Dark personalities and their sympathies towards state-sponsored extremism

ABSTRACT

There is little empirical research conducted on state-sponsored extremism (SSE). Yet, SSE can be seen in society from policy to policing that can lead to discrimination and aggression. The purpose of the study was to investigate if those higher on the Dark Tetrad (D4; Machiavellianism, narcissism, psychopathy, and sadism) and right-wing authoritarianism (RWA) would demonstrate more support for SSE. We hypothesized those who were higher on the D4 and RWA would show more sympathy towards SSE. We also investigated if cognitive suppression use could predict more support for SSE scenarios. In a cross-sectional design with vignettes, participants indicated their level of agreement with four real-world examples of SSE. A total of 398 participants (287 women, 103 men, 6 non-binary, and 2 who did not indicate gender) from 43 countries were recruited online. We ran a series of hierarchical regressions to develop predictive models to test our hypotheses. Being higher on RWA was a significant predictor that explained the most variance in nearly all models. Being higher on the D4 also predicted more support for SSE in select circumstances. Our findings contribute to the understanding of dispositional dynamics and interpersonal differences in the legitimization of SSE.

KEYWORDS:

• State-sponsored extremism
• Dark Triad
• Dark Tetrad
• Right-wing authoritarianism

There has been a global movement towards right-wing populist style governments over the past decade. Right-wing populism, as seen in Brexit and recent Turkish and Polish elections, tends to be anti-globalization and anti-immigration while being more nationalistic and authoritarian (Merelli, 2019). Right-wing governments tend to have more aggressive leadership and policies to preserve or re-instate their nationalistic ‘pride’ (Blee & Creasap, 2010) with antagonistic rhetoric (Lammers & Baldwin, 2020). These governments often target already marginalized segments of the population, increasing tensions, divisiveness, and conflict, often leading to an increase in violence and state-sponsored extremism (SSE; Sandel, 2018). There are no universally accepted definitions of SSE (see Sotlar, 2004 for a discussion on the difficulties of defining extremism). For this paper, SSE is defined as a situation where a state officially supports or partakes in violent and/or non-violent actions that attempt to harm (physically or psychologically) another state or its own residents or citizens but outside formally declared war.

There are numerous forms of extremism from ecoterrorism, right-wing radicalization, political extremism, to genocide. Previous research has found perpetrators (i.e. active actors) and sympathizers (i.e. passive actors who tolerate/support the perpetrators) of one form of violent extremism have attributes that will differ from perpetrators of another form (Boylan, 2015; Clemmow et al., 2020). For example, in a Flemish sample, those who engaged in political violence were found to be less socially integrated in that society (De Waele & Pauwels, 2014). The focus of this paper is attitudinal support for SSE and traits that may explain individual differences in support for SSE.

Dark Tetrad and extremism

There have been decades of research that have shown that those high on the Dark Tetrad (D4; Machiavellianism, narcissism, psychopathy, and sadism) are more likely to aggress when a situation presents itself albeit with different triggers (Buckels et al., 2013; Jones & Paulhus, 2011, 2010; Jonason & Webster, 2010. See Paulhus & Williams, 2002 for an overview of the etiology of the D4 and how the traits differ from clinical diagnoses). While research has shown the D4 share a conceptual resemblance, each trait has distinct facets and adds to the construct of the so-called ‘dark personality’ (Buckels et al., 2013; Jones & Paulhus, 2011). Those high on Machiavellianism tend to manipulate in a cold, callous manner and only aggress if they perceive the benefits to outweigh the costs with no fear of being caught or retaliated against (see Tetreault et al., 2021 for a table that outlines the key attributes of the D4 traits and the triggers). Core traits for those who are higher on narcissism demonstrate a sense of superiority and entitlement. While they only aggress in response to an ego threat (e.g. criticism of their work or social identity; Jones & Neria, 2015), they are hyperaware of ego threats and can perceive innocuous encounters as slights, triggering an aggressive response. People who are high on psychopathy demonstrate anti-social behavior and high impulsivity with low anxiety and empathy. They use violence instrumentally (i.e. act aggressively to achieve a goal) when it is low cost to themselves (Buckels et al., 2013). Lastly, the main core trait for those high on sadism is enjoying others’ suffering (schadenfreude). Those high on sadism aggress in any opportunity even if unprovoked or at a high cost to themselves, showing aggression is often the chosen behavior (Jones & Paulhus, 2010). Studies have found that those higher on sadistic traits, for example, show an increase in positive affect while aggressing and a decrease in positive affect when they were not (Buckels et al., 2013), suggesting they enjoy aggressing. The D4’s attributes potentially increase the likelihood of engaging in and sympathizing with perpetrators in differing forms of violent extremism, including SSE. In the current study, we hypothesize that those higher on the D4 traits will report more support for SSE.

Studies have investigated the D4 and the relationship between these traits and extremism, but the findings have been mixed. For example, in a French undergraduate sample of women, using cluster analysis, the psychopathic, sadistic, and Machiavellianistic trait cluster was found to be more associated with more radicalized cognitions and behavior (Chabrol et al., 2020). Morgades-Bamba et al. (2020) found in a study of radicalized women that the D4, in particular being high on sadism, was associated with having radicalized cognition whereas being high on narcissism was associated with radicalized behavior. Machiavellianism and psychopathy have been associated with more direct anti-social behavior and justifications for it (Klimstra et al., 2014), narcissism and sadism have been directly associated with negative attitudes towards immigrants (Petrović, 2019), and all D4 traits have been associated with a devalorization of others (Chabrol et al., 2020).

McGregor et al. (2015) reviewed how being high on the D4 traits may predispose people to violent radicalization. For Machiavellianistic traits, the authors postulated the core traits of low morality coupled with the high drive for power and authority could lead to radicalization. Narcissistic traits of feelings of being unjustly treated, aggressive reactions to perceived slights or discrimination, and devaluing others may contribute to radicalized beliefs and sympathies towards perpetrators. The remaining two psychopathic (low empathy, anti-social behavior, and callousness) and sadistic (enjoying other’s suffering) traits of the D4 could cause people high on the traits to engage in certain forms of violent radicalization and have positive regard for perpetrators of radicalization.

These findings point to the potential value of conducting further empirical research into the relationship of the D4 traits and extremism, and this is an aim of the current study. As we are investigating attitudes and sympathies towards those who support SSE in the current study, we hypothesize that the D4 will show more agreement. While previous literature has shown that the D4 do link to various forms of extremism, there are gaps in the literature that we aim to address as well. There is a tendency to rely on student samples, and these samples may not be reflective of society at large. Some have used other samples, such as soldiers (e.g. Lindén et al., 2019), but they are not common. In extremism research, there is more focus on individual perpetration with very little on SSE, which is in part why we focused on SSE. Lastly, some researchers still exclude sadism in their research and focus on the Dark Triad. Research has demonstrated that those high on sadism are particularly aggressive and even work to inflict pain and violence because they enjoy it (Chester et al., 2019). We suggest sadism could be an important trait that is associated with more sympathies towards SSE and, as such, have included it in the current study.

Right-wing authoritarianism and extremism

Right-wing authoritarianism (RWA) is a concept originating from the 1950s (Adorno et al., 1950). RWA has received a lot of attention in research, and it has developed into a stable construct for individual differences and attitudinal beliefs (Bizumic et al., 2009). Being high on RWA is associated with favoring traditionally socially conservative values, adherence to laws, and punitive punishments for breaking conservative values and norms (Duckitt, 2010). Those high on RWA lean towards ‘Right’ conservative parties, favoring strict immigration policies, nationalism, and ethnocentrism. In contrast, those low on RWA lean towards ‘Left’ liberal and progressive parties, favoring diversity and more lenient immigration policies while disagreeing with nationalism and militarism. Additionally, those high on RWA tend to be more likely to view the world through in-group/out-group eyes, where out-groups are seen as threatening to their way of life (Altemeyer, 1998; Whitley, 1999). Furthermore, those high on RWA tend to view their way of life and their in-group as morally superior, which justifies in their eyes expressions of prejudice against an out-group who is either deemed directly by them or authorities as being or behaving as inferior (Altemeyer, 1998; Whitley, 1999). Such findings concord with studies that suggest that people who are higher on RWA may have more negative attitudes against same-sex couples (Whitley, 1999) and immigrants (Craig & Richeson, 2014), while simultaneously showing tolerance and support of unethical war behavior (Lindén et al., 2016) and aggressive behavior from authorities (Altemeyer, 1998). All these factors could culminate into an RWA disposition that would demonstrate sympathies towards the SSE we are investigating in the current study.

RWA has some characteristics that are also found in the D4 (e.g. lower empathy, want of power), yet RWA and the D4 also tap into different facets of dark traits (Lindén et al., 2019; Zeigler-Hill et al., 2020). For instance, narcissism was the only D4 trait found to be weakly, positively associated with RWA (r = .17 or .26, p < .001) in two different samples: one an American undergraduate sample and the other a mix of an American undergraduate sample and adults recruited through Mechanical Turk (Zeigler-Hill et al., 2020). However, in a study of Swedish soldiers, Machiavellianism (r = .27, p < .001) and psychopathy (r = .31, p < .001) were positively correlated with RWA (Lindén et al., 2019). In that study, those high on the Dark Triad and RWA along with the dominance aspect of social dominance orientation contributed to creating a latent ‘core of darkness’ personality that predicted both unethical warzone behavior and support for unethical behavior within the military organization. These findings also suggest that while there are shared features between RWA and the D4, there are distinctive aspects that need to be further understood in how they relate with extremist thoughts, support, and/or behavior in order to counteract them. Thus, the current study seeks to explore the relationship between the facets of the D4 and RWA, and their predictive power of support for SSE.

Dark Tetrad, right-wing authoritarianism, and suppression

The research on the D4 has demonstrated that the D4 tend to ‘release’ their aggressive tendencies in a variety of situations, such as killing bugs (Buckels et al., 2013), blasting white noise (Chester et al., 2019), and intimate partner violence (Tetreault et al., 2021). We are proposing based on this tendency to ‘release’ aggressive behavior in select situations according to the aforementioned unique triggers that there is an underlying theoretical supposition for those who are high on the D4 to engage in suppression regularly. This habitual suppression of aggressive behavior then acts as a coping strategy that allows people higher on D4 traits to function more easily within society than their clinical counterparts. While suppression has been shown to be a useful short-term coping strategy, it is not an effective long-term strategy (Geraerts et al., 2006; Mund & Mitte, 2012) as we discuss more below. No research to the best of our knowledge has been conducted to determine if those higher on the D4 use suppression. We aim to fill this gap.

For RWA, there is no theoretical background on the use of suppression in the same manner as the D4, but there have been a few studies that investigated suppression and RWA. One study found that RWA was not correlated with either externally or internally motivated suppression (Webster et al., 2014). They did not examine suppression and RWA in their other analyses. However, other research found that more anti-social attitudes emerged for those who were higher on RWA when external influences (e.g. threatening situations) mediated by beliefs the world was dangerous were present (Duckitt & Fisher, 2003). This could indicate that those higher on RWA may be suppressing anti-social thoughts, feelings, beliefs, or behavior in certain atmospheres and not in others. The study did not investigate if these effects were retained long term nor if there were any negative effects of the suppression.

If people higher on the D4 and RWA do consistently suppress their behavior to cope and fit in better with societal norms, we believe this could lead to rebound effects through hyperawareness. Rebound effects, an increase in the thoughts, feelings, or behaviors one is actively trying to suppress, have been shown to be a long-term effect of suppression (Wegner et al., 1987). We have theorized that long-term suppression of thoughts, feelings, or behaviors creates a hyperawareness to find situations to ‘release’ what is trying to be suppressed (i.e. rebound). The hyperawareness could lead to seeking out situations to aggress or support aggression. We argue that this hyperawareness results from the incongruence between their authentic selves (e.g. wanting to aggress) and the behavior they are ‘forced’ to suppress to align with societal norms. Thus, suppression leads to a sense of incongruence between how one wants to behave and how one feels they can behave, leading to rebound effects by creating a hyperawareness to find a way to release their natural propensities. In other words, people who are more susceptible to behaving or responding aggressively, such as those higher on the D4 or RWA, will look for opportunities to do so or interpret ambiguous situations as times to react aggressively. While aggressing, those high on the dark personalities are no longer suppressing, which could dissipate the effects from the rebound effect (i.e. their behavior and thoughts are congruent). For example, someone who is high on sadism may become hyper vigilant to find opportunities to inflict pain, such as cyberbullying (van Geel et al., 2017). We examined if those higher on the D4 and RWA engaged in suppression and if suppression could predict more support for SSE (i.e. rebound), as this is missing from the literature.

Research questions and hypotheses

There were two main research questions: (RQ1) which factors and traits could predict passive support of SSE? (RQ2) Could those who score higher on the D4 and RWA be at greater risk of sympathizing and showing more support for SSE? There are four main overarching hypotheses: Machiavellianism and narcissism will be positively correlated with RWA; whereas psychopathy and sadism will be negatively correlated with RWA (H1). Higher scores on the D4 and RWA will be associated with higher use of suppression (H2). Higher scores on RWA and the D4 will predict more agreement with SSE with each vignette garnering support from people higher with different traits (H3; see Table 1 for sub-hypotheses for each SSE scenario). More suppression use will predict more agreement to SSE (H4). While we did not hypothesize age, gender identification, education level, or whether currently living in the same country as one grew up would be significant predictors of support for any SSE, they were important to include. Research has shown that younger males tend to be more aggressive against others (Moffitt, 2001). Additionally, as some of the vignettes included SSE that specifically targeted immigrants, it was important to include whether or not participants were living in the country they grew up in case this aspect influenced how they indicated support (or not). The only demographic variable we hypothesized to be a significant predictor was self-placement on the political spectrum. We hypothesized more right-leaning on the political spectrum would show support in the same pattern as those higher on RWA.

Table 1. H3 Sub-hypotheses for support of the four SSE scenarios.

	SSE 1 (violence against protesters)	SSE 2 (use of arms against another country)	SSE 3 (reversed, volunteers saving refugees)	SSE 4 (law forcing indoctrination)
Higher Machiavellianism	Will predict more support. Rationale: due to the want of power with little risk of being caught	Will predict more support. Rationale: due to the want of power	Not hypothesized to be associated	Not hypothesized to be associated
Higher narcissism	Not hypothesized to be associated	Not hypothesized to be associated	Not hypothesized to be associated	Will predict more support. Rationale: not assimilating could be perceived as a threat to identity, which could trigger a response
Higher psychopathy	Will predict more support. Rationale: due to the use of physical instrumental violence	Will predict more support. Rationale: due to the use of physical instrumental violence	Not hypothesized to be associated	Not hypothesized to be associated
Higher sadism	Will predict more support. Rationale: due to the fact people would be harmed	Will predict more support. Rationale: due to the fact people would be harmed	Not hypothesized to be associated	Not hypothesized to be associated
Higher RWA/more right-leaning on the political spectrum	Will predict more support. Rationale: favor more traditional or aggressive actions against out-groups	Not hypothesized to be associated	Will predict less support (or more against). Rationale: volunteers assisting ‘illegal’ immigration and higher RWA/right leaning tend to be anti-immigration	Will predict more support. Rationale: tend to be less open about immigration and immigrants especially if immigrants are perceived as not assimilating
More suppression	Will predict more support	Will predict more support	Will predict more support	Will predict more support
Demographics (age, gender, education, living in same country as grew up)	Not hypothesized to be associated	Not hypothesized to be associated	Not hypothesized to be associated	Not hypothesized to be associated

Methods

Participants and procedure

In this cross-sectional study, participants provided demographic information, then read four vignettes on SSE with two to three sub-questions each (see Appendix). The vignette presentation was randomized to avoid any order effects. The next section contained the Short Dark Triad (SD3) and Short Sadistic Impulse Scale (SSIS) with all items being randomized. The Right-wing Authoritarianism scale and White Bear Suppression Inventory (WBSI) were the last two sections. Within each of those, the items were randomized. Participants were asked to place themselves on the left-right political spectrum on an 11-point Likert scale with 1 being completely liberal, 6 being neither left nor right, and 11 being completely conservative (based on Kroh’s [2007] and Vegetti & Širinić’s [2019] recommendations). Participants were lastly asked if they wished to receive course credit or enter a draw for a gift card. Full ethical approval was granted by the Research Ethics Committee at NUIG before online data collection through social media and the human subject pool commenced.

There were a total of 404 participants with 398 participants¹ (287 women, 103 men, 6 non-binary, and 2 who preferred not to indicate how they identified) being retained for analyses. There were numerous univariate and multivariate outliers in our data. For instance, selecting a 2 on the 10-point Likert scale for SSEs was enough statistically to have participants identified as extreme outliers. Upon inspection of the standardized residuals of all variables, we determined the residuals were within the expected distribution parameters, and no participants were removed for being outliers.

The average age of the participants was 32.75 (SD = 12.52, range 18–78). The majority identified as being White (85.2%); other, including mixed race (7.8%); Asian (5.5%); Black (1.3%); or White Irish Traveler (.3%). Participants grew up in 43 different countries with Ireland (n = 198), Canada (n = 73), and United States (n = 39) representing the three most indicated.² The participants were currently living in 24 different countries with Ireland (n = 257), Canada (n = 43), United States (n = 30), and Sweden (n = 28) being the top four. For the highest level of completed education: 5.5% had completed a PhD, 33.4% a Master’s degree, 32.4% an undergraduate degree, 27.5% high school, and 1.0% some high school. In terms of placement on the political spectrum, the mean was 3.79 (SD = 1.97).

Measures

To measure participants’ level of agreement to SSE, participants were asked to read four vignettes that were closely based on real-world examples (see Appendix) and indicate their level of agreement on 10 sub-questions on a 10-point Likert scale (1 – not at all to 10 – completely). The four vignettes investigated four types of SSE: violent SSE, financially supporting SSE, non-violent SSE, and political policies that could instigate SSE. The first vignette was inspired by the 2020 US election and other countries who have used their military or police forces to ensure the incumbent government is re-elected, regardless of how much violence is necessary to examine support for violent SSE. Vignette 2 was Canada’s arms’ deal with Saudi Arabia, where Canada is selling arms to Saudi Arabia for large financial gains while Saudi Arabia is using these arms to target Yemen to gain control over the area (Guardian, 2020), investigating support for intergovernmental agreements that financially support SSE. Vignette 3 was regarding Italian volunteers who go into the Mediterranean to rescue migrants from ships that are in distress and take them to larger ships that then take them either to Greece or Italy (Trilling, 2020), examining support for non-violent SSE (reverse scored). Vignette 4 was Denmark’s law (LOV nr 1322 af 27/11/2018) that forces babies of immigrants who are living in ‘ghettos’ starting at the age of 1 to attend 25 waking hours per week of assimilation education to ensure they adopt Danish culture and values (Barry & Sorensen, 2018) to explore political policies that could instigate SSE. The countries’ names were removed to prevent any pre-existing bias.

We used two scales to measure the D4 personalities. The first was the SD3 (Jones & Paulhus, 2013), a 27-item scale, consisting of 3 subscales, Machiavellianism, narcissism, psychopathy, each containing nine items. Participants were asked to indicate their level of agreement to each statement on a 5-point Likert scale, ranging from 1 – disagree strongly to 5 – agree strongly. Some examples are: ‘I like to use clever manipulation to get my way,’ (Machiavellianism subscale, α = .75); ‘I insist on getting the respect I deserve’ (the narcissism subscale, α = .66); and ‘Payback needs to be quick and nasty’ (psychopathy subscale, α = .63). We used the SSIS (O’Meara et al., 2011, α = .83), which consists of 10 items (e.g. ‘I enjoy seeing people hurt’), scored from 1 (strongly disagree) to 5 (strongly agree).

The Short-version Right-wing Authoritarianism (Rattazzi et al., 2007, α = .84) is a 14-item scale, asking participants to rate their agreement on a 7-point Likert scale with −3 (totally disagree) to +3 (totally agree). ‘Obedience and respect for authority are the most important values children should learn’ is an example item. To assess people’s natural suppression tendency, the WBSI (Wegner & Zanakos, 1994, α = .93) was used. Participants rated their level of agreement on a 5-point Likert scale (1 – strongly disagree to 5 – strongly agree) on 15 items, such as ‘My thoughts frequently return to one idea.’

Results

In bivariate correlations, we found the D4 all significantly positively correlated (ranging from weak to strong) with one another (see Table 2). RWA was weakly positively correlated with Machiavellianism and narcissism. Suppression was also weakly positively correlated only to Machiavellianism, psychopathy, and sadism.

State-sponsored extremism

We are presenting the analysis for four sub-questions (see supplemental material for the others), one from each vignette. Unfortunately, there were too few participants who identified as non-binary or who preferred not to indicate their gender to include them in the regression analyses. RWA and where people would place themselves on the political spectrum were strongly positively correlated (r = .58, p < .001). However, they were not strongly correlated enough to cause multicollinearity issues in the regression analyses, and as such, both were included in the models (Tabachnick et al., 2007).

Our DVs were not normally distributed. While transformations improved the skewness and kurtosis, they did not normalize the data. There were no obvious patterns in the residuals to suggest homoscedasticity violations. To ensure no data points were having undue influence, we examined Cook’s distance. All were deemed acceptable (i.e. <1.0; Tabachnick et al., 2007). No data points had both high leverage and Cook’s distance, thus, we used the raw data.

We ran a series of hierarchical regressions to determine the variables that predict support for the scenarios. In model 1, the demographic and background variables: gender identification, age, education level, and living in the country where they grew up or not, and the political spectrum placement were included. Model 2 added Machiavellianism, narcissism, psychopathy, sadism, and RWA. In the final model, suppression (measured by the WBSI) was added. The model construction is in line with the theoretical background and our hypotheses.

Vignette 1 – agreement with the army’s actions to use force to end political protests

All three models had significant changes, F(5, 384) = 10.215, p < .001 (model 1); F(5, 379) = 6.113, p < .001 (model 2); and F(1, 378) = 6.603, p = .011, (model 3; see Table 3). In the final model, 17.4% (adjusted R²) of the variance was explained with being younger and more right-leaning on the political spectrum, higher sadism and RWA, and less suppression being significant predictors for support of the army’s use of violence against the protesters.

Table 2. Correlations between the Dark Tetrad, RWA, and suppression tendencies.

	Mach.	Narcissism	Psychopathy	Sadism	RWA
Narcissism	.34**
Psychopathy	.47**	.25**
Sadism	.45**	.20**	.64**
RWA	.15*	.15*	.06	.09
Suppression	.25**	.00	.25**	.24**	−.03

*p < .01 **p < .001 Correlation is significant at the .01 level (2-tailed). Mach. = Machiavellianism.

Vignette 2 – Middle Eastern country’s use of arms to control neighboring area

In the regression, models 1 and 2 had significant changes, F(5, 384) = 4.051, p = .001; F(5, 379) = 3.712, p = .003; and F(1, 378) = .170, p = .68 for model 1, 2, and 3 respectively (see Table 4). More right leaning and higher narcissism were the only significant predictors in every model in which they were included, explaining 6.9% of the variance in the final model. This is the only DV where RWA was not a significant predictor though it neared significance.

Table 3. Coefficients for agreement with army’s actions to use force to end political protests.

Model		1*				2*				3**
Predictor	B	SE	beta	sig	B	SE	beta	sig	B	SE	beta	sig
Gender ID	−.44	.17	−.12	.01	−.13	.18	−.04	.48	−.01	.18	.00	.95
Age	−.01	.01	−.10	.05	−.01	.01	−.10	.04	−.02	.07	−.14	.01
Education	−.16	.09	−.09	.07	−.09	.08	−.05	.31	−.10	.08	−.06	.24
Same country	.15	.17	.04	.39	.15	.17	.04	.37	.15	.17	.04	.36
Pol. Spectrum	.23	.04	.28	<.001	.16	.05	.20	.001	.16	.05	.20	.001
Mach.					−.01	.02	−.04	.51	−.01	.02	−.02	.73
Narcissism					−.02	.02	−.07	.19	−.03	.02	−.08	.10
Psychopathy					.03	.02	.09	.18	.04	.02	.11	.09
Sadism					.07	.02	.20	.002	.08	.02	.23	<.001
RWA					.02	.01	.15	.01	.02	.01	.15	.01
Suppression									−.02	.01	−.14	.01
R^2		.117				.183				.197
Adjusted R^2		.106				.162				.174

*p = .001 **p < .01.

Vignette 3 – disagreement with volunteers rescuing migrants in the Mediterranean

The models with significant changes were model 1, F(5, 384) = 11.028, p < .001, and model 2, F(5, 379) = 8.897, p < .003 (see Table 5). Model 3 was non-significant, F(1, 378) = .075, p = .785. A total of 19.5% (adjusted R²) of the variance was explained. The significant predictors were higher education, Machiavellianism, and RWA.

Table 4. Coefficients for agreement with Middle Eastern country’s use of arms.

Model		1*				2**				3
Predictor	B	SE	beta	sig	B	SE	beta	sig	B	SE	beta	sig
Gender ID	−.13	.13	−.05	.30	.02	.13	.01	.89	.03	.14	.01	.81
Age	.00	.01	−.03	.59	.00	.01	−.01	.83	.00	.01	−.02	.75
Education	−.03	.06	−.02	.68	.00	.06	.00	.99	.00	.06	.00	.97
Same country	−.10	.13	−.04	.43	−.04	.13	−.02	.75	−.04	.13	−.02	.75
Pol. Spectrum	.12	.03	.21	<.001	.07	.04	.13	.04	.07	.04	.13	.04
Mach.					.00	.01	.02	.78	.00	.01	.02	.74
Narcissism					.03	.01	.11	.03	.03	.01	.11	.04
Psychopathy					.02	.02	.06	.35	.02	.02	.07	.33
Sadism					.02	.02	.07	.32	.02	.02	.07	.30
RWA					.01	.01	.11	.06	.01	.01	.11	.07
Suppression									.00	.01	−.02	.68
R^2		.050				.094				.095
Adjusted R^2		.038				.071				.069

*p = .001 **p = .003.

Vignette 4 – agreement with ‘educating’ babies to European values being implemented

For this analysis, model 1 (F[5, 384] = 11.931, p < .001) and 2 (F[5, 379] = 7.313, p < .001), had significant changes (see Table 6). Model 3 did not F(5, 378) = 2.216, p = .14. The explained variance was 19.2%. The three significant predictors were more right-leaning on the political spectrum, higher psychopathy, and higher RWA.

Table 5. Coefficients for disagreement with volunteers rescuing migrants in the Mediterranean.

Model		1*				2*				3
Predictor	B	SE	beta	sig	B	SE	beta	sig	B	SE	beta	sig
Gender ID	−.53	.24	−.10	.03	−.28	.25	−.06	.26	−.27	.26	−.05	.30
Age	.01	.01	.04	.42	.01	.01	.07	.12	.01	.01	.07	.16
Education	.32	.12	.13	.01	.42	.12	.17	<.001	.41	.12	.17	<.001
Same country	−.26	.24	−.05	.29	−.15	.23	−.03	.53	−.15	.23	−.03	.53
Pol. Spectrum	.33	.05	.29	<.001	.11	.06	.09	.10	.11	.06	.09	.10
Mach.					.09	.02	.20	<.001	.09	.02	.20	<.001
Narcissism					.01	.02	.02	.68	.01	.02	.02	.71
Psychopathy					−.03	.03	−.06	.38	−.03	.03	−.05	.41
Sadism					.02	.03	.05	.43	.02	.03	.05	.42
RWA					.05	.01	.27	<.001	.05	.01	.27	<.001
Suppression									.00	.01	−.01	.79
R^2		.126				.217				.218
Adjusted R^2		.114				.197				.195

*p < .001.

Table 6. Coefficients for agreement with ‘educating’ babies to European values being implemented.

Model		1*				2*				3
Predictor	B	SE	beta	sig	B	SE	beta	sig	B	SE	beta	sig
Gender ID	−.65	.18	−.17	<.001	−.35	.19	−.09	.06	−.28	.19	−.08	.14
Age	−.01	.01	−.08	.11	−.01	.01	−.05	.28	−.01	.01	−.07	.14
Education	−.08	.09	−.04	.38	.00	.09	.001	.99	−.01	.09	.00	.95
Same country	−.28	.18	−.08	.12	−.20	.17	−.05	.25	−.20	.17	−.05	.25
Pol. Spectrum	.25	.04	.30	<.001	.14	.05	.16	.004	.14	.05	.16	.004
Mach.					.02	.02	.07	.25	.02	.02	.08	.19
Narcissism					.01	.02	.04	.46	.01	.02	.03	.58
Psychopathy					.06	.02	.15	.02	.06	.02	.17	.01
Sadism					.00	.02	.01	.85	.01	.02	.03	.67
RWA					.03	.01	.22	<.001	.03	.01	.22	<.001
Suppression									−.01	.01	−.08	.14
R^2		.134				.211				.215
Adjusted R^2		.123				.190				.192

*p < .001.

Discussion

The aim of the current study was to investigate the predictive power of the D4 and RWA traits and suppression on different types of real-world examples of SSE. While the dark personalities (Tetreault et al., 2021) and RWA (Lindén et al., 2016) have been linked to aggression and extremism, we cannot assume they will predict SSE in the same manner.

The bivariate correlations (H1) support the previous research that demonstrated that the D4 are correlated to varying degrees, highlighting they have overlap yet remain distinct traits (Buckels et al., 2013; Jones & Paulhus, 2013). In our sample, there were two correlations that are noteworthy, the weaker correlation between narcissism and psychopathy (r = .25 in our sample versus .40 or .37) and the strong correlation between psychopathy and sadism (r = .64 in our sample versus .44 or .48), as they seem to differ from previous findings (Buckels et al., 2013; Pineda et al., 2021, respectively for each correlation). This needs further investigation; however, one explanation could be due to our sample being more of a multinational sample compared to undergraduate university student samples that are prevalent in previous research. This raises the question if there are potential cultural influences in the measurement of everyday sadism and psychopathy. The vast majority of research using the SD3 and SSIS have been used on North American and European samples, and even between North America and Europe the internal reliability of the D4 scales has varied (Tetreault et al., 2021; Tetreault & Hoff, 2019). Trait manifestations are not always generalizable between cultures, and scales could require re-standardization to detect the different ways in which those traits present themselves in each culture (Cooke et al., 2005). This needs further investigation.

The hypothesis (H1) that RWA would be correlated to the D4 was partially supported. Machiavellianism was found to be in line with Lindén et al.’s (2019) finding that it was positively correlated with RWA, but psychopathy and sadism were not negatively correlated as predicted. In our sample, narcissism and RWA were also positively correlated; whereas in previous research there have been mixed results with some finding a correlation (Zeigler-Hill et al., 2020) and others not (Lindén et al., 2019). A core trait of psychopathy is loyalty only to oneself, schadenfreude for sadism’s core trait, and RWA’s core is adherence to conservative norms. Those who are higher on psychopathy and sadism may not care whether society is conservative or liberal, and as such, this could explain why no correlations were found between psychopathy, sadism, and RWA in our sample. At this stage in the research, the inconsistent findings between the D4 and RWA need more investigation. The discrepancies could be due to RWA being assessed by different scales, for instance. It would be good to investigate if there are cultural aspects to measuring RWA that could also account for some of the inconsistencies.

One of the theoretical underpinnings of the D4 is that those who are higher on those traits are able to function in society by suppressing their aggressive tendencies until they are triggered by specific situations. There is no previous research investigating suppression and the D4 and very little on RWA. In terms of suppression, our hypothesis (H2) was partially supported with Machiavellianism, psychopathy, and sadism being weakly, yet significantly, positively correlated with suppression proclivity. With items such as, ‘Many group activities tend to be dull without me’ on the narcissism scale, it could demonstrate that those higher on narcissism do not suppress their thoughts or feelings. Additionally, a pervasiveness of self-enhancement while lacking personal insight are key elements for those higher on narcissism (Paulhus & Williams, 2002). This could indicate people with higher narcissism are not prone to or are unaware of suppressive tendencies. For those who are higher on RWA, with the prevalence and use of social media, researchers have found far-right posts have more engagement than center- or left-leaning posts (Edelson et al., 2021). This could mean that those high on RWA are emboldened and have no reason to engage in suppression under the current social and political climate.

Overall findings for predictors in regression analyses

Prior to discussing the specifics of each SSE scenario and examining explanations for our results, we will discuss some of the overall findings. We did not hypothesize there would be demographic or background variables that would be significant predictors of SSE. In the regression models, placing oneself on the political spectrum was the only regularly significant predictor. Self-placement on the political spectrum can be challenging to analyze and interpret in a multinational sample, as there is no objectivity to the scale. For instance, a 6 (being neither left nor right) would be perceived differently between Irish, Canadians, and Americans. While interpretation of the scale is not straightforward and conclusions need to be drawn with caution, we followed best practices based on previous researchers’ findings to investigate the left-right spectrum. There were some models that had age (two models), gender (one model), and education (two models) as significant predictors. This is in line with previous research that has shown these variables often only explain little of the variance if any at all (Zmigrod et al., 2021).

Overall, we did find that those higher on the D4, but especially RWA, were more likely to support SSE in the varying scenarios (H3). However, as we see in our results for the individual vignettes discussed below, the D4 did not always show support in ways that were consistent with their triggers as hypothesized. This leads to the question of if SSE elicits different reactions in the D4 compared to violence and aggression in laboratory and other situations. This needs further exploration. However, our models do explain more of the variance than other models that have investigated extremism (Zmigrod et al., 2021), indicating that the dark personalities are more susceptible to supporting SSE. These results then add to the existing literature and should be considered in interventional work. While showing more sympathy/support towards SSE does not measure actual extremist behavior, it could demonstrate that those higher on the dark personalities have more vulnerability to governmental justifications for violence, which could lead to more passive acceptance of the violence.

One aspect of these findings that needs further investigation is if not being triggered to support SSE is equivalent to suppressing aggressive propensities, as this could impact how interventions could be tailored to help people higher on the dark traits. Interventions could be aimed at preventing the expression of aggression or redirecting people towards more pro-social outlets. Even though suppression and some of the D4 were correlated (H2), suppression was not a significant predictor in our models save one, which was not in line with our hypothesis (H4). As this was a cross-sectional study, we measured suppression proclivities and not if suppression was actually used while the SSE scenarios were read. This then raises the question of whether suppression dispositions are sufficient in and of themselves to influence support for SSE by creating hyperawareness through incongruence to find situations to show support for aggression as hypothesized.

We created and ran analysis on an aggregate score for the SSEs (see Supplementary Table 7), as it has the potential to assist with future research. However, it is important to note that the aggregate score removes the nuances of each type of SSE (support for non-violent SSE, violent SSE, intergovernmental agreements that financially support SSE, and political policies that instigate SSE). Based on the regression results we will discuss next, there are patterns of support (or not) for SSE. However, there are different predictors with varying explained variances in each vignette and subquestion, highlighting these nuances could be crucial in understanding each SSE form.

Support for using military to stop election protesters

We had hypothesized that those higher on psychopathy (due to the use of instrumental violence), sadism (due to the tendency to enjoy harming others), RWA (due to the support of ‘traditional’ governments), and Machiavellianism (due to the want of power) would also be more likely to have higher sympathies towards the army’s actions (H3). As such, our findings partially support our hypotheses. Being higher on sadism and RWA were in fact significant predictors. While psychopathy is characterized by the use of instrumental violence (Jones & Paulhus, 2010), this scenario may not have directly related to them, thereby not being sufficient to garner support. It is surprising that those higher on Machiavellianism did not demonstrate support, especially considering the violent methods were successful, as they tend to hold the end justifies the means ideology (Buckels et al., 2013). This finding needs further examination. We had not predicted demographic and background variables, save placement on the political spectrum, to be significant predictors. While younger people are often more likely to act aggressively (Moffitt, 2001), there has been stronger political movements by younger people (e.g. Hong Kong, Myanmar) fighting for democracy. Thus, younger people being in support of the use of violence by a military against political protesters was unexpected and needs further investigation.

Support for using weapons to regain control of a neighboring country

We had predicted that those who were higher on D4 traits, except narcissism, would be more likely to indicate support for the use of weapons against a neighboring country because of the use of physical aggression and the desire to gain power (H3). However, it was those who were higher on narcissism that were in fact more likely to predict being in support of this SSE scenario. It could be that those that are higher on narcissism are in support because of the ‘regain’ aspect, whereby the previous loss of control (i.e. self-entitlement) elicited the reactive support. We did not hypothesize that RWA would be a significant contributor to the model, and this is the only scenario where RWA was not significant. Interestingly, being more right leaning on the political spectrum did contribute significantly to the model. As we mentioned previously, participants’ placement on the political spectrum and RWA were significantly moderately correlated, however, not enough to cause multicollinearity issues. This could explain why the political spectrum placement is significant if it is tapping into only some aspects of RWA but not all of it. This needs further investigation to determine which parts of participants’ indications are correlating exactly with which aspects of RWA.

Not supporting volunteers who rescued migrants in the Mediterranean

For those indicating less support for the volunteers who rescued migrants on boats, we found partial support for our hypotheses (H3) with those being higher on RWA, Machiavellianism, and education being significant predictors. Those who are lower on RWA would be more open to relaxed immigration policies (Duckitt, 2010), thus indicating more support for volunteers who are rescuing migrants from drowning is in line with the defining characteristics and our hypothesis. Being higher on Machiavellianism and having more education showing less support for the volunteers were unexpected findings. In terms of Machiavellianism, it could be those higher on the trait indicated less support for the volunteers’ actions if it had been perceived as weakening the power of the government. A core trait of Machiavellianism is a want of power, including political power, and these findings could be aligned with that (Paulhus & Williams, 2002). One possible explanation for a higher level education showing less support for the volunteers could be that the vast majority of our sample had at least an undergraduate degree, indicating that the sample was still highly educated. These findings need more investigation to understand fully why less support for the volunteers was indicated for those with higher education.

Support for babies being forced into ‘education’ classes in your country

We had expected that those who were higher on narcissism and RWA would indicate more agreement for the northern European government’s methods to be implemented in their own country (H3). While we did not anticipate those being higher on psychopathy to significantly contribute to this model, as they tend to engage and/or support actions that are goal-oriented and low-cost to themselves, it is logical that they would indicate support for this type of program. It is noteworthy, yet not unexpected, that being higher on RWA and being more right leaning on the political spectrum were significant predictors. This type of policy is the crux of right-leaning political parties around the world as well as being anti-immigration, which would also appeal to those higher on RWA (Međedović & Bulut, 2018). We had hypothesized that those higher on narcissism would also indicate more support, as they tend to be in favor of acts when there is a (perceived) slight and not assimilating could be seen as a slight (H3). However, this was not the case. Those higher on collective narcissism, ‘an emotional investment in an unrealistic belief about the greatness of an in-group’ (de Zavala & Cichocka, 2011, p. 213), rather than the individual narcissistic trait of the D4 may be better at predicting support for this type of SSE, and future research should consider using collective narcissism in these scenarios.

These types of laws could incite extremist beliefs. In previous toxic situations, legislative actions to segregate and isolate different groups within a society led to citizens targeting the new outgroup with violence (Snyder, 1981). Most toxic situations are top-down (government initiated) and bottom-up (citizen initiated) when they are the most destructive and violent. These legislative moves of forcing babies into indoctrination schools are the beginning of a top-down approach (i.e. socially isolating and stigmatizing a specific marginalized group, creating an us versus them situation) that could lead to bottom-up extremism (i.e. aggressive discrimination or targeted hate). Attitude shifts precede behavioral ones, and as such this could be an initial step in extremist attitude by a government.

While the legislation in Denmark is aimed at assimilating poor immigrants, the reverse could happen, where those having to give up their babies for 25 hours a week will feel less integrated into society, more aware of the discrimination, further isolating them, and perhaps leading to more extremist opinions/behavior. There is also evidence that demonstrates second generation immigrants partake in more anti-social behavior than the first generation (Vaughn et al., 2014). This type of legislation could be one of the first steps in leading to extremism for those who are being targeted by this program.

Limitations

This study is cross-sectional, and as such, causation cannot be drawn. While attitudes and intention do lead to behaviors, having positive attitudes towards extremism is not sufficient to lead to participating in violent extremism. Thus, with our data, it is not possible to determine if showing support for SSE is a predictor for extremist behavior. However, knowing that being higher on the D4 and/or RWA, especially, is associated with more support for SSE highlights a potential susceptibility, which could bring a different approach to interventions. While participants were from numerous countries, there were not enough participants per country to analyze country specific data.

Conclusion

Much of the focus in the literature and research on extremism tends to focus on individuals or groups committing violent extremist acts. There is less research on SSE, and with this study, we wished to investigate the D4’s (Machiavellianism, narcissism, psychopathy, and sadism) and RWA’s association with varying types of SSE. Overall, RWA was the best predictor and explained the most variance in the support indicated for the SSE. Those high on the D4 also demonstrated support but not in the same SSE scenarios. Our findings also add to the literature in that we found those who were higher on Machiavellianism, psychopathy, and sadism were also more likely to engage in suppression in their daily lives, which to the best of our knowledge has not been conducted previously. In terms of narcissism, one area that should be investigated is collective narcissism, as it may be better associated with SSE support rather than the individualistic narcissism trait. It could also prove beneficial to investigate how collective narcissism relates to RWA to investigate factors that could indicate susceptibility to SSE justifications. Our findings, in their totality, could contribute to how dispositional differences and the acceptance of violence could inform counter-extremism interventions.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was supported by the Hardiman Scholarship at the National University of Ireland, Galway.

Notes on contributors

Christie Tetreault

Christie Tetreault is a PhD candidate and Hardiman Scholar in the School of Psychology at the National University of Ireland, Galway. Her current research focusses on resilience and susceptibility to extremism. Christie also has research interests in forensic and social psychology with a special interest in aggression and group dynamics.

Kiran M. Sarma

Kiran M. Sarma, Dr, is a Senior Lecturer in Psychology at the National University of Ireland, Galway and a Chartered Forensic Psychologist. He leads the Risky and Extreme Behaviour Research Group with a special interest in violence, aggression, and risk taking.

Notes

1 Six participants were removed from analysis for providing nonsensical answers (e.g. creating a country name).

2 For the participants who indicated they grew up in several countries, we used the one they listed first.

Appendix

Table

Vignette	Sub-question	Analysis
1. A leader of a western country sent out their army to end peaceful protests. The army was instructed to use as much force as necessary as long as the protests ended. This made the protests extremely violent, but the leader was re-elected after the protests were shut down.	a. To what extent to do you agree with the leader’s decision?	Supplementary Material
	b. To what extent do you agree with the army’s actions?	Manuscript
2. A western country and a Middle East country have an arms’ deal worth billions of euros yearly for military equipment, including heavy machine guns and artillery systems. The Middle Eastern country is using these guns and artillery systems specifically to target and kill civilians in a neighboring country to regain control over the area. Academics and policy advisors have asked the western government to stop selling the arms, but that government has refused saying ending the contract is too costly.	a. To what extent do you agree with the arms’ deal?	Supplementary Material
	b. To what extent do you agree with the Middle Eastern country’s use of the arms to regain control over the area?	Manuscript
	c. To what extent do you agree with the western country keeping the arms’ contract?	Supplementary Material
3. Volunteers have been using a small rescue boat to save refugees who are stuck on ships that are in distress in the Mediterranean trying to land in Europe. The volunteer’s boat is used to keep the refugees safe while a bigger rescue boat comes to take them to mainland Europe. European authorities seized the volunteer’s boat and are investigating the volunteers for human trafficking. An independent investigation determined that the volunteers were only saving lives, but the European prosecutor is refusing to give them back the boat and is still investigating them as criminals. Without these volunteers, more refugees are dying in the Mediterranean.	a. To what extent do you agree with the volunteers?	Manuscript
	b. To what extent do you agree with the prosecutor seizing the boat?	Supplementary Material
	c. To what extent do you agree with the prosecutor still investigating the volunteers as criminals?	Supplementary Material
4. Throughout a lot of the Western world, there has been an influx of immigrants. Some are concerned that Western values will disappear. One method to look at this came out of one northern European country. This country decided that immigrants in low-income ‘ghettos’ were not becoming European enough, so they created a law to ensure all babies become European in culture and attitude. All babies born to low-income non-European parents at the age of 1 will be separated from their families for a minimum of 25 hours a week, not including nap time. During this mandatory separation, the children get lessons in ‘European values.’ Babies born to European parents do not have to send their children to school until the age of 6.	a. To what extent do you agree with this northern European government’s methods?	Supplementary Material
	b. To what extent would you agree with this type of program being done in your country?	Manuscript