How Preventive Counter-Radicalization Interventions Can Stigmatize Their Target Groups
Jaap van der Veen August, 2017
Master Thesis Student number: 11083301 Research Master Psychology
University of Amsterdam
Word count (main body): 9,642
Author note
For their inspiring guidance, I would like to thank my supervisors: Dr. Dianne van Hemert, Prof. Dr. Bertjan Doosje, and Dr. Allard Feddes. I would also like to thank many supportive colleagues and friends who helped me throughout this project. I think especially of Dr. Helma van den Berg, Dr. Anke van Gorp, Aiden Hoyle, and Suzyong de Bueger.
Correspondence regarding this article should be addressed to Jaap van der Veen. E-mail: hi@jaapvanderveen.com.
Abstract
We investigated whether preventive counter-radicalization interventions can stigmatize their target groups as an unwanted side-effect. Specifically, we examined the effects of a number of characteristics of these interventions. To my knowledge, this is the first study to specifically investigate stigmatization by approaching professionals involved in such interventions. We employed a multi-method approach. First, in qualitative Q-sort interviews, participants (N = 3) rated intervention characteristics on their contribution to effectiveness and potential to stigmatize. Second, in quantitative online surveys, participants (N = 88) indicated to what extent they deemed a selection of intervention characteristics likely to contribute to
stigmatization. The results from the Q-sort interviews and online surveys suggest that how an intervention is shaped can influence the extent to which it stigmatizes the target group. This study hopes to contribute to preventing what is now often a negative side effect of preventive counter-radicalization interventions: stigmatization.
How Preventive Counter-Radicalization Interventions Can Stigmatize Their Target Groups
If you keep pushing something off on a person, pretty soon that person stops fighting the reputation and becomes everything he is accused of being. . . . You start to think, “Fuck them. If that’s what they think I am, and I have to bear that cross, I got nothin’ to lose in being all they think I am.” (Emmons, 1986, p. 47–48).
This quote by Charles Manson, former cult leader and serial murderer, illustrates how stigma can cause individuals to change their self-concept and behavior. In this study, we explored how preventive counter-radicalization interventions can stigmatize their target groups as an unwanted side-effect.
Terrorism can be preceded by violent radicalization. Doosje et al. (2016) define violent radicalization as “a process through which people become increasingly motivated to use violent means against members of an out-group or symbolic targets to achieve behavioral change and political goals” (p. 79). Many interventions therefore aim to prevent violent radicalization (for an extensive overview, see Feddes & Gallucci, 2015). However, such interventions may also have unintended consequences. For instance, it has been argued that the British counter-radicalization intervention “Prevent” not only jeopardizes programs that attempt to increase social cohesion, but also stigmatizes Muslim communities, possibly leading to more radical behavior (Bhui, Hicks, Lashley, & Jones, 2012; Husband & Alam, 2011; McDonald & Mir, 2011).
Through a mixed-method approach, we investigated how certain preventive counter-radicalization interventions can stigmatize their target groups. To my knowledge, no study has specifically investigated stigmatization by interviewing and surveying professionals who
actually design, facilitate, and carry out preventive counter-radicalization interventions. Ultimately, this study could help determine which interventions are deemed successful
without contributing to what they set out to prevent: stigmatization, polarization, and possibly even radical behavior.
We focus on the first stage of radicalization: the vulnerability phase (Doosje et al., 2016). Individuals on this “ground floor of the staircase to terrorism” (Moghaddam, 2005) are those who may, for instance, perceive the government or another group to treat them unfairly, deprive them, or threaten their way of life. Preventive interventions are per definition most effective in the vulnerability phase, as they aim to “prevent social or political radicalization, [or] violent radicalization” (Bhui et al., 2012, p. 2). We focus on a wide variety of such interventions in the vulnerability phase, ranging from hotlines to education programs for youngster and resilience trainings. We will argue that certain characteristics of these interventions can cause unintended consequences, in particular stigmatization. Stigmatization
Corrigan and Rao (2012) define stigmatization as “negative attitudes held by members of the public about people with devalued characteristics” (p. 464). Because stigmatization has not been studied extensively in the radicalization domain, we draw on the criminology
literature. There, researchers have proposed a construct identical to stigmatization: labeling. In this context, labels are names that individuals give to those who break the law, such as “criminal” or “activist”. Labeling theory posits that lawbreakers feel labeled after the criminal justice system has processed them. Subsequently, these individuals perceive how society now reacts to them differently (Atchison & Heide, 2011). In turn, individuals can change their self-concept and act on these changes in their self-self-concept (Lopes, Krohn, Lizotte, Schmidt, & Edward, 2012; Ward, Krohn, & Gibson, 2014).
In line with Feddes, Nickolson, and Doosje (2015), we argue that there are similarities between the terrorism and criminology domains. In some cases, similar processes and trigger factors may force individuals towards terrorism and criminality (Weenink, 2015; De Waele & Pauwels, 2014). Individuals may also move from criminality to terrorism and vice versa (Bakker, 2006; Weenink, 2015), even to such an extent that some authors speak of a “crime-terror nexus” (Gallagher, 2016; Ljujic, Van Prooijen, & Weerman, 2015; see also Bovenkerk & Chakra, 2005; Bovenkerk, 2010).
Given the similarities between the terrorism and criminology domains, it is plausible that certain counter-radicalization interventions might also label and stigmatize their target group. Indeed, Powell (2016) describes such consequences of the aforementioned Prevent intervention: British Muslims felt stigmatized by the intervention’s rhetoric (O’Toole, Meer, DeHanas, Jones, & Modood, 2015) and wide focus on their communities (Communities and Local Government Committee, 2010). Kundnani (2015) notes that a related British
intervention, “Channel”, “widens the definition of suspicion beyond criminality to much vaguer notions” (p. 15). As a result, the intervention stigmatizes Muslim communities as “risky” and “at risk” (Heath-Kelly, 2012, as cited in Blackbourn, Dexter, Dhanda, & Miller, 2012).
More generally, Briggs, Fieschi, and Lownsbrough (2006) argue that interventions focused on groups of individuals often do not take into account differences within such groups. As a result, these interventions can stigmatize members of a minority groups who are not involved in extremism. In fact, stigmatization might even contribute to violent
radicalization. For instance, Moghaddam (2005) explains how a lack of fairness and just treatment can motivate individuals to “march along alternative paths, even desperate and radical ones, to address their grievances” (p. 163). Similarly, the Open Society Justice Initiative (2016) has interviewed individuals targeted by Prevent. These individuals often
indicated how “their experience of Prevent could have drawn them towards terrorism” (p. 18) rather than away from it.
Intervention Characteristics
In this study, we focus on preventive counter-radicalization interventions. We
investigate how these interventions can stigmatize their target groups. In line with previous studies (e.g., SAFIRE, 2011; Bartlett, Birdwell, & King, 2010), we examine the underlying principles of interventions. These “intervention characteristics” describe how an intervention is set up or carried out. For instance, intervention characteristics could be that the intervention is carried out by the police or that the intervention takes place on a small scale. In the
overwhelming multitude of interventions in existence, we argue that these interventions characteristics allow us to draw generalizable conclusions about the potential of interventions to stigmatize their target groups. In the next section, we specify how a selection of
interventions characteristics might cause stigmatization (see Table 1).
Table 1. A selection of characteristics of preventive counter-radicalization interventions.
Category Intervention characteristics
Target group The intervention can focus on 1) individuals with a specific ethnicity, religion, or ideology, 2) individuals deemed “at risk” for radicalization, regardless of their ethnicity, religion, or ideology, or 3) individuals in general, regardless of their ethnicity, religion, or ideology and whether they are deemed “at risk” for radicalization.
Initiation 1) The intervention approaches individuals, 2) individuals
approach the intervention, or 3) the intervention approaches key figures in a community, who in turn report signals of radicalization to the intervention.
Scale The intervention takes place on a 1) small scale or 2) large
Focus The intervention targets religion or ideology 1) directly, in that it aims to reduce radical interpretations, or 2) indirectly, in that it aims to achieve other goals, such as proper identity development.
Executor The intervention is carried out by 1) the police, 2) the municipality, or 3) a non-governmental organization.
Normative The intervention is 1) normative, in that it considers certain
norms, values, and beliefs superior and teaches them to participants, or 2) non-normative, in that it does not consider certain norms, values, and beliefs superior and discusses them (and others) with participants.
Cultural aim The intervention teaches 1) differences in groups, namely
multiculturalism, in that it stimulates participants to respect different cultural, ethnic, and religious communities, or 2) differences in individuals, namely hybrid identities, in that it stimulates participants to realize that they belong to multiple communities.
Matching The intervention 1) matches first-line professionals to
participants on the basis of background, ethnicity, or other characteristics, or 2) does not match first-line professionals to participants.
Target Group. Interventions can focus on target groups with a specific ethnicity, religion, or ideology. For instance, Prevent focuses almost solely on Muslim communities (Thomas, 2016). Per definition, such a specific focus on a single group stigmatizes: it sets the group apart from others and signals that members of the group are “risky” or “at risk”. As a report by the Open Society Justice Initiative (2016) points out, such a specific focus “creates a risk of discrimination” (p. 17). As mentioned before, it may even work counterproductive: individuals in their case studies describe how “their experience of Prevent could have drawn them towards terrorism, and not away from it” (p.18).
Interventions can also focus on individuals deemed “at risk” for radicalization, regardless of their ethnicity, religion, or ideology. An example is “BOUNCEyoung”, a
2014). Such interventions do not focus on a target group with a specific ethnicity, religion, or ideology, but still label their target group as (potential) radicals.
Finally, interventions can focus on individuals in general, regardless of their ethnicity, religion, or ideology and whether they are deemed radical. For instance, the “Fort of
Democracy” is an interactive educational exposition on democracy that is open to the general public (Van Ooijen, 2011). As Van Gorp and Roosendaal (2013) point out, it might be “stigmatizing to point to a specific group that needs guidance or support,” and therefore, “in principle, programmes should be open for every adolescent” (p. 3). In sum, we hypothesize that interventions that focus on a target group with a specific ethnicity, religion, or ideology stigmatize strongest, followed by interventions that focus on individuals deemed “at risk” for radicalization, regardless of their ethnicity, religion, or ideology. Finally, we hypothesize that interventions stigmatize least when they focus on individuals in general, regardless of their ethnicity, religion, or ideology and whether they are deemed “at risk” for radicalization.
Initiation. Some interventions actively approach individuals. For instance, “street coaches” approach youth in the neighborhoods of Amsterdam (Van Wijk, 2011). However, as the Open Society Justice Initiative (2016) points out, approaches and referrals can be
“inherently stigmatising and intensely intimidating” and lead individuals to fear that others may “view them as ‘extremists’ in the future” (p. 16).
There are also interventions that approach key figures in a community. In turn, these key figures report signals of radicalization in their community to the intervention. For instance, “Allies” brings together representatives from the community, municipality, and police (Hermens, Van Kapel, Van Wonderen, & Booijink, 2016). At first sight, such
interventions seem to empower the community. However, they may actually have “direct and negative effects on […] community cohesion” (Husband & Alam, 2011, p. 1), because they require community members to monitor and possibly even report each other. Additionally,
community members cooperating with the intervention may be associated with it; as a result, “their standing within their own communities could be put in jeopardy” (Husband & Alam, 2011, p. 3).
Other interventions wait for individuals to approach them. An example is the “Hotline support parents with radicalization”, a helpline for parents with concerns about their children (Samenwerkingsverband van Marokkaanse Nederlanders, 2017). Such interventions only assist when called upon and do not label individuals “risky” or “at risk” by approaching or reporting them. To summarize, we hypothesize that interventions that approach individuals stigmatize strongly, followed by interventions that approach key figures in a community, who in turn report signals of radicalization in their community. We also hypothesize that
interventions that wait for individuals to approach them stigmatize least.
Scale. Interventions can take place on a small scale, such as in a single neighborhood, school, or religious venue. An example is the empowerment training “Me, my identity, my choices, my chances” for a small group of young women in Amsterdam (Gemeente
Amsterdam, 2016). On the one hand, small-scale interventions may stigmatize more, because they single out a very specific target group as “risky” or “at risk”. On the other hand, such interventions may stigmatize less: because the target group is small, they are able to approach a larger proportion of the target group and are therefore less selective. In contrast,
interventions can also take place on a large scale. For instance, “ADFYWIAD” was a multi-faceted program implemented in Wales (Sheikh, Sarwar, & King, 2012). Such large-scale interventions may stigmatize less, because they focus on a more general group. Then again, such interventions may also stigmatize more, because they approach only a relatively small proportion of the target group. Because we are unsure which intervention characteristic may stigmatize more, we investigate them exploratively.
Focus. Some interventions target religion or ideology directly and aim to reduce radical interpretations. For instance, “Extreme Dialogue” is a series of Canadian documentary films that aims to counter the appeal of radical interpretations of Islam (Radicalization
Awareness Network, 2015). Unfortunately, such interventions may “significantly [fail] to address the deeper factors that lie within the wider context” (Powell, 2016, p. 51). In the case of Prevent, Bodi (2014) furthermore argues that “the state as some kind of high priest” (p. 10), deciding what are acceptable beliefs and what not, can lead Muslim communities to mistrust the government. Importantly, he fears that such interventions will inevitably stigmatize Muslim communities as “an inevitable source of terrorism” (p. 10).
Other interventions target religion or ideology indirectly and aim to prevent radical behavior through other aims, such as identity development, self-esteem training, or finding a job. An example is “Diamond”, an intervention that aims to increase participants’ resilience through identity development (Feddes, Mann, & Doosje, 2015). Such interventions do not stigmatize individuals by challenging their beliefs, but offer them skills that are generally useful in life. As such, we hypothesize that interventions that focus directly on religion or ideology stigmatize more than interventions that do so indirectly.
Executor. Interventions can be carried out by the police. For instance, the police can engage in “personal disruption”, including stop and searches, surveillance, and approaching family members of suspects (Lousberg, Van Hemert, & Langelaan, 2009). As argued above, police contact can have undesirable consequences, such as labeling and stigmatization (Atchison & Heide, 2011). In turn, individuals can change their self-concept and act on these changes (Ward et al., 2014; Lopes et al., 2012). Interventions can also be carried out by the municipality. Examples are both the aforementioned street coaches and the “Expertise training for mosque teachers” program, which aims to improve the educational skills of teachers in religious venues (Hermens et al., 2016). As Husband and Alam (2011) point out,
local authorities can have “impressive local knowledge, including sensitivity to local
community histories and their implications when dealing with current difficult issues” (p. 3). However, municipalities often communicate with the police, such as in the aforementioned “Allies” intervention, and ultimately stand outside the communities.
Finally, interventions can be carried out by a non-governmental organization. Exemplary is the “Hotline radicalization Assadaaka”, which supports individuals who are concerned with signs of radical behavior in their friends or family. Assadaaka takes special care not to be associated with the municipality or police, because otherwise their callers could fear that criminal procedures may follow. Importantly, callers also often “do not want to be associated with the topic ‘radicalization’” (Hermens et al., 2016, p. 27). My hypothesis dovetails nicely with these insights, namely that interventions carried out by the police stigmatize strongly. We further hypothesize that interventions carried out by the municipality stigmatize less, whereas interventions carried out by non-governmental organization are least likely to stigmatize.
Normative. Some interventions consider certain norms, values, and beliefs superior and teach them to participants. A clear example is the aforementioned Prevent intervention, which defines extremism “as the vocal or active opposition to fundamental British values, including democracy, the rule of law, individual liberty and mutual respect and tolerance of different faiths and beliefs” (HM Government, 2011, p. 62). Many have pointed out the ridiculously inclusive nature of these criteria (Open Society Justice Initiative, 2016; “Ken Macdonald QC launches Orwell Prize 2016,” 2015), yet Prevent’s values are intended to be taught actively. For instance, PACEY, the Professional Association for Childcare and Early Years, gives clear advice to teachers on how the values can be effectively taught to young children (Turner, 2015). The Office for Standards in Education (Ofsted, 2015) furthermore
clearly states that “schools will struggle to get a decent rating if they fail to deliver […] 'British values'” (Citizenship Foundation, 2017).
Other interventions do not consider certain norms, values, and beliefs superior and aim not to teach them, but instead discuss them (and others) with participants. Exemplary is “Project Safe Space”, in which individuals from different backgrounds openly discuss politics (UK Youth Parliament, 2009). As Thomas (2016) puts it, it is “better to enable young people to voice their feelings, even supposedly extremist or racist ones, so that they can be
interrogated and considered” (p. 182). In such interventions, participants may not feel judged or stigmatized a priori for their own norms, values, and beliefs. As such, we hypothesize that interventions that consider certain norms, values, and beliefs superior stigmatize strongly, whereas interventions that do not consider certain norms, values, and beliefs superior do so less.
Cultural aim. Interventions can aim to teach multiculturalism, namely the idea that different cultural, ethnic, and religious communities should be respected equally. For instance, the “Strengthening of intercultural contacts” intervention aims to increase understanding between different groups (KplusV, 2010). Such interventions foster respect between communities, but still emphasize the differences between them. Interventions can also aim to teach hybrid identities, namely the idea that individuals belong to multiple communities. Exemplary is the aforementioned Diamond intervention, which empowers individuals to strengthen their dual identities (Feddes, Mann, et al., 2015). According to Davies (2008), such interventions transcend simple multiculturalism: “[Interventions] should enhance the resistance to such simple labels and characterisations and give children status in the uniqueness and multiplicity of their hybrids” (p. 33). In summary, we hypothesize that interventions that aim to teach multiculturalism stigmatize weakly, whereas interventions that aim to teach hybrid identities stigmatize even less.
Matching. Some interventions match first-line professionals to participants on the basis of background, ethnicity, or other characteristics. For instance, the intervention “Training identity & resilience” employs trainers who have similar backgrounds as
participants (Gielen, 2009). Indeed, sometimes we should ask “who works?” instead of “what works?” That is, professionals report that matching can contribute significantly to the
effectiveness of interventions (IMPACT Europe, 2014). As an extension, participants of interventions can also feel less stigmatized when matching is done properly (Somers, De Wever, Bonte, & Creemers, 2013; Hurdle, 1991). Other interventions do not match first-line professionals to participants. In fact, a stunning 71% of interventions in a recent sample did not (IMPACT Europe, 2014). For instance, Dutch “community officers” are sometimes required to report signals of radicalization in their communities, but they often have different backgrounds than the individuals they are supposed to monitor (Dijkshoorn, 2016).
Differences between first-line professionals and participants can have negative effects on the relationship between the two and ultimately the effectiveness of the intervention (Lousberg et al., 2009). In conclusion, we hypothesize that interventions that match first-line professionals to participants stigmatize less than interventions that do not.
Current Study
Through a mixed-method approach, we intend to determine whether the
aforementioned intervention characteristics can stigmatize the target group. As such, we interviewed and surveyed individuals who are involved in preventive counter-radicalization interventions (see Figure 1). A number of researchers have studied reports and government documents to answer similar questions (e.g., Powell, 2016; Thomas, 2016), but to my
knowledge few have actually engaged with professionals who design, facilitate, and carry out preventive counter-radicalization interventions.
Figure 1. The mixed-method approach, namely Q-sort interviews followed by online surveys.
Study 1
In Study 1, we did not only ask participants about the potentially stigmatizing effects of intervention characteristics, but also about their effectiveness. We employed Q methodology (Shinebourne, 2009; Ellingsen, Størksen, & Stephens, 2010), which we will explain below. In sum, participants first rated intervention characteristics on their effectiveness and then on their potential to stigmatize. In a confirmatory fashion, Study 1 allowed for a preliminary test of the hypotheses. In an exploratory fashion, it allowed participants to suggest conditions, contexts, and so forth that might mediate or moderate the effects of these intervention characteristics.
Method
Participants. As Q does not aim to determine causality or generalize to a population, few participants suffice (Ellingsen et al., 2010). Q ideally engages individuals with different backgrounds to yield “viewpoints, helpful comments, and additional statements from a variety of perspectives” (Shinebourne, 2009, p. 94). Additionally, it is important to focus on
individuals actually involved in interventions, because their consequences can only be
“deduced from study of how [the intervention] has been understood and operationalized at the ground level” (Thomas, 2016, p. 175). In The Netherlands, we therefore interviewed two
policy advisors for multiple municipalities and one director and first-line professional of a non-governmental organization that carries out interventions.
Materials. We developed a Q sample, namely a set of 19 cards in 8 categories that each featured an intervention characteristic (see Appendix A). For Target group, one card read “The intervention is aimed at a target group on the basis of ethnicity, religion, ideology, etc. For example, a group of devout youth with a Turkish ethnicity or a group of youngsters who voice their opinions on animal rights.”; for Initiation, one card read “The intervention approaches individuals (and not the other way around).”; for Scale, one card read “The intervention takes place on a small scale, such as a neighborhood, school, or mosque. The intervention is therefore aimed at a small target group. Because the intervention is aimed at such a small group, it can select a relatively large proportion of this group to participate.”; for
Focus, one card read “The intervention is directly aimed at the religion, ideology, or
worldview of individuals. The intervention aims to reduce radical interpretations of a religion, ideology, or worldview.”; for Executor, one card read “The police carry out the intervention. For example, the local police or community officers.”; for Normative, one card read “The intervention is normative. That is, she deems certain norms, values, and beliefs per se superior in comparison to others. As such, she teaches these norms, values, and beliefs to
participants.”; for Cultural aim, one card read “The intervention teaches multiculturalism: different cultural, ethnic, and religious communities are of equal worth. The intervention stimulates participants to respect other communities.”; and for Matching, one card read “The intervention matches first-line professionals to individuals of the target group. Matching takes place on the basis of, for example, background, ethnicity, religion, or ideology.”
We also created two distribution grids, namely one for the effectiveness of the
intervention characteristics and the other for their potential to stigmatize (see Figure 2). The grids function much like Likert scales. That is, participants order all the cards on the grids
depending on the extent to which they deem the interventions characteristics on the cards not effective or not stigmatizing (left) or effective or stigmatizing (right). The grids have quasi-normal shapes in order to stimulate participants to differentiate between the cards.
Figure 2. Two distribution grids for Study 1, namely one for the effectiveness of the
Procedure. After signing an informed consent, participants carried out two Q-sorting tasks. First, participants ranked the cards with the intervention characteristics from -4 (“Not effective”) to 0 (“Neutral”) to +4 (“Effective”). Second, participants ranked the same cards from -4 (“Not stigmatizing”) to 0 (“Neutral”) to +4 (“Stigmatizing”). In post-sorting
interviews, we asked participants to comment on the cards. We audio-recorded the Q-sorting tasks and post-sorting interviews.
Results and Discussion
Target group. According to interviewee 1, there are two reasons why interventions that focus on individuals with a specific ethnicity, religion, or ideology could be effective. First, specific target groups struggle with specific challenges. For instance, young individuals with an Islamic background may struggle with the meaning of haram, halal, etc. These individuals can most effectively deal with such challenges together with individuals with a similar background. Second, interventions can only empower individuals effectively when these individuals feel safe, which again requires a group of individuals with similar
backgrounds. For instance, young individuals with an Islamic background may not feel comfortable discussing their views on gender relations with individuals who hold radical left-wing views. Although a specific focus may therefore be effective, interviewee 1 pointed out one reason why it might also stigmatize: “When you clearly select a group on the basis of their religion, ethnicity, or beliefs, then you do impose the ‘radicalization and extremism frame’ on them,” which might stigmatize.
Interviewees 1 and 2 rated interventions that focus on individuals deemed radical, regardless of their ethnicity, religion, or ideology, as generally effective and not stigmatizing, except for two reservations. First, religion and ideology can be important to individuals. As such, interventions would do well to take them into account, regardless of how they select their participants. Second, interventions that focus on individuals in general, regardless of
their ethnicity, religion, or ideology, can still stigmatize depending on how they define radicalization. As interviewee 2 put it: “When professionals only think that violent Jihadism is a risk, then they will still focus only on individuals with a Muslim background.”
Finally, all interviewees agreed that the last option was least stigmatizing, namely interventions that focus on individuals in general, regardless of their ethnicity, religion, or ideology and whether they are deemed radical. Although such a focus might prevent stigmatization, interviewees 1 and 2 pointed out that it is impractical to work with such a large target group: “It is hard to focus on those individuals who need it most, […] which might make it less effective.” Considering these three ways to select a target group, interviewee 2 argued that when you aim for behavioral change, resilience, or identity development, then you need to focus on a specific group. In contrast, when you aim to raise awareness or foster understanding between groups, then you can focus on more general groups. In conclusion, preventive counter-radicalization interventions with a general focus may stigmatize less, but they may also be less effective than interventions that focus on a more specific target group.
Initiation. All interviewees agreed that it is preferable to have individuals approach the intervention. Not only does this respect the individuals’ agency, but as interviewee 1 put it: “It would also increase their involvement.” However, preventive counter-radicalization
interventions can have many different goals, and some individuals “do not want any
assistance, so you will have to approach them.” Unfortunately, it may also be stigmatizing to approach individuals, as they might wonder whether they are approached “because of a certain label.”
Opinions were divided about interventions that approach key figures in a community and ask them to report signals of radicalization. On the one hand, such interventions were deemed effective and popular. On the other hand, interviewees 1, 2, and 3 identified three
risks. First, the key figures can be seen as informants not only by the intervention, but also by the community. In the words of interviewee 3: “It is important that they do not become some sort of extension of the [secret service]”. Second, the current interventions are all
institutionalized, which can lead key figures to experience loyalty conflicts. More specifically, the municipality may require too much of them, which leads their fellow
community members to brand them as traitors. Interviewee 3 had seen “key figures who were damaged and seen as traitors.” Third, some interventions only approached key figures with an Islamic background, which could have a stigmatizing effect.
The interviewees proposed conditions to improve the key-figure interventions. For one, the tasks of the key figures should not include reporting to the authorities unless absolutely necessary. Rather, key figures should be able to request training, support, and advice without providing personal details about community members. Furthermore, interviewee 1 pointed out that it is important to “distinguish professional key figures from informal ones.”
Especially with informal key figures, namely members of a community who do not work for an organization, the municipality should take extreme caution to prevent loyalty conflicts. In sum, it is preferable for participants to approach the intervention, although this is not always possible. Furthermore, key-figure interventions can be carried out, but the intervention should be careful with the key figures.
Scale. Interviewees then compared interventions that take place on a small versus large scale. All interviewees agreed that small-scale interventions are not only more practical than large-scale interventions, but small-scale interventions can also be adapted to different contexts. As interviewee 2 put it, small-scale interventions can be used for “a certain neighborhood or school, for instance because more signals of radicalization have been
reported there,” and “tailored to fit the needs of the target group.” Importantly, when first-line professionals work at a small scale, it may be preferable to involve someone on a higher level
for supervision and quality control. Unfortunately, a focus on a small group may also
stigmatize more than a focus on large group. For instance, a certain neighborhood can become known as a “radicalization neighborhood”. Regarding stigmatization, interviewee 3 pointed out that the goal of the intervention is important: “Do you want to empower the community [or] protect it from radicalization?” That is, an intervention may stigmatize less when it aims to empower the community and this is communicated towards the community.
Focus. Interviewees agreed that when interventions aim to deradicalize participants, then it can be effective to aim directly at the religion or ideology of individuals. In contrast, when interventions aim to prevent radicalization, interviewees deemed it better to not directly address the religion or ideology of individuals. In the words of interviewee 1: “If you’re talking about young people in phase 1, susceptible and sensitive, you should focus on other things, such as jobs, skills, and identity development.”
A direct focus on religion or ideology can also stigmatize. Interventions should therefore “not focus solely on ideology.” In practice, interviewees 1 and 2 pointed out that most Dutch interventions do not focus directly on religion or ideology. Moreover, these interventions often do not even focus directly on radicalization. That is, most interventions aim to achieve other goals, such as community building. To summarize, all interviewees thought it was better for preventive counter-radicalization interventions to focus on other goals than on religion or ideology.
Executor. Interventions can be carried out by the police. Interviewee 2 pointed out that “it can work well, but it can also be counterproductive.” Preventive interventions carried out by the police were deemed more likely to be effective when the police were not involved as the leading party. In addition, all interviewees argued that community officers were more suitable than other police, because “they play an important role in a neighborhood, know the young people there, and are closely tied to youth workers, teachers, and the municipality.” In
terms of stigmatization, interviewee 2 argued that “a confrontation with the authorities can cause stigma, because you are approached from a security perspective.”
Interventions can also be carried out by the municipality. Interviewees deemed this generally more effective than police involvement, because, as interviewee 3 put it, the municipality “is close to the target group, in the neighborhoods or schools.” Interviewees 1 and 2 stressed the importance of street coaches, who operate more informally than youth workers, can connect individuals to the system, and “have many contacts, as they are perceived as trusted sources.” Interviewees did not deem interventions carried out by the municipality particularly stigmatizing.
Finally, interventions can be carried out by a non-governmental organization. All interviewees deemed this preferable over the police or municipality, because organizations not related to the government are closest to the community. They emphasized that such interventions could benefit from support from the municipality or central government. However, when the municipality or central government only supports non-governmental organizations with a certain background, such as Islamic organizations, then that could set certain groups apart from others and stigmatize them as a result. Thus, interventions might be least stigmatizing when they are carried out by an organization not related to the government, followed by the respectively the municipality and police.
Normative. All interviewees agreed that interventions that teach norms, values, and beliefs can be effective. For instance, interviewee 1 pointed out that some individuals “have the feeling that their beliefs are superior, so you need to make them doubt those beliefs.” Interviewee 2 argued that non-normative interventions result in relativism, which is ultimately ineffective. Preventive counter-radicalization interventions might even be inherently
normative, because “you try to draw participants away from radical ideologies.” However, when interventions are too normative, then they can also lead to reactance and stigmatization.
As interviewee 2 put it: “Say there is a young individual who makes a rather radical statement the day after a terrorist attack took place; if you react very normatively, this individual may feel judged and singled out.” In conclusion, there is a dilemma, as indicated by interviewee 2: “On the one hand you want to stay in contact with these young individuals, to discuss with them and not judge too quickly, but on the other hand you want to set limits.”
Cultural aim. On the one hand, interviewees 1 and 2 argued that interventions that teach multiculturalism may “emphasize the community, which can be a buffer against
radicalization.” On the other hand, such interventions can also “stress the differences between communities,” which can cause individuals “to entrench in their own identities.” Similarly, interventions that teach multiculturalism can also stigmatize individuals, as these
interventions assume “clearly defined identities.” In contrast, interventions that teach hybrid identities can show that there “is more than just one truth, one religion, or one ideology.” Indeed, when individuals learn about hybrid identities, then interviewee 3 pointed out that they may “become more tolerant towards individuals with other identities.” All interviewees agreed that teaching hybrid identities does not stigmatize.
Matching. Interventions can match first-line professionals to individuals of the target group, for instance on the basis of background, ethnicity, religion, or ideology. According to all interviewees, matching is not always necessary, although it can be effective. First, it can be “important to the trainer to have the same background, in order for participants to feel safe,” as interviewee 3 put it. Second, individuals who are attracted to a certain ideology may identify themselves with others who do so as well. In that case, such individuals may connect more easily with first-line professionals who belong, or have belonged, to a similar group. Interviewees 1 and 2 thought that matching could potentially stigmatize, because individuals from the target group might think: “Why are we assigned someone from a similar
background?” In sum, matching might make the intervention more effective, although there is a chance that it can also stigmatize the participant.
Study 2
The Q-sort interviews gave insight into which intervention characteristics were deemed likely to be effective and stigmatize. More specifically, interviewees examined 19
intervention characteristics in 8 categories. In each of the 8 categories, the interviewees found differences in effectiveness and potential to stigmatize between the intervention
characteristics.
In Study 2, we examined a selection of these characteristics quantitatively and in a larger sample. More specifically, we selected three categories with seven intervention characteristics, namely Focus, Executor, and Matching. We selected specifically these
categories, because the interviewees indicated these as the most relevant and important. First, interviewees thought that the category Focus was relevant, because an intervention could stigmatize severely when it focused only on religion or ideology. Second, given the potentially severe consequences of police contact and the positive potential of
non-governmental organizations, the category Executor was deemed important. Third, given that interviewees were unsure whether Matching could stigmatize or not, the online survey may provide a more decisive answer.
We administered an online survey to professionals who design, facilitate, and carry out preventive counter-radicalization interventions. In short, participants indicated to what extent they deemed a selection of intervention characteristics likely to stigmatize. Participants also indicated the sectors they worked in, the levels in their organization they occupied, and what ideologies they mainly focused on. Study 2 allowed for a confirmatory test of the hypotheses.
Method
Participants. We approached individuals involved in preventive counter-radicalization interventions in The Netherlands. Because we anticipated difficulties in approaching a sufficient number of these individuals, we approached students of the social sciences as well. Although these students may have little knowledge of interventions, they are familiar with social processes such as stigmatization. A power analysis with G*Power (Faul, Erdfelder, Buchner, & Lang, 2009) suggested 90 participants for the paired samples t-tests, based on a two-tailed small-to-medium effect size of .3, conventional alpha of .05, and common power of .8.
88 participants completed the questionnaire. When asked to indicate what sectors they were active in, where multiple answers were possible, 31 indicated research, 15 indicated policy and advice (e.g., consultancy, central government), 8 indicated police, 32 indicated first-line professionals (e.g., social worker, education), and 26 indicated student. Participants also indicated at what level within an organization they were active (see Figure 3). Again, multiple answers were possible. First, 26 participants indicated to work as policy maker, researcher, or consultant. Second, 26 participants indicated to work as a professional without direct contact with the target group, such as train the trainer or research and development. Third, 23 participants indicated to work as professional with direct contact with the target group, such as trainers or mentors. Finally, when asked to indicate what types of extremism participants mainly worked with, where again multiple answers were possible, 34 indicated religious extremism, 23 indicated right-wing extremism, and 11 indicated left-wing
Figure 3. Sectors that participants were active in. (Multiple answers were possible.)
Materials. From the 19 intervention characteristics in the Q-sort interviews, we selected seven based on their importance and relevance. As explained before, we selected specifically these intervention characteristics because the interviewees indicated these as the most relevant and important.
Vignettes. For each of the seven interventions characteristics, we developed a vignette
(see Appendix B). In the category Focus, interventions could target religion or ideology 1) directly, in that they aim to reduce radical interpretations, or 2) indirectly, in that they aim to achieve other goals, such as proper identity development. In the category Executor,
interventions could be carried out by 3) the police, 4) the municipality, or 5) a
non-governmental organization. In the category Matching, interventions could 6) match first-line professionals to participants on the basis of background, or 7) not match first-line
Description of the intervention:
• The intervention focuses directly on religion or ideology. • And attempts to reduce radical interpretations.
Example 1
Youth workers talk about the Quran with youngsters.
Example 2
School coaches have discussions with youngsters who hold extreme right-wing beliefs.
Scales. For each of the seven interventions characteristics, participants answered four
items to indicate the extent to which they deemed the intervention characteristics likely to stigmatize. Two items were adapted from the labeling and discrimination subscales of the STIG-9 scale (Adeyemi, Abiola, & Solomon, 2015), namely “I think that most people…” followed by “Consider a participant of this intervention to be dangerous” and “Feel uneasy when a participant of this intervention moves into the neighborhood.” Two other items were adapted from the Anticipation of Rejection Scale (Link, Wells, Phelan, & Yang, 2015), namely “I think that a participant of this intervention…” followed by “Feels that most people would look down on him/her because he/she participates in this intervention” and “Worries that employers might not hire him/her if they knew he/she participated in this intervention.” Participants answered the items on a scale of 1 (“Strongly disagree”) to 4 (“Neutral”) to 7 (“Strongly agree”). For each of the seven interventions characteristics, all four items were combined into a single scale with good to excellent Cronbach’s alpha reliabilities ranging from .80 to .93 (see Appendix B).
Procedure. Participants were approached via e-mail. The e-mail contained a link to an online Qualtrics survey (see www.qualtrics.com). After signing an informed consent,
vignette, participants answered four items to indicate the extent to which they deemed the intervention characteristic likely to stigmatize. The categories were presented to the participants in a fixed order, but the intervention characteristics within the categories were randomized (see Figure 4). Participants then indicated which sectors they were active in, at which level within an organization they worked, and which types of extremism they mainly dealt with. Here, multiple answers were possible. Finally, participants could participate in a lottery for two times €25. On average, participants took 11.6 minutes (SD = 5.4 minutes) to complete the survey.
Figure 4. Randomization of the vignettes of intervention characteristics.
Results and Discussion
Given that a sufficient number of professionals involved in preventive counter-radicalization interventions filled out the survey (N = 62), we excluded student participants from the analysis. Moreover, to correct for six multiple comparisons, we applied the Bonferroni correction and set the significance level α to .008 (Schaffer, 1995).
Focus. We hypothesized that interventions that focus directly on religion or ideology stigmatize more strongly than interventions that do so indirectly and aim to prevent radical behavior through other aims, such as identity development. In a paired samples t-test, we compared to what extent participants rated these intervention characteristics as stigmatizing
(see Figure 5). In line with my hypothesis, participants rated interventions that target religion or ideology directly (M = 4.02, SD = 1.11) as more stigmatizing than interventions that do so indirectly (M = 2.77, SD = 1.26), at a highly significant level (t(61) = 6.87, p < .001, d = .87).1
Executor. We furthermore hypothesized that interventions carried out by the police stigmatize strongest, followed by respectively the municipality and a non-governmental organization. We employed a one-way within subjects (or repeated measures) ANOVA to compare whether participants rated these intervention characteristics differently (see Figure 5). As hypothesized, there was a highly significant effect between the intervention
characteristics (F(2, 122) = 81.16, p < .001, f = 1.15).
We carried out three paired samples t-tests to make post hoc comparisons between the intervention characteristics. These tests suggest that interventions carried out by the police (M = 5.42, SD = 1.01) stigmatize significantly stronger than both interventions carried out by the municipality (t(61) = 9.92, p < .001, d = 1.26) and a non-governmental organization (t(61) = 11.55, p < .001, d = 1.47). In addition, participants rated interventions that are carried out by the municipality (M = 3.79, SD = 1.34) as more stigmatizing than those carried out by a non-governmental organization (M = 3.33, SD = 1.21), albeit at a marginally significant level when correcting for multiple comparisons (t(61) = 6.87, p = .010, d = .34).2
Matching. Finally, we hypothesized that interventions that match first-line
professionals to participants on the basis of background, ethnicity, or other characteristics stigmatize less strongly than interventions that do not match first-line professionals to participants. In a paired samples t-test, we compared to what extent participants rated these
1 If we included the student participants in the analysis (N = 88), the results of the paired samples t-test were
similar (t(87) = 8.54, p < .001).
2 If we included the student participants in the analysis (N = 88), the outcome of the one-way within subjects (or
repeated measures) ANOVA remained similar (F(2, 174) = 127.65, p < .001). In addition, the three post hoc comparisons show similar results between interventions carried out by the police and both the municipality (t(87) = 12.47, p < .001) and a non-governmental organization (t(87) = 14.28, p < .001), and between the latter two (t(87) = 3.13, p = .002).
intervention characteristics as stigmatizing (see Figure 5). Contrary to my hypothesis, participants rated interventions that match first-line professionals to participants (M = 3.82,
SD = 1.08) as more stigmatizing than interventions that do not (M = 3.28, SD = 1.31), at a
highly significant level (t(61) = 3.41, p = .001, d = .43).3
Figure 5. For the intervention characteristics, their means, 95% confidence intervals, and
significant differences between them (N = 62; * is significant at p < .01; † is marginally significant when correcting for multiple comparisons).
Finally, participants also indicated the sectors they worked in, the levels in their organization they occupied, and which ideologies they mainly focused on. The results provide
3 If we included the student participants in the analysis (N = 88), the results of the paired samples t-test were
little to no evidence that participants rated interventions characteristics differently based on what sectors, levels, or ideologies they indicated.4
General Discussion
Many interventions aim to prevent radicalization. Such preventive
counter-radicalization interventions may have unintended consequences, such as stigmatization (Van Gorp & Roosendaal, 2013; Bodi, 2014). Through a mixed-method approach, I therefore investigated how preventive counter-radicalization interventions might stigmatize their target groups. To my knowledge, no study has specifically investigated stigmatization by
interviewing and surveying professionals who actually design, facilitate, and carry out preventive counter-radicalization interventions. Ultimately, this study could help determine which interventions are deemed successful without contributing to what they set out to prevent: stigmatization, polarization, and possibly even radical behavior.
Intervention Characteristics and Stigmatization
In the Q-sort interviews of Study 1, interviewees evaluated intervention characteristics in terms of effectiveness and stigmatization. In the category Target group, interviewees deemed interventions that focus on a group with a specific ethnicity, religion, or ideology as potentially effective, because participants may feel safer in a group with a similar background
4 For these analyses, we only included participants who indicated a single sector, level, and ideology to prevent
overlapping samples. In addition, to correct for 14 multiple comparisons, we applied the Bonferroni correction and set the significance level α to .003 (Schaffer, 1995). Between the five sectors that participants worked in (N = 58), one-way within subjects (or repeated measures) ANOVAs suggested no differences for the extent to which they rated the following intervention characteristics as stigmatizing: intervention with a direct (F(4, 53) = .46, p = .77) or indirect (F(4, 53) = 2.92, p = .03) focus on religion or ideology; intervention carried out by the police (F(4, 53) = .88, p = .48), the municipality (F(4, 53) = 1.11, p = .36), a non-governmental organization (F(4, 53) = 1.26, p = .30); and interventions with (F(4, 53) = .71, p = .59) or without (F(4, 53) = 1.30, p = .28) matching. Similarly, between the three levels in their organizations that participants occupied (N = 52), one-way within subjects (or repeated measures) ANOVAs did not suggest differences for the extent to which they rated the following intervention characteristics as stigmatizing: intervention with direct (F(2, 49) = 1.38, p = .26) or indirect (F(2, 49) = .52, p = .60) focus on religion or ideology; intervention carried out by the police (F(2, 49) = .06, p = .94), the municipality (F(2, 49) = .42, p = .66), or a non-governmental organization (F(2, 49) = .41, p = .67); and interventions with (F(2, 49) = 1.00, p = .38) or without (F(2, 49) = .08, p = .92) matching. Finally, for the three ideologies that participants could focus on, we could not test the differences due to strongly
overlapping samples. In other words, when participants indicated that they focused on a certain ideology, then they were likely to also indicate one or both of the other ideologies.
and may also be better able to deal with group-specific challenges together. In contrast, interventions can also focus on individuals who are deemed “at risk” for radicalization,
regardless of their background. A drawback of such interventions is that they can hardly focus on religion or ideology, whereas these may be important to participants. The same holds for interventions that focus on individuals in general, regardless of their background and whether they are deemed “at risk” for radicalization, although these interventions have the added disadvantage that it can be impractical to work with large groups.
Interviewees further argued that interventions would stigmatize more when they focus on a specific target group, because selecting a specific group sets it apart from others and may label it “risky” or “at risk” (Heath-Kelly, 2012, as cited in Blackbourn et al., 2012). A similar conundrum arises in the category Scale. On the one hand, the interviewees agreed that small-scale interventions are not only more practical than large-small-scale interventions, but small-small-scale interventions can also be adapted to different contexts. On the other hand, small-scale
interventions with a specific focus are more likely to stigmatize their target groups than large-scale interventions.
With regard to the category Initiation, interviewees agreed that it is preferable to have individuals approach the intervention, because this increases their involvement. However, “sometimes interventions must approach individuals,” as interviewee 1 put it, although interviewees admitted this could be “inherently stigmatizing” (Open Society Justice Initiative, 2016, p.16). With regard to key-figure approaches, interviewees expressed concern about the key figures: they could be seen as informants by the community, which could jeopardize “their standing within their own communities” (Husband & Alam, 2011, p. 3). As such, interviewees thought that key figures should only have to report to the authorities when absolutely necessary. Furthermore, key figures should be able to receive advice and support without disclosing fellow community members. Finally, interviewees 1 and 2 pointed out that
interventions should not limit themselves to recruiting key figures with an Islamic background, because that could label a group as “risky” or “at risk”.
In the category Normative, interviewees agreed that interventions may have to
consider certain norms, values, and beliefs superior and teach them to participants, because, in the words of interviewee 1, participants may “have the feeling that their beliefs are superior, so you need to make them doubt those beliefs.” However, when interventions are strongly normative, they can also lead to reactance and stigmatization. Indeed, such interventions may even be counterproductive: “It is better to enable young people to voice their feelings, even supposedly extremist or racist ones, so that they can be interrogated and considered”
(Thomas, 2016, p. 182). As such, interviewees acknowledged a dilemma: You want to discuss sensitive topics with participants and not stigmatize them, but you have to set boundaries.
With regard to the category Cultural aim, some interventions teach multiculturalism, namely the idea that different cultural, ethnic, and religious communities should be respected equally. As interviewees 1 and 2 pointed out, these interventions can “emphasize the
community, which can be a buffer against radicalization.” Yet they can also emphasize the differences between such communities, which can force individuals “to entrench in their own identities” and stigmatize them. Other interventions teach hybrid identities, namely the idea that individuals belong to multiple communities. These interventions may teach participants to resist simple labels. As a result, participants may become more tolerant towards individuals with other identities (see also Davies, 2008).
Three categories were investigated in both the Q-sort interviews and online surveys. With regard to the category Focus, participants in the online surveys rated interventions aimed directly at religion or ideology as more stigmatizing than interventions that do so indirectly. As one participant put it: “Intervening on religion or ideology will have little effect; I think that the crux of extremism lies in other issues.” Indeed, researchers agree that
interventions aimed at religion or ideology directly may fail “to address the deeper factors that lie within the wider context” (Powell, 2016, p. 51) and stigmatize communities as a “source of terrorism” (Bodi, 2014, p. 10). Relatedly, the Q-sort interviews suggest a possible moderator. When interventions aim to deradicalize participants, then it can be effective to aim directly at the religion or ideology of individuals. In contrast, when interventions aim to prevent radicalization, interviewees deemed it preferable to focus on other goals, such as community building, personal skills, identity development.
In the category Executor, participants in the online surveys rated interventions carried out by the police as more stigmatizing than those carried out by either the municipality or a non-governmental organization. In fact, police presence gave participants “a sense of unrest, as if something were very wrong.” This dovetails with the literature, where authors have reported that police contact can cause labeling (Atchison & Heide, 2011; Ward et al., 2014; Lopes et al., 2012). In the Q-sort interviews, interviewees agreed that contact with the police can stigmatize, because individuals “are approached from a security perspective.” Yet, the interviewees preferred community officers over other police, because they are closer to the target group. It should be noted that the police generally carry out different interventions than other executors. Although this could have confounded the results, the literature and Q-sort interviews suggest that the police can truly stigmatize more than municipalities and non-governmental organizations.
In the online surveys, participants rated interventions carried out by the municipality as more stigmatizing than those carried out by a non-governmental organization. However, it must be noted that this difference was only marginally significant when correcting for
multiple comparisons. Regardless, some participants remarked that municipalities can play an important role in preventing radicalization, as they may be ”close to the target group, such as in neighborhoods and schools” (see also Husband & Alam, 2011). However, municipalities
are ultimately not part of the community and they may coordinate with the police. It is therefore not surprising that Assadaaka, a Dutch radicalization hotline, takes special care not to be associated with the municipality or police, as their callers could fear criminal procedures or to be associated with the label “radicalization” (Hermens et al., 2016). In the Q-sort
interviews, interviewees deemed non-governmental organizations preferable over
municipalities for the simple reason that “organizations not related to the government are closest to the community,” as interviewee 2 put it. However, interviewees 1 and 2 emphasized that non-governmental organizations could benefit from support from the municipality or central government.
Finally, with regard to the category Matching, interviewees in the Q-sort interviews indicated that it can be important to match first-line professionals to participants on the basis of background, ethnicity, or other characteristics. Not only may participants connect more easily with first-line professionals who belong, or have belonged, to a similar group, but they may also feel safer. Relatedly, authors have stated that matching can contribute significantly to the effectiveness of interventions (IMPACT Europe, 2014; Lousberg et al., 2009). Somers et al. (2013) and Hurdle (1991) even argue that proper matching can reduce stigmatization.
Yet, the quantitative results of the online surveys show that participants rated interventions that match first-line professionals to participants as more stigmatizing than interventions that do not. The Q-sort interviews point to a possible explanation. Two
interviewees believed that matching could potentially stigmatize, because individuals might wonder why they are assigned a first-line professional with a similar background, ethnicity, religion, or other characteristic. Matching then tells the participant one thing: apparently this aspect of them is important to the intervention. Regardless of whether these participants thought so themselves, they are now treated differently on the basis of their background,
ethnicity, religion, or other characteristic. As such, matching could potentially stigmatize participants.
Limitations and Future Directions
Below, I address a number of criticisms. For one, I interviewed and surveyed professionals who design, facilitate, and carry out preventive counter-radicalization
interventions. Yet these are not the individuals who can potentially experience stigmatization as a result of preventive counter-radicalization interventions. That is, whereas my study aims to draw conclusions about individuals targeted by interventions, I have not consulted these individuals themselves. This is mostly due to practical considerations: not only is it difficult to reach individuals targeted by interventions, but it is also challenging to guarantee their privacy and not stigmatize these individuals by approaching them (Van Gorp, 2013; SAFIRE, 2013). Future studies could approach individuals targeted by interventions directly, although they might face the same challenges. A first step would be to set up an extensive ethics framework beforehand and not make the study about “radicalization” but rather about social issues in general. I hope that my intervention characteristics, materials, and results and future directions may provide a springboard for such studies.
Moreover, the online surveys may have measured intuition rather than expert
judgment. First, the interventions in the vignettes were described in general terms (e.g., “The police carry out the intervention”). Participants may therefore have answered abstractly and intuitively. One participant’s remark is exemplary: “I cannot answer these questions; there are multiple factors at play, […] much depends on recruitment, presentation, social skills, etc.” Similarly, many participants found it difficult to judge intervention characteristics when the target group was not clearly defined. As one participant put it: “The examples concern
religious and extreme-right backgrounds; but those are very different, which makes it difficult to answer the items.” Second, participants indicated the sectors they worked in and levels
within their organizations they occupied. The results provide little to no evidence that
professionals answered the items differently than students. Moreover, the results show little to no differences between the answers of professionals working in different sectors and at
different levels. As such, the question arises whether the online surveys may have measured intuitions rather than expert judgments.
In line with previous studies (e.g., SAFIRE, 2011; Bartlett, Birdwell, & King, 2010), I chose to investigate the underlying principles of interventions. These “intervention
characteristics” describe how an intervention is set up or carried out. In order to draw generalizable conclusions about the potential of interventions to stigmatize, it was therefore necessary to formulate the vignettes in general terms and leave the target groups unspecified. Still, I hope that future research will evaluate the stigmatizing effects of specific interventions in specific contexts and with specific target groups. Previous research has proved that
intuition can make way for expert opinions in such cases (e.g., O’Toole et al., 2015; Powell, 2016; Thomas, 2016).
Finally, an interesting alternative explanation for the results was put forward by several participants in the online surveys, namely the visibility of the intervention and its goals. In the words of one participant: “Stigmatization could occur only when the
participants’ friends and family know that the intervention aims to prevent radicalization.” This suggests that when an intervention is not visibly aimed at radicalization, it may not stigmatize: “A belief does not have to be visible, so how should their neighborhood know that he or she is an extreme-right person?” Future research could therefore examine how
interventions can prevent stigmatization by managing visibility and properly communicating its goals.
In addition to the recommendations outlined above, we propose three more future directions. First, every vignette featured two examples and participants indicated that
sometimes they had wished to answer differently for each example. As such, future studies could use only one example or none at all. Second, the target groups differed between the examples on the vignettes. For instance, one example focused on young people, whereas another focused on their parents. In order to properly compare intervention characteristics, future studies would do well to hold the target groups constant. Third, additional intervention characteristics could be important. For instance, the current study did not examine peer-to-peer approaches (see, for instance, www.plattelandsjongeren.nl). Furthermore, some communities set up interventions that are based on transformative justice and community accountability (see, for instance, www.phillystandsup.wordpress.com). Such interventions are interesting avenues for future research.
Implications
The current study adds to the literature in two ways. First, I conceptualized 19 intervention characteristics in 8 categories on the basis of the literature. Second, I used a subset of items from two scales about stigmatization (Adeyemi et al., 2015; Link et al., 2015) in a new domain, namely the evaluation of preventive counter-radicalization interventions. In terms of validation, the scales performed well with good to excellent reliabilities.
Practically, this study hopes to inform professionals who design, facilitate, and carry out preventive counter-radicalization interventions. Most convincing are the findings from the multi-method approach, namely Q-sort interviews followed by online surveys. More
specifically, in order to prevent stigmatization, professionals should aim for interventions that focus on other goals than reducing radical interpretations of religion or ideology, such as community building, identity development, and resolving moral dilemmas. Moreover, non-governmental organizations should be empowered to carry out interventions rather than the police. Similarly, an intervention carried out by the municipality is preferable over the police. In fact, this is in line with recommendations of a recent toolkit for police (TERRA, 2017).
