptibility to radical or extremist beliefs (see also Doosje,

Loseman, and van den Bos 2013; LaFree and Ackerman 2009; cf. McGilloway et al.

2015).

Moral and Legal Neutralization of Violence

Scholarship on violent extremism has documented extensively how those who support or

engage in violent extremism and terrorism disengage from moral, legal, and religious

standards in order to justify the use of violence against civilians (Aly, Taylor, and

Karnovsky 2014; Kruglanski and Fishman 2006; LaFree and Ackerman 2009; Pauwels

and De Waele 2014; Schils and Pauwels 2014; Slootman and Tille 2006). Psychologically

these mechanisms serve to overcome barriers to harming others and present an internal

moral justification for violence. In criminology, such mechanisms are known as

neutralization processes or cognitive distortions (Ribeaud and Eisner 2010; Sykes and

Matza, 1957).

Two such neutralization mechanisms are particularly relevant here, namely moral

neutralization and legal neutralization. The best-known version of moral neutralization

theory is moral disengagement theory by Albert Bandura. Bandura (1986, 1999)

developed a theory to explain engagement in and support for atrocities and violence on

behalf of a group. The theory predicts that engagement in harmful behavior requires

disengagement from moral self-sanctions against harmful behavior against others.

Disengagement processes may “center on redefining harmful conduct as honourable by

moral justification, exonerating social comparison and sanitising language” (Bandura

2002: 102). Substantial empirical evidence supports the link between moral

disengagement and aggressive behavior more generally (Fritsche 2005; Gini, Pozzoli, and

Hymel 2014; Ribeaud and Eisner 2015), as well as between moral disengagement and

support for political extremism (see Aly et al. 2014; Hafez 2006; Pauwels and De Waele

2014; Schils and Pauwels 2014; Slootman and Tille 2006).

8

A related but conceptually distinct mechanism refers to the disengagement from

the inner obligation to comply with the law, or what Sampson and Bartusch (1998) called

“legal cynicism.” Legal cynicism refers to attitudes that deny the binding nature of laws

and that ratify acting in ways that are “outside” of law and social norms (Sampson and

Bartusch 1998; Nivette et al. 2015). Legal cynicism researchers argue that these attitudes

arise as an adaptation to persistent experiences of injustice, disadvantage, and alienation

(Kirk and Papachristos 2011; Sampson and Bartusch 1998). This cynicism “frames” the

way individuals interpret the law (Kirk and Papachristos 2011) and on the individual level

can act as a justification for rule-breaking behavior, or legal neutralization (Nivette et al.

2015). Similar to moral disengagement processes, legal cynicism thus serves as a

mechanism to delegitimize legal sanctions against violent behaviors. Indeed, there is

evidence to suggest that legal cynicism is correlated with crime and violence (Fagan and

Piquero 2007; Jackson et al. 2012; Kirk and Papachristos 2011; Nivette et al. 2015;

Sampson and Bartusch 1998; Reisig, Wolfe, and Holtfreter 2011).

Legal cynicism has also been linked to the use of extra-legal violence to support

political and ideological goals (Hagan, Kaiser, and Hanson 2016; Rattner and Yagil 2004).

Hagan et al. (2016) explored the role of legal cynicism in justifying the use of violent

attacks against state and U.S./Coalition forces in post-invasion Iraq. They argue that

“cynicism can amplify group experiences and beliefs” which “can lead groups to form

violent responses to the dilemmas imposed by defeats – whether, for example, these

defeats follow from concentrated poverty, state repression, or both” (Hagan et al. 2016:

319). Controlling for other forms of violence, they find that legal cynicism was directly

related to the use of violence among Arab Sunnis against U.S./Coalition and Iraqi state

forces.

The interaction between strain and the moral and legal neutralization of violence

Not all who experience strain cope with crime. Rather, GST specifies several factors that

condition the effect of strain on criminal coping. This includes, amongst others,

mechanisms of inner control such as perceived moral and legal restraints or personality

characteristics such as self-control (Agnew et al. 2002; Hagan et al. 1995; Hobfoll et al.

2009; Mazerolle and Maahs 2000). For example, Mazerolle and Maahs (2000) found that

9

the effects of strain were stronger among individuals with more delinquent peers, high

propensity to commit crime, and low moral beliefs (see also Agnew and White 1992).

Similarly, Agnew et al. (2002: 64) found support for the notion that negative emotionality

and low constraint condition the impact of strain on criminal behavior. An individual’s

moral constraints and perceptions of legal boundaries and legitimacy can act as internal

controls to buffer the effects of collective strain and prevent the adoption of extremist

attitudes. Conversely, mechanisms of moral and legal neutralization may work to

minimize internal controls and amplify the effects of strain.

THE CURRENT STUDY

This paper seeks to examine the effects of vicarious exposure to collective strain on

support for violent extremism. Research suggests that collective strain generates negative

emotions, such as anger, which in turn fosters support for violence used to alleviate the

strain or “right” the perceived wrong. Although Agnew (2010) has outlined a clear

theoretical framework, no study has yet empirically tested the direct and conditional

effects of collective strain on support for violent extremism. This study begins to fill this

gap by investigating the impact of vicarious collective strain on adolescents’ violent

extremist attitudes in Zurich, Switzerland. Specifically, we explore two theoretical claims

made by Agnew (2010): first, we examine the proposition that exposure to collective

strain is associated with higher support for violent extremism. Given our current sample

of native and second-generation immigrant adolescents in Zurich, we focus on the impact

of vicarious collective strain on extremist beliefs. Second, we test the extent to which the

effect of collective strain is conditional on inner controls, namely one’s perceptions of

moral and legal constraints. While there are other possible conditional factors (e.g.

disposition, personality, delinquent peers), we focus our study on moral and legal

conditional effects based on the apparent importance of these factors in prior research on

both crime and extremism (Aly et al. 2014; Bandura 1999; Hafez 2009; Hagan et al. 2016;

Mazerolle and Maahs 2000; Rattner and Yagil 2004; Slootman and Tille 2006).

DATA AND METHODS

10

This study examines the direct and conditional effects of collective strain on adolescent

support for violent extremism using data from two waves of the Zurich Project on the

Social Development of Children and Youths (z-proso), an ongoing prospective

longitudinal study of a cohort of children that entered 1 of 56 primary schools in the City

of Zurich in 2004 (see Eisner, Malti, and Ribeaud 2011). The initial sample of schools was

randomly selected using a stratified random sampling procedure that over-sampled

disadvantaged school districts, resulting in 1,675 children from 56 primary schools

(Eisner and Ribeaud 2005). This study comprises seven waves of child interviews at ages

7, 8, 9, 11, 13, 15, and 17. In wave 5 (age 13), the participating youths were legally old

enough to give the active consent to participate on their own, while their parents received

an information letter that allowed them to proscribe their child’s participation (passive

consent procedure).

Support for violent extremism was measured in wave 7 (age 17), whereas

explanatory variables are drawn from wave 6 (age 15) or are retrospectively measured in

wave 7 (ages 16-17) to distinguish the temporal order between predictors and outcome.

The sample was restricted to all who participated in waves 6 and 7 (n=1,288), and for

whom complete information was available, resulting in 1214 respondents.

Measures

Violent extremist attitudes scale

There is no consensus on how to best measure attitudes in support for violent extremism.

Some studies have attempted to measure support for violent extremism with one single

item, while other scales are developed to measure support for a particular extremist

ideology or group. For example, In the 2009/10 UK Citizenship Survey, attitudes towards

violent extremism were measured with four items, wherein each item measured approval

of the use of violence for one specific political motivation such as “using violence to

protect animals,” “encourage violence towards different ethnic groups,” or use “violent

extremism, in the name of religion, to protest or achieve a goal” (Department for

Communities and Local Government and Ipsos MORI 2011). In our view, the selective

11

presentation of some, but not other motivations to use violence as well as the use of the

term “violent extremism” in two out of the four questions limit the utility of the

instrument.

In light of these limitations, a new scale was developed for this study. The

instrument aims to measure generic support for violent extremism defined as attitudes

that “encourage, endorse, condone, justify, or support the commission of a violent

criminal act to achieve political, ideological, religious, social, or economic goals” (ICAP

2014) . Four items were constructed so that each measures a different aspect of using

violence for collective goals. This includes using violence to fight against injustice, to

defend the values, convictions, or religious beliefs of a group, to support groups that use

violence, and to fight for a better world by using violence, committing attacks or

kidnapping people.

Responses were given on a 4-point Likert scale that ranged from “fully untrue”

(1) to “fully true” (4). The reliability was good with a Cronbach’s alpha of .80. The scale

has a positive skew (.618) reflecting that a minority of young people endorse violent

extremist attitudes. Table 1 reports the breakdown of responses on the Likert scale for

each item.

[Table 1 about here]

Independent variables

Collective strain. There are many potential sources of collective strain, including political,

cultural, and economic discrimination, systematic exclusion, and exposure to war and

conflict. Notably, Agnew (2006, 2010) argues that strain (collective or individual) is likely

to have the highest impact when it is high in magnitude, unjust, and chronic or persistent.

Thus we operationalized collective strain in a way that aims to capture all of these

characteristics, so as to maximize the likelihood of detecting an effect. An adolescent’s

experience of collective strain was measured using an average of the 2010 to 2015 Fragile

State Index (Fund for Peace [FFP], 2016), a composite score reflecting a country’s

stability on 12 political, social, and economic indicators. The average index covers events

and data for the years 2009 to 2014. To construct each indicator, a “mixed method”

12

approach is used to collect, triangulate, and integrate data from online documents,

quantitative databases, as well as qualitative input (Messner et al. 2015: 16). Social

indicators include demographic pressures (e.g. natural disasters, population growth, water

scarcity), refugees and internally displaced persons (e.g. displacement, refugee camps),

group grievances (e.g. discrimination, powerlessness, ethnic, communal, or religious

violence), and human flight and brain drain (e.g. migration per capita, emigration).

Economic indicators include uneven economic development (e.g. GINI coefficient, slum

population) and poverty and economic decline (e.g. economic deficit, unemployment,

inflation). Political and military indicators include state legitimacy (e.g. corruption,

government effectiveness, political participation), public services (e.g. provision of

policing, education, and healthcare, criminality, literacy), human rights and rule of law

(e.g. civil liberties, political freedoms, religious persecution, torture), security apparatus

(e.g. internal conflict, riots and protests, coups, fatalities from conflict), factionalized elites

(e.g. power struggles, flawed elections), and external intervention (e.g. presence of

peacekeepers, foreign military intervention, sanctions). Taken together, the overall index

reflects the degree to which residents of a country are exposed to significant collective

strain, including discrimination, repression, exclusion, and conflict.

Second generation immigrants may experience vicarious strain due to ongoing

strife in their parent’s country of birth due to the magnitude, unjust nature, and often

protracted length of the conflict or instability (Agnew 2002). In addition, collective strains

are likely to affect these adolescents through their sense of shared identity with their

national or ethnic background. As such, we assigned the relevant Fragile States Index

score according to adolescents’ parents’ country of origin. In cases where participants had

parents from two different countries, we kept the highest score. Scores ranged from 22.6

(Switzerland) to 113.9 (Somalia). Figure 1 displays the distribution of Fragile States Index

scores according to parental background. Given that the index is highly positively skewed,

we constructed a binary variable to distinguish adolescents experiencing high levels of

collective strain. Adolescents with a score equaling the median (55.1) or above are

exposed to high levels of collective strain and are coded as 1. All others are coded as 0.

Countries with scores over the median reflect a range of countries with histories of

protracted conflict and civil war (e.g. Bosnia and Herzegovina, Serbia and Montenegro,

Sri Lanka), as well as countries vulnerable to instability, group conflicts, or insecurity (e.g.

13

Turkey, Angola, Morocco). We expect that a higher score on the Fragile States Index

indicates greater and more varied vicarious exposure to collective strain.

[Figure 1 – distribution of FSI scores]

Personal strain. In addition to collective strain, we include a composite measure of personal

strain. In contrast to collective strain, personal strains are experienced on the individual

level. According to Agnew (2006), these strains can include negative school experiences,

negative encounters with the criminal justice system, violent victimization, death in the

family, or family instability. Personal strain was measured using a summary score of

negative life events measured retrospectively at wave 7, covering ages 15-17. The scale

includes 10 events similar to those identified by Agnew as significant individual stressors

(2006): received censure or punishment at school, repeated a grade, broke up with a

significant other, parent lost their job, parent died, sibling died, stayed at a mental

hospital, violent victimization, and negative encounter with police. The scale ranged from

0 to 6 events.

Moral neutralization/disengagement. Moral disengagement or neutralization reflects cognitive

processes and distortions by which deviant beliefs and behaviors become justifiable

within one’s moral landscape (Ribeaud and Eisner, 2010). Moral disengagement is

measured using a 18-item scale derived from overlapping theoretical sources, including

moral disengagement (Bandura et al., 1996), neutralization theory (Sykes and Matza 1957;

Huizinga et al. 2003), and self-serving cognitive distortions (Barriga and Gibbs 1996).

Four mechanisms of moral disengagement and neutralization are included in the scale:

cognitive restructuring (8 items), blaming the victim (3 items), distorting negative impact

(3 items), assuming the worst (2 items) and minimizing own agency (2 items). Agreement

with each item is measured using a 4 point Likert scale. Moral neutralization was

measured in wave 6 (age 15; alpha=.89)

14

Legal cynicism. Legal cynicism is measured using six items derived from Karstedt and

Farrall (2006) and Sampson and Bartusch’s (1998) original scale. Items include “It is okay

to do whatever you want as long as you don’t hurt anyone,” “Laws were made to be

broken,” and “Sometimes it’s necessary to ignore rules and laws to do what you want.”

Agreement with each item is measured using a 4 point Likert scale. Legal cynicism was

measured in wave 6 (age 15) and is reliable (alpha = .72).

Generalized trust. Generalized trust refers to the perception that unfamiliar others in society

can be relied upon (Delhey et al. 2011; Smith 2010). An adolescent who generally trusts

others is expected to be more attached and embedded in wider societal norms and

relations. Generalized trust is measured using three items adapted from the World Values

Survey Questionnaires.1 Participants were asked whether they agreed with the statements,

“most people can be trusted,” “people usually try to help other people,” and “most

people try to be fair” using a 4 point Likert-type scale. The scale was measured at wave 6

(age 15). The reliability was good with a Cronbach’s alpha of .78.

Parental involvement. Parental involvement reflects the extent to which parents are involved

in an adolescent’s everyday life. Parenting items were adapted from the Alabama

Parenting Questionnaire (Shelton, Frick, and Wootton 1996) and the Parenting Scale

from the Criminological Research Institute of Lower Saxony (KFN). The scale consists of

six items measuring how often a child’s parents engage with them and help with their

problems on a scale from 1 “never” to 5 “very often”. Items include e.g. “your parents

show interest in what you do” and “when you have problems, you can go to your

parents.” Parental involvement was measured in wave 6 (age 15) and is reliable (alpha =

.76).

Conflict coping skills. Individuals who are able to competently cope with conflict and

negative encounters or situations are less likely to be affected by collective or personal

strain (Agnew 2006, 2010). Conflict coping skills is measured using 4 items. Agreement is

1 Available online at http://www.worldvaluessurvey.org/index_html.

15

measured on a 5-point Likert scale ranging from “never” to “very often.” Items include

“I listen very carefully so that there are no misunderstandings,” “I try to put myself in the

position of the other person, to try and understand him/her,” and “I try to control my

anger.” Conflict coping skills were measured in wave 6 (age 15, alpha = .71).

Additional Measures

We include a range of additional variables that bear on theoretically relevant domains,

including personality and dispositional characteristics and social learning perspectives.

Personality and dispositional characteristics, such as low self-control and prior aggression,

reflect latent tendencies to support rule-breaking and antisocial behavior, including

violent extremism (Gottfredson and Hirschi 1990; Simi et al. 2016). Social learning

perspectives contend that support for violent extremism and related behaviors must be

learned from peers, family, or the media (Akers and Silverman 2004). Thus we include

two sources from which adolescents can be exposed to crime and violence for imitation

and adoption of beliefs: belonging to a deviant peer group and consumption of violent

media. In addition, we control for three key socio-demographic characteristics: gender,

socio-economic status, and religious denomination.

Low self-control. Low self-control is measured using 10 items adapted from Grasmick et al. (1993), incorporating five subdimensions of self-control: impulsivity, self-centeredness,

risk-seeking, preference for physical activities, and short temper. Agreement was coded

on a 4-point Likert scale, and is reliable (alpha = .75). Low self-control was measured in

wave 6 (age 15).

Aggression. Aggression was measured using the relevant subscales of the Social Behavior

Questionnaire [SBQ] (Tremblay et al. 1991). Three items refer to physical aggression (e.g.,

“you kicked, bit, or hit other people”), three items refer to proactive/instrumental

aggression (e.g., “you threatened other people to get something from them”), and three

items refer to reactive aggression (e.g., “you got very angry when someone teased or

16

irritated you”). Item response were provided on a five point Likert scale from never to

very often. The reliability and validity of the SBQ has been supported in previous

research (e.g., Tremblay et al. 1991; Tremblay et al., 1992). Overall aggression was

measured at age 15, and has good reliability with a Cronbach’s alpha of .83.

Deviant peer group. An adolescent’s exposure to deviant norms and delinquent peers was measured using a binary variable indicating whether or not an individual is a member of a

deviant peer group in wave 6 (age 15). Those who identified as part of a deviant peer

group were coded as 1, whereas those who identified as part of a non-deviant peer group

or were not part of a group were coded as 0 (Mean=.21).

Violent media consumption. Participants’ violent media consumption was measured with five items, including “watching horror movies suitable for ages 18 and older (18+)”,

“watching thriller or action movies 18+”, “searching for, and watching violent content on

the internet, watching videos with violent content on your cell phone, and sharing them

with friends”, and “playing action-packed 18+ computer or video games, which contain

intense and/or realistic portrayals of violence and killing (e.g. first person shooters)”.

These items were derived from a scale developed by the KFN (Mössle et al. 2007)

Questions were answered on a 7-point Likert scale ranging from 1 (never) to 7 (daily).

Violent media consumption was measured at age 15, and has good reliability

(Chronbach’s alpha=.80).

Sociodemographic background. Three sociodemographic variables were included: gender, SES,

and religious denomination. Gender was coded 0 for females and 1 for males

(Mean=.50). SES was measured based on the primary caregiver’s current occupation, and

the codes were transformed into an International Socioeconomic Index of occupational

status (ISEI) score (Ganzenboom, de Graaf, and Treiman 1992). The ISEI scores reflect

the relationship between education and income, with higher scores indicating higher SES.

An adolescent’s SES score was based on the highest ISEI recorded for each household. If

information from wave 6 was missing, we used the most recent high score from previous

17

waves (Mean=49.82). Given the attention on Islamic violent extremism in recent years,

we created a dummy variable for adolescents who identify as Muslim to examine whether

this particular religious background is associated with higher support for violent

extremism compared to other religious or non-religious backgrounds. Individuals who

identified as Muslim (Sunni, Shiite, Alevi, Alawi) in wave 5 or 6 (age 13/15) were coded as

1, whereas all other religious or non-religious backgrounds (i.e. Christian (Protestant,

Catholic, or Orthodox), Jewish, Buddhist, Hindu, None) were coded as 0 (Mean=.19).

Analytical Procedure

This study uses ordinary least squares regression to examine the direct and conditional

effects of collective strain on support for violent extremism. The analysis was conducted

in two parts. First, we examined direct effects by regressing support for violent extremism

on strain, moral and legal restraint variables, as well as additional and control measures.

Second, conditional effects were tested by creating an interaction term for collective strain

and moral disengagement and legal cynicism, respectively. Interactions were estimated

and reported separately. Continuous interaction variables were centered at their means in

order to facilitate the interpretation of the main effects. Due to heteroscedasticity, all

models were estimated using robust standard errors.

The percentage of missing values among the variables was low, with the highest

number of missing values found for SES (3%, n=40). As such, all primary analyses were

conducted using listwise deletion. As a robustness check, full models were reestimated

using multiple imputation (see Results section).

RESULTS

Tables 2 and 3 present the descriptive characteristics and bivariate correlations for all

study variables, respectively. The bivariate relationship between collective strain and

support for violent extremism is moderate (r = .13, p<.001). The strongest correlates of

18

support for violent extremism is moral disengagement (r = .43, p<.001), consumption of

violent media (r = .34, p<.001), and aggressive behavior (r = .31, p<.001). Exposure to

collective strain is, albeit weakly, associated with higher moral disengagement (r = .18,

p<.001), lower generalized trust (r = -.10, p<.001), lower parental involvement (r = -.22,

p<.001), lower coping skills (r = -.09, p<.01), and higher levels of aggressive behavior (r =

.17, p<.001).

[Tables 2 and 3 here]

In order to examine direct effects, we estimated three regression models. Table 4

presents the standardized coefficients (β), t values, and significance levels for each

coefficient. The first model estimates the relationship between personal and collective

strain and support for violent extremism, excluding all other study variables. Model 2

incorporates moral and legal neutralization variables, and Model 3 estimates the effects of

key variables independent of social, dispositional, and socio-demographic factors.

Model 1 shows that both personal (β=.13, p<.001) and collective strain (β =.12,

p<.001) are associated with significantly higher support for violent extremism. However

the proportion of variance explained is small at 3 percent. In Model 2, adolescents who

espouse high levels of moral disengagement (β =.36, p<.001) and legal cynicism (β =.09,

p<.01) were significantly more likely to support violent extremism. With the addition of

moral and legal neutralization, the model explained 19 percent of the variance in violent

extremist attitudes. When controls were added in Model 3, the relationship between

collective strain and violent extremist attitudes dropped to non-significance, whereas the

strongest predictor remained moral disengagement (β =.25, p<.001). However, most

social, dispositional, and socio-demographic characteristics were not significantly related

to violent extremist attitudes. Notably, those who reported competent coping skills were

less likely to support violent extremism (β =-.09, p<.01) and, in line with broader research

on violence, males were more likely than females to support violent extremism (β = .15,

p<.001). The full model explained 24 percent of the variance.

19

[Table 4 here]

Next, we examined the conditioning influences of moral and legal constraints on

the relationship between collective strain and support for violent extremism. Table 5

presents the results separately for each interaction term. Model 4 tests the conditional

influence of moral disengagement on collective strain. The significant interaction term

indicates that the effect of collective strain depends on the degree to which adolescents

employ cognitive techniques to neutralize moral constraints against the use of violence (β

=.07, p<.05). To interpret the effect, we estimated the marginal means for support for

violent extremism by exposure to collective strain and level of moral neutralization

holding all other variables at their means, and plotted the values (see Figure 2). Figure 2

illustrates that the effect of collective strain is highest at high levels of moral

neutralization: the estimate of the slope under conditions of high collective strain is .39

(p<.001), and .26 (p<.001) under conditions of low collective strain.

[Table 5 here]

[Figure 2 here]

Model 5 shows that legal cynicism conditions the effect of collective strain on

violent extremist attitudes (β = .09, p<.05). Again, we explored the marginal means for

support for violent extremism by the interaction variables while holding all other variables

at their means. Figure 3 shows that at low levels of legal cynicism, there is little difference

in attitudes about violent extremism between adolescents exposed to high or low levels of

collective strain. However, high levels of legal cynicism amplify the effect of collective

strain on support for violent extremism. That is, adolescents exposed to high collective

strain and who hold cynical attitudes towards the law are more susceptible to violent

extremist attitudes than those who have not experienced such strain, but who are

comparably cynical. The estimate of the slope under conditions of high collective strain is

significant and positive at .21 (p<.001), whereas the slope for low collective strain is non-

significant at .06 (p=.28).

20

[Figure 3 here]

In order to assess whether the results were affected by listwise deletion, values for

SES were imputed using the remaining variables in the analysis and the regression-based

multiple imputation technique. The full models (models 3, 4, and 5) were reestimated

using imputed values for SES (n=1,249). Substantive results (not shown, but available

from the authors by request) for Models 3 (direct effects) and 5 (conditional effects of

legal cynicism) remained the same. However, upon reestimation using imputed values for

Model 4 (conditional effects of moral disengagement), the coefficient for the interaction

term dropped to non-significance (b = .13, p = .057). This suggests that results for

conditional effects regarding moral disengagement were sensitive to the inclusion of cases

for which SES information was missing.

DISCUSSION

Strain perspectives have long been considered important to our understanding of violent

extremism and terrorism, yet few studies have empirically examined this relationship.

Drawing on research by Agnew (2010) as well as terrorism and extremism research more

broadly, this study aimed to examine the direct and conditional influences of strain on

support for violent extremism among an ethnically and religiously mixed sample of Swiss

adolescents. According to Agnew (2010), the type of strain most likely to influence

support for collective violence is collective strain, particularly if it is high in magnitude,

considered unjust, and inflicted by more powerful “others.” Collective strain is expected

to foster negative emotions, such as anger, which in turn encourage corrective action to

reduce, escape, or seek revenge on the source of the strain. In order to capture vicarious

exposure to collective strain, we used an average of the 2010-2015 Fragile States Index

scores which reflects the degree to which residents of a country are exposed to social,

political, and economic strife, including discrimination, repression, exclusion, and conflict.

Children whose parents are from countries with high levels of ongoing strife were

21

expected to be experiencing vicarious collective strain. The results show that collective

strain is associated with a marginal increase in support for violent extremism, however

this effect disappears when other social and individual variables are included in the model.

Males, those with high levels of moral and legal disengagement or neutralization, and

those with poor coping skills are more likely to support violent extremism. The results for

conditioning influences suggest that the degree to which individuals neutralize moral and

legal constraints amplifies the impact of collective strain on violent extremist attitudes.

However, the results for the conditioning influence of moral disengagement were not

robust.

Specifically, our results shed light on the direct, indirect, and conditional effects of

strain on support for violent extremism among adolescents, as well as predictors of

support more broadly. First, vicarious collective strain does not have a direct effect on

support for violent extremism once other variables are controlled. This is generally in line

with previous research that has found small direct or only indirect effects of strain on

crime (see Agnew, 2006). Agnew (2006) argues that, in addition to generating negative

emotions, exposure to strain can impact social, developmental, and situational variables

that in turn affect deviant attitudes and behaviors. In line with this, we find that at age 15,

exposure to collective strain was associated with higher moral disengagement, lower trust

and parental involvement, and poor coping skills. High levels of collective strain may

therefore weaken internal moral controls, social bonds, and attachments and encourage

adolescents to seek out negative peer relations or media. Prior research on radicalization

processes and extremism has documented how collective, external experiences can affect

family and social bonds and motivate extremist sympathy and activity (Pape 2005; Weine

et al. 2009, but see Bhui et al. 2014).

It is important to note that the impact of vicarious strain on negative emotions

and coping responses depends on several factors, including proximity to the source of the

strain, whether or not the strain has been resolved, and perceived contagiousness of the

strain (Agnew 2002). In relation to collective strain, the most important factor is perhaps

the degree to which individuals identify with the affected collectivity. In other words,

adolescents who do not readily identify with their parents’ ethnic or national background

are unlikely to be adversely affected by the ongoing civil strife. Given that we were not

22

able to include a measure of self-perceived ethnic or national identity, the average effect

of collective strain may be underestimated. Further research is needed to determine the

extent to which self-identification moderates the relationship between collective strain

and support for violent extremism.

In addition, the results revealed that support for violent extremism is strongly

associated with low moral and legal constraints. Adolescents who justify the use of

violence more generally and who dismiss the “bindingness” and legitimacy of the law are

more likely to support the use of violence to achieve political, social, or other ideological

goals. This finding contributes to the growing body of theoretical and empirical research

suggesting that disengagement from moral and legal norms is an important social-

psychological process that precedes and facilitates the adoption of extremist beliefs (Aly

et al. 2014; Bandura 1986; Kruglanski and Fishman 2006; LaFree and Ackerman 2009;

Pauwels and De Waele 2014; Schils and Pauwels 2014; Slootman and Tille 2006). We

argue that such processes are cognitive “tools” used by actors to overcome conventional

moral standards on the use of violence and legitimacy of legal institutions. As such,

criminological knowledge on moral and legal disengagement more generally as part of a

process of moral and legal socialization can contribute to our understanding of support

for violent extremism (Fagan and Tyler 2005).

For example, social learning perspectives can shed light on the mechanisms by

which individuals acquire deviant and violent extremist beliefs (Akers and Silverman

2004; Hagan et al. 1995; Huesmann et al. in press). Applying social learning theory to

terrorism, Akers and Silverman (2004: 27) state:

As part of their subcultural identities, terrorists learn an ideology that the ends

justify the means; violence for political ends is accepted and rewarded. These

function as definitions favorable to violence. […] In essence, the “framing” of the

conflict teaches the terrorists definitions of the situation and when, where, and

how often, it is morally right or justified to engage in political violence.

23

In addition, social media is considered to be a source in which these learning mechanisms

operate in regard to both traditional crime and violent extremism (Decker and Pyrooz

2011; Pauwels and Schils 2016). Notably, research on radicalization and pathways into

violent extremism reveal that political or ideological motivations are not necessarily

prerequisites for differential association with extremist groups and the adoption of pro-

violent beliefs (Gill et al. 2014; Simi et al. 2016). Simi and colleagues (2016: 15) found that

“the importance of ideology primarily follows rather than precedes entry” into violent

extremist groups, and that initial contact and involvement was based on existing informal

(nonideological) criminal networks (see also Freilich et al. 1999; Horgan 2009;

McGilloway et al. 2015; Schafer et al. 2014: 176). Thus, while we did not find a direct

effect of belonging to a deviant peer group or violent media consumption on support for

violent extremism, it is likely that these factors work indirectly to influence the

justification of violence and neutralization of legal norms more generally.

Finally, a key tenant of GST states that not all of those who experience strain

respond with anger and violence, and that there are certain conditions under which

violence is more likely to result from experiences of strain (Mazerolle and Piquero 1997).

In particular, an individual’s internal controls are important to regulating responses to

strain and shaping the pathways with which to cope (Wikström and Bouhana in press).

Consistent with this perspective (Angew et al. 2002; Mazerolle and Maahs 2000;

Mazerolle and Piquero 1997), we find that low moral and legal constraints, as

operationalized here by moral disengagement and legal cynicism, condition the effect of

collective strain on support for violent extremism. In other words, adolescents who

employed cognitive distortions to neutralize their moral beliefs and deny the

“bindingness” of the law were more likely to respond to collective strain with violent

extremist attitudes. Those who already espouse justifications for violence and rule-

breaking are more vulnerable to extremist violent pathways, particularly when vicariously

exposed to conditions of collective social and economic strife, conflict, and repression.

Furthermore, Kirk and Matsuda (2011; see also Kirk and Papachristos 2011)

argue that legal cynicism stems from social alienation, perceived injustice, and experiences

of misconduct and harassment by criminal justice agents. Thus we may interpret the

result as an indication that an individual’s perceived embeddedness and attachment to

24

social and legal institutions can buffer the negative effects of collective strain and prevent

the adoption of violent extremist attitudes. As adolescents encounter persistent injustices

their cynicism increases and perceptions of the legitimacy of the law diminish,

subsequently increasing susceptibility to violent extremist attitudes. This risk is

particularly heightened for those experiencing collective strain.

Limitations and Future Research

There are notable strengths to this study. First, it is one of few studies on violent

extremism that examines theoretically relevant putative mechanisms predictive of

extremist pro-violence attitudes using a representative sample of adolescents. This study

includes a wide range of indicators from relevant theoretical frameworks such as social

bonds and control, personality characteristics and predispositions, and social learning

perspectives. Second, to our best knowledge this is currently the only study that uses

prospective longitudinal data to examine support for violent extremism, allowing for the

plausible distinction of temporal order between predictors and the outcome.

This study also has several limitations. First, while the Fragile States Index is a

reflection of the cumulative exposure to ongoing collective strife in a parent’s country of

birth and therefore a plausible proxy for vicarious strain, the validity of this claim relies

on the assumptions that adolescents are first knowledgeable about the presence of

collective strain and second identify with their affected ethnic or national background. In

order to address these limitations, future studies should include a measure of self-

identification, and adapt Agnew’s (2002) measures of experienced, vicarious, and

anticipated victimization to capture subjective and objective exposure to collective strains,

such as group discrimination, repression, injustice, and physical victimization.

Second, this study did not formally analyze the mediating mechanisms according

to GST (Agnew 2006), and therefore provides only suggestive evidence that collective

strain indirectly affects support for violent extremism. Most notably, we were not able to

assess the mediating role of negative emotions, in particular anger, in generating support

for extremist violence. Previous research suggests that anger is a strong justification and

25

motivator for violence (Agnew 1992; Mazerolle and Piquero 1997; Mazerolle et al. 2003).

Collective strain may affect negative emotions like anger on two dimensions: prolonged

exposure to collective strain can lead to the development of negative emotional traits,

which reflect one’s propensity to react to stressful situations in a negative way, and/or

exposure to collective strain can generate negative emotional states, which reflects the

experience of an emotion (Agnew 2006). For example, collective strain such as exposure

to conflict has been shown to adversely affect a child’s emotional traits and coping skills

both directly (e.g. by increasing anxiety, stress, and normalizing violence) and indirectly

(e.g. by generating family stress and poor maternal health, increasing the risk of child

psychological and behavioral problems) (Merrilees et al. 2011; Muldoon 2013). In order to

distinguish the mediating effects of negative emotional traits compared to emotional

states, Agnew (2006) recommends the use of vignettes. Vignettes can portray a range of

potential situations and sources of collective strain while measuring respondents’

emotional reactions and subsequent responses to cope with the strain.

Finally, the measure of support for violent extremism used here was designed to

measure a general support for violence to achieve political, ideological, religious, social, or

economic goals. It did not measure support for specific violent ideologies or the extent to

which subjects may personally consider engaging in extremist activities. Future studies

should incorporate items that measure support for specific extremist movements as well

as measures of action intent.

26

REFERENCES

Agnew, Robert. 2010. “A General Strain Theory of Terrorism.” Theoretical Criminology 14: 131-153. Agnew, Robert. 2006. “General Strain Theory: Current Status and Directions for Further

Research.” Pp. 101-123 in Taking Stock: The Status of Criminological Theory, Advances in Criminological Theory, Vol. 15, edited by F.T. Cullen, J.P. Wright, and M. Coleman. New Brunswick, NJ: Transaction.

Agnew, Robert. 2002. “Experienced, Vicarious, and Anticipated Strain: An Exploratory Study on Victimization and Delinquency.” Justice Quarterly 19: 603-632.

Agnew, Robert. 1992. “Foundation for a General Strain Theory of Crime and Delinquency.” Criminology 30: 47–87.

Agnew, Robert, Timothy Brezina, John Paul Wright, and Francis T. Cullen. 2002. “Strain, Personality Traits, and Delinquency: Extending General Strain Theory.” Criminology 40: 43-72.

Agnew, Robert and Helene Raskin White. 1992. “An Empirical Test of General Strain Theory.” Criminology 30: 475-500.

Akers, Ronald L. and Adam L. Silverman. 2004. “Toward a Social Learning Theory of Violence and Terrorism.” Pp. 19-35 in Violence: From Theory to Research, edited by M.A. Zahn, H.H. Brownstein, and S.L. Jackson. Dayton, OH: Anderson Publishing.

Aly, Anne, Elisabeth Taylor, and Saul Karnovsky. 2014. “Moral Disengagement and Building Resilience to Violent Extremism: An Education Intervention.” Studies in Conflict & Terrorism 37: 369-385.

Bandura, Albert. 2002. “Selective Moral Disengagement in the Exercise of Moral Agency.” Journal of Moral Education 31: 101-119.

Bandura, Albert. 1999. “Moral Disengagement in the Perpetration of Inhumanities.” Personality and Social Psychology Review 3: 193-209.

Bandura, Albert. 1986. Social Foundations of Thought and Action: A Social Cognitive Theory. Englewood Cliffs, NJ: Prentice Hall.

Bandura, Albert, Claudio Barbaranelli, Gian Vittorio Caprara, and Concetta Pastorelli. 1996. “Mechanisms of Moral Disengagement in the Exercise of Moral Agency.” Journal of Personality and Social Psychology 71: 364-374.

Barriga, Alvaro Q. and John C. Gibbs. 1996. ‘‘Measuring Cognitive Distortion in Antisocial Youth: Development and Preliminary Validation of the ‘‘How I Think’’ Questionnaire.’’ Aggressive Behavior 22:333-43.

Bhui, Kamaldeep, Nasir Warfa, and Edgar Jones. 2014. “Is Violent Radicalisation Associated with Poverty, Migration, Poor Self-Reported Health and Common Mental Disorders?” PLOS One 9: e90718. doi:10.1371/journal.pone.0090718.

Borum, Randy. 2011a. “Radicalization into Violent Extremism I: A Review of Social Science Theories.” Journal of Strategic Security 4: 7-36.

Borum, Randy. 2011b. “Radicalization into Violent Extremism II: A Review of Conceptual Models and Empirical Research.” Journal of Strategic Security 4: 37-62.

Canetti-Nisim, Daphna, Brian J. Hall, Carmit Rapaport, and Carly Wayne. “Exposure to Political Violence and Political Extremism: A Stress-Based Process.” European Psychologist 18: 263- 272.

Comer, Jonathan S. and Philip C. Kendall. “Terrorism: The Psychological Impact on Youth.” Clinical Psychology: Science and Practice 14: 179–212.

Dalgaard-Nielsen, Anja. 2010. “Violent Radicalization in Europe: What we Know and What we do not Know. Studies in Conflict & Terrorism 33: 797-814.

27

Decker, Scott and David Pyrooz. 2011. “Gangs, Terrorism, and Radicalization.” Journal of Strategic Security 4: 151-166.

Delhey, Jan, Kenneth Newton, and Christian Welzel. 2011. “How General is Trust in ‘Most People’? Solving the Radius of Trust Problem.” American Sociological Review 76: 786-807.

Department for Communities and Local Government and Ipsos MORI. 2011. Citizenship Survey, 2009-2010, [data collection]. UK Data Service, Accessed 01 July 2016. SN: 6733, http://dx.doi.org/10.5255/UKDA-SN-6733-1.

Doosje, Bertjan, Annemarie Loseman, and Kees van den Bos. 2013. “Determinants of Radicalization of Islamic Youth in the Netherlands: Personal Uncertainty, Perceived Injustice, and Perceived Group Threat. Journal of Social Issues 69: 586-604.

Eisner, Manuel, Tina Malti, and Denis Ribeaud. 2011. ‘‘Large-scale Criminological Field Experiments: The Zurich Project on the Social Development of Children.’’ Pp. 410-24 in The SAGE Handbook of Criminological Research Methods, edited by D. Gadd, S. Karstedt, and S. Messner. London, UK: Sage.

Eisner, Manuel and Denis Ribeaud. 2005. ‘‘A Randomised Field Experiment to Prevent Violence: The Zurich Intervention and Prevention Project at Schools, ZIPPS.’’ European Journal of Crime, Criminal Law and Criminal Justice 13:27-43.

Fagan, Jeffrey and Alex R. Piquero. 2007. ‘‘Rational Choice and Developmental Influences on Recidivism among Adolescent Felony Offenders.’’ Journal of Empirical Legal Studies 4:715- 48.

Fagan, Jeffrey and Tom R. Tyler. 2005. ‘‘Legal Socialization of Children and Adolescents.’’ Social Justice Research 18: 217-41

Freilich, Joshua D., Adamczyk, Amy, Chermak, Steven M., Boyd, Katharine A., and William S. Parkin. 2015. “Investigating the Applicability of Macro-Level Criminology Theory to Terrorism: A Country-Level Analysis.” Journal of Quantitative Criminology 31: 383-411.

Freilich, Joshua D. and Gary LaFree. 2015. “Criminology Theory and Terrorism: Introduction to the Special Issue.” Terrorism and Political Violence 27: 1-8.

Fritsche, Immo. 2005. “Predicting Deviant Behavior by Neutralization: Myths and Findings.” Deviant Behavior 26: 483-510.

Fund for Peace. 2016. “Fragile States Index.” Washington, D.C.: The Fund for Peace. Accessed 01 March 2015. http://fsi.fundforpeace.org/.

Ganzeboom, Harry, Paul De Graaf, and Donald Treiman. 1992. ‘‘A Standard International Socio- economic Index of Occupational Status.’’ Social Science Research 21:1-56.

Garbarino, James and Kathleen Kostelny. 1996. “The Effects of Political Violence on Palestinian Children’s Behavior Problems: A Risk Accumulation Model.” Child Development 67: 33-45.

Gill, Paul, John Horgan, and Paige Deckert. 2014. “Bombing Alone: Tracing the Motivations and Antecedent Behaviors of Lone-Actor Terrorists.” Journal of Forensic Sciences 59: 425-435.

Gini, Gianluca, Tiziana Pozzoli, and Shelley Hymel. 2014. “Moral Disengagement among Children and Youth: A Meta-Analytic Review of Links to Aggressive Behavior.” Aggressive Behavior 40: 56-68.

Goli, Marco and Shahamak Rezaei. 2010. “House of War: Islamic Radicalization in Denmark.” Centre for Studies in Islamism and Radicalisation. Denmark: Aarhus University.

Gottfredson, Michael R. and Travis Hirschi. 1990. A General Theory of Crime. Stanford, CA: Stanford University Press.

Grasmick, Harold G., Charles R. Tittle, Robert J. Bursik, Jr., and Bruce J. Arneklev. 1993. ‘‘Testing the Core Empirical Implications of Gottfredson and Hirschi’s General Theory of Crime.’’ Journal of Research in Crime and Delinquency 30:5-29.

28

Hafez, Mohammed. M 2006. “Moral Agents, Immoral Violence: Mechanisms of Moral Disengagement in Palestinian Suicide Terrorism.” Pp. 292-307, in Tangled Roots: Social and Psychological Factors in the Origins of Terrorism, edited by J. Victoroff. Washington, D.C.: IOS Press.

Hagan, John, Joshua Kaiser, and Anna Hanson. 2016. “The Theory of Legal Cynicism and Sunni Insurgent Violence in Post-Invasion Iraq.” American Sociological Review 81:316-346.

Hagan, John, Hans Merkens, and Klaus Boehnke. 1995. “Delinquency and Disdain: Social Capital and the Control of Right-Wing Extremism among East and West Berlin Youth.” American Journal of Sociology 100: 1028-1052.

Heath, Nicole M., Daphna Canetti, Brian J. Hall, and Stevan E. Hobfoll. 2013. “Exposure to Political Violence, Psychological Distress, Resource Loss, and Benefit Finding as Predictors of Domestic Violence among Palestinians.” Psychological Trauma: Theory, Research, Practice, and Policy 5: 366-376.

Hirsch-Hoefler, Sivan, Daphna Canetti, Carmit Rapaport, and Stevan E. Hobfoll. 2014. “Conflict will Harden your Heart: Exposure to Violence, Psychological Distress, and Peace Barriers in Israel and Palestine.” British Journal of Political Science 46: 845-859.

Hobfoll, Stevan E., Patrick A. Palmieri, Robert J. Johnson, Daphna Canetti-Nisim, Brian J. Hall, and Sandro Galea. 2009. “Trajectories of Resilience, Resistance, and Distress During Ongoing Terrorism: The Case of Jews and Arabs in Israel.” Journal of Consulting and Clinical Psychology 77: 138-148.

Horgan, John. 2008. “From Profiles to Pathways and Roots to Routes: Perspectives from Psychology on Radicalization into Terrorism.” Annals of the American Academy of Political and Social Science 618: 80-94.

Huesmann, L. Rowell, Eric F. Dubow, Paul Boxer, Simha F. Landau, Shira Dvir-Gvirsman, Khalil Shikaki, and Jeremy Ginges. In press. “Children's Exposure to Violent Political Conflict Simulates their Subsequent Aggression at Peers by Increasing their Normative Beliefs Favoring Aggression.” Development and Psychopathology.

Huizinga, David, Anne Wylie Weiher, Rachele Espiritu, and Finn Esbensen. 2003. “Delinquency and Crime: Some Highlights from the Denver Youth Survey.” Pp. 47-91, in Taking Stock of Delinquency, edited by T.P. Thornberry and M.D. Krohn. New York: Kluwer Academic and Plenum.

International Association of Chiefs of Police. 2014. “Homegrown Violent Extremism. Awareness Brief.” Washington, DC: Office of Community Oriented Policing Services.

Jackson, Jonathan, Ben Bradford, Mike Hough, Andy Myhill, Paul Quinton, and Tom R. Tyler. 2012. ‘‘Why Do People Comply with the Law? Legitimacy and the Influence of Legal Institutions.’’ British Journal of Criminology 52:1051-71.

Kalmakis, Karen A. and Genevieve E. Chandler. 2015. “Health Consequences of Adverse Childhood Experiences: A Systematic Review.” Journal of the American Association of Nurse Practitioners 27: 457-465.

Karstedt, Susanne and Stephen Farrall. 2006. ‘‘The Moral Economy of Everyday Crime: Markets, Consumers and Citizens.’’ British Journal of Criminology 46: 1011-36.

Kirk, David S. and Mauri Matsuda. 2011. ‘‘Legal Cynicism, Collective Efficacy, and the Ecology of Arrest.’’ Criminology 49:443-72.

Kirk, David S. and Andrew V. Papachristos. 2011. ‘‘Cultural Mechanisms and the Persistence of Neighborhood Violence.’’ American Journal of Sociology 116: 1190-233.

Kruglanski, Arie W. and Shira Fishman. 2006. “The Psychology of Terrorism: ‘Syndrome’ versus ‘Tool’ Perspectives.” Terrorism and Political Violence 18: 193-215.

29

LaFree, Gary and Gary Ackerman. 2009. “The Empirical Study of Terrorism: Social and Legal Research.” Annual Review of Law and Social Science 5: 347-374.

Lyons-Padilla, Sarah, Michele J. Gelfand, Hedieh Mirahmadi, Mehreen Farooq, and Marieke van Egmond. 2015. “Belonging Nowhere: Marginalization and Radicalization Risk among Muslim Immigrants.” Behavioral Science and Policy 1: 1-12.

Mazerolle, Paul and Jeff Maahs. 2000. “General Strain and Delinquency: An Alternative Examination of Conditioning Influences." Justice Quarterly 17: 753-778.

Mazerolle, Paul and Alex Piquero. 1997. “Violent Responses to Strain: An Examination of Conditioning Influences.” Violence and Victims 12: 323-343.

Mazerolle, Paul, Velmer S. Burton, Jr., Francis T. Cullen, T. David Evans, Gary L. Payne. 2000. “Strain, Anger, and Delinquent Adaptations: Specifying General Strain Theory.” Journal of Criminal Justice 28: 89-101.

McCauley, Clark and Sophia Moskalenko. 2008. “Mechanisms of Political Radicalization: Pathways Toward Terrorism.” Terrorism and Political Violence 20: 415-433.

McGilloway, Angela, Priyo Ghosh, and Kamaldeep Bhui. 2015. “A Systematic Review of Pathways to and Processes Associated with Radicalization and Extremism amongst Muslims in Western Societies.” International Review of Psychiatry 27: 39-50.

Merton, Robert. 1968. Social Theory and Social Structure. New York: Free Press. Merrilees, Christine E., Ed Cairns, Marcie C. Goeke-Morey, Alice Cl. Schermerhorn, Peter

Shirlow, and E. Mark Cummings. 2011. “Associations between Mothers’ Experience with The Troubles in Northern Ireland and Mothers’ and Children’s Psychological Functioning: The Moderating Role of Social Identity.” Journal of Community Psychology 39: 60-75.

Messner, J.J., Nate Haken, Patricia Taft, Hannah Blyth, Kendall Lawrence, Sebastian Pavlou, Felipe Umaña. 2015. “Fragile States Index 2015.” Washington, D.C.: Fund for Peace.

Mössle, Thomas, Matthias Kleimann, and Florian Rehbein. 2007. Bildschirmmedien im Alltag von Kindern und Jugendlichen: Problematische Mediennutzungsmuster und ihr Zusammenhang mit Schulleistungen und Aggressivität. Interdisziplinäre Beiträge zur kriminologischen Forschung. Vol. 33. Baden-Baden: Nomos.

Muldoon, Orla T. 2013. “Understanding the Impact of Political Violence in Childhood: A Theoretical Review Using a Social Identity Approach.” Clinical Psychology Review 33: 929- 939.

Nivette, Amy E., Manuel Eisner, Tina Malti, and Denis Ribeaud. 2015. “The Social and Developmental Antecedents of Legal Cynicism.” Journal of Research in Crime and Delinquency 52: 270-298.

Pape, Robert A. 2005. Dying to Win: The Strategic Logic of Suicide Terrorism. New York: Random House.

Pauwels, Lieven and Nele Schils. 2016. “Differential Online Exposure to Extremist Content and Political Violence: Testing the Relative Strength of Social Learning and Competing Perspectives.” Terrorism and Political Violence 28: 1-29.

Pauwels, Lieven and Maarten de Waele. 2014. “Youth Involvement in Politically Motivated Violence: Why do Social Integration, Perceived Legitimacy, and Perceived Discrimination Matter?” International Journal of Conflict and Violence 8: 134-153.

Pedersen, Duncan. 2002. “Political Violence, Ethnic Conflict, and Contemporary Wars: Broad Implications for Health and Social Well-Being.” Social Science & Medicine 55: 175-190.

Piazza, James A. 2012. “Types of Minority Discrimination and Terrorism.” Conflict Management and Peace Science 29: 521-546.

30

Rattner, Arye and Dana Yagil. 2004. “Taking the Law into One’s Own Hands on Ideological Grounds.” International Journal of Sociology of Law 32: 85-102.

Reisig, Michael D., Scott E. Wolfe, and Kristy Holtfreter. 2011. ‘‘Legal Cynicism, Legitimacy, and Criminal Offending: The Nonconfounding Effect of Low Self-control.’’ Criminal Justice and Behavior 38:1265-79.

Ribeaud, Denis and Manuel Eisner. 2015. “The Nature of the Association between Moral Neutralization and Aggression: A Systematic Test of Causality in Early Adolescence.” Merrill-Palmer Quarterly 61: 68-84.

Ribeaud, Denis and Manuel Eisner. 2010. ‘‘Are Moral Disengagement, Neutralization Techniques, and Self-serving Cognitive Distortions the Same? Developing a Unified Scale of Moral Neutralization of Aggression.’’ International Journal of Conflict and Violence 4:298- 315.

Sampson, Robert J. and Dawn Jeglum Bartusch. 1998. ‘‘Legal Cynicism and (Subcultural?) Tolerance of Deviance: The Neighborhood Context of Racial Differences.’’ Law and Society Review 32:777-804.

Schafer, Joseph A., Christopher W. Mullins, and Stephanie Box. 2014. “Awakenings: The Emergence of White Supremacist Ideologies.” Deviant Behavior 35: 173-196.

Schils, Nele and Lieven Pauwels. 2014. “Explaining Violent Extremism for Subgroups by Gender and Immigrant Background, Using SAT as a Framework.” Journal of Strategic Security 7: 27- 47.

Shelton, Karen K., Paul J. Frick, and Jane Wootton. 1996. “Assessment of Parenting Practices in Families of Elementary School-Age Children.” Journal of Clinical Child Psychology 25: 317- 329.

Simi, Pete, Karyn Sporer, and Bryan F. Bubolz. 2016. “Narratives of Childhood Adversity and Adolescent Misconduct as Precursors to Violent Extremism: A Life-Course Criminological Approach.” Journal of Research in Crime and Delinquency 53: 536-563.

Slootman, Marieke and Jean Tille. 2006. “Processes of Radicalisation: Why Some Amsterdam Muslims become Radicals.” Institute for Migration and Ethnic Studies. Amsterdam, the Netherlands: University of Amsterdam.

Smith, Sandra Susan. 2010. “Race and Trust.” Annual Review of Sociology 36: 453-475. Sykes, Gresham and David Matza. 1957. ‘‘Techniques of Neutralization: A Theory of

Delinquency.’’ American Journal of Sociology 22:664-70. Tremblay, Richard E., Benoit Mâsse, D Perron, Matt LeBlanc, Alex E. Schwartzman, Jane E.

Ledingham. 1992. “Early Disruptive Behavior, Poor School Achievement, Delinquent Behavior and Delinquent Personality: Longitudinal Analyses.” Journal of Consulting and Clinical Psychology 60: 64–72.

Tremblay, Richard E., Rolf Loeber, Claude Gagnon, Pierre Charlebois, Serge Larivée, and Marc LeBlanc. 1991. “Disruptive Boys with Stable and Unstable High Fighting Behavior Patterns During Junior Elementary School.” Journal of Abnormal Child Psychology 19: 285– 300.

Weine, Stevan, John Horgan, Cheryl Robertson, Sana Loue, Amin Mohamed, and Sahra Noor. 2009. “Community and Family Approaches to Combating the Radicalization and Recruitment of Somali-American Youth and Young Adults: A Psychosocial Perspective.” Dynamics of Asymmetric Conflict: Pathways Toward Terrorism and Genocide 2: 181-200.

Wikström, Per-Olaf H. and Noemie Bouhana. In press. "Analyzing Radicalization and Terrorism: A Situational Action Theory." In Handbook of the Criminology of Terrorism, edited by G. LaFree and J.D. Freilich. Oxford, UK: Wiley Blackwell.

31

Table 1 – Percentage of respondents agreeing with statements supporting violent extremism

Item Fully

untrue Somewhat

untrue Somewhat

true Fully true

It’s sometimes necessary to use violence to fight against things that are very unjust

29.5% 36.3% 27.8% 6.3%

Sometimes people have to resort to violence to defend their values, convictions, or religious beliefs.

44.0% 31.5% 19.8% 4.8%

It's OK to support groups that use violence to fight injustices.

43.0% 33.1% 19.7% 4.2%

It's sometimes necessary to use violence, commit attacks or kidnap people to fight for a better world.

65.2% 22.8% 9.7% 2.3%

32

Table 2 – Means and standard deviations for all variables in the analysis (n=1,214).

Variables Mean SD

1 Violent extremist attitudes 1.81 0.67 2 Personal strain 0.83 0.99 3 Collective strain 0.48 0.50 4 Moral disengagement 2.06 0.51 5 Legal cynicism 2.19 0.55 6 Generalized trust 2.43 0.58 7 Parental involvement 3.02 0.62 8 Coping skills 3.37 0.80 9 Low self-control 2.27 0.43

10 Aggression 1.67 0.54 11 Deviant peer group 0.21 0.41 12 Violent media consumption 2.30 1.19 13 Gender 0.50 0.50 14 SES 49.82 19.17 15 Religion 0.19 0.39

Note. SD = Standard deviation

33

Table 3 – Bivariate correlations between variables included in the analyses (n = 1,214).

Variables 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

1 Violent extremist attitudes

1

2 Personal strain .13*** 1 3 Collective strain .13*** .06* 1

4 Moral

disengagement .43*** .19*** .18*** 1

5 Legal cynicism .30*** .18*** .05 .53*** 1

6 Generalized trust -.08** -.16*** -.10*** -.18*** -.16*** 1 7 Parental involvement -.18*** -.09** -.22*** -.29*** -.26*** .17*** 1 8 Coping skills -.22*** -.09** -.09** -.31*** -.26*** .14*** .22*** 1

9 Low self-control .21*** .22*** .02 .49*** .51*** -.20*** -.23*** -.38*** 1 10 Aggression .31*** .20*** .17*** .60*** .41*** -.17*** -.27*** -.35*** .49*** 1

11 Deviant peer

group .17*** .20*** -.05 .24*** .29*** -.05 -.07* -.10*** .25*** .26*** 1

12 Violent media consumption .34*** .23*** .10*** .51*** .31*** -.15*** -.18*** -.23*** .32*** .45*** .30*** 1 13 Gender .29*** .08** -.03 .34*** .11*** .03 -.10*** -.09** .08** .21*** .15*** .62*** 1

14 SES -.11*** -.01 -.37*** -.14*** -.04 .04 .23*** .16*** -.04 -.15*** .09** -.12*** .05 1 15 Religion .09** -.06* .49*** .15*** .00 -.09*** -.13*** -.05 .04 .13*** -.06* .06* -.05 -.33*** 1 Notes. *p<.05; **p<.01; ***p<.001

34

Table 4 – Ordinary least squares regression of support for violent extremism (age 17) on strain, moral and legal neutralization, and control variables (ages 15-17) (n = 1,214).

Model 1 Model 2 Model 3

β t-value β t-value β t-value Personal strain .13*** 4.09 .04 1.55 .04 1.21 Collective strain (1 = High) .12*** 4.31 .06* 2.19 .04 1.31 Moral disengagement .36*** 10.64 .25*** 6.25 Legal cynicism .09** 2.82 .11** 3.04 Generalized trust .01 0.29 Parental involvement -.02 -0.78 Coping skills -.09** -2.97 Low self-control -.06 -1.74 Aggression .03 0.78 Deviant peer group (1 = yes) .04 1.33 Violent media consumption .04 0.94 Gender (1 = male) .15*** 4.42 SES -.04 -1.30 Religion (1 = Muslim) .02 0.59 Constant 1.66*** 58.64 0.53*** 6.41 1.16*** 4.99 F-value 19.47*** 67.28*** 27.08*** R2 .03 .19 .24 Notes. All models are estimated using robust standard errors. *p<.05; **p<.01; ***p<.001.

35

Table 5 – Conditional effects of moral and legal constraints on the effect of collective strain (n=1,214).

Model 4 Model 5

β t-value β t-value Personal strain .03 1.20 .04 1.24 Collective strain (1 = High) .04 1.38 .04 1.26 Moral disengagement .20*** 4.36 .25*** 6.30 Legal cynicism .11** 3.10 .05 1.09 Collective strain x Moral disengagement .07* 1.99 Collective strain x Legal cynicism .09* 2.35 Generalized trust .01 0.25 .004 0.15 Parental involvement -.02 -0.78 -.02 -0.83 Coping skills -.09** -3.05 -.09** -3.06 Low self-control -.06 -1.75 -.07 -1.83 Aggression .03 0.67 .02 0.56 Deviant peer group (1 = yes) .04 1.48 .04 1.51 Violent media consumption .04 0.89 .04 0.87 Gender (1 = male) .15*** 4.55 .15*** 4.48 SES -.04 -1.38 -.04 -1.31 Religion (1 = Muslim) .02 0.48 .02 0.70 Constant 1.85*** 7.63 1.49*** 6.25 F-value 25.63*** 25.59*** R2 .24 .24 Notes. All models are estimated using robust standard errors. Moral disengagement and legal cynicism are mean centered. *p<.05; **p<.01; ***p<.001.

36

Figure 1 – Frequency of Fragile States Index (2015) scores matched to parent country of origin

0

50

100

150

200

250

300

350 22

.6 24

.8 26

.9 27

.8 32

.2 33

.7 34 35

.7 41

.3 43

.1 44

.8 48

.7 49

.8 55

.1 59

.3 65 65

.9 67

.4 69 71

.5 73

.5 74

.1 74

.3 75

.5 77

.5 78

.1 78

.5 79

.1 79

.3 80 81

.7 83

.9 84

.9 86 88

.4 89

.4 93

.3 94

.3 99 10

2. 1

10 4.

5 10

5. 4

11 0.

2 11

3. 9

F re

qu en

cy

Fragile State Index Score (2010-2015)

Low collective strain

High collective strain

Switzerland

Germany

Serbia and Montenegro

TurkeyItaly Bosnia and Herzegovina

Sri Lanka

Somalia

37

Figure 2 – Estimated support for violent extremism by level of moral neutralization and exposure to collective strain

1

1.5

2

2.5

3

3.5

4

1 1.5 2 2.5 3 3.5 4

V io

le nt

e xt

re m

is t a

tt it

ud es

( pr

ed ic

te d)

Moral neutralization

High collective strain

Low collective strain

38

Figure 3 – Estimated support for violent extremism by level of legal cynicism and exposure to collective strain

1

1.2

1.4

1.6

1.8

2

2.2

2.4

2.6

2.8

1 1.5 2 2.5 3 3.5 4

V io

le nt

e xt

re m

is t a

tt it

ud es

( pr

ed ic

te d)

Legal cynicism

High collective strain

Low collective strain