Population distribution of extremism

We are not aware of a previous population study that has investigated the distribution of extremist attitudes and associations with psychiatric morbidity. We showed that identification with British national or other cultural identity did not determine direction of support or opposition to the war but were associated with holding strong opinions. Men could have a strong British identity yet oppose the war, and vice versa. However, a combination of strong British identity and support for the war were associated with willingness to fight in the army; other cultural identity and opposition to the war with willingness to fight against Britain. We next investigated the distribution of these attitudes and confirmed a pyramidal structure within this population, corresponding to a postulated pyramidal model of distribution of support for terrorism. ^{Reference McCauley and Moskalenko4,Reference Leuprecht, Hataley, Moskalenko and McCauley9} A large number of men held neutral views at the base of this structure, progressing through increasing levels of opposition to the war and non-British identity, with willingness to fight against the British army at the apex (Fig.1). However, when we attempted to apply this structure to the larger group of men who supported their country, the pyramidal distribution (inverted) was not closely replicated. Despite declining support for the war by 2011, 1 in 6 young adult British men still reported they would fight in the army.

The stronger their pro-British attitudes, the more likely men were White and UK-born, but less likely to report higher educational qualifications or religious affiliation. They were more likely to report violence in the past 5 years, excitement from violence, criminal convictions and imprisonment. With increasing anti-British attitudes, men were more likely to be non-UK born, living in socioeconomically deprived geographical areas, from BME subgroups and religious. Nevertheless, they were similar to those with pro-British attitudes in being more likely to report violence and imprisonment.

Externalising and antisocial behaviour are not uncommon among applicants for military service in the USA and European countries. ^{Reference MacManus, Dean, Jones, Rona, Greenberg and Hull17,Reference Rosellini, Monahan, Street, Heeringa, Hill and Petukhova18} Our findings correspond with recent trends in recruitment to terrorist organisations, with decline in education and socioeconomic status, ^{Reference Christmann6} and need for those prepared to engage in open armed conflict rather than covert operations requiring specialist skills and subterfuge. They also correspond to successful targeting of prison populations for radicalisation and recruitment. ^{Reference Neumann19}

Extremism and depression

The key finding was that men at the bottom of this structure, with neutral or undecided views, (Fig. 1) were more likely to be depressed. Anti-British extremist views may have offered protection against depression, specifically among men of Pakistani origin. These findings correspond to the hypothesis that lack of personal identity and meaning, with unfulfilled need for belonging, create psychological vulnerability both to extremism ^{Reference Borum8,Reference Borum20–Reference Borum22} and anxiety and depression. ^{Reference Fisher, Overholser, Ridley, Braden and Rosoff23} Within this theoretical framework, attributing blame, identifying responsible perpetrators, ^{Reference Silber and Bhatt24} strong national or other cultural identity, and active support for or opposition to a cause, may protect against depression. For some men, depression may be a precursor to ‘mobilisation’, leading to active support for and consideration of involvement in terrorism or armed conflict along a pathway of radicalisation. ^{Reference Borum8} Lack of identity and uncertainty, together with depression, may contribute to a vulnerable state in which personal crisis can act as a trigger, resulting in an opening for new beliefs and values, ^{Reference Lotz, Baumert, Schlösser, Gresser and Fetchenhauer25–Reference Wilner and Dubouloz27} encouraged by people holding similar values that legitimise violence. ^{Reference Ranstorp28} Relatives' and friends' experiences of social exclusion, including poverty and reported experiences of racism, may have influenced these individuals to take a more active position. Factors such as turning to religion or new political beliefs triggered by a war (against people with similar cultural and religious characteristics) could result in a protective sense of empowerment involving new meaning, belief systems and identity ^{Reference Lotz, Baumert, Schlösser, Gresser and Fetchenhauer25} along a pathway ultimately leading to violent action. However, since we cannot determine the direction of association in this cross-sectional survey, respondents with depression may simply have been less likely to fight for or against their country or to hold extreme views because of their depression.

The other key finding observed towards opposite extremes in Fig. 1 was a linear association with antisocial personality disorder and violence. Decisions to fight for or against a country may involve a process of selection. Those with pre-existing antisocial and pro-violent dispositions may simply be more likely to hold extreme attitudes and be more suited for military action than those whose attitudes are neutral or who are depressed.

Religion

As expected, religious affiliation and practice was protective against externalising disorders such as alcohol dependence, drug misuse and antisocial personality disorder, in line with religious teachings. ^{Reference Herrenkohl, Tajima, Whitney and Huang29,Reference Kim-Spoon, Longo and McCullough30} Increasing anti-British extremist views were linearly associated with religion and religiosity, whereas increasing support for British intervention in Afghanistan was negatively associated. This polarisation of religious views may reflect ongoing processes of mutual rejection and an exclusionary circle in the British population, currently evident among many Muslim and non-Muslim communities. ^{Reference Pratt31} Within these communities, particularly those with multiple risk factors for radicalisation, extremism may result from perception of an aggressive, oppressive non-Muslim culture. A four-stage model of Al-Qaeda-influenced radicalisation proposes a pre-radicalisation stage of vulnerability followed by early exploration of Salafi Islam and gradual gravitation away from old identity, association with like-minded individuals and adoption of ideology. ^{Reference Silber and Bhatt24} This is followed by progressive adoption of jihadi-salafi ideology and finally jihadisation. ^{Reference Christmann6} Further investigation is needed into whether depression is intensified among Muslims by this mutual exclusionary process and whether radicalisation, with new sense of belonging and identity, ^{Reference Silke32} and willingness to take violent action are associated with protection and symptomatic improvement.

Experimental studies of support for suicide attacks have found that attendance at religious services corresponds as a function of collective religious activities, but differed from our findings in that frequency of praying does not. ^{Reference Ginges, Hansen and Norenzayan33} We did not investigate support for suicidal attacks. Furthermore, these studies found inconsistencies across different countries and that these specific associations with religious practice were strongest among Palestinians and Israeli settlers. In the UK, needs for excitement through violent action, camaraderie and strong masculine identity may, for some men, be more important proximal factors than religion, politics or cultural identity. Nevertheless, the latter may legitimise and ultimately determine targets of violence.

Limitations

Our survey had several limitations, including our method of defining identity where we combined ethnicity, cultural, religious and political beliefs. Although we defined support for armed conflict as ‘extremism’ for the purpose of our study, it could not be ascertained whether men reporting willingness to fight against the British Army actually intended to do so. Violent behaviour within the past 5 years was self-reported, without objective information such as data on arrest or convictions. Use of self-report captures more violence than comes to the attention of the authorities but may still have underestimated true prevalence. Socially undesirable behaviours may be underreported because of religious beliefs and cultural expectations.

Clinical presentations were derived from self-report questionnaires and not confirmed by clinical interview. However, clinician assessments compare favourably with self-report instruments. Given the cross-sectional nature of the data, we could not identify whether these measures were associated with changes in cultural, ethnic, religious and political views held by the respondent. Nonetheless, the community-based design and large sample size allowed examination of associations between different population subgroups holding different beliefs and mental disorder categories. Furthermore, sample size provided sufficient statistical power to test complex models and control for confounding.

Random location sampling does not provide detailed information on number of young men who declined to participate. However, because the method is based on the national census, participants were identified and included according to representative strata and their actual frequency in the population. Because young adult men of lower social class are more likely to decline participation in household surveys, this method has considerable advantages when investigating antisocial behaviour. The alternative would be to rely on a method requiring statistical weighting to adjust for attrition, which may be particularly high among this subgroup.

The finding that anti-British extremist views were negatively associated with depression among Pakistani men but not Black African men may be as a result of heterogeneity among countries included in the latter combined category. The BME boost sample, although large and representative, was underpowered to examine men from specific African countries with large Muslim populations and ongoing insurgencies. It was of interest that men of Bangladeshi origin were no more likely to hold anti-British extremist views or to have depression than White UK-born men. However, it would be necessary to conduct further studies with larger BME samples to investigate the roles of differing experiences of racial discrimination, social mobility and ethnic density effects across certain BME subgroups. It would also be necessary to investigate the effects of right-wing and anti-immigrant views on exacerbation of mutual rejection and an exclusionary circle. ^{Reference Pratt31}

Implications

Since withdrawal of coalition forces from Afghanistan, concern has grown over future support for terrorism within BME communities, particularly in Europe. Young men who travel to fight for Islamic State in Syria and Iraq, follow radical interpretations of Islam, and oppose political and economic interests of their host countries may return to threaten national security through insurgency. Religious identity may have become an increasingly important unifying and mobilising factor since 2011, when shared language, geographical and cultural origin, and blood ties additionally contributed to willingness to fight against the British army in southern Afghanistan among young men of Pakistani origin. ^{Reference Siddique34} Our findings indicated considerable heterogeneity across the population among those expressing extremist views, with deep divisions in personal attitudes and identity corresponding to processes of mutual rejection and exclusion. Depressive symptoms are more prevalent among minority populations in Europe. ^{Reference Missinne and Bracke35} Separation with maintenance of ethnic culture but no participation in local culture is associated with mood and anxiety disorders. ^{Reference Ünlü Ince, Fassaert, de Wit, Cuijpers, Smit and Ruwaard36} Further research is needed into interventions that reduce discrimination, exclusion and extremism in these communities and associated depression.

Extremism can spread to those with moderate views through epidemic transmission when culture, religious and political values shift quickly. ^{Reference Lafferty, Smith, Madin, Sagarin and Taylor37} Views and opinions initially considered extreme may become the norm. ^{Reference Johnson, Madin, Sagarin and Taylor38} Simulated models suggest that when a large part of the population hold moderate views or are uncertain, extreme views tend to prevail. This leads to either convergence on a single extreme or to bipolarisation. ^{Reference Deffuant, Amblard, Weisbuch and Faure39} Bipolarisation implies strongly opposing views with neither achieving dominance.

Endorsement of willingness to fight for one's country would not be interpreted by most UK people as ‘extremism’ but willingness, if necessary, to make an extreme sacrifice for the benefit of, and to defend, the majority. Fusion theory would suggest that the asymmetrical distribution in Fig. 1 was explained by more UK men being strongly ‘fused’ through their personal and social identity with UK values to make such sacrifices. ^{Reference Swann, Buhrmester, Gómez, Jetten, Bastian and Vázquez40} However, this process is thought to depend on whether members of a group share the same core characteristics and values. Individuals in the group are more likely to fight and die for family members, or those with whom they perceive close ties similar to those in families within small social groups than for their country. Promoting shared values and perceived similarity in personal identity may therefore be important in preventing processes that lead to bipolarisation. Our study may be a starting point for more sophisticated investigation into whether these processes are failing in European countries, whether bipolarisation has accelerated and intensified within vulnerable communities with high levels of depression, leading to convergence on a single extreme, and increasing numbers of radicalised young men prepared to engage in armed conflict.