Violent Conflict and Hostility Towards Ethnoreligious Outgroups in Nigeria

Tuki, Daniel Article — Published Version Violent Conflict and Hostility Towards Ethnoreligious Outgroups in Nigeria Terrorism and Political Violence Provided in Cooperation with: WZB Berlin Social Science Center Suggested Citation: Tuki, Daniel (2025) : Violent Conflict and Hostility Towards Ethnoreligious Outgroups in Nigeria, Terrorism and Political Violence, ISSN 1556-1836, Taylor & Francis, London, Vol. 37, Iss. 2, pp. 239-261, https://doi.org/10.1080/09546553.2023.2285939 This Version is available at: https://hdl.handle.net/10419/307778.2 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. http://creativecommons.org/licenses/by/4.0/ Violent Conflict and Hostility Towards Ethnoreligious Outgroups in Nigeria Daniel Tuki Migration, Integration and Transnationalization Research Unit, WZB Berlin Social Science Center, Berlin, Germany ABSTRACT This study examined the effect of exposure to violent conflict on hostility towards ethnic and religious outgroups among Nigeria’s population and among its two major religious groups (i.e., Christians and Muslims). Violent conflict had a robust positive effect on outgroup hostility among the Nigerian population and among Christians. A plausible mechanism behind this finding is that the threat posed by violent conflict strengthens ingroup cohesion, erodes trust in outgroup members, and makes intergroup boundaries salient. This is especially so when the opposite party to the conflict constitutes a distinct cultural outgroup. The main conflict affecting Christians involves nomadic pastoralists of Fulani ethnicity, who are Muslims. Among Muslims, violent conflict rather had a weak positive effect on outgroup hostility that was not robust to alternative operationalizations of outgroup hostility. The null effect might be because the main conflict affecting Muslims —the Boko Haram insurgency—does not involve Christians. A significant number of Muslims are also affected by conflicts involving nomadic Fulani pastoralists. KEYWORDS Violent conflict; conflict exposure; outgroup hostility; ethnicity; religion; Nigeria Introduction A cursory look at Nigeria reveals that it has a dyadic structure comprising a predominantly Christian Southern Region and a predominantly Muslim Northern Region. Although there are some overlaps between the two regions, the contrast between them is quite stark. The overlap between religion and ethnicity makes the fault line between the two regions even more salient. 1 This North-South bifurcation is apparent when one looks at Nigeria through the lens of the nine civilizations into which Samuel Huntington divided the world: Nigeria’s Northern Region was associated with Islamic civilization, while the Southern Region was associated with African civilization. 2 This cultural divide has historical roots. Islam first came to Northern Nigeria between the eleventh and fourteenth centuries through the trans-Sahara trade between the Hausa people of Northern Nigeria and merchants from the Maghreb states. Besides the exchange of tangible commodities, there was also a diffusion of cultural and religious values. 3 Islam gained a stronger foothold in the region between 1804 to 1808, when a cleric of Fulani ethnicity, Usman dan Fodio, launched a jihad against the rulers of the Hausa kingdoms. The jihad led to the establishment of the Sokoto Caliphate, which consisted of several emirates. The caliphate was in existence for a century until its conquest by British forces at the beginning of the twentieth century. 4 Although Christianity in Nigeria can be traced to the fifteenth century when Portuguese slave traders visited Nigeria’s southernmost parts, it was not until the 1840s that the religion started to gain a foothold, propagated by freed slaves from Sierra Leone and missionaries from the West. 5 CONTACT Daniel Tuki [email protected] Migration, Integration and Transnationalization Research Unit, WZB Berlin Social Science Center, Reichpietschufer 50, Berlin 10785, Germany TERRORISM AND POLITICAL VIOLENCE 2025, VOL. 37, NO. 2, 239–261 https://doi.org/10.1080/09546553.2023.2285939 © 2024 The Author(s). Published with license by Taylor & Francis Group, LLC. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The terms on which this article has been published allow the posting of the Accepted Manuscript in a repository by the author(s) or with their consent. Christian missionary evangelization was concentrated in Southern Nigeria because the Muslim rulers in the Northern Region, in an effort to preserve their religious way of life, forbade Christian proselytization in the region. 6 The British government did not change much in Northern Nigeria after capturing it. They appropriated the existing institutions and even used the local Hausa language in administering the Northern protectorate. Conversely, the policies of Westernization and Christianization were pursued fervently in the Southern Protectorate because its population was more open to Western influence. 7 After Nigeria’s independence from British colonial rule in 1960, it remained divided along ethnic and religious lines. Commenting on the Northern-Southern dichotomy, Coleman observed: “Certain basic underlying differences in history, culture, temperament, and levels of development and acculturation provided the classical setting for intergroup friction.” 8 Nigeria’s historical timeline is punctuated by ethnoreligious conflicts as evidenced by the HausaIgbo riots of 1945, 9 the Kano riots of 1953 between Northerners and Southerners, 10 the pogroms of 1966 against members of the Igbo ethnic group, which led to the Biafran War from 1967 to 1970, 11 the Kafanchan riots in Kaduna between Christians and Muslims in 1987, 12 the Kano riots of 1991 between Christians and Muslims, 13 the Shariah Crisis in Kaduna between Christians and Muslims in 2000, 14 the 2011 post-election violence, which had a religious undertone, 15 and the recurrent clashes between Christians and Muslims in Jos, 16 amongst others. Ethnoreligious conflicts are not peculiar to Nigeria; they occur in several countries around the world. Examples include the conflict between members of the Sinhalese and Tamil ethnic groups in Sri Lanka, 17 the conflict between Muslims and Buddhists in Myanmar, 18 the violent clashes between Hindus and Muslims 19 and those between the Assamese and Bengalis, 20 both in India, the Nagorno-Karabakh conflict between Armenians and Azerbaijanis, 21 the Malay-Chinese conflicts 22 and the disputes between Muslims and non-Muslims, 23 both in Malaysia, amongst others. Ethnoreligious conflicts can lead to segregation, which in turn solidifies intergroup boundaries. 24 Some studies have recommended segregation as a strategy for mitigating interethnic conflicts because separating the conflicting groups eliminates the existing threat. 25 Bollig has conducted a study among nomadic tribes in Kenya where he finds that interethnic ties via marriage and friendships does not lead to conflicting loyalties, but neither does it attenuate the risk of interethnic conflict. 26 Using a game theoretic approach, Larson has shown that interethnic cooperation could increase the risk of conflict, especially when cooperation is underpinned by the threat of retaliation. 27 She points out that the speed of retaliation depends on the density of a group’s network: Information flows quickly within dense networks, and this makes it easy for retaliation to be meted out for misbehavior, which in turn makes the threats made by groups with dense networks credible. Conversely, groups characterized by sparse networks where information spreads slowly, find it hard to make credible threats because the slow diffusion of information within the network constrains group mobilization. This is problematic because it makes conflicts intractable. For instance, when a peace agreement between conflicting groups has been reached, it might take longer for this development to spread through a sparse network; attacks after the agreement could be interpreted as unwarranted, which then triggers a new wave of violence. She cautions that “uncareful efforts to promote peace by imposing crossgroup ties can do more harm than good, especially if they are aimed at the most peripheral members of both groups.” (470). Some studies argue that intergroup contact is crucial in reducing ethnic conflicts. For instance, Eke has conducted a qualitative study in the city of Jos in Nigeria’s Middlebelt Region where he finds that when mutual distrust is present and ethnic groups perceive each other as threats, interethnic violence is likely to erupt even when both groups are completely segregated. 28 This is because segregation “eliminate[s] opportunities for post-conflict reconciliation” and hinders the rebuilding of interethnic trust. Furthermore, he asserts that even though partial segregation does not entirely eliminate the perceived threat from the outgroup, it nonetheless creates avenues for contact between the rival groups 240 D. TUKI which serve as conduits for establishing trust and averting future conflict. Another key finding of his study is that when the leaders of the different groups engage in dialogue, this signals to the group members that the ethnic outgroup is not so threatening and disputes can be resolved amicably without recourse to violence. Rydgren et al. have conducted a study in Iraq where they find that people who spend time in ethnic heterogenous spaces are more likely to develop friendship ties across ethnic boundaries, are more tolerant towards people of other ethnic groups, and report higher levels of interethnic trust. 29 Similarly, Kanas et al., relying upon large-N survey data collected from Muslim and Christian students in the Philippines and Indonesia, have found that interreligious friendships reduce hostility towards religious outgroups. 30 These findings are congruent with the argument of Allport who contends that “separateness” heightens the risk of conflict because it leads to the exaggeration of intergroup differences. 31 Present-day Nigeria remains polarized along ethnic and religious lines. 32 Nigerians define their identity “by affiliation to religious and ethnic groups rather than the Nigerian state.” 33 The two major conflicts that have ravaged Nigeria during the past two decades—i.e., the Boko Haram insurgency and the violent clashes between nomadic pastoralists and resident communities—have taken a religious turn because of the distinct ethnoreligious identities of the conflict actors. Despite the persistence of violent conflicts in Nigeria, no study, to the best of my knowledge, has examined how these conflicts influence hostility towards ethnic and religious outgroups using representative survey data for Nigeria’s population and econometric techniques. Moreover, no study has examined the heterogenous effects of violent conflict on outgroup hostility among Nigeria’s two major religious groups—i.e., Christians and Muslims. This study does so. To measure outgroup hostility, I developed an additive indicator by combining the responses to two survey items probing the respondents’ willingness to have people from a different religion and people from a different ethnic group as neighbors. To measure exposure to violent conflict, I drew buffers with a radius of thirty kilometers around the respondents’ dwellings using QGIS software and counted the total number of violent conflicts within them. I was able to do that because I relied on data obtained from Afrobarometer 34 and the Armed Conflict Location and Events Database (ACLED), 35 both of which are georeferenced. Causal identification stemmed from instrumenting conflict exposure with forest cover. The regression results show that among the Nigerian population and among Christians, exposure to violent conflict has a positive effect on outgroup hostility. A plausible mechanism behind this finding is that the threat of violent conflict strengthens ingroup cohesion, erodes trust in outgroup members, and makes intergroup boundaries salient. This is especially so when the opposite party to the conflict constitutes a distinct cultural outgroup. The main conflict affecting Christians involves nomadic pastoralists of Fulani ethnicity, who are Muslims. Among Muslims, violent conflict had a weak positive effect on outgroup hostility that was not robust to alternative operationalizations of outgroup hostility. A possible reason for the null effect among Muslims is that the main conflict affecting them—the Boko Haram insurgency—does not involve Christians. Many Muslims have also been affected by conflicts involving nomadic Fulani pastoralists. This study contributes to the broader literature on intergroup relations in the shadow of violent conflict. 36 The subsequent sections are organized as follows: In the second section, I discuss the trend of violent conflicts in Nigeria, after which I review the literature on the nexus between conflict and social cohesion. Next, I operationalize the variables that will be used to estimate the regression models and discuss the empirical strategy; I then present the regression results and discuss them, after which I summarize the paper and conclude. Violent conflicts in Nigeria Nigeria has witnessed a lot of violent conflicts during the past two decades. Data from ACLED 37 shows that Nigeria had a total of 18,781 incidents between 1997 to 2022, which makes it the country with the third highest incidence of violent conflict in Africa. 38 Only Somalia and the Democratic Republic of Congo performed worse. These incidents caused 98,877 fatalities. The distribution of violent conflict TERRORISM AND POLITICAL VIOLENCE 241 incidents varies across Nigeria’s two regions: 68 percent of them occurred in Northern Nigeria while the remaining 32 percent occurred in the Southern Region. The conflicts are also spread unevenly across the years, with 9 percent of them occurring between 1997 to 2008, and the remaining 91 percent occurring between 2009 to 2022. The two major conflicts affecting Nigeria are the Boko Haram insurgency and the violent clashes between Muslim nomadic pastoralists of Fulani ethnicity and resident communities (especially those involved in crop cultivation). A report by the Institute of Economics and Peace noted: “In Nigeria, terrorist activity is dominated by Fulani extremists and Boko Haram. Together, they account for 78 percent of terror-related incidents and 86 percent of deaths from terrorism.” 39 (p. 21) The incidence of violent conflict in Nigeria can roughly be broken down into two epochs: preand post-Boko Haram eras. The pre-Boko Haram era covers the period from 1997 to 2008 before the radical Islamist group, Boko Haram, started its insurgency. The post-Boko Haram era covers the years from 2009 onwards after Boko Haram launched its first attack. The Boko Haram insurgency ushered Nigeria into a phase of violence it had never witnessed. The ACLED data shows that between 2009 to 2022, there were 4,776 incidents where at least one of the parties to the conflict was Boko Haram. These incidents caused a total of 43,019 fatalities. Because Boko Haram attacks are concentrated in Northeastern Nigeria where the population is predominantly Muslim (see Figure 1), most of the fatalities from these attacks are Muslims. Figure 1. Incidents involving Boko Haram and nomadic Fulani pastoralists (1997–2022). The figure shows the administrative boundaries of the states that constitute Nigeria’s Northern and Southern Regions. The red dots show the geolocations of conflicts where at least one of the actors is Boko Haram. The blue dots show the geolocations of conflicts where at least one of the actors is a “Pastoralist” or belongs to the “Fulani” ethnic group. Virtually all the actors defined as pastoralists in the ACLED dataset are identified as “Fulani Ethnic militia,” which makes the two terms almost synonymous. Although Northern Nigeria has a predominantly Muslim population, there are a few states there like Benue and Plateau, where the population is predominantly Christian and Muslims constitute a minority. These two states, which were not captured by the Muslim jihadists in the early nineteenth century, have the highest incidence of conflicts involving nomadic Fulani pastoralists. The shapefiles containing Nigeria’s administrative boundaries was developed by UNOCHA. 242 D. TUKI Nigerians tend to associate Muslims with extremism. The Round 7 Afrobarometer survey conducted in 2017, and which is representative for Nigeria’s population, had a question where respondents were asked about the degree to which they thought Muslims supported extremist groups. 40 26 percent of them chose the “none” response category, 37 percent chose the “some of them” response category, 24 percent chose the “most of them” response category, 7 percent chose the “all of them” response category, while the remaining 6 percent refused to answer the question. This suggests that 68 percent of Nigerians associate Muslims with extremism at least to some degree. Disaggregating the data based on religious affiliation revealed that compared to Muslims, Christians are more likely to associate Muslims with extremism: 84 and 48 percent of Christians and Muslims respectively associated Muslims with extremism at least to some degree. The violent clashes between nomadic Fulani pastoralists and resident communities are the second major conflict affecting Nigeria. This conflict, which is primarily caused by increased competition over land and water resources due to droughts, has quickly taken a religious turn because of the distinct ethnic and religious identities of the opposing parties. Some reports have portrayed conflicts involving pastoralists as attacks on Christians by Muslims because the pastoralists are Muslims and most of the communities where these conflicts are concentrated have predominantly Christian populations. 41 Relying on large-N survey data collected from Kaduna, the state with the third highest incidence of farmer-pastoralist conflicts in Nigeria, Tuki found that Christians and Muslims view the conflict differently: 52 percent of Christians agree that farmer-pastoralist conflicts are caused by religion; only 17 percent of Muslims hold this view. 42 The ACLED data shows that between 1997 to 2022, there were 2,416 violent conflicts where at least one of the actors was a pastoralist or belonged to the Fulani ethnic group. These incidents caused a total of 15,333 fatalities. As shown in Figure 2, incidents involving nomadic Fulani pastoralists, unlike Boko Haram attacks, are spread across all of Nigeria’s thirty-six states. This is due to the migratory nature of pastoralists in search of pasture for their livestock. Theoretical considerations Some studies have shown that exposure to violent conflict could foster social cohesion among ingroup members. In a study conducted in Nepal, Gilligan et al. found that communities exposed to violent conflict had higher levels of ingroup trust and prosocial behavior than those that were not. 43 The mechanism behind this finding was that community members who were not socially oriented fled the conflict zone leaving behind those who were more socially oriented. Moreover, the common threat posed by conflict prompted community members to band together so they could better cope. Calvo et al. conducted a study in Mali where they found that conflict exposure had a positive effect on prosocial behavior. 44 Although they acknowledged that social cohesion could foster post-conflict recovery, they pointed out that in the case of Mali this was problematic because increased social participation was observed only in family and ethnically homogenous associations—i.e., “inward-looking associations.” This reinforced kinship ties, made ethnic fault lines salient, and heightened the risk of further conflict. Rohner et al. had a similar finding in a study conducted in Uganda where they found that conflict exposure strengthened cohesion within ethnic ingroups. 45 Conflict has also been found to erode social cohesion. Weidmann and Zürcher (3) found that violent conflict fostered divisions in Afghan communities because it “could introduce shifting loyalties to the fighting parties and thus introduce new internal cleavages.” 46 Relying on survey data collected from members of the Tamil ethnic group in Sri Lanka, Greiner and Filsinger found that men who had been victims of sexual violence during the Sri Lankan Civil War were distrustful of both members of their ethnic group and the ethnic outgroup—i.e., the Sinhalese. 47 Conversely, women who had been victims of sexual violence were distrustful of their ethnic ingroup and had higher levels of trust in the ethnic outgroup. They explained the erosion of ingroup trust on the grounds that “the conflict was characterized by a climate of distrust due to denunciations and betrayal within Tamil communities with harmful consequences for in-group cohesion” (2). Using representative survey data for Pakistan, TERRORISM AND POLITICAL VIOLENCE 243 Ahmad and Rehman found that exposure to terrorist attacks negatively correlated with interpersonal trust. 48 Rohner et al. had a similar finding in Uganda where they found that conflict exposure reduced generalized social trust. 49 In a study conducted in Nigeria, Tuki showed that exposure to conflicts involving nomadic Fulani pastoralists led to distrust in both members of the Fulani ethnic group and Muslims. 50 This was because the Fulani pastoralists were Muslims and the population conflated Fulani ethnicity with being Muslim. Similarly, Kanas et al., in a study conducted among Muslim and Christian students in Indonesia and the Philippines found that the experience of interreligious violence leads to hostility towards religious outgroups. 51 When the perpetrators of violence belong to a distinct cultural outgroup (e.g., based on ethnicity or religion), ingroup members might associate the entire outgroup with violence even if only a few of them were involved in the act, a phenomenon that Hall et al. referred to as the “better safe than sorry approach.” 52 This is associated with the concept of prejudice which Allport (7) defined as “an aversive or hostile attitude towards a person or group, simply because he belongs to that group, and is therefore presumed to have the objectionable qualities ascribed to the group.” 53 In a similar vein, Lickel et al. developed a theory to explain the psychological mechanisms underlying retributive violence. 54 Vicarious retribution, they observed, “occurs when a member of a group commits an act of aggression toward members of an outgroup for an assault or provocation that had no personal consequences for him or her, but did harm a fellow ingroup member.” 55 They pointed out that when an act of aggression occurred, people who were not directly involved in the conflict tried to make sense of it by construing it in terms of the broader ingroup-outgroup dichotomy between the conflict actors. If an ingroup-outgroup distinction was salient, they would then interpret the event in a way that was favorable toward their ingroup and encouraged retaliation against members of the outgroup. However, when ingroup-outgroup distinctions could not be extrapolated from the initial act of aggression, people were likely to interpret it as a personal dispute between two individuals. This reduced the likelihood of retaliation. Ahmed has shown how the terrorist attack that occurred in the U.S. on September 11, 2001 altered perceptions towards British Muslims in the U.K. 56 The ensuing “War on Terror” policy shifted the British government’s focus from the diverse Asian identity of British Muslims to their religious identity, which portrayed them as a “suspect community” and associated them with terrorism. As she concisely put it, “it is the Muslim in British Muslim which now shapes the concrete policies which govern British Muslims.” 57 Ferwerda et al. conducted an experimental study in the U.S. where they found that the association of Muslim refugees with terrorism reduced support for refugee resettlement both within the U.S. and within the communities where the participants resided. 58 Their analysis also showed that exposing subjects to counter frames that challenged the portrayal of refugees as threats had no statistically significant effect on support for refugee resettlement. This indicates that negative attitudes towards cultural outgroups, once formed, tend to persist. In a study conducted in Kenya, Schutte et al. found that indiscriminate violence caused fear of religious outgroups, strengthened ingroup cohesion, and led to increased calls for residential segregation along religious lines. 59 Moreover, they found that attacks perpetrated by Islamist insurgents led to distrust in Muslims. In another study conducted in India, Schutte et al. found that conflict not only caused prejudice towards religious outgroups and strengthened ingroup cohesion, but also increased support for extremist activities perpetrated by ingroup members. 60 Using experiments, Obaidi et al. have shown that the perceived cultural threat posed by Muslims leads to increased support for the persecution of the Muslim outgroup among the Swedish and Danish populations. 61 They also found a similar effect among Muslims who view Western culture as decadent and a threat to Islam. Conversely, Whitt et al. conducted an experimental study in Syria, Bosnia and Kosovo where they found that hostile attitudes towards outgroups tend to change following productive interactions between the two groups. 62 This is consistent with the premise of the contact hypothesis put forth by Allport, which asserts that intergroup contact, conditional upon cooperation towards a common goal and equality between the groups, reduces prejudice. 63 244 D. TUKI Returning to the Nigerian case, I expect conflict exposure to have a positive effect on outgroup hostility, especially because of how polarized the country’s population is along ethnic and religious lines, coupled with the huge importance that Nigerians attach to their ethnoreligious identities. This facilitates the construction of ingroups and outgroups. However, there might be heterogenous effects among Christians and Muslims: Among Muslims, it is likely that exposure to violent conflict would have no effect on hostility towards ethnoreligious outgroups. This is because the main conflict affecting Muslims—i.e., the Boko Haram insurgency—does not involve Christians. Moreover, a significant number of Muslims are affected by the violent clashes involving nomadic Fulani pastoralists who are also Muslims. The common religion of Islam between the conflict actors thus makes it difficult for Muslims to establish ingroup-outgroup distinctions. Among Christians, however, conflict exposure is likely to have a positive effect on outgroup hostility because the major conflict affecting them involves nomadic Fulani pastoralists who are Muslims. Because the conflict actors belong to different religious groups, it becomes easy to establish ingroup-outgroup distinctions. Moreover, nomadic Fulani pastoralists tend to be perceived as a “suspect community” with a high predisposition toward violence. 64 I will test the following hypotheses: Hypothesis 1: Among Nigerians, conflict exposure leads to hostility towards ethnoreligious outgroups. Hypothesis 2: Among Christians, conflict exposure leads to hostility towards ethnoreligious outgroups. Hypothesis 3: Among Muslims, conflict exposure has no effect on hostility towards ethnoreligious outgroups. Data and methodology This study relies on the Round 7 Afrobarometer survey data 65 collected in 2017. 66 The dataset consists of 1,600 observations and is representative for Nigeria’s population. Respondents were drawn from each of Nigeria’s thirty-six states and the federal capital territory—Abuja. Of Nigeria’s 774 local government areas (LGAs) (i.e., municipalities), data were collected from 147 of them. Respondents were at least eighteen years old, with males and females equally represented in the sample. Table A1 in the appendix reports the summary statistics of the variables that were used to estimate the regression models. Operationalization of the variables Dependent variable Outgroup hostility. This is an additive indicator that measures the respondents’ willingness to have people from other religions and other ethnic groups as neighbors. It was derived by combining the responses to the following two questions: “For each of the following types of people, please tell me whether you would like having people from this group as neighbors, dislike it, or not care: (a) People of a different religion? (b) People from other ethnic groups?” The responses were measured on a five-point ordinal scale ranging from “1 = strongly like,” to “5 = strongly dislike.” The additive indicator ranges from 2 to 10, with higher values denoting a higher level of outgroup hostility and vice versa. 67 I treated the “don’t know” and “refused to answer” responses as missing observations. I applied this rule to all variables derived from the Afrobarometer survey. TERRORISM AND POLITICAL VIOLENCE 245 The two survey items had a Cronbach Alpha statistic of 0.84, which shows internal reliability. The two items also had a correlation of 0.72, which highlights the close association between ethnicity and religion in Nigeria. As shown in the first two bar charts from the top of Figure 2, Nigerians have a slightly higher level of hostility towards religious outgroups than ethnic outgroups. Christians are slightly more hostile towards people of a different ethnic group than Muslims. Muslims are slightly more hostile towards people of a different religion than Christians. Both Christians and Muslims are more hostile towards people of a different religion than people of a different ethnic group. Explanatory variable Violent conflict. This measures the total number of violent conflict incidents within the thirtykilometer buffer around the respondents’ dwellings. I developed the buffers using QGIS software. This was possible because I relied upon data obtained from Afrobarometer 68 and ACLED, 69 both of which are georeferenced. Based on the ACLED dataset, I define a violent conflict as any incident that falls under any of the following three categories: Battles, Violence against civilians, and Explosions/Remote violence. 70 Although the ACELD dataset is available starting from 1997 and is updated in real time, I excluded conflict incidents that occurred after 2016. This lags the explanatory variable since the dependent variable is measured in 2017. I considered all the conflict incidents within the buffer from 1997 to 2016 because I am particularly interested in the cumulative effect of violent conflict. Some studies have shown that memories from past conflicts tend to persist and could shape action in the present. 71 Buffers are a more efficient way of measuring exposure to violent conflict than the LGA administrative boundaries. This is because the spatial area occupied by each buffer is unique for each respondent and allows for more variation in the conflict exposure variable. If I had measured conflict exposure at the LGA level, I would have associated all the respondents residing within a particular LGA with the total number of conflict incidents there, which presumes that all respondents residing within a particular LGA are exposed to the same level of violent conflict. This would have been inefficient because incidents in a contiguous LGA might be nearer to a respondent’s dwelling than those in the particular LGA where he/she resides. As shown in Figure 3, the respondent resides in Asa LGA, yet conflicts in Moro, Olorunsogo, and Ori Ire LGAs are closer to his/her dwelling than some incidents in Asa LGA. Another challenge that comes along with working with Nigeria’s administrative boundaries (especially those at the lower levels) is that they are not clearly defined. In fact, there were a few observations where respondents residing close to Nigeria’s national border were more exposed to conflicts in the contiguous Figure 2. Hostility towards ethnic and religious outgroups. The y-axis shows the total number of respondents in the full sample, and the number of Muslim/Christian respondents who had answered the relevant questions regarding their willingness to have people from a different religion and ethnic group as neighbors. The x-axis shows the percentage of respondents who chose a particular response category. 246 D. TUKI Wu-Hausman statistics were both insignificant, which suggests that endogeneity was not present and the use of an instrumental variable approach to estimate the model was inappropriate. I thus reestimated the model using OLS regression. As shown in model 3, violent conflict remained statistically insignificant. This suggests that among Muslims, conflict exposure has no effect on hostility towards people of a different religion. This is consistent with Hypothesis 3. In models 4, 5, and 6, the dependent variable measures hostility towards ethnic outgroups only. In model 4 which was estimated using the Christian subsample of respondents, violent conflict carried the expected positive sign and was significant at the one percent level, which indicates that among Christians, exposure to violent conflict has a positive effect on hostility towards ethnic outgroups. The size of the coefficient does not differ much from that in model 1, which further highlights the close association between ethnicity and religion in Nigeria. In model 5, which was estimated using the Muslim subsample of respondents, violent conflict was statistically insignificant. Because the Durbin and Wu-Hausman statistics were both statistically insignificant, I re-estimated the model using OLS regression. As shown in model 6, violent conflict remained statistically insignificant, which indicates that among Muslims, conflict exposure has no effect on hostility towards ethnic outgroups. This supports Hypothesis 3. Conclusion This study examined the effect of exposure to violent conflict on hostility towards ethnoreligious outgroups among the Nigerian population and among its two major religious groups (i.e., Christians and Muslims). Causal identification stemmed from instrumenting conflict Table 3. Effect of violent conflict on outgroup hostility II (Full sample) Outgroup hostility ϕ Religion Ethnicity (1) (2) (3) (4) Violent conflict # 0.002** 0.014*** 0.002** 0.011*** (0.001) (0.004) (0.001) (0.004) Nighttime light # −0.236*** −0.194*** (0.066) (0.059) Prevalence of stunting # 2.149*** 1.618*** (0.687) (0.62) Household deprivation 0.007 0.015* (0.009) (0.008) Log Population size # −0.006 −0.00 (0.076) (0.068) Educational level −0.098*** −0.092*** (0.022) (0.02) Religious affiliation 0.176 0.22* (0.132) (0.119) Gender −0.217*** −0.124* (0.073) (0.066) Age −0.006** −0.003 (0.003) (0.003) Constant 2.552*** 1.98** 2.286*** 1.738** (0.072) (0.977) (0.065) (0.882) Estimation method 2SLS 2SLS 2SLS 2SLS Observations 1439 1408 1438 1407 R-squared 0.116 0.118 Durbin statistic 9.083*** 12.01*** 6.578** 9.91*** Wu-Hausman statistic 8.95*** 11.787*** 6.470** 9.711*** Sargan statistic 0.317 0.05 0.018 0.163 Basmann statistic 0.311 0.049 0.018 0.158 ϕ is the dependent variable, # denotes variables measured using buffers with a radius of 30 kilometers, standard errors are in parenthesis, *** p < 0.01, ** p < 0.05, * p < 0.10. All models are estimated using twostage least squares (2SLS) regression. All models contain fixed effects for the respondents’ ethnic groups. Only the second-stage regressions are reported here. TERRORISM AND POLITICAL VIOLENCE 253 exposure with forest cover. The regression results showed that among the Nigerian population and Christians, conflict exposure had a robust positive effect on outgroup hostility. A plausible explanation for this finding is that the threat of violent conflict fosters cohesion within ingroup members, erodes trust in outgroup members, and makes intergroup boundaries salient. This is especially so when the opposite party to the conflict constitutes a distinct cultural outgroup. Because the main conflict affecting Christians involves nomadic Fulani pastoralists who are Muslims, it becomes easy for ingroup-outgroup distinctions to be established. Among Muslims, violent conflict had a weak positive effect on outgroup hostility that was robust to different operationalizations of outgroup hostility. This null effect among Muslims is likely because the main conflict affecting them—i.e., the Boko Haram insurgency —does not involve Christians. Moreover, a significant number of Muslims are affected by conflicts involving nomadic Fulani pastoralists. The common religion of Islam shared by the parties makes the establishment of ingroup-outgroup boundaries arduous, and makes it illogical for Muslims to be hostile towards Christians. The regression results also showed that religion is closely associated with ethnicity in Nigeria, and the population tends to conflate the two. This is problematic because it makes intergroup boundaries more salient, which in turn heightens the risk of conflict. If the Nigerian government intends to reduce violent conflict and outgroup hostility, it would have to adopt a policy that tackles these two factors simultaneously because each one reinforces the other. For instance, the government could reduce the incidence of violent conflict by equipping its security agencies with the requisite skills and equipment needed to respond promptly and effectively to conflict situations, while simultaneously pursuing policies that foster social cohesion and elevate a shared national identity over ethnic and religious Table 4. Effect of violent conflict on outgroup hostility III (Religious subsamples) Outgroup hostility ϕ Religion Ethnicity (1) (2) (3) (4) (5) (6) (Xtian) (Muslim) (Muslim) (Xtian) (Muslim) (Muslim) Violent conflict # 0.025*** 0.004 0.001 0.021*** 0.003 0.001 (0.007) (0.005) (0.001) (0.006) (0.005) (0.001) Nighttime light # −0.464*** −0.05 −0.02 −0.393*** −0.051 −0.018 (0.113) (0.067) (0.022) (0.105) (0.061) (0.02) Prevalence of stunting # −2.17** 2.738** 2.163*** −0.836 1.807 1.182** (1.044) (1.356) (0.623) (0.972) (1.235) (0.565) Household deprivation −0.001 0.022* 0.023* 0.015 0.019* 0.019* (0.012) (0.013) (0.013) (0.011) (0.011) (0.012) Log Population size # 0.111 −0.039 0.018 0.134 −0.027 0.035 (0.115) (0.143) (0.079) (0.108) (0.13) (0.072) Educational level 0.019 −0.109*** −0.102*** −0.02 −0.089*** −0.081*** (0.031) (0.029) (0.025) (0.029) (0.026) (0.022) Gender −0.019 −0.514*** −0.517*** 0.062 −0.39*** −0.393*** (0.097) (0.101) (0.102) (0.09) (0.092) (0.092) Age −0.005 −0.005 −0.004 −0.003 −0.00 0.001 (0.004) (0.004) (0.004) (0.004) (0.004) (0.004) Constant 1.262 2.269 1.777* 0.729 2.079 1.546 (1.866) (1.458) (1.041) (1.737) (1.326) (0.945) Estimation method 2SLS 2SLS OLS 2SLS 2SLS OLS Observations 807 601 601 806 601 601 R-squared 0.187 0.198 0.008 0.124 0.142 Durbin statistic 14.998*** 0.23 13.32*** 0.329 Wu-Hausman statistic 14.639*** 0.221 12.973*** 0.316 Sargan statistic 1.232 4.236** 1.958 1.142 Basmann statistic 1.182 4.103** 1.88 1.1 ϕ is the dependent variable, # denotes variables measured using buffers with a radius of thirty kilometers, standard errors are in parenthesis, *** p < 0.01, ** p < 0.05, * p < 0.10. All models are estimated using two-stage least squares (2SLS) regression, except for models 3 and 6 which are estimated using ordinary least squares (OLS) regression. All models contain fixed effects for the respondents’ ethnic groups. Only the second-stage regressions are reported here. 254 D. TUKI identities, e.g., by encouraging inter-ethnic and inter-religious dialogue. However, I must also acknowledge that the latter recommendation might be difficult to achieve because it is not uncommon for the Nigerian elites to exploit the ethnic and religious divisions among the population for political gain. Acknowledgments An early version of this paper was presented at a colloquium organized by the Migration Integration and Transnationalization Research Unit at the WZB Berlin Social Science Center, Germany. I thank the participants for their feedback. Thanks to the handling editor and two anonymous referees for their feedback. Thanks to Roisin Cronin for editorial assistance. Disclosure statement No potential conflict of interest was reported by the author(s). Data availability statement The data underlying this study are available in the Harvard Dataverse. https://doi.org/10.7910/DVN/0WOK6X Notes on contributor Daniel Tuki is a Research Fellow at the WZB Berlin Social Science Center, Germany. His research focuses on conflict studies and economic development. ORCID Daniel Tuki http://orcid.org/0000-0003-1097-3845 Notes 1. Although Nigeria has 250 ethnic groups, it has three major ones: The Hausa/Fulani who are predominantly Muslim and mainly reside in Northern Nigeria. The Igbo and the Yoruba constitute the major ethnic groups in Southern Nigeria. The Igbos are predominantly Christian, while the Yoruba is evenly split between Muslims and Christians (D. D. Laitin, Hegemony and Culture: Politics and Religious Change Among the Yoruba (Chicago: University of Chicago Press, 1986). 2. P. S. Huntington, The Clash Of Civilizations and the Remaking of World Order (New York: Simon & Schuster, 1996). 3. T. Falola and M. M. Heaton, A History of Nigeria (Cambridge: Cambridge University Press, 2008). 4. A. H. M. Kirk-Greene, The Principles of Native Administration in Nigeria (London: Oxford University Press, 1965). 5. Falola and Heaton, A History of Nigeria; A. F. Ogunsola, Legislation and Education in Northern Nigeria (Ibadan: Oxford University Press, 1974). 6. I. O. Albert, “Ethnic Residential Segregation in Kano, Nigeria and its Antecedents,” African Study Monographs 17, no. 2 (1996): 85–100. 7. J. Campbell and M. T. Page, Nigeria: What Everyone Needs to Know (New York: Oxford University Press, 2018); L. Diamond, Class, Ethnicity and Democracy in Nigeria: The Failure of the First Republic (London: Macmillan, 1988); J.S. Coleman, Nigeria: Background to Nationalism (Berkeley: University of California Press, 1958). 8. Coleman, Nigeria: Background to Nationalism. 9. L. Plotnicov, “An Early Nigerian Civil Disturbance: The 1945 Hausa-Ibo Riot in Jos,” Journal of Modern African Studies 9, no. 2 (1971): 297–305. 10. I. O. Albert, “Violence in Metropolitan Kano: A Historical Perspective,” in IFRA, Urban Violence in Africa (Ibadan: Institut Français de Recherche en Afrique, 1994), 111–138, https://books.openedition.org/ifra/788? lang=en (accessed October 1, 2023). TERRORISM AND POLITICAL VIOLENCE 255 11. C. Achebe, There was a Country: A Personal History of Biafra (London: Penguin Books, 2012); H. Ekwe-Ekwe, Conflict and Intervention in Africa: Nigeria, Angola, Zaire (New York: Palgrave Macmillan, 1990). 12. J. Ibrahim, “The Politics of Religion in Nigeria: The Parameters of the 1987 Crisis in Kaduna State,” Review of African Political Economy 1, no. 45–46 (1989): 65–82. 13. Albert, “Ethnic Residential Segregation in Kano, Nigeria and its Antecedents.” 14. H. Angerbrandt, “Political Decentralisation and Conflict: The Sharia Crisis in Kaduna, Nigeria,” Journal of Contemporary African Studies 29, no. 1 (2011): 15–31; Human Rights Watch, “The Miss World Riots: Continued Impunity for Killing in Kaduna,” Human Rights Watch, 2003, https://www.hrw.org/reports/2003/nigeria0703/2. html (accessed August 4, 2023). 15. H. Angerbrandt, “Deadly Elections: Post-Election Violence in Nigeria,” The Journal of Modern African Studies 56, no. 1 (2018): 143–167. 16. S. Eke, “Why does Segregation Prevent Conflict in Some Regions but not Others? Interrogating Social Distance Amid Ethnic Conflicts in Jos, Nigeria,” Third World Quarterly 43, no. 9 (2022): 2209–2224; J. Krause, “Gender Dimensions of (non) Violence in Communal Conflict: The Case of Jos, Nigeria,” Comparative Political Studies 52, no. 10 (2019): 1466–1499; A.R. Mustapha, A. Higazi, J. Lar, and K. Chromy, “Jos: Fear and Violence in Central Nigeria,” in Creed and Grievance: Muslim-Christian Relations and Conflict Resolution in Northern Nigeria, eds. A. R. Mustapha and D. Ehrhardt (New York: James Currey, 2018), 225–268; Human Rights Watch, They do not Own this Place: Government Discrimination Against “Non-Indigenes” in Nigeria (New York: Human Rights Watch, 2006), https://www.hrw.org/sites/default/files/reports/nigeria0406webwcover.pdf (accessed October 11, 2023); Human Rights Watch, Jos: A city Torn Apart (New York: Human Rights Watch, 2001), https://www.hrw.org/ sites/default/files/reports/nigeria1201_1.pdf (accessed October 11, 2023). 17. N. DeVotta, “From Ethnic Outbidding to Ethnic Conflict: The Institutional Bases for Sri Lanka’s Separatist War,” Nations and Nationalism 11, no. 1 (2005): 141–159; R. Ganguly, “Sri Lanka’s Ethnic Conflict: At a Crossroad Between Peace and War,” Third World Quarterly 25, no. 5 (2004): 903–918; G. Pradhan, “Economic cost of Sri Lanka’s Ethnic Conflict,” Journal of Contemporary Asia 31, no. 3 (2001): 375–384. 18. G. Van Klinken and S. M. T. Aung, “The Contentious Politics of Anti-Muslim Scapegoating in Myanmar,” Journal of Contemporary Asia 47, no. 3 (2017): 353–375; M. Schissler, M. J. Walton, and P. P. Thi, “Reconciling Contradictions: Buddhist-Muslim Violence, Narrative Making and Memory in Myanmar,” Journal of Contemporary Asia 47, no. 3 (2017): 376–395; N. Kipgen, “Conflict in Rakhine State in Myanmar: Rohingya Muslims’ Conundrum,” Journal of Muslim Minority Affairs 33, no. 2 (2013): 298–310. 19. A. Mitra and D. Ray, “Implications of an Economic Theory of Conflict: Hindu-Muslim Violence in India,” Journal of Political Economy 122, no. 4 (2014): 719–765; Z. Kausar, “Communal Riots in India: Hindu–Muslim Conflict and Resolution,” Journal of Muslim Minority Affairs 26, no. 3 (2006): 353–370. 20. C. K. Sharma, “The Immigration Issue in Assam and Conflicts Around it,” Asian Ethnicity 13, no. 3 (2012): 287– 309; A. T. Darnell and S. Parikh, “Religion, Ethnicity, and the Role of the State: Explaining Conflict in Assam,” Ethnic and Racial Studies 11, no. 3 (1988): 263–281; S. Goswami, “Ethnic Conflict in Assam,” The Indian Journal of Political Science 62, no. 1 (2001): 123–137. 21. T. German, “The Nagorno-Karabakh Conflict between Azerbaijan and Armenia: Security Issues in the Caucasus,” Journal of Muslim Minority Affairs 32, no. 2 (2012): 216–229; B. Özkan, “Who Gains from the ‘No War No Peace’ Situation? A Critical Analysis of the Nagorno-Karabakh Conflict,” Geopolitics 13, no. 3 (2008): 572–599. 22. Y. L. Fong and M. S. A. Ishak, “Framing Interethnic Conflict in Malaysia: A Comparative Analysis of Newspapers Coverage on the Keris Polemics,” Ethnicities 14, no. 2 (2014): 252–278; K. K. Soong, “Racial Conflict in Malaysia: Against the Official History,” Race & Class 49, no. 3 (2008): 33–53; E. K. B. Tan, “From Sojourners to Citizens: Managing the Ethnic Chinese Minority in Indonesia and Malaysia,” Ethnic and Racial Studies 24, no. 6 (2001): 949–978. 23. Y. L. Fong and M. S. A. Ishak, “Framing Inter-Religious Dispute: A Comparative Analysis of Chinese-, English-, and Malay-Language Newspapers in Peninsular Malaysia,” Chinese Journal of Communication 9, no. 2 (2016): 103–123. 24. M. Schaub, “Solidarity with a Sharp Edge: Communal Conflict and Local Collective Action in Rural Nigeria,” Afrobarometer, Working Paper No. 149, 2014; A. Scacco and S. S. Warren, “Christian-Muslim Relations in the Shadow of Conflict: Insights from Kaduna, Nigeria,” in The Oxford Handbook of Politics in Muslim Societies, eds. M. Cammett and P. Jones (New York: Oxford University Press, 2021),539–560; Albert, “Ethnic Residential Segregation in Kano, Nigeria and its Antecedents.” 25. A. B. Downes, “More Borders, Less Conflict? Partition as a Solution to Ethnic Civil Wars,” SAIS Review 26, no. 1 (2006): 9–61; J. Tir, “Dividing Countries to Promote Peace: Prospects for Long-Term Success of Partitions,” Journal of Peace Research 42, no. 5 (2005): 545–562. 26. M. Bollig, “Intra-and Interethnic Conflict in Northwest Kenya: A Multicausal Analysis of Conflict Behaviour,” Anthropos 88, no. 1/3 (1993): 176–184. 27. J. M. Larson, “Interethnic Conflict and the Potential Dangers of Cross-Group Ties,” Journal of Peace Research 53, no. 3 (2016): 459–471. 256 D. TUKI 28. S. Eke, “Why does Segregation Prevent Conflict in Some Regions but not Others? 29. J. Rydgren, D. Sofi, and M. Hällsten, “Interethnic Friendship, Trust, and Tolerance: Findings from Two North Iraqi cities,” American Journal of Sociology 118, no. 6 (2013): 1650–1694. 30. A. Kanas, P. Scheepers, and C. Sterkens, “Positive and Negative Contact and Attitudes Towards the Religious Out-Group: Testing the Contact Hypothesis in Conflict and Non-Conflict Regions of Indonesia and the Philippines,” Social Science Research 63 (2017): 95–110. 31. G. Allport, The Nature of Prejudice (Boston: The Beacon Press, 1954). 32. D. Tuki, “Pastoral Conflicts and (dis)Trust: Evidence from Nigeria using an Instrumental Variable Approach,” WZB Berlin Social Science Center, Discussion Paper SP VI 2023–101, 2023; D. E. Agbiboa, “Ethno-Religious Conflicts and the Elusive Quest for National Identity in Nigeria,” Journal of Black Studies 44, no, 1 (2013): 3–30. 33. D. E. Agbiboa and B. Maiangwa, “Boko Haram, Religious Violence, and the Crisis of National Identity in Nigeria: Towards a Non-Killing Approach,” Journal of Developing Societies 29, no. 4 (2013): 379–403. 34. A. BenYishay, R. Rotberg, J. Wells, Z. Lv, S. Goodman, L. Kovacevic, and D. Runfola, “Geocoding Afrobarometer Rounds,” (2017) 1–6: Methodology & Data Quality, “AidData,” https://www.aiddata.org/publications/geocod ing-afrobarometer-rounds-1-6-methodology-data-quality. 35. C. Raleigh, A. Linke, H. Hegre, and J. Karlsen, “Introducing ACLED—Armed Conflict Location and Event Data,” Journal of Peace Research 47, no. 5 (2010): 651–660. 36. S. Schutte, C. Ruhe, and N. Sahool, “How Fear of Violence Drives Intergroup Conflict: Evidence from a Panel Survey in India,” Terrorism and Political Violence 35, no. 2 (2023): 229–247; S. Schutte, C. Ruhe, and A. M. Linke, “How Indiscriminate Violence Fuels Conflicts Between Groups: Evidence from Kenya,” Social Science Research 103 (2022): 102,653; D. Tuki, “Pastoral Conflicts and (dis)Trust: Evidence from Nigeria Using an Instrumental Variable Approach, WZB Berlin Social Science Center,” Discussion Paper SP VI 2023–101, 2023; S. Whitt, R. K. Wilson, and V. Mironova, “Inter-Group Contact and Out-Group Altruism After Violence,” Journal of Economic Psychology 86 (2021): 1–19; T. Calvo, E. Lavallée, M. Razafindrakoto, and F. Roubaud, “Fear not for Man? Armed Conflict and Social Capital in Malim,” Journal of Comparative Economics 48, no. 2 (2020): 251–276; J. Ferwerda, D. J. Flynn, and Y. Horiuchi, “Explaining Opposition to Refugee Resettlement: The role of NIMBYism and Perceived Threats,” Science Advances 3, no. 9 (2017): 1–7. 37. C. Raleigh, A. Linke, H. Hegre, and J. Karlsen, “Introducing ACLED—Armed Conflict Location and Event Data,” Journal of Peace Research 47, no. 5 (2010): 651–660. 38. Based on the ACLED dataset, I define violent conflicts as incidents categorized under any of the following three categories: Battles, Violence against civilians, and Explosions/Remote violence. This implies that I have excluded incidents categorized as Riots, Protests, and Strategic developments. 39. Institute for Economics and Peace, Global Terrorism Index 2019: Measuring Peace in a Complex World (Sydney: IEP, 2019), http://visionofhumanity.org/resources (accessed July 5, 2023). 40. BenYishay et al., “Geocoding Afrobarometer Rounds.” 41. Christian Association of Nigeria, “Nigerian Church Protest Killing of Christians,” CAN Nigeria, 2018, https://canng. org/news-and-events/news/173-nigerian-church-protest-killing-of-christians (accessed July 5, 2023); Christian Association of Nigeria, “Political Opposition Sponsorship of Killings in Nigeria: Questions Buhari Government Must Answer,” CAN Nigeria, 2018a, July 11, https://canng.org/news-andevents/news/181-political-oppositionsponsorship-of-killings-in-nigeria-questions-buharigovernment-must-answer (accessed August 4, 2023). 42. D. Tuki, “Pastoral Conflicts and (dis)Trust.” 43. M. J. Gilligan, B. J. Pasquale, and C. Samii, “Civil War and Social Cohesion: Lab‐in‐the‐Field Evidence from Nepal,” American Journal of Political Science 58, no. 3 (2014): 604–619. 44. T. Calvo, E. Lavallée, M. Razafindrakoto, and F. Roubaud, “Fear Not for Man? Armed Conflict and Social Capital in Mali,” Journal of Comparative Economics 48, no. 2 (2020): 251–276. 45. D. Rohner, M. Thoenig, and F. Zilibotti, “Seeds of Distrust: Conflict in Uganda,” Journal of Economic Growth 18, no. 3 (2013): 217–252. 46. N.B. Weidmann and C. Zürcher, “How Wartime Violence Affects Social Cohesion: The Spatial-Temporal Gravity Model,” Civil Wars 15, no. 1 (2013): 1–18. 47. A. Greiner and M. Filsinger, “(Dis) Trust in the Aftermath of Sexual Violence: Evidence from Sri Lanka,” Households in Conflict Network, Working Paper No. 377, 2022. 48. N. Ahmad and F. U. Rehman, “Does Terrorism Reduce Trust? Empirical Evidence from Pakistan,” Defence and Peace Economics 33, no. 8 (2022): 993–1009. 49. D. Rohner, M. Thoenig, and F. Zilibotti, “Seeds of Distrust: Conflict in Uganda,” Journal of Economic Growth 18, no. 3 (2013): 217–252. 50. D. Tuki, “Pastoral Conflicts and (dis)Trust.” 51. Kanas, Scheepers, and Sterkens, “Positive and Negative Contact and Attitudes Towards the Religious Out-Group.” 52. J. Hall, D. T. Kahn, E. Skoog, and M. Öberg, “War Exposure, Altruism and the Recalibration of Welfare Tradeoffs Towards Threatening Social Categories,” Journal of Experimental Social Psychology 94 (2021): 1–9. 53. Allport, The Nature of Prejudice. TERRORISM AND POLITICAL VIOLENCE 257 54. B. Lickel, N. Miller, D. M. Stenstrom, T. F. Denson, and T. Schmader, “Vicarious Retribution: The Role of Collective Blame in Intergroup Aggression,” Personality and Social Psychology Review 10, no. 4 (2006): 372–390. 55. Lickel, Miller, Stenstrom, Denson, and Schmader, “Vicarious retribution.” 56. S. Ahmed, “British Muslims Perceptions of Social Cohesion: From Multiculturalism to Community Cohesion and the ‘War on Terror,’” Crime, Law and Social Change 71, no. 5 (2019): 581–595. 57. Ahmed, “British Muslims Perceptions of Social Cohesion.” 58. J. Ferwerda, D. J. Flynn, & Y. Horiuchi, “Explaining Opposition to Refugee Resettlement: The Role of NIMBYism and Perceived Threats,” Science Advances 3, no. 9 (2017): 1–7. 59. S. Schutte, C. Ruhe, and A. M. Linke, “How Indiscriminate Violence Fuels Conflicts between Groups: Evidence from Kenya,” Social Science Research 103 (2022): 102653. 60. Schutte, Ruhe, and Sahoo, “How Fear of Violence Drives Intergroup Conflict.” 61. M. Obaidi, L. Thomsen, and R. Bergh, “They Think We are a Threat to their Culture”: Meta-Cultural Threat Fuels Willingness and Endorsement of Extremist Violence against the cultural outgroup. International Journal of Conflict and Violence 12 (2018): 1–13. 62. S. Whitt, R. K. Wilson, and V. Mironova, “Inter-Group Contact and Out-Group Altruism after Violence,” Journal of Economic Psychology 86 (2021): 1–19. 63. See note 53 above. 64. D. Tuki, “You’re not Welcome! Violence and Support for a Grazing Ban Policy in Kaduna, Nigeria.” Households in Conflict Network, Working Paper, No. 397, 2023; P. F. Ejiofor, “‘Fulanis are Foreign Terrorists’: The Social Construction of a Suspect Community in the Sahel,” Critical Studies on Terrorism 15, no. 2 (2022): 333–355; S. Eke, “‘Nomad Savage’ and Herder-Farmer Conflicts in Nigeria: The (Un)making of an Ancient Myth,” Third World Quarterly 41, no 5 (2020): 745–763. 65. See note 34 above. 66. To access the Afrobarometer dataset and the survey questionnaire visit: https://www.afrobarometer.org/. 67. In the original Afrobarometer dataset, higher ordinal values denote a lower level of outgroup hostility and vice versa. For easy interpretation of the regression results, I inverted the ordinal values assigned to the response categories by subtracting each of them from 6; this allows higher (lower) values to denote a higher (lower) level of outgroup hostility. 68. See note 34 above. 69. C L. Raleigh, A. Inke, H. Hegre, and J. Karlsen, “Introducing ACLED—Armed Conflict Location and Event Data,” Journal of Peace Research 47, no. 5 (2010): 651–660. 70. To access the ACLED dataset visit: https://acleddata.com/. 71. B. Wagoner, and I. Brescó, “Conflict and Memory: The Past in the Present,” Journal of Peace Psychology 22, no. 1 (2016): 3–4; B. Tint, “History, Memory, and Intractable Conflict,” Conflict Resolution Quarterly, 27, no. 3 (2010): 239–256. 72. J. Herbst, States and Power in Africa: Comparative Lessons in Authority and Control (Princeton University Press, 2000); P. Collier and A. Hoeffler, “Greed and Grievance in Civil War,” World Bank, Policy Research Working Paper No. 2355, 2000. 73. C. Xu, “Effects of Urbanization on Trust: Evidence from an Experiment in the Field,” Journal of Economic Psychology 87 (2021): 1–8; J. Delhey, and K. Newton, “Predicting Cross-National Levels of Social Trust: Global Pattern or Nordic Exceptionalism? European Sociological Review 21, no. 4 (2005): 311–327. 74. To access the population dataset visit: https://www.worldpop.org/. 75. T. Ghosh, K. E. Baugh, C. D. Elvidge, M. Zhizhin, A. Poyda, and F. C. Hsu, “Extending the DMSP Nighttime Lights Time Series Beyond 2013,” Remote Sensing 13, no. 5004 (2021): 1–19. 76. R. Florida, C. Mellander, and T. Gulden, “Global Metropolis: Assessing Economic Activity in Urban Centers based on Nighttime Satellite Images,” The Professional Geographer 64, no. 2 (2012): 178–187; C. Mellander, J. Lobo, K. Stolarick, and Z. Matheson, “Night-time Light Data: A Good Proxy Measure for Economic Activity?” PloS One 10, no. 10 (2015): 1–18; N.B. Weidmann, and S. Schutte, “Using Night Light Emissions for the Prediction of Local Wealth,” Journal of Peace Research 54, no. 2 (2017): 125–140. 77. P. Collier, The Bottom Billion: Why the Poorest Countries are Failing and What Can be Done About It (New York: Oxford University Press, 2008). 78. A. J. Stewart, N. McCarty, and J. J. Bryson, “Polarization under Rising Inequality and Economic Decline,” Science Advances 6, no. 50 (2020): 1–9; J. Delhey, and K. Newton, “Predicting Cross-National Levels of Social Trust: Global Pattern or Nordic Exceptionalism?” European Sociological Review 21, no. 4 (2005): 311–327. 79. To access the nighttime light dataset visit: https://eogdata.mines.edu/products/dmsp/. 80. C. Bosco, V. Alegana, T. Bird, C. Pezzulo, L. Bengtsson, A. Sorichetta, and A. J. Tatem, “Exploring the Highresolution Mapping of Gender-disaggregated Development Indicators,” Journal of The Royal Society Interface 14, no. 129 (2017): 1–12. 81. N. Von Uexkull, “Sustained Drought, Vulnerability and Civil Conflict in Sub-Saharan Africa,” Political Geography 43 (2014): 16–26; Q. T. Do and L. Iyer, “Geography, Poverty and Conflict in Nepal,” Journal of 258 D. TUKI Peace Research 47, no. 6 (2010): 735–748; P. Collier, The Bottom billion: Why the Poorest Countries are Failing and What can be Done About It (New York: Oxford University Press, 2008). 82. J. Gereke, M. Schaub, and D. Baldassarri, “Ethnic Diversity, Poverty and Social Trust in Germany: Evidence from a Behavioral Measure of Trust,” PloS One 13, no. 7 (2018): 1–15; A. Alessina, and E. La Ferrara, “The Determinants of Trust, National Bureau of Economic Research,” Working Paper No. 7621, 2000. 83. S. Gordon, “Xenophobia Across the Class Divide: South African Attitudes Towards Foreigners 2003–2012,” Journal of Contemporary African Studies 33, no. 4 (2015): 494–509; D. Everatt, “Xenophobia, State and Society in South Africa, 2008–2010,” Politikon 38, no. 1 (2011): 7–36; M. Choane, L. S. Shulika, and M. Mthombeni, “An Analysis of the Causes, Effects and Ramifications of Xenophobia in South Africa,” Insight on Africa 3, no. 2 (2011): 129–142. 84. To access the prevalence of stunting dataset visit: https://hub.worldpop.org/geodata/summary?id=1268. 85. Including “prevalence of stunting” and “household deprivation” in the same model did not lead to multicollinearity because the correlation between was −0.06. 86. Ferwerda et al., “Explaining Opposition to Refugee Resettlement”; A. T. Jenssen and H. Engesbak, “The Many Faces of Education: Why are People with Lower Education more Hostile Towards Immigrants than People with Higher Education?” Scandinavian Journal of Educational Research 38, no. 1 (1994): 33–50. 87. P. Collier, and A. Hoeffler, “Greed and Grievance in Civil War,” World Bank, Policy Research Working Paper No. 2355, 2000. 88. M. Schaub, and D. Auer, “Rebel Recruitment and Migration: Theory and Evidence from Southern Senegal,” Journal of Conflict Resolution (2022): 1–28; Q. T. Do and L. Iyer, “Geography, Poverty and Conflict in Nepal,” Journal of Peace Research 47, no. 6 (2010): 735–748. 89. B. Kayode, “Inside Nigeria’s Sambisa Forest, the Boko Haram Hideout where Kidnapped School Girls are Believed to be Held,” The Guardian, April 29, 2014, https://www.theguardian.com/world/2014/apr/29/nigeriasambisa-forest-boko-haram-hideout-kidnapped-school-girls-believed-to-be-held (accessed July 5, 2023); B. Grill, & T. Selander, “The Devil in Nigeria: Boko Haram’s Reign of Terror,” Der Spiegel, May 30, 2014, https://www. spiegel.de/international/world/boko-haram-continues-to-terrorize-northern-nigeria-a-972282.html (accessed July 5, 2023). 90. E. Akinwotu, “Almost 300 Schoolgirls Kidnapped in Nigeria are Free, says State Governor,” The Guardian, March 2, 2021, https://www.theguardian.com/world/2021/mar/02/almost-300-schoolgirls-kidnapped-in-nigeriaare-free-says-state-governor (accessed July 5, 2023). 91. To access the GlobCover dataset and the codebook/validation report visit: http://due.esrin.esa.int/page_glob cover.php; S. Bontemps, P. Defourny, E. Van Bogaert, O. Arino, V. Kalogirou, and J.R. Perez, GLOBCOVER 2009 Product Description and Validation Report (Paris and Louvain-la-Neuve: European Space Agency and Université Catholique de Louvain, 2011). 92. W. Chow and E. Han, “Rugged Terrain, Forest Coverage, and Insurgency in Myanmar,” Conflict Management and Peace Science, (2023): 1–20. 93. I also estimated some 2SLS regression models where I used either forest cover or its square as instrumental variables. The second-stage regression results were consistent with those reported in this paper. 94. Table A2 in the appendix shows the ethnic distribution of the respondents in the Afrobarometer dataset. 95. M. Parsons, “Militant Religion, not Just Climate Change, is Fuelling Violence in Nigeria,” The Critic, 2023, https://thecritic.co.uk/nigerias-climate-of-terror/ (accessed July 8, 2023); Christian Association of Nigeria, “Nigerian Church Protest Killing of Christians,” CAN Nigeria, 2018, https://canng.org/news-and -events/news/173-nigerian-church-protest-killing-of-christians (accessed July 5, 2023); Christian Association of Nigeria, “Political Opposition Sponsorship of Killings in Nigeria: Questions Buhari Government must Answer,” CAN Nigeria, 2018a, July 11, https://canng.org/news-andevents/news/181political-opposition-sponsorship-of-killings-in-nigeria-questions-buharigovernment-must-answer (accessed August 4, 2023). TERRORISM AND POLITICAL VIOLENCE 259 Appendix Table A2. Ethnic distribution of respondents Ethnic group Frequency (n) Percent Hausa 322 22.24 Igbo 251 17.33 Yoruba 328 22.65 Fulani 49 3.38 Ibibio 35 2.42 Kanuri 35 2.42 Ijaw 33 2.28 Tiv 26 1.80 Ikwere 25 1.73 Efik 24 1.66 Ebira 20 1.38 Idoma 19 1.31 Nupe 18 1.24 Igala 16 1.10 Isoko 10 0.69 Edo 10 0.69 Gwari 9 0.62 Kalabari 9 0.62 Jukun 7 0.48 Urhobo 4 0.28 Birom 3 0.21 Shuwa-Arab 1 0.07 Others 194 13.41 Total 1,448 100.00 Based on the Round 7 Afrobarometer survey data collected in 2017. Table A1. Descriptive statistics Variable Obs. Mean Std. Dev. Min Max Outgroup hostility ϕ 1437 4.389 2.351 2 10 Outgroup hostility (religion) 1439 2.261 1.331 1 5 Outgroup hostility (ethnicity) 1438 2.13 1.202 1 5 Violent conflict # 1592 67.886 106.807 0 475 Nighttime light # 1592 3.1 5.191 0 20.104 Prevalence of stunting # 1592 0.331 0.147 0.136 0.634 Household deprivation 1440 4.935 4.11 0 20 Log Population size # 1592 14.032 1.116 11.536 16.39 Educational level 1445 4.513 2.155 0 9 Religious affiliation 1428 0.569 0.495 0 1 Gender 1448 0.501 0.5 0 1 Age 1447 32.658 12.428 18 80 Forest cover 1592 0.599 0.223 0.056 0.985 (Forest cover) 2 1592 0.408 0.26 0.003 0.969 ϕ is the dependent variable which is derived by adding “Outgroup hostility (religion)” and “Outgroup hostility (ethnicity),” # denotes variables measured using buffers with a radius of thirty kilometers. Although the Afrobarometer dataset has 1,600 potential observations, the variables in the table contain fewer observations because not all respondents were asked the relevant questions. Also, I treated “don’t know” and “refused to answer” responses as missing observations which may have exacerbated the problem of listwise deletion. 260 D. TUKI Table A3. Correlation between the variables Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (1) Outgroup hostility 1.000 (2) Outgroup hostility (religion) 0.938 1.000 (3) Outgroup hostility (ethnicity) 0.925 0.736 1.000 (4) Violent conflict −0.118 −0.126 −0.092 1.000 (5) Nighttime light −0.168 −0.172 −0.140 0.837 1.000 (6) Prevalence of stunting 0.098 0.109 0.072 −0.374 −0.397 1.000 (7) Household deprivation 0.049 0.030 0.062 −0.048 −0.068 −0.066 1.000 (8) log population size −0.078 −0.086 −0.059 0.643 0.768 −0.500 −0.094 1.000 (9) Educational level −0.158 −0.150 −0.144 0.250 0.270 −0.378 −0.102 0.332 1.000 (10) Religious affiliation −0.038 −0.045 −0.024 0.141 0.207 −0.649 0.058 0.358 0.369 1.000 (11) Gender −0.074 −0.083 −0.052 −0.002 −0.011 0.011 0.028 −0.012 0.096 −0.042 1.000 (12) Age −0.007 −0.025 0.013 0.013 −0.017 0.009 0.040 0.007 −0.175 −0.028 0.124 1.000 (13) Forest cover −0.100 −0.094 −0.093 −0.094 −0.084 −0.318 0.060 0.054 0.079 0.298 0.003 0.056 1.000 (14) (Forest cover) 2 −0.113 −0.107 −0.103 −0.120 −0.096 −0.301 0.042 0.075 0.062 0.289 0.000 0.069 0.982 1.000 TERRORISM AND POLITICAL VIOLENCE 261