Empirical perspectives on religion and violence

The ‘religion as cause’ argument implies that religious faiths are more inherently prone to violence than ideologies that are secular. Following an evaluation of the scientific literature on religion and violence, we argue that wherever evidence links specific aspects of religion with aggression and violence, these aspects are not unique to religion. Rather, these aspects are religious variants of more general psychological processes. Further, there are numerous aspects of religion that buffer against aggression and violence among its adherents. The most distinct feature of religion, supernaturalism, is not often the focus of researchers of religion and violence. Despite this, the paucity of research that has been conducted on this key feature suggests that supernaturalism is associated with reduced aggression and violence. There appears to be very little support for the notion that there is something uniquely religious that causes violence among followers.

ideologies and institutions that are identified as 'secular'" (Cavanaugh, 2007;p. 1). We examine two general questions in this review. First, to what extent are religion and violence related? Second, is this connection between religion and violence specific to some feature of religion or part of a more general social psychology of group identity, which also applies to the secular. Our interest is in a scientific examination of these questions. We are interested in systematically collected data that are both reproducible and objective. As such, we examine correlational, experimental, and longitudinal research in the science of religion and violence. Through this review we argue that, despite evidence linking specific aspects of religion with aggression, these aspects are not unique to religion. These aspects are religious variants of more general psychological processes. The key feature that distinguishes religious groups from secular groups-supernaturalism-has been ignored in studies of religion and aggression. Ultimately, the literature reveals the role of general psychological processes inherent to aggression and violence but provides no evidence of a unique role of religion.

Religion and intergroup relations
Religion follows general social psychological principles of group behavior. The existence of groups and related social processes cause ingroup solidarity and outgroup derogation under conditions in which these group differences become salient. Additionally, strong beliefs are often related to greater defense of beliefs when threatened.

Ingroup solidarity and outgroup derogation
Social identity theory (SIT) (Tajfel and Turner, 1979;Tajfel and Turner, 1986) developed out of realistic group conflict theory (RCT) (Campbell, 1965), which argued that conflict emerged over competition for resources. The foundation of SIT argued that RCT ignored discussions about how group identity developed and how group identity could be maintained or changed (Tajfel and Turner, 1986). Social identity refers to "those aspects of an individual's self-image that derive from the social categories to which he perceives himself as belonging" (Tajfel and Turner, 1986;p. 283). In describing the religious identity and violence link, James Wellman and Kyoko Tokuno state, "The symbolic and social boundaries of religion (no matter how fluid or porous) mobilize individual and group identity into conflict, and sometimes violence, within and between groups" (2004, p. 291). The minimal group paradigm (Sherif et al., 1961) suggests that even minimal similarity between individuals is a sufficient mechanism for ingroup solidarity to form and for antagonistic behavior toward outgroups to develop. SIT posits that a primary goal of group behavior is to maintain positive distinctiveness (Tajfel and Turner, 1986). When this positive group distinctiveness is threatened, conflict may result. Terrorism in the Islamic world may be a result of a loss of personal identity and the development of a collective identity through joining a terrorist group, and the expression of "a chronic orientation to intergroup conflict" stemming from minority status and collective identity salience (Taylor and Louis, 2004, p. 180). Collective identity salience and perceived threats to that collective identity may enhance intergroup conflict and cause retaliatory behaviors (Fischer, Haslam and Smith, 2010;Wright, 2017Wright, , 2015.

Threats to social and religious identity
Before the Charlie Hebdo murders, the paper had just published provocative cartoons of the Islamic prophet, Muhammad (Bilefsky and de la Baume, 2015). They did the same thing in 2011, for which the magazine was bombed (Jolly, 2011). These and similar instances can be viewed within this social identity approach wherein perceptions of threat, which can lead to reduced positive distinctiveness, can result in reactive aggression as a mechanism of retaining positive distinctiveness (Wright, 2015).
Threats against the collective may imply a threat to the self, especially if the collective identity of the target is salient . This effect is succinctly stated, "When individuals experience an event as threatening their identity, they may react emotionally and may be easily mobilized by leaders to react violently" (Beit-Hallahmi, 2015;p. 54). Wright and Young (2017) demonstrated that the experimental manipulation of religious identity salience affected the extent to which anger and hostility were increased in response to a threat towards one's religious identity.
Aggressive responses to threats exist as ways of bolstering the positive value of one's social identity when the group's value is impeded in some way (Tajfel and Turner, 1986). Threats to the positive value of a social identity can lead to high identifiers responding to threats with outgroup derogation, perceived ingroup homogeneity, and increased self-stereotyping (Branscombe et al., 1999). Branscombe et al. (1999) further suggest that threats to the value of social identity will lead to defensive reactions. Researchers demonstrated this effect in a sample of Christian and Muslim students . Participants were subjected to a manipulation intended to threaten the participant's religious group, followed by assessments of emotions and action intentions. Stronger religious identity was related to increased anger in response to threat and greater intentions to engage in confrontation.
An analysis of perceptions of threat among Protestants, Catholics, Jews and Muslims in the United States suggested that greater perception of threats toward one's religious identity was linked to more negative outgroup attitudes among Protestants, Catholics, and Jews (Pasek and Cook, 2018). Threats from Muslims toward Norwegian and American Christian identities have been linked to support for collective action against Muslim immigrants (Study 1 and Study 2; Obaidi et al., 2018) and threats toward Muslim identity were associated with anti-Western hostilities (Study 3 and 4; Obaidi et al., 2018). Similar results were found in a study assessing reactions to threats towards one's gender and national identities , suggesting that this process is not an inherent aspect of religion, but an aspect of social identities in general. In one study, participants engaged in a conversation where either a participants' national identity or religious identity was denigrated, or participants had a neutral conversation. Denigration of participants' national identity resulted in greater aggression relative to both denigration of religion and the neutral control (Wright, Agterberg, and Esses, 2019).
Following the terrorist attacks in New York City on 11 September 2001, researchers had individuals report the extent to which they perceived the terrorist attacks to be a violation of sacred values (Mahoney et al., 2002). They found that the greater this perception was, the more individuals endorsed the use of nuclear and biological weapons as a response to the terrorist attacks on 9/11. Similarly, the more Christians perceive Jews as desecrators of Christianity, the more prejudice they exhibit toward Jews (Pargament et al, 2007). More government restrictions on religious groups are associated with more religious hostilities and the most violent regions in the world reflect the most government restrictions toward religion (Institute for Economics and Peace, 2013). However, research on identity fusion (i.e. a family-like bond) suggests that fusion with Judaism, rather than fusion with Israel, was linked to greater endorsement of extreme retaliatory actions against the Palestinians following the 2015 Intifada (Fredman, Bastian and Swann Jr., 2017). This would seem to indicate a specific effect of religion within the Israeli-Palestinian case, except other work similarly links secular identity fusion with extreme pro-group behaviors (Bortolini et al., 2018) and violence (Newson, 2017), suggesting that the effect is not religiously specific.
Additionally, some research casts doubt that threat will always result in ingroup solidarity and outgroup hate.  found evidence that an experimental condition of threat caused increased positive feelings toward the ingroup but had no effect on feelings toward the outgroup in a sample of Christians and Atheists from Britain. Despite this, the consensus across the literature appears to be that, under condition of threat, hostility toward outgroups is generally the result. Karen Armstrong states, "Every fundamentalist movement that I have studied in Judaism, Christianity and Islam is rooted in a profound fear of annihilation, convinced that the liberal or secular establishment is determined to destroy their way of life" (2014, p. 1). While threat perceptions toward individuals' religious identities may institute aggressive or violent responses, these effects are a product of a general social psychological process of group behavior, rather than anything inherent to religion. Is there uniqueness to religious identity that leads to violence beyond an overall effect of group differentiation, ingroup solidarity and outgroup derogation that is shared by secular social groups?

Religious fundamentalism
Following the Charlie Hebdo massacre, and as a response to Nicholas Kristof's question, Wright (2016b) examined why Islam is overrepresented in religious terrorism (Ben-Dor and Pedahzur, 2003;Piazza, 2009). In a study of over 52,000 religious individuals across 59 countries, Muslims reported more fundamentalist beliefs than did adherents of other religions including Protestants, Catholics, Hindus and Jews (Wright, 2016b). The link between fundamentalism and hostility is well established: the higher the level of fundamentalism, the more outgroup animosity is exhibited (Altemeyer and Hunsberger, 1992;Henderson-King et al., 2004;Hunsberger, 1996;Johnson, Rowatt and LaBouff, 2012;Rothschild, Abdollahi and Pyszczynski, 2009). It is clear that majority Muslim countries have seen a rapid rise in the overall share of civil wars, peaking at one hundred percent of all civil wars in 2012 (Gleditsch and Rudolfsen, 2016). However, is fundamentalism a direct cause or might fundamentalism simply exacerbate responses to threats toward one's social identity? This is one argument described by Wright (2016b), who suggested that Muslims might be more susceptible to religious calls to violence via this heightened fundamentalism. However, studies have yet to empirically examine this idea. Secular groups can also exhibit fundamentalist beliefs and react violently, in part, due to these beliefs. For example, fundamentalist thinking is exhibited in both extremes of the political spectrum (Toner et al., 2013) and eco-terrorism is not linked to a religious ideology (Eagan, 1996). Dogmatism (lack of openness in beliefs) is associated with hostility and aggression (Heyman, 1977) and dogmatic atheists act with prejudice against value violating outgroups (Kossowska et al., 2017). In fact, nonbelievers may actually be more dogmatic than believers (Uzarevic, Saroglou and Clobert, 2017). In part, religious violence in Muslim countries may be a result of pressures from secular nationalism (Juergensmeyer, 2010) and reactance to interventions by major powers (Gleditsch and Rudolfsen, 2016), rather than some unique fundamentalist character of religion.

Conflicts over religious values and religious beliefs
Perhaps one underlying difference between religious groups and others are the values that form out of a belief in the supernatural and the practices developed to express these beliefs and values. Hasenclever and Rittberger (2000) describe three reasons that conflicts over values are more likely to be violent than conflicts over interests. First, conflicts over values place one's deeply held values in threat of being extinguished and replaced by outgroup's values. Second, values are related to morality and the defense of values may be seen as morally justified. Third, religious values passed on by supernatural authorities are absolute, meaning that compromises are not possible without implying a non-absolute nature to these values. We examine three areas in which empirical research has addressed potential links between beliefs/values and violence: the relationship between religious practices and aggression, the relationship between religious orientation and aggression, and through cognitive priming studies.

Religious involvement and aggression
Blanket statements about religion causing violence assume that religious people are equivocal in their engagement with their religion, in the importance of their religion and in their interpretation of religion. Yet, different religions place different emphases on different aspects of religious practice (e.g. prayer, service attendance, reading of scripture) and highlight different values and beliefs to different degrees (Wright, 2016b). For example, a greater percentage of Jews report attending religious services at least once per week relative to Christians, Muslims, and Hindus (Wright, 2016b). In the United States, Jehovah's witnesses pray more often than any other religious group, including Protestant and Catholic Christians, Muslims, and Jews (Rosentiel, 2009). We noted at the outset that the quantitative research in this area has failed to account for differences in content expressed in different houses of worship within the same religious group. While some researchers have investigated the impact of religious content on aggression (Bushman et al., 2007), this work has not been incorporated into studies of frequency of involvement. Greer et al. (2005) reveal the complexities in evaluating religious practices to aggressive behavior. They tested whether church attendance, engagement in church activities, or donation pattern affected self-reported vengeance in Christian participants, finding that greater frequency of church attendance and greater frequency of engagement in church activities was associated with less selfreported vengeance (Greer et al., 2005). However, a more consistent donation pattern was related to greater self-reported vengeance. Based upon the church attendance and prejudice relationship (Allport and Ross, 1967), there appeared to be a relational inconsistency between frequency of church attendance and the outcomes of vengeance compared to prejudice. Greer et al. (2005) explained this finding as a problem of church attendance reflecting social pressures of parents in their sample, which they argued made donation pattern and engagement in church activities better indicators of religious involvement. However, in a similar study, American Christian participants were measured on frequency of prayer and frequency of church attendance, followed by an assessment of attitudes towards violence against terrorists in the Middle East (Shaw, Quezada and Zárate, 2011). Prayer was unrelated to attitudes towards violence but did strengthen the relationship between moral certainty (belief that one's morals are absolute) and attitudes towards violence against terrorists in the Middle East. The same effect was found for church attendance (Shaw et al., 2011).
In a study of 600 men in the Arkansas correctional system, Benda and Toombs (2000) found a combined measure of religiosity (frequency of prayer, bible study, church activity, talking about religion and attempts to convert others) related to lower self-reported acts of actual violent behavior over one's lifetime. A negative relationship between frequency of church involvement and number of violent crimes committed nationally in Sweden has also been documented (Pettersson, 1991). Longitudinal work confirms the relationship between greater involvement in religious activities and less aggressive behavior across the lifespan (Huesmann, Dubow, and Boxer, 2011).
Another way to evaluate the link between religion and violence would be to compare religious versus non-religious people rather than relating degree of religiousness across religious people to these outcome variables. Landau et al. (2002) compared aggression between religious and secular Israeli Jewish school children. Religious Jewish school children engaged in significantly less indirect aggression than secular Jewish school children; however, religion did not affect physical or verbal aggression. A study in Indonesia indicated that religious practice (measured as a function of mandated prayer, optional prayer, fasting and religious activities) was associated with less agreement with violent jihad (Muluk et al., 2013). This finding is in line with the argument that many religious Islamist radicals are recent converts, not steeped in their religious tradition (Stern, 2010). Other results, however, suggest that religious involvement is unrelated to aggression (Hardy et al., 2012).
Religious involvement is comprised of a number of unique factors and lumping various indices of religious involvement together may lead to inaccurate findings. Ginges, Hansen and Norenzayan (2009) indicated that frequency of Mosque attendance was associated with Palestinian Muslims supporting suicide attacks against Israeli Jews, but prayer frequency was not. In a similar manner, Wright (2014) found that the inverse relationship between a combined religious involvement measure and anger could be independently attributed to frequency of reading scripture. Under condition of cognitive activation of religious identity, high frequency of scriptural reading related to decreased anger in response to a threat to religious identity. Likewise, the inverse relationship between religious involvement and hostility could be independently attributed to frequency of prayer (Wright, 2014). In this case, when religious identity was cognitively activated, high prayer frequency related to decreased hostility in response to a threat toward one's religious identity.
Aspects of the religious experience are each unique and can have independent, and even opposite, effects on aggression and violence. This is reiterated in a cross-national sample, which suggests that higher frequency of religious service attendance may be the only religious involvement variable associated with greater agreement that violence against others is justified (Wright, 2016a). This supports the coalitional commitment hypothesis suggested by others Ginges, Hansen and Norenzayan, 2010). In general, frequency of attendance at religious services has been most associated with increased hostility toward outgroups and greater intentions to harm outgroup members. However, this variable represents a general commitment to a coalitional identity and is not religious specific. The link between religious service attendance and aggression or violence may be a byproduct of collective ritual Ginges et al., 2010), which can exist outside of religious contexts Masters, 2006, Jones, 2000). Additionally, activation of religious identity may play a role in moderating the effects of involvement (Wright and Young, 2017). This activation is accomplished through cognitive priming.

Cognitive priming
Cognitive priming is utilized to activate a particular identity-based response. It is assumed that attitudes and behaviors are guided by the salient identity. In one study, religious and non-religious participants were primed with either a passage framed as coming from the bible or framed as coming from an ancient scroll (Bushman et al., 2007). Furthermore, participants either read an inserted two sentences, in which God sanctioned the violence in the passage, or did not. Results indicated that exposure to violent scripture that is also supported by God can cause readers to act aggressively in a later competitive reaction time game (Bushman et al., 2007). This finding occurred in both religious and non-religious participants, suggesting that the priming effect of scriptural violence paired with priming a deity's justification of this violence can induce aggression even in non-believers. Additionally, this effect mirrors the broader effect of engaging in harm when this harm is sanctioned by an authority figure (Burger, 2009). It matters not whether this authority is religious or secular.
Other work on the religiosity-aggression link utilizes priming of specific religious practices (e.g. prayer). Leach, Berman and Eubanks (2008) found no differences in aggression across groups engaged in different religious practices. Participants engaged in either reading of biblical scripture, meditation, or secular reading for five minutes. Participants who engaged in biblical reading or meditation/prayer showed similar aggression levels in a later competitive reaction time game. Similarly,  utilized a synagogue priming condition, a prayer priming condition (asking participants about their frequency of prayer), and a no prime control to determine that Israeli Jewish participants in the synagogue prime condition were more likely than those in both the prayer prime and the no prime condition to support suicide attacks against Palestinian Muslims. Ginges and colleagues (2010) argued that the perceived relationship between religion and violence may be due to the role religion plays in binding communities of believers and facilitating parochial altruism (Choi and Bowles, 2007;Ginges and Atran, 2009;Graham and Haidt, 2010).
The studies of Leach et al. (2008) and  indicate that different aspects of religion may have different behavioral correlates and priming one aspect of religion does not necessarily activate religious identity as a general construct. The activation of religious identity may be context specific and while priming prayer may activate connectedness with God, priming service attendance may activate coalitional commitment and parochial altruism.
Activating one's religious identity through an experimental manipulation moderates the relationship between religious involvement and anger, as well as hostility, such that religious involvement is inversely related to anger and hostility when one's religious identity is activated (Wright and Young, 2017). The opposite effect appears to occur with religious commitment, whereby religious commitment relates to increased anger when one's religious identity is activated and religious identity threat is present (Wright and Young, 2017). However, some work questions whether religious identification and threat necessarily promote intergroup hostility . These researchers manipulated group-based threat by either exposing Christian participants to negative images (Dawkins's book, The God Delusion) or exposing participants to positive images hands praying and candles burning). Results indicated that among those exposed to the groupbased threat images, positive feelings toward the in-group increased. However, negative feelings toward the out-group (i.e. atheists) were unaffected .
Religious primes may actually act as moderators between an experience (e.g. perceived threat or social exclusion) and aggression against others. Aydin, Fischer and Frey, (2010) assessed whether priming participants' religious identities acted as a buffer against aggressive behavior as a response to social exclusion. Participants were primed on social inclusion or exclusion via a hypothetical vignette (Aydin et al., 2010). Afterward, participants either described their perceptions of religiousness and how religion affects their own lives or described their perceptions of environmental protections and how environmental protection affects their own lives. Participants then completed the ice water aggression paradigm, in which participants allocate how long another "participant" must insert their hand in ice-cold water as part of a separate research study. Exclusion primed participants were more aggressive toward the other participant than inclusion primed participants. More importantly, priming religious identity effectively buffered exclusion primed participants from any increased aggressive behavior (Aydin et al., 2010).
These results support arguments that religious beliefs generally promote peace in human relations (Nepstad, 2004;Shepperd, Miller and Smith, 2015) and corroborate other empirical research (Bremner, Koole and Bushman, 2011;Schumann et al., 2014). For example, Bremner et al. (2011) suggests that praying for others results in decreased anger and aggression toward an insulting stranger. In Study 1 of Bremner et al., (2011), participants wrote a five-minute essay, which was rated poorly by another participant. Following this, participants read a story about a student who suffered from neuroblastoma and were asked to either think about the student or to pray for the student. Participants in the prayer condition reported less self-reported anger relative to those in the "thinking" condition. In Study 2 of Bremner et al., (2011), participants either wrote an essay about an event that made them angry or wrote a neutral essay. As in Study 1 of Bremner et al., (2011), participants were provided either negative feedback or positive feedback Participants then either engaged in prayer for the partner who graded their essay or simply thought about their partner. Participants in the prayer prime exhibited lower aggression compared to participants in the thinking prime (Bremner et al., 2011).
That prayer reduces aggression in response to provocation may be because religious beliefs are generally non-hostile and promote peace and forgiveness, not violence. Schumann et al. describe religious beliefs as magnanimous and demonstrate that priming religious beliefs buffers negative effects of mortality salience; participants have reduced accessibility of revenge-related words and exhibit less vengeful reactions to an offending group (2014, p. 432). However, participants' endorsements of revenge against an offending group depend upon the tone of a religious prime. If the prime enhances the eye for an eye mantra, then revenge is enhanced; if the prime enhances the turn the other cheek mantra, then revenge is reduced (Schumann et al., 2014).
Specific beliefs about heaven and hell have been divergently related to national crime rates such that higher belief in hell at the national level is related to lower national crime rates (Shariff and Rhemtulla, 2012). Beliefs about whether religion serves purposes within the physical life or the afterlife also moderate the relationship between religious service attendance and the extent that believers justify violence against others (Wright, 2016a). Among people who see religion as a supernatural meaning system, the association between greater religious service attendance and increased justification of violence against others is attenuated (Wright, 2016a).
Supernatural punishment theory suggests that divergent beliefs about a punishing versus a forgiving God should affect both prosocial and aggressive behavior (Johnson, 2011;Johnson and Krüger, 2004). This has been examined via priming specific God concepts, such as authoritarian images of God or benevolent images of God (Johnson et al., 2013). Participants were randomly assigned to one of four priming conditions, in which an image of one of the following was shown: an angry God (Authoritarian), Jesus (Benevolent), symbols of the spirit (Benevolent), or abstract art (Control). Increased aggression toward a hypothetical other only emerged for those primed with the authoritarian God image. This was further moderated by religion, in which the effects only applied to non-Catholic Christians. These results are important in light of evolving and infinite conceptualizations of God (Johnson et al., 2013). In contrast to a generic "religion as cause" perspective, priming benevolent conceptualizations of God did not induce aggression relative to the secular abstract art condition (Johnson et al., 2013).
The cognitive activation of religious identity is not unitary. Rather, the content, context and specificity of priming religious identity have important implications for assessing hostile cognitions and aggressive and violent behavior. Another avenue of research on the religion and violence link addresses the motivations of people who engage in religious activities.

Religious orientation and aggression
People have different motivations for engaging in religious behaviors and conceptualizations of extrinsic versus intrinsic religiosity were developed to differentiate between those who engage in religion for personal gain (e.g. to feel good about oneself) versus those who are motivated to live out their faith in their daily lives (Allport and Ross, 1967). To explain the curvilinear relationship [1] between church attendance and prejudice (Streuning, 1963), Gordon Allport and J. Michael Ross suggested that high frequency attenders (i.e. more than 11times per month) are less prejudiced than both non-frequent attenders and non-attenders because these persons "receive something of special ideological and experiential meaning" (1967, p. 434). Casual members regard religion and religious contacts as less binding in their personal lives, while frequent attenders see religion as intricately connected to their daily lives, including their behaviors, and self-conception. Thus one's motivations for attending religious services may be more explanatory than frequency of involvement. C. Daniel Batson (1976) developed a third motivation, quest religiosity, which involves "religion as an endless process of probing and questioning generated by tensions, contradictions, and tragedies…" (1976, p. 32). Quest believers are not necessarily aligned with particular religious creeds, but continually question the nature and structure of religion and life itself. Thus there is no absolute truth in religion for quest believers. In other words, there is no fundamentalism. Greer et al. (2005) explored the relationship between religious orientation and retaliation. Participants were paired with a partner in order to compete in a reaction time game. Participants completed self-report measures of extrinsic and intrinsic religious orientation, quest orientation and vengeance. Subjects were told they were playing a reaction time game against a "partner" (but wins and losses were predetermined). Participants selected a shock level before each trial and the loser received the selected shock. Shock levels increased throughout the experiment and topped at a level double the participant's maximum tolerance. The highest-level shock was never delivered, because the participant would always win this trial, but acted as provocation. Retaliation was operationalized as the average of the last two participants' assigned shock levels (i.e. those following the attempted delivery of the strongest shock by their "partner"). Extrinsic orientation was related to more self-reported vengeance and unrelated to retaliation. Intrinsic orientation was related to less self-reported vengeance and unrelated to retaliation. Only quest orientation was related to less self-reported vengeance and retaliation (Greer et al., 2005). Other studies have corroborated the negative relationship between intrinsic orientation and self-reported aggression (Storch and Storch, 2002;Abel-Cooper, 2001). (2008) extended these findings by investigating how religious orientation affects aggression, both via self-report and a behavioral task. Participants completed self-reported aggression and religious orientation, followed by the reaction time game. Intrinsic orientation was associated with less self-reported aggression but unrelated to behavioral aggression. Extending on the findings of Greer et al., Leach and colleagues divided extrinsic orientation into two subscales: extrinsic personal (EP) and extrinsic social (ES). The former focuses on personal well-being and protection (e.g. "I pray mainly to get relief and protection"), while the latter focuses on social benefits and rewards (e.g. "I go to church mainly to spend time with my friends"). While evidence had previously shown a positive relationship between extrinsic orientation and self-reported aggression, Leach et al. (2008) demonstrated that this relationship is due to the extrinsic personal subscale, which related positively to both self-reported aggression and behavioral aggression.

Leach, Berman and Eubanks
The evidence suggests that intrinsic orientation relates to less self-reported aggression, while at least one aspect of extrinsic orientation relates to more self-reported aggression. In addition, both do not appear directly related to behavioral aggression in laboratory tasks (Greer et al., 2005;Leach et al., 2008). Only quest orientation was related to less actual and self-reported aggression. Also instructive are the relationships between religious orientations and prejudice and between religious orientations and pro-sociality. Intrinsic and quest orientations are linked with less overt prejudice; [1] A curvilinear relationship describes a relationship in which one variable increases as the other variable also increases up to a certain point before the relationship reverses. Subsequently, as one variable increases the other decreases. however, intrinsic orientation has been found unrelated to covert prejudice while quest orientation has been found negatively related (Batson, 1978;Batson, 1986). No significant relationship was found between extrinsic orientation and prejudice. The relationships between religious orientations and prejudice partially mirror the relationships between religious orientation and self-reported aggression. Further, the literature suggests that intrinsically oriented individuals score higher on self-reported altruism and are more likely to engage in prosocial behavior than extrinsically oriented individuals (Batson and Florey, 1990;Watson et al., and 1984;Chau et al, and 1990;Hunsberger and Platonow, 1986;Benson et al., 1980). Additionally, quest-oriented individuals are more engaged in prosocial behavior although their approach may differ from intrinsically oriented individuals (Batson, 1976).

Religion as post-hoc justification
Here we focus on research addressing whether religion acts as a post-hoc justification for violence caused by other factors. Large international datasets provide some way of getting at these types of questions. Religious grievances do not appear related to mobilization of religious groups for rebellion, although they relate to mobilization of religious groups for protest (Fox, 1999). Fox further notes that political discriminations and grievances over autonomy must be present for a positive relationship to exist between religious institutions and mobilization for rebellion. He summarizes that "religious institutions do, in fact, support both violent and quietist tendencies, depending upon the situation" (1999, p. 130). When no other factors are considered, religious institutions are related to lower levels of protest and unrelated to rebellion, indicating that secular factors motivate religious institutions to mobilize for rebellion. Some empirical laboratory work has been carried out to investigate the use of religion as a legitimizer of violence. Bushman et al. (2007) investigated whether scriptural violence and God-supported violence increased believers' aggression in a competitive reaction time task. Participants who read a passage framed as coming from the Bible delivered a greater number of noise bursts at the highest level than did participants who read a violent passage framed as coming from an ancient scroll. Participants who read the violent passage with an inserted section about God sanctioning violence allocated a greater number of the highest-level noise bursts compared to those who read the passage without the inserted section about God sanctioning the violence. In a second study that included non-believers, noise allocation was increased in both believers and non-believers when they read the passage in which God sanctioned the violence, although this effect was still greater in believers. This series of studies suggests that scripture and the presentation of God as sanctioning violence may increase aggressive behavior in both religious and secular individuals. A similar study suggested that both religious and secular individuals administered increased shock levels in a competitive reaction time task if the fictitious opponent was not a member of the participant's own group (Dor-Shav, Friedman and Tcherbonogura, 1978). Selengut (2008) argues that doctrine is a cause of violence in the case of Palestine and Israel and concepts such as just war theory and liberation theology seem to use doctrine in order to justify violence in some cases (Hasenclever and Rittberger, 2000). In Christian theology, the book of Revelation is a rather violent collection of passages describing the end of the world when God comes to judge the world. Munson (2005) argues that passages such as this indicate the causal effect of religion on violence, and yet empirical evidence suggest that belief in hell (i.e. evidence of a punishing God) actually relates to drastic decreases in national crime rates (Shariff and Rhemtulla, 2012). Likewise, a multilevel and cross-national investigation indicated that high aggregated national level ratings of importance of God strengthen the negative relationship between individual level importance of God and the extent to which people justify violence against others (Wright, 2016a). People who support the contention that religious belief causes violence rely on selected application of religious texts and ignore empirical investigations of this relationship. As Ginges et al. state, "Insurgent political violence has been conducted in the name of any number of beliefs, including religion, human rights, freedom, and preferred forms of economic organization. That these various rationales are employed does not tell us whether the beliefs themselves-the tenets of religion, democracy, freedom-actually cause people to go to war, or whether they are merely epiphenomenal post hoc rationales to justify intergroup violence" (2010, p. 347).
People may utilize religion as a justification in order to ameliorate negative attributes of oneself that stem from engaging in socially unacceptable or morally questionable behavior.

Alternative explanations for religion and violence
Thus far, the major findings in the religion and violence literature do not support the notion that violence is inherent to religion or that there is something unique to religion that causes violence. Additionally, there are a multitude of well-supported alternative explanations to violence even when religious people or religious groups are involved in violent encounters.

Deprivation and marginalization
In considering alternative explanations for a link between religion and political violence, Zaidise, Canetti-Nisim and Pedahzur (2007) found that socio-economic deprivations underscored Muslim's support for political violence in Israel. Additionally, higher religiosity was associated with less support for political violence in both Muslims and Jews. Canetti et al. (2010) found Muslims to be more religious and more supportive of political violence. They also are subject to more discrimination in comparison to Jews in Israel. The relationships between religion being Muslim and support for political violence and between religiosity and support for political violence are mediated by factors of psychological and economic loss (Canetti et al., 2010). These same researchers found that loss of psychological resources has the greatest impact on support for political violence. Furthermore, feelings of marginalization and insignificance and perceptions of discrimination increase support for radicalism among American Muslims (Lyons-Padilla et al., 2015) and feelings of being treated unfairly is linked to support for violent jihad (Muluk, Sumaktoyo and Ruth, 2013). Terrorists consistently reveal profound experiences of humiliation or shame brought upon them by external forces (Jones, 2010). Avalos's (2005) theory of religion and violence reiterates realistic conflict theory (Campbell, 1965), in which "scarce resources, real or perceived, are a major factor in violence" (p. 93). Avalos's contribution is a series of resources that he claims are religiously specific and that religious groups engage in conflict over. His inscripturation is nothing more than conflict over ideas, sacred space as conflict over territory through hierarchical structures (i.e. social dominance theory; Pratto et al., and 1994), and group privileging as social identity theory (Tajfel and Turner, 1986). These resources fit well within the category of cultural and symbolic resources incorporated in other models (Esses, Jackson and Armstrong, 1998), which apply to both secular and religious groups. Avalos's concept of conflict over access to salvation is undermined by research suggesting that peoples' motives for violent collective action are bound by moral commitment to collective interests, not individual incentives (Ginges and Atran, 2009). Avalos's claim that religions, as opposed to secular groups, are "inherently prone to violence" (2005, p. 347) relies predominantly on evidence of basic group processes, not any inherent aspect of religion. Canetti et al. conclude, "It is likely that the extent to which people suffer resource loss, predominantly psychological resource loss, explains support for political violence" (2010, p. 583), and this resource loss is not specific to religious groups.

Mobilization hypothesis
A second alternative mechanism to any inherent nature of religion that connects religion and violence is mobilization and politicization of religious factors (Basedau et al., 2011). If the mobilization hypothesis has merit, we should find that regions with diverse religious groups or rapid large-scale changes in religious groups, should inherently have more conflict than those that do not, and yet evidence suggests that this is not the case (Fox, 1999;Russett, Oneal and Cox, 2000;Tusicisny, 2004). However, some findings in conflict prone African countries suggest that the dominance of one religious group and an overlap between religious and ethnic identities may be two factors that directly contribute to increased risk of armed conflict, without any necessary politicization (Basedau et al., 2011). These researchers summarize that "one theoretical conclusion may be that socio-psychological intergroup processes are more important than religious ideas and their use by leaders" (Basedau et al., 2011, p. 766).

Moral certainty/moral identity
A third mechanism by which religion may relate to increased violence apart from a claim of inherency is through morality. All religions dictate a system of moral values and since supernatural authority creates morality in religious traditions, followers can feel certain that these morals are absolute. This can satisfy a person's need to feel moral and lessens any necessary concern for consequences of violence so long as the violence can be placed within the moral tradition passed down by the supernatural authority (Shaw et al., 2011).
Being certain about one's moral principles is related to greater support for violent warfare (Shaw et al., 2011). Additionally, religiosity variables moderate this relationship. For example, in those who pray more often, stronger moral certainty is related to greater support for violent warfare. This study also identified that greater moral certainty related to greater support for violent warfare primarily in conflicts that are framed as religious rather than geopolitical. It may be that religious identity enhances support for violent conflict only if a person's religious identity is connected to greater moral certainty. Additionally, there is no theoretical reason why this effect could not also exist in non-religious persons, if moral certainty is created outside the bounds of religion. Even atheists develop their own moral foundations (Simpson and Rios, 2016) and can express dogmatism in defense of their moral beliefs (Kossowska et al., 2017).
The binding motivations of loyalty, authority, and purity appear to underlie religious terrorist groups (Hahn, Tamborini, Novotny, Grall, and Klebig 2018). The PPT-US (Profiles of Perpetrators of Terrorism in the United States) dataset provides detailed information on all known terrorist organizations to have committed a terrorist attack in the United States from 1970-2016, including their philosophy statements (Miller and Smarick, 2012). Hahn et al. (2018) coded the most salient moral foundations depicted in the philosophy statements of these terrorist organizations and found that the binding motivations of loyalty, authority, and purity are the most common among religious terrorist groups. Purification rituals of sacrifice can expel the guilty and unclean, maintaining the purity of the community (Jones, 2008). The importance of authority can be seen through the centralized structure and dynamic leadership of these terrorist organizations (Whitehouse, 2002). However, these same binding motivations underlie secular right-wing and separatist movements (Hahn et al., 2018), suggesting that they are not exclusively, or even primarily, religious.
In related work, moral identity is considered the extent to which people consider being a moral person as important to defining themselves (Hardy et al., 2012). Some research has found that intrinsic religiosity, but not extrinsic religiosity, positively relates to moral identity (Vitell et al., 2009). While moral disengagement strategies can be used to increase support for warfare, this effect may be eliminated when people place a high importance on moral identity or when moral identity is experimentally primed (Aquino et al., 2007). Hardy et al. (2012) found that greater religious commitment indirectly related to lower aggression by way of moral identity. In general, moral certainty relates to greater support for violence and aggression but holding a strong moral identity relates to less support for violence and aggression.

When religion merges with other identities
Real-world conflicts are often confounded by interacting identities. Two important identities in intergroup conflict are nationality and ethnicity, both of which may affect how religion relates to conflict. Historically, the indication that the crusades were religious wars is an over-simplification of a complex interplay between identification with the state and identification with one's religion. As Munson states, "National and religious identity tended to be intertwined, as remains true in much of the world today" (2005, p. 228). The predominant modern region in which an apparent national-religious identity has developed is the Middle East. Some explanations for the violence, dictatorship, and oppression in the Middle East focus on the theocratic nature of Islam.
But are Middle East conflicts more religious? The Middle East appears to have a high proportion of religiously differentiated groups, which should cause increased and more violent conflict than elsewhere (Fox, 1999). Furthermore, most groups involved in conflict seem to involve religious discrimination of a group and the discriminated group demanding more religious rights (Fox, 2001). While religion is important in the conflicts in the Middle East, they are similar to conflicts raging in other parts of the world that do not involve these religious dimensions. Fox (2001) finds that conflicts in the Middle East involve similar levels of political discrimination, economic discrimination, cultural discrimination, repression, desire for autonomy, and terrorism or rebellion as non-religious conflicts in other parts of the world. He summarizes, "It is easy to assume that the prevalence of religious conflict in the Middle East is due to the region's Islamic and autocratic character. It is also easy to assume that the region's high concentration of autocracy is due to the region's Islamic character. Yet neither of these assumptions appears to be correct" (Fox, 2001, p. 39).
It is important to remember that "When religious boundaries roughly align with boundaries between nation-states, religious competition may become the grist on which international conflict is ground…" (Hall, 2003, p. 15). These conflicts cannot be viewed as stemming entirely from one or the other identity. In sub-Saharan Africa, overlapping ethnic and religious identities are particularly influential in mobilizing groups for armed-conflict (Basedau et al., 2011) and former European colonies may be eager to use shared religion as a spark to form new governments based upon the indigenous ethnic culture (Juergensmeyer, 1996). In general, secular motivations for conflict are far more common than religious motivations (Institute for Economics and Peace, 2013), yet "Secular loyalty to nations and movements, leading to death on the battlefield, raises fewer questions than the loss of life for religious ideals" (Beit-Hallahmi, 2015;p. 151).
In the prior sections we have detailed the various links between religion and violence. This revealed that there are many aspects of religion that are associated with reduced aggression and violence. In addition, when aspects of religion have been associated with increased aggression and violence, these aspects do not appear unique to religion. These aspects are religious variants of more general psychological processes, such as social identity, coalitional commitment, dogmatism, and obedience to authority. The key feature that distinguishes religious groups from secular groupssupernaturalism-has been mostly ignored in studies of religion and aggression. Ultimately, the literature reveals the role of general psychological processes inherent to aggression and violence but provides no evidence of a unique role of religion. In the rare case that supernaturalism has been the focus of researchers' attention, supernatural beliefs in hell and the afterlife have been associated with reductions in violence (Shariff and Rhemtulla, 2012;Wright, 2016a). We now describe supernaturalism as the fundamental distinguishing feature between religion and secular groups and why this should be the focus of researchers' attention if they wish to make claims of the inherency of violence to religion.

Supernaturalism in religion
Religion entails many facets (e.g. Alatas, 1977;Sterelny, 2017). It has the dual function of a social identity and a system of beliefs (Ysseldyk, Matheson and Anisman, 2010). It involves communal experiences (Beit-Hallahmi, 1984), with shared beliefs, values, rituals, histories, stories, and sacrifice (Beit-Hallahmi, 2015). But these elements are not unique to religious groups if detached from their supernatural elements. In defining religion's distinctive quality that differentiates it from the secular, we find this to indispensably be the belief in the supernatural. In psychology, "religion is a belief system, which includes the notion of a supernatural, invisible world, inhabited by gods, human souls, angels, demons, and other conscious spirit entities" (Beit-Hallahmi, 2015; p. 3). Beit-Hallahmi further suggests that belief systems are religious only when those committed to them make reference to the supernatural. Other scholars adopt this same essentialist feature. Schaffalitzky de Muckadell (2014) defines religion as a system of beliefs and practices that incorporate supernatural elements and where these beliefs are internalized by the believer and have normative implications. Sterelny (2017) refers to these supernatural elements as "hidden agents or forces". Without these elements, religious groups would mirror secular social groups, such as fraternities and sports teams, which involve social identification, systems of beliefs, sets of rituals, communal histories, and stories (Goody, 1961;Jackson and Masters, 2006;Jones, 2000;Workman, 2001). Many religious and secular actions are similar but religious actions "appeal to superhuman or supernatural agency" (Barrett, 2007;p. 179).
Our adoption of religion as a belief system distinct through belief in the supernatural, whether constituted by supernatural beings or supernatural powers, is in line with the history of the empirical psychological tradition (Beit-Hallahmi, 1984;Bloom, 2007). It applies to the world's largest religious groups (Hackett and McClendon, 2017), to obscure traditions (Baum, 2009;Conklin, 2001), to religious groups across time (Sterelny, 2017), and is adopted by cognitive science, anthropology and sociology alike (Avalos, 2005;Barrett, 2017;Goody, 1961;Hall, 2013;Stewart and Strathern, 2013). The degree of supernaturalist elements may vary across religious traditions and traditions may account for supernatural entities in different ways (Boyer, 2001). For example, supernaturalism may be more important to Christianity, which requires a belief in an omnipotent, omnipresent God, relative to Theravada Buddhists, but supernaturalism is not relegated solely to theism. Theravada Buddhists believe in rebirth, supernatural power, and dryads (Jerryson, 2017;Nugteren, 2005) and engage in rituals to please the spirits (Collin, 1990). If researchers are genuinely interested in making claims of the inherency of violence to religion, then studying the essential feature of religion, which distinguishes it from secular social groups, is necessary.

Limitations to our approach
We took an empirical scientific approach-we are interested in data that are objective and reproducible. We explicitly examined aggression and violence as potential outcomes of religion, whether as a result of beliefs, values, supernatural authority, identity, ritual, or some other aspect of religious groups. We dealt almost exclusively with Abrahamic religions by the necessity of the available scientific research (Hood, Hill and Spilka, 2009). This relegates our examination to doctrinal forms of religion and religious expression (Whitehouse, 2002). As can be seen throughout the review, operational definitions of religious constructs are study specific. For example, quantitatively examining ritual can be seen through measuring prayer or meditation frequency, reading of scripture, or through frequency of attending religious services. Beliefs have been examined through assessing agreement with fundamentalist doctrine, the importance of a set of institutional beliefs to one's life, or through manipulating passages of scripture. Even twenty years ago, there were already well over 100 measures in use within the psychology of religion (Hill and Hood, 1999). We largely took a nomothetic approach, which we believe is enlivened by the idiothetic approach elsewhere (Hood, Hill and Spilka, 2009). We examined believers in the contemporary world while avoiding the textual approach (Munson, 2005) because religious believers often know little about their scriptures (Bakker, 2006;Pew Research Center, 2010;Schuurman, Grol and Flower, 2016;Wright, 2016). These boundary conditions to our review reveal the limits of the scientific literature primarily published in psychology of religion and social psychology journals.

Concluding remarks on religious identity
The Chinese government oppresses Christians, Buddhists, and Muslims (Hernández, 2018;Kuo, 2019Vanderklippe, 2017. Islamic countries are invaded by secular Western powers (Brands, 1994;Freedman, 2009;Smith, 1992). In Houston, TX, Mayor Annise Parker promoted legal constraints on biblical sermons (Driessen and Morris, 2014). In the Charlie Hebdo incident, retaliatory murders were the result of perceived threats to the sanctity of Islam (Bilefsky and de la Baume, 2015). The evidence discussed in this article suggests that these kinds of pressures on religious freedoms and religious identity have and will result in increases in aggression and violence as a result of religious groups defending their sacred identities. This is the general social psychological mechanism that protects ingroup identification and the positive distinctiveness that is derived from it. Religious groups can exhibit violence in defense of religious freedoms and in defense of a threatened group identity, but religion can also be used to reduce violence. Hasenclever and Rittberger (2000) conclude succinctly, "differences in religious creed are rarely, if ever, genuine causes of violent clashes in and between nations" and that "religious communities usually live in peace-understood as the absence of civil unrest or war-as long as the society as a whole prospers" (Hasenclever and Rittberger, 2000, p. 658).
Many aspects of religion appear to reduce aggression and violence. These include prayer and reading of scripture, which appear to activate moral beliefs and values (Bremner et al., 2011). Even the priming of religious identification more generally can buffer aggressive responses to exclusion (Aydin et al., 2010). Supernatural beliefs in hell and the afterlife appear to reduce crime rates (Shariff and Rhemtulla, 2012) and buffer the link between coalitional commitment and willingness to justify violence against others (Wright, 2016a). In cases where aspects of religion are associated with aggression and violence, these aspects have direct secular counterparts and cannot be said to be unique features of religion. For example, the link between fundamentalism and outgroup hostility is the function of a more general process of moral certainty and dogmatism (Kossowska et al., 2017;Shaw et al., 2011;Uzarevic et al. 2017). Religious involvement reflects a more general process of coalitional commitment (Ginges and Atran, 2009;Ginges et al., 2010;Wright, 2016a), and the general role of social identity is not confined to religion (Tajfel and Turner, 1986). Beit-Hallahmi states, "religion does not involve unique processes, but a unique content" (2005, p. 236). Its processes are equivocal to all social groups. If social group processes are the primary ones engaged in the violence debate, then religion is not unique in its connection to violence.
What makes religion unique apart from generic discussions of social identity is a belief in the supernatural, the meaning of this belief to the individual and the group, and the internalization and integration of religious identity to the individual. Direct study of supernaturalism is noticeably absent in the literature. Where it does appear, belief in the supernatural appears to be associated with reduced violence. If we are genuinely interested in studying the effects of religion on violence, we must take Beit-Hallahmi seriously when he writes "Religion is studied by talking to believers, not reading scriptures, and the psychology of religion is the reality of believers and their beliefs" (2005, p. 234). Ignoring the reality of believers by examining religious texts or general social psychological processes do not move us further to answering the question, "is there something inherent to religion that leads to violence"? At least for now, the answer appears to be "no".

About the authors
Joshua D. Wright received his PhD in social psychology from the University of Western Ontario and is currently a postdoctoral fellow and lecturer at Simon Fraser University. He was awarded the Sukemune-Bain Early Career Research Award from the International Council of Psychologists for his research. His research spans the areas of religion and violence, political psychology, and intergroup behavior.
Yuelee Khoo graduated from the University of Western Ontario with his BSc in psychology. His research interests combine immigration and religion. He has examined how religiosity changes as a function of time spent in the host country and has examined the role of spirituality in immigrants' access to mental healthcare.