Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Retribution or Reconciliation? Post‐Conflict Attitudes toward Enemy Collaborators

Retribution or Reconciliation? Post‐Conflict Attitudes toward Enemy Collaborators Verification MaterialsThe data and materials required to verify the computational reproducibility of the results, procedures, and analyses in this article are available on the American Journal of Political Science Dataverse within the Harvard Dataverse Network, at: https://doi.org/10.7910/DVN/0ONBF6.After violent conflicts over territory end, there is often a widespread belief that those who lived under the rule of an armed group were “collaborators”1Collaboration is a widely used term in conflict research, but it is not often defined. We use collaborator to mean a person who supports an armed group, whether voluntarily or under coercion. and are therefore complicit in crimes perpetrated by the enemy. As an Iraqi interviewee attests: “People assume that everyone who stayed in Mosul is an Islamic State supporter or member, but many of us were victims.”2Author interview with “Khaled” (38, accountant) in Mosul, Iraq (April 2017). To ensure anonymity, interviewees are identified by pseudonyms. Generally, policy makers, scholars, and civilians living outside of enemy‐held territory take for granted “unlimited and unwavering support of the population for the political actor who claims to represent it,” treating individuals involved in conflict as constituting an undifferentiated and monolithic “entity that must be ‘won’ by political actors” (Kalyvas 2006, 6–7). In reality, an enemy collaborator can be both a victim and a perpetrator, or lie somewhere on a continuum between the two. Many residents of territory captured by an armed enemy group are victims of its violence and only comply with its policies in order to stay alive. Others do so willingly, or even enthusiastically.The territorial defeat of the Islamic State (IS, also known by its Arabic acronym, Daesh), a Sunni armed group that captured large swaths of territory in Iraq and Syria, provides a unique opportunity to collect data on public attitudes toward enemy collaborators at a time when peace processes and accountability mechanisms are still at an early stage in development. Our fieldwork in Iraq indicates that war‐torn communities perceive significant variation in the culpability of different types of accused enemy collaborators. At the time of our survey in Mosul, IS's de facto capital from 2014 to 2017, many former civilian employees of IS's administrative and service‐providing institutions were still living in the city without fear of prosecution or reprisals. One municipal worker (“Zyad”) explained, “No one blames us for keeping our jobs when IS arrived because we needed to feed our families, and we continued doing the same work we had done before, just with new bosses. Besides, quitting was not an option because it would have been an act of rebellion, which would have put me and my family at risk for severe punishment.”3Author interview with “Zyad” (35, municipal services) in Mosul (April 2017). Yet, Iraqis affiliated with IS in other ways were not being forgiven so easily. Widows of IS fighters reported that they would rather remain indefinitely in camps for internally displaced persons because they fear for their safety and that of their children if they return to their former hometowns. For example, “Laila,” whose brother's house was attacked with grenades as a result of the family's ties to IS, said, “I am afraid that if I return, my neighbors would kill me in my sleep.”4Author interview with “Laila” (40) in Hajj Ali IDP Camp (December 2017). What explains the stark difference between the cases of these two collaborators? One (“Zyad”) is perceived as innocent while the other (“Laila”) is facing death threats.To assess variation in attitudes toward the punishment and forgiveness of enemy collaborators, we employ an experiment embedded in an original survey of 1,458 Sunni Arab residents of Mosul (Moslawis). We evaluate the effects of randomly varied identity attributes of hypothetical IS collaborators and a range of collaborative acts on preferences for punishment and willingness to forgive. Although attitudes toward reconciliation are necessarily context‐specific, our research design can be applied in a wide variety of post‐conflict settings—including civil wars, foreign occupations, and post‐authoritarian transitions—to further our understanding of the prospects for legitimate justice and reconciliation at the microlevel. By widening our analytical lens to consider a broad spectrum of enemy collaboration, this study challenges a false dichotomy between victims and perpetrators found in much of the existing research on transitional justice (Tabak 2011). Our results demonstrate that variation in the type of enemy collaboration is an important determinant of preferences for post‐conflict punishment and forgiveness. This finding is very strong, remaining robust even in interactions with the identity attributes (e.g., age, gender, tribe) manipulated in our experiment as well as with respondent identity characteristics (e.g., age, gender, educational background). We argue that the predominant focus of existing transitional justice work on ingroup versus outgroup dynamics has distracted scholars from considering a potentially stronger determinant of preferences for punishment and forgiveness: variation in individual culpability.Observationally, we explore several mechanisms that may mediate the effects of collaborator culpability. Our data provide some support for the “revenge hypothesis”—that victimization at the hands of an enemy group leads to an increased desire for retribution—although the effects are substantively small by comparison. Instead, a key finding is that attitudes regarding punishment and forgiveness are strongly associated with perceived volition behind the act of collaboration, a mechanism that is understudied in the empirical literature and should receive greater attention in models of attitudes toward transitional justice. In emphasizing this mechanism, we contribute to a growing literature on the dynamics of civilian agency during conflict (e.g., Arjona 2016; Wood 2003) by empirically evaluating how it shapes prospects for reconciliation after conflict ends. This study provides uniquely fine‐grained data on the factors that shape perceptions of individual enemy collaborators, which have not been systematically tested or theorized by scholars thus far. Moreover, it offers a replicable research design and an expandable theoretical framework for furthering research on perceived collaborators in other post‐conflict settings.Why Study Public Perceptions of Former Enemy Collaborators?When conflicts end, national and international elites determine transitional justice policies, often paying scant attention to the opinions of local populations (Jones, Parmentier, and Weitekamp 2012). Similarly, studies of post‐conflict public opinion focus on attitudes toward different state‐imposed justice mechanisms, including lustration laws, truth commissions, or peace processes (Fabbe, Hazlett, and Smmazdemir 2019; Hall et al. 2018; Hirsch‐Hoefler et al. 2016; Tellez 2019). Although these studies make important contributions, the microfoundations of reconciliation with individual enemy collaborators remain poorly understood. As a result, there is increasing concern that transitional justice processes do not adequately address the concerns and needs of victims on the ground, whose support for and belief in the legitimacy of these processes is necessary for the reintegration of individual perpetrators into their local communities (e.g., Shaw, Waldorf, and Hazan 2010), without which sustainable peace cannot be achieved. To fill this gap, our study takes a bottom‐up, microlevel, and victim‐centered approach that gives voice to the people whose attitudes matter most for conflict resolution and sustainable peace.Rich descriptive work provides vivid accounts of the wide range of engagement options available to civilians, from defiance to full support of armed groups (e.g., Arjona 2016; Kalyvas 2006). Petersen (2001, chap. 1) develops a scaled spectrum of roles for defectors versus collaborators, placing participation in violence at the extremes, to theorize about when and why individuals decide to take on these different roles. The various types of collaboration that civilians engage in have been shown to shape the internal organization of enemy groups, in addition to the establishment or collapse of their rule (Arjona 2016; Arjona, Kasfir, and Mampilly 2015; Weinstein 2006). We posit that variation in the type of collaboration also matters for the design of post‐conflict transitional justice processes. Individuals develop attitudes toward collaborators based on their actions, which inform perceptions of their culpability and preferences for punishment.Previous quantitative research has largely overlooked this important variation in the different roles and varying culpability of enemy collaborators, gauging attitudes toward armed groups as a whole (e.g., Blair, Imai, and Lyall 2014; Dyrstad and Binningsbø 2019) or the most egregious type of enemy collaboration—participation in violence (e.g., David 2014; Lyall, Blair, and Imai 2013; Samii 2013). Empirically, we know much less about public opinion toward civilian collaborators despite their importance. Parties to conflicts over territorial control rely heavily on civilians to obtain food, water, shelter, labor, and information (Arjona 2016; Kalyvas 2006; Weinstein 2006; Wood 2003).A handful of studies randomize some features of enemy combatants and ask about justice mechanisms to examine the effects of ingroup versus outgroup identities. David (2014) finds that social identity matters for popular perceptions of justice among Serbs and Croats, with outgroup members receiving harsher scrutiny (David 2014, 489). In South Africa, ex‐combatant leaders are more likely to be blamed for their actions than their subordinates, and those who were pro‐apartheid in particular, although these outcomes are mediated by respondent race (Gibson and Gouws 1999). Another study asks respondents about forgiveness of “people of other nationalities for the violence they have committed in the last fifteen years,” finding that personal experience with violence and its effects, not ethnic hostility toward outgroup members, negatively correlates with willingness to forgive (Bakke, O'Loughlin, and Ward 2009, 1017). Notably, this type of experimental design does not allow us to distinguish whether it is the reminder of past violence, the various events that have taken place within the last 15 years, the many different outgroups invoked by the prompt, or combinations of some or all of these factors that drive attitudes.A common thread linking these studies—and informing ours—is the core idea that characteristics of enemy collaborators shape attitudes toward justice and reconciliation. Yet, in line with other work on attitudes toward enemy combatants (e.g., Lyall, Blair, and Imai 2013; Samii 2013), these studies do not examine collaboration by civilians, nor do they consider variation in individual identity characteristics of collaborators such as gender or age. In sum, the existing literature fails to explore more nuanced understandings of guilt and commensurate justice for individuals engaged in lesser crimes. Our research design experimentally manipulates the identities and actions of hypothetical enemy collaborators in order to make causal inferences about microlevel determinants of attitudes toward justice and forgiveness.Our work also has important and timely policy implications. Although public opinion should never be the sole basis for the design of transitional justice policies because of the risk of “mob justice” (Daly 2001, 383), it should be taken into consideration by policy makers to facilitate durable peace after conflict. Social psychologists have found that sensitivity to popular concerns about the fairness of legal institutions is an important determinant of trust in the police and courts, warning that failure to consider public opinion may increase the likelihood of noncompliance with laws and state authorities (e.g., Tyler 2003). Work in criminology finds that justice processes are more likely to lead to reconciliation between adversarial groups and a reduction of criminal recidivism when they take into account the voices of victims (Latimer, Dowden, and Muise 2005). In communities where victims of violence are deeply dissatisfied with the state's official response, they may take matters into their own hands, resulting in extrajudicial revenge (Human Rights Watch 2017). Thus, there is a real need for safeguards to ensure that transitional justice processes protect the fundamental rights of perpetrators—including the right to due process in trials (United Nations 2006, 36)—and that these processes are acceptable to the society at‐large.Conducting Multimethod Research in Mosul, IraqWe conducted an original survey of 1,458 Mosul residents from March to April 2018 with an Iraqi research firm.5The company was the Independent Institute for Administration and Civil Society Studies. A team of enumerators recruited from Mosul conducted face‐to‐face surveys with tablets. In addition to the quantitative data, our research draws on qualitative evidence from fieldwork in Mosul and other areas of northern Iraq in 2017 (supporting information [SI], pp. 17–20). This fieldwork enabled us to ensure the appropriateness of our survey questions for the context and to validate the realism of the experiment. It included visits to public institutions that were previously administered by IS, as well as observations of trials of alleged IS members. We carried out interviews with 61 individuals from areas previously controlled by IS, as well as 17 lawyers, judges, and experts involved in prosecuting and defending suspected IS collaborators.Case SelectionCivilian collaboration with enemy rule and governance is a widespread phenomenon. Focusing solely on civil wars, Stewart (2018) estimates that one‐third of all rebel groups active from 1945 to 2003 engaged in governance of civilians. Scholarly accounts are replete with descriptions of how collaboration is the lifeblood of any armed campaign to control territory (e.g., Kalyvas 2006; Petersen 2001).6For a long list of armed groups that provided public services, see Arjona (2016, chaps. 1–3). Throughout history and around the world, armed groups such as the Revolutionary Armed Forces (FARC) in Colombia (Arjona 2016), the Taliban in Afghanistan (Terpstra 2020), the Liberation Tigers of Tamil Eelam (LTTE) in Sri Lanka (Mampilly 2011), the Bakonozos of Northern Uganda (Kasfir 2005), and the Nazis in occupied territories during the Second World War (Wistrich 2013) have relied heavily on a range of collaborators to capture and hold territory. The spectrum of collaboration in these cases ranged from horrific acts—such as murder, kidnapping, torture, rape, and even genocide—to much less severe transgressions that did not involve violence; civilian collaborators in these cases performed a wide range of nonmilitary functions and services, including tax collection, sanitation, healthcare, education, and supplying or otherwise supporting combatants.IS rule in Mosul is just one example of the broader phenomenon of enemy rule and governance. Across a wide variety of settings and cases, rebel governance, defined as “the set of actions rebels engage in to regulate the social, political, and economic life of non‐combatants during war” (Arjona, Kasfir, and Mampilly 2015, 3), has become more common with the increase in civil wars since the end of the Cold War (Kalyvas and Balcells 2010). Likewise, the broader phenomenon of enemy governance is characterized by significant civilian participation—whether voluntary or coerced—in the enemy's governing institutions. IS set up a government, established its own rules of conduct, and maintained control of substantial territory for 3 years. The group operated a variety of institutions that provided public goods and basic services, which necessitated a civilian bureaucracy staffed by employees who generally did not engage in violence. Moslawis witnessed a wide range of collaboration, including (1) compliance with taxation (Revkin 2020b); (2) social integration with the group through marriage to its members and the enrollment of children in IS‐controlled schools; (3) employment in IS's civilian workforce as teachers, doctors, or cooks; and (4) recruitment as fighters (Revkin 2020a). Since Moslawis confronted these different types of collaboration on a daily basis, they can recognize and imagine the scenarios described in our experiment.Residents of Mosul are now grappling with the question of what to do with former collaborators and under what conditions they might be forgiven for their transgressions. Surveying people in the place where these transgressions occurred provides important insights into the microfoundations of attitudes toward reconciliation in a context where there is an urgent need to reduce the risk of conflict recurrence. Thus, Mosul is a particularly relevant setting in which to collect data on attitudes toward former enemy collaborators.Our sample includes only Sunni Arabs living in Mosul in June 2014—when IS arrived—and therefore having had some exposure to IS. Sunnis made up more than 97% of the city's population at the time of the survey. The sample is not representative of Iraq as a whole. Including representative samples of other identity groups in Iraq (e.g., Shias) not only would have been prohibitively costly, but it also would have shifted the focus of this study to the very salient ethno‐sectarian divides in Iraq. We expect that Shias would be much less forgiving of former IS collaborators than the Sunni respondents in our sample given IS's particularly harsh treatment of Shias. As our goal was instead to focus on the effects of varying degrees of enemy collaboration on the prospects for post‐conflict reintegration and reconciliation, we intentionally held this very salient social identity constant in our study. It is also important to note here that the conflict with IS was not fought purely along ingroup versus outgroup lines: IS killed more Sunni Muslims than any other religious or ethnic group (Verini 2016). In sum, our sample includes both IS collaborators and victims, identities that are not mutually exclusive. It is representative of populations living in territories that have experienced enemy rule and governance and are now grappling with the question of how to assess the culpability of the enemy collaborators among them. Our findings would not generalize to cases where a defeated enemy did not seek to govern, such as armed groups whose sole purpose is economic predation of a territory's resources or where chaos ensued. Such cases are rare, however, as even “roving bandits” have incentives to establish a social contract with civilian populations to profit over the long term (Arjona 2016; Olson 1993).We were also motivated by the immediate policy implications that such a study in Mosul could have for post‐transitional justice in Iraq. At the height of its expansion in 2014, IS governed millions of people. When IS was militarily defeated in 2017, it left behind a population that is now widely perceived as collectively complicit in the group's crimes. The government is currently facing the monumental challenge of reintegrating this population back into their local communities, but authorities have taken a heavy‐handed approach that fails to differentiate between voluntary and involuntary collaboration, and more serious crimes and lesser offenses. Iraq's Anti‐Terrorism Law criminalizes membership in any terrorist group without requiring proof of a specific criminal act; anyone with a plausible connection to the group can easily be sentenced to life in prison, the minimum punishment allowed by the law.7Law No. 13 (2005). More than 8,000 accused IS collaborators have been convicted in trials that are often decided in under 30 minutes, with a conviction rate of around 98%,8Author observations of trials in Tel Kaif (December 2017). and more than 3,000 have been sentenced to death (Abdul‐Zahra and George 2018).This one‐punishment‐fits‐all approach—which is widely perceived as collective punishment of Sunnis—is generating new grievances that could fuel the emergence of an “IS 2.0” (Revkin 2018). A correlation between repression and radicalization has been documented in many contexts (Davenport and Inman 2012). And it has been argued that the rise of IS—which emerged from the remnants of al Qaeda in Iraq—was fueled by resentment over the collective punishment of Sunnis through de‐Baathification and incarceration (Sly 2015).In addition to its direct and immediate implications for the case of Iraq, this work speaks to a broader set of ongoing as well as future cases of post‐conflict transitional justice processes. In all cases of conflict involving territorial sovereignty, victors face the challenge of walking a fine line between under‐ and overpunishment of former enemy collaborators. When punishment of a perpetrator falls short of what the victims believe is commensurate to the crime committed, the resulting perception of an “injustice gap” increases the likelihood of victims being dissatisfied with the outcomes (Worthington 2006). Yet, excessive punishment may be perceived as victors’ justice and delegitimize transitional justice efforts (de Greiff 2014, 18). Popular opinion data can help policy makers find the middle ground.Theorizing Post‐Conflict Justice PreferencesWe develop an original theoretical framework for analyzing how individuals in conflict‐affected societies form preferences for punishment and forgiveness of other community members who collaborated to varying extents with an enemy. As noted, previous research has established that social identity is an important determinant of reconciliation in intergroup conflicts (e.g., David and Yuk‐Ping 2005; Gibson and Gouws 1999; Samii 2013). Our theoretical contribution is to highlight the importance of an additional factor that has been understudied in the conflict literature: culpability. We posit that perceived culpability varies depending on an individual collaborator's physical proximity to and social intimacy with the enemy. In this section, we develop and test several hypotheses about the expected effects of variation in individual collaborator attributes and acts of enemy collaboration during conflict on preferences for accountability and prospects for reconciliation.9We preregistered these hypotheses and secondary expectations prior to data collection. See Kao, Kristen and Mara Revkin. “Pre‐Analysis Plan: ‘Reintegration of Enemy Collaborators: Survey Experiments in Mosul,’” Evidence in Governance and Politics (March 29, 2018), https://egap.org/registration/4395.Variation in Types of Collaboration and Perceived CulpabilityWe argue that information about the type of collaboration, which is closely linked with the perceived culpability of the collaborator, is an important determinant of preferences for punishment. Studies in social psychology posit that more severe transgressions in personal relationships are more difficult to forgive (e.g., Boon and Sulsky 1997). Work in political science indicates that enemy fighters from more abusive units are less likely to be reintegrated back into society (Humphreys and Weinstein 2007) and that commanders are held more responsible than subordinates (Gibson and Gouws 1999). Considerable evidence across a variety of contexts suggests that individuals see violent behaviors as the most serious offenses (Stylianou 2003).To explore variation in perceptions of culpability behind collaboration, we chose collaborator roles that vary in their proximity to violence. In addition to the role of combatants, we examine a spectrum of nonmilitary roles that collaborators—both men and women—may perform. The specific roles included in the experiment are based on our fieldwork in Iraq as well as common patterns of enemy collaboration in other conflicts. In particular, we hypothesized that fighters would be most harshly punished and least likely to be forgiven due to the violent nature of their collaboration. Additionally, we expected that respondents would prefer harsher punishments for collaborators who were physically closest to enemy combatants (cooks for fighters) and those who were the most intimate with them (wives of fighters) in contrast with collaborators in civilian roles for institutions that provided services to other civilians (janitors working in IS's department of municipal services) or those who financially supported the insurgency (taxpayers).10We validated our assumption that different acts of collaboration are associated with varying levels of severity. We found that 98% of the sample agrees that being a fighter is the most condemnable transgression, followed by civilians directly involved with fighters. Those not directly involved with fighters are ranked as the least condemnable (SI, p. 5).Identity TraitsAlthough not the main focus of this study, we believed that the individual identity characteristics of collaborators would have either direct or moderating effects on attitudes toward punishment and forgiveness. Some identities are seen as less agentic and therefore less culpable than others. Shared identities between the respondent and a collaborator may trigger empathy and forgiveness.Based on previous research in criminology, we expected respondents to prefer more lenient punishments for and be more forgiving of younger collaborators. Juveniles are generally assumed to be less agentic than adults because they are easily influenced by those around them and they may not be able to distinguish between right and wrong (Scott et al. 2006). Similar to most penal codes around the world, Iraqi law requires reduced punishments for children, taking into consideration their age and the stage of their mental development at the time of the offense.11Iraq's Penal Code (Act No. 111 of 1969), Articles 67–78.We also expected respondents to prefer more lenient punishments for women than for men. Previous studies find that women in other contexts are perceived as being less responsible when they cause harm, and they receive lighter punishments compared to men, ceteris paribus (Honey 2017). Interviews in Iraq support this expectation. “Fadila” explained that when her husband decided to join IS and she expressed misgivings, he replied, “You can leave and I will keep the kids.”12Author interview with “Fadila” (35, wife of an IS fighter) in Ninewa (December 2017). Anecdotal evidence suggests that female collaborators should be perceived as less culpable than men because, as “Fadila” put it, “We did not have a choice.”Research demonstrates that members of the same group tend to favor one another and punish outsiders (e.g., David and Yuk‐Ping 2005; Samii 2013). In conflict settings in particular, ingroup biases become more pronounced (Hewstone, Rubin, and Willis 2002). In social psychology, McCullough, Fincham, and Tsang (2003) also suggest that shared social identity encourages empathy with and increased benevolence toward a transgressor. In the context of the tribal society of Iraq, these previous findings concerning ingroup biases led us to expect that people should be more empathetic, lenient, and forgiving of transgressions by members of their own tribe (ingroup) in comparison with members of other tribes (outgroups).VictimizationPreferences for punishment of enemy collaborators may be affected by whether an individual was victimized by the enemy group. Many studies find that exposure to violence decreases willingness to forgive, reconcile, and cooperate with transgressors (e.g., Bakke, O'Loughlin, and Ward 2009; Hall et al. 2018; Hirsch‐Hoefler et al. 2016). Based on this previous research, we expected that individuals who have been victimized by an enemy group—as measured by the death or injury of family members or property destruction—would prefer harsher punishments for enemy collaborators than those who did not experience such a personal loss.Volition of CollaborationWe expected the perceived volition behind the acts of collaboration to be associated with preferences for punishment of collaboration. Assessing volition behind acts of transgression, while difficult, is a key component of most psychological models of blame and responsibility (e.g., Alicke 2000, 57). Experimental research finds that belief in intention is a major driver of preferences for punishment of criminals (Aharoni and Fridlund 2011), and that more intentional transgressions are more difficult to forgive (e.g., Boon and Sulsky 1997). We test whether these findings generalize to a post‐conflict setting.Perceived volition in engagement might be expected to have stronger effects for lesser collaborative acts; as the act increases in condemnability, intent may not be as impactful on attitudes toward retribution or reconciliation. Volition is a particularly important factor in contexts where the collaborator may be acting under considerable economic or physical duress, as is the case in wartime environments where enemy groups exercise coercive control over territory and people. For instance, IS required all residents of its territory—except for the extremely poor—to pay taxes. Yet, some IS supporters may have voluntarily paid taxes, seeing their payment as an appropriate payment for services that the group was providing. Ten percent of our sample openly agreed that the fees collected by IS were fair in exchange for the services that IS was providing.13Another 74% disagreed and 16% refused to answer. Altogether, 54% of the sample admitted paying taxes to IS in some form. Analyses of engagement in this act of collaboration on our outcomes are shown in the SI (p. 13).There was also considerable variation in the voluntariness of marriage to IS fighters. Some women were already married when their husbands decided to join the group. Others married fighters for the purpose of gaining social status. As one interviewee explains: “In poor neighborhoods, some women believed that they could become princesses by marrying IS emirs.”14Author interview with “Salem” in Mosul (April 2017). Emir means “prince,” which IS used to refer to high‐ranking officials. Still others were coerced into marrying fighters through social pressure, economic duress, or physical threats.Punishment, Forgiveness, and ReconciliationThis work measures two interrelated outcomes: forgiveness and punishment. We adopted a widely cited definition of forgiveness: the lessening of negative feelings, thoughts, and behaviors toward transgressors (McCullough, Fincham, and Tsang 2003). Forgiveness is “one of the most important processes in the restoration of interpersonal relationships after conflict” (Hill 2001). Based on previous research, we also assumed that benevolence toward transgressors is generally conditional upon the belief that justice has been served (Enright 1991, 128). Yet, although punishment may facilitate forgiveness in some cases, it is not established that punishment is necessary—which we allow for in our design.Our experimental prompt is designed to trigger attitudes toward reintegration and reconciliation by priming respondents with the information that the collaborators want to move back into their neighborhood (and could therefore interact with them in the future). We ask respondents whether—after selecting the punishment that they feel is appropriate for a given hypothetical collaborator—they would be willing to forgive these persons. Although we cannot identify the causal effect of punishment on forgiveness, this second dependent variable nonetheless enables us to examine the potential for reconciliation through correlational analysis.Experimental DesignTo evaluate respondents’ beliefs about the type of justice deserved by former collaborators, we implement a ratings‐based conjoint experiment. The design significantly reduces the number of participants needed while maintaining sufficient power to test multiple hypotheses by randomizing each potential driver of outcomes independently of others. We included a follow‐up question on willingness to forgive after punishment, which we use to assess prospects for reconciliation.The experiment randomizes enemy collaborators’ identities and the nature of their collaboration with IS. As identity characteristics, we included gender,15Though the majority of IS fighters were male, IS had some female combatants. age, and whether or not the collaborator is a member of the respondent's tribe to serve as identity cues.16We did not vary ethno‐religious identities because IS was predominantly a Sunni Arab organization. Almost all Moslawis identify with a tribe. We also randomize acts of collaboration. The five collaboration roles specified in the experiment are (1) fighting for IS, (2) working as a cook for IS fighters, (3) being married to an IS fighter, (4) working as a janitor for the IS municipality, and (5) paying taxes to IS.Every respondent evaluated three separate profiles that were generated by randomizing the attributes listed in Table 1. The total sample of evaluated profiles was 4,275. Before the enumerator reads the descriptions of the hypothetical collaborators, the respondent is told, “I am going to read you some hypothetical scenarios about people from Mosul who are being prosecuted for their past cooperation with Daesh (IS). These people now want to move back into your neighborhood. I would like you to choose the type of punishment that you view as appropriate for this person. The person is a [insert profile].”1TableRandomized Enemy AttributesDimensionAttributesGender– Man– WomanAge– 15– 35Tribal member– Respondent's tribal group– Member of another tribeType of collaboration– IS fighter– Cook for IS fighters– Married to an IS fighter (limited to female candidates)– Janitor at the municipality employed by IS's government– Resident of Mosul who paid taxes to ISAfter the respondent is read a collaborator profile, she is told: “A thorough investigation concluded that this is the only act of collaboration that the person committed. I have ordered the following punishments from least harsh to most harsh. I would like you to choose the type of punishment you deem appropriate for this former Daesh collaborator, who now wants to move back into your neighborhood.” The responses to this question make up our dependent variables and include the following options: no punishment necessary (least harsh), mandatory community service (e.g., picking up trash, rebuilding homes) for six months, imprisonment for 3 years, imprisonment for 15 years, and capital punishment (most harsh).This first sentence in the prompt serves two purposes. First, it helps address the concern that respondents might impute other types of collaboration to the profile by specifying that the stated act is the only one committed. Second, it encourages respondents to view the five punishment types as an ordered scale from least to most harsh. To validate our ranked scale, we ran a pilot study in which we asked 100 Moslawis to rank the punishments from least to most harsh. We also ran a post‐experiment validation check of this question (SI, p. 11).17Over 90% of those asked agreed with our ranking in the pilot and full sample; the 10% that do not simply add noise to our findings. Only 4% of the sample preferred a different type of punishment, suggesting the scale of punishments we offered was well‐aligned with actual preferences.Research EthicsResearch in conflict areas raises unique ethical challenges (Wood 2006), as well as security concerns for researchers and their subjects. Asking respondents in post‐conflict settings to share their experiences and attitudes about a recent conflict in which they and their families may have experienced several personal harms has the potential to trigger retraumatization (Revkin and Wood 2020). Another concern is that such questions could increase polarization, particularly in multiethnic communities. We took these concerns very seriously and conducted semi‐structured interviews with Iraqis, as well as our enumerators from Mosul, to assess whether any of the questions would be offensive or disturbing to respondents.The overwhelming feedback was that the questions would not be upsetting to respondents because the aftermath of the very recent conflict was still, tragically, a constant presence in the lives of residents of Mosul in 2018. Conversations about IS collaborators, their culpability, and justice mechanisms were commonplace in 2018 and therefore less sensitive than might be expected. For example, one Moslawi expressed frustration about the lack of accountability: “Family members of IS fighters, who were beneficiaries of IS and its crimes, are living among us, and no one is holding them accountable.”18Author interview with “Walid” (33, store clerk) in Mosul (April 2017). Many of our interlocutors were eager to discuss their experiences with IS and preferences for justice, apparently viewing this study as an opportunity to tell their stories to a broader audience than they could otherwise reach.We found similar enthusiasm for participation in the household survey, with only 15% of potential respondents declining. Within the survey, the response rate for arguably the most sensitive question—“During the first six months of Daesh rule, did you believe that Daesh was doing a better job of governing Mosul than the Iraqi government did previously?”—was even higher: Only six respondents (4% of the sample) said they did not know or refused to answer.Given the extent to which IS collaboration was already being publicly debated by Iraqis at the time of the survey, as well as an informed consent procedure that allowed all respondents to opt out of the survey at any time, we do not believe that the survey exposed respondents to significant risk.19We also designed a list experiment to assess sensitivity about preference for IS governance. The results suggest that our results were not significantly affected by social desirability bias (SI, pp. 9–10). The SI (pp. 15–20) discusses other potential risks to survey respondents and the steps taken to minimize those risks.20The Yale University Institutional Review Board approved the survey instrument (Protocol #2000022022), observations of trials (Protocol #2000021840), and interviews with Iraqis (Protocol #1506016040).Analyses and ResultsExamining the distribution of the dependent variable across all types of collaboration, the two most frequently selected options were no punishment (28%) and capital punishment (33%), indicating that there is considerable variation in the preferences of Moslawis concerning justice. Table 2 shows that IS fighters and those who were most closely associated with fighters (cooks for and wives of fighters) receive consistently harsher punishments than those less closely associated with fighters (janitors who worked for the IS municipality and taxpayers). More than three‐quarters of the sample sought capital punishment for IS fighters, whereas a similar proportion did not think any punishment was necessary for taxpayers. About one‐third of the sample considered death to be an appropriate punishment for cooks and wives of fighters, although for married women half of the sample sought less than 15 years in prison. The largest gaps in perceptions of appropriate punishment are between those who are intimately involved with violence in contrast to those who are more distant from it. Importantly, more than two‐thirds of the sample sought a restorative punishment (community service) or no punishment for janitors at the IS municipality and for taxpayers.2TablePunishments Preferred for Types of Collaboration (Percentage of Cases)ActNo PunishmentCommunity Service3 Years in Prison15 Years in PrisonCapital PunishmentIS fighter1251378Cook for fighters314222536Married a fighter1717161831Janitor at municipality412711615Paid taxes747549Total2814121433We employ ordinary least squares regression (OLS) to assess the average marginal component effect (AMCE) of each of the profile attributes, pooling across all respondents and tasks.21The results are robust to ordinal logistical analysis. Following Hainmueller, Hopkins, and Yamamoto (2013), we expect OLS regression to be a consistent estimator of the AMCE. This allows us to estimate the effects of profile attributes on degree of punishment through the following equation:Punishmentik=θ0+θ1Genderik+θ2Ageik+θ3Tribeik+θ4Collaborationik+εik,$$\begin{eqnarray*} Punishmen{t_{ik}} &=& {\theta _0} + {\theta _1}Gende{r_{ik}} + {\theta _2}Ag{e_{ik}}\\ && + \ {\theta _3}Trib{e_{ik}} + {\theta _4}\textit{Collaboration}_{ik} + {\varepsilon _{ik}}, \end{eqnarray*}$$where i denotes the respondent and k denotes which round of three rounds each respondent completes. Punishmentik is the outcome on the scale of least to most severe punishment. The analysis is run with robust standard errors clustered at the level of the respondent to account for within‐respondent correlation across the rounds. The error term εik refers to any random variation and, importantly, the effects of any additional determinants of preferences for punishment not accounted for in our model. The point estimates from the OLS regression are displayed in Figure 1. The dependent variable is the 5‐point scale of punishment, in which 1 is no punishment, 2 is six months of community service, 3 is three years in prison, 4 is 15 years in prison, and 5 is the death penalty.1FigureEffects of Collaborator Identity and Type of Act on PunishmentNote: The figure depicts point estimates (circles) with 95% confidence intervals (horizontal lines) and robust standard errors clustered at the individual. The circles on the vertical line at 0 denote the reference category for each attribute.Former IS fighters receive punishments that are higher than all other acts of collaboration to a statistically significant level (see Figure 1). On average, former IS taxpayers receive punishments that are 2.97 points lower than IS fighters (shown on the vertical dashed line at 0), accounting for approximately 59% of the entire 5‐point scale, with a standard error (SE) of 0.06. In other words, punishments for taxpayers were nearly three levels less harsh than for fighters, which on our 5‐point scale is the difference between 6 months of community service and capital punishment.22Additional analyses reveal that taxpayers receive significantly less harsh punishments than former collaborators who worked as cooks or janitors for IS or were married to IS fighters.We also find that civilian collaborators who were directly involved with fighters (e.g., women married to fighters and cooks for fighters) receive harsher punishments than those who did not work directly with fighters (e.g., janitors working for the IS municipality). On average, cooks receive punishments that are 0.87 points (SE = 0.05) lower than fighters—a difference of 17% of the 5‐point scale. Women married to fighters and janitors receive, respectively, punishments that are on average 1.36 (SE = 0.07, 27% of the scale) and 2.37 (SE = 0.07, 47% of the scale) points less harsh than former fighters.Contrary to expectations, respondents prefer harsher punishments for members of their own tribe (0.09 points equivalent to 2% of the scale, SE = 0.04). This finding suggests that respondents may hold members of their own tribe to a higher moral standard than members of other tribes, consistent with a theory of “ingroup policing” (Fearon and Laitin 1996). However, this effect is substantively quite small. Additionally, more lenient punishments are selected for younger collaborators (15 years old) than for older ones (35 years old) by 0.07 points (SE = 0.04), although this outcome is only significant at the p < .10 level and is substantively small. We lack support for the expectation that female collaborators would receive more lenient punishments than male collaborators, finding instead that women and men are seen as equally culpable for their actions.Identity characteristics of collaborators included in our experiment do not seem to have a substantial effect on perceptions of culpability, not even in interaction with the different acts.23There are a few exceptions, but the effects are substantively small. See the SI (pp. 6–7). These results underscore the notion that variation in the type of collaboration is an important, overlooked determinant of preferences for justice.ForgivenessWe ask a post‐treatment question for each profile to better understand the implications of variation in collaborator acts and identity characteristics on forgiveness: “Given the punishment you have selected, would you forgive this person?” This question sheds light on whether respondents are able to forgive and are potentially open to the related objectives of reconciliation and reintegration. Overall, a high proportion of respondents (59%) who did not choose the death penalty24Those who chose the death penalty, who make up 34% of the sample, were not asked this question. were willing to forgive collaboration with IS after punishment. Within this sample, 29% who were presented with profiles of IS fighters were willing to forgive them, about one‐third of those who were presented with profiles of cooks were willing to forgive them, while 42%, 72%, and 85% were willing to forgive women married to fighters, janitors for the IS municipality, and taxpayers, respectively.We employ OLS regression to analyze this outcome, coding those who chose the death penalty as not being willing to forgive collaborators.25The results are robust to logit regression. The findings are also largely robust to dropping out respondents who chose the death penalty (SI, p. 13). Compared to fighters, cooks are 15 percentage points more likely to be forgiven on a 0–1 scale (SE = 0.02). Likewise, respondents are significantly more likely to forgive wives of fighters (by 24 percentage points, SE = 0.02) and janitors working for the IS municipality (by 55 percentage points, SE = 0.02) than they are to forgive fighters. Taxpayers are 71 percentage points (SE = 0.02) more likely to be forgiven than fighters.Notably, women are less likely to be forgiven by about 3 percentage points (SE = 0.014). Although this finding is statistically significant, the effect is small in magnitude. Yet, it is in line with qualitative research documenting the intense stigma surrounding female enemy collaborators (e.g., McKay and Mazurana 2004). Age and shared tribal membership are insignificant. Figure 2 shows that the actions of former collaborators matter more than their identities for reconciliation.2FigureEffects of Identity and Act on Forgiveness of Former Enemy CollaboratorsNote: The figure depicts point estimates (circles) with 95% confidence intervals (horizontal lines) and robust standard errors clustered at the individual. The circles on the vertical line at 0 denote the reference category for each attribute.These results mirror those for the punishment outcome presented above, which suggests that severity of the act may drive both punitive preferences and willingness to forgive. If we add punishment to the forgiveness analysis, harsher punishments are negatively correlated with forgiveness, even when holding the act constant in subsamples. It is also notable that 90% of those who chose no punishment for a collaborator were willing to forgive them. These outcomes indicate that those who are already more punitive are less forgiving, and by contrast those who are less retributive are also more open to reconciliation (SI Table B4, p. 12). Further research should examine how punishment relates to forgiveness.Correlates of Punishment and ForgivenessWe employed observational analyses to explore the effects of respondent characteristics on forgiveness along with two additional preregistered hypotheses. The first hypothesis predicts that personal victimization should lead to a hardening of hearts and a desire for retribution against collaborators, which we refer to as the “revenge hypothesis.” The second hypothesis examines perceived agency behind the act of collaboration, an empirically unexplored potential mechanism explaining why different types of collaboration are met with different levels of punishment or forgiveness, which we refer to as the “volition hypothesis.” Our analyses did not find substantively significant differences among respondents of different ages, genders, and tribal identities.The Revenge HypothesisFollowing other studies on attitudes toward transitional justice mechanisms (Hall et al. 2018; Hirsch‐Hoefler et al. 2016), we examine whether personal victimization by an enemy group is related to our results. We compare respondents who had their residence seriously damaged or confiscated or had a member of their household injured or killed and hold IS responsible for at least one of these grievances (60% of the sample) with those who did not experience such harms. Moslawis with these grievances against IS are about 0.26 points (SE = 0.05, about 5% of the full 5‐point scale) harsher in their assessment of appropriate punishment; they are also 6 percentage points (SE = 0.017) less likely to forgive. Thus, we find some support for the revenge hypothesis, but the effect is substantively small.In interaction with type of act committed, our measure of personal victimization loses statistical significance; only the type of act remains a significant driver of punishment. Average marginal effects reveal that victimization is significantly associated with harsher punishments for janitors (by 0.33 points, SE = 0.13, or 9% of the scale) and less forgiveness for taxpayers by 9 percentage points (SE = 0.38; SI, p. 15).26Measuring victimization as those who experienced the death or injury of a family member reveals a 0.19‐point (SE = 0.06, 5% of the full scale) increase in harshness of punishment; this specification is not significantly correlated with forgiveness.The Volition HypothesisTable 3 explores the relationship between perceptions of different types of collaboration as voluntary and preferences for punishment and forgiveness. Respondents overwhelmingly perceived paying taxes to IS as an involuntary act (92%) in contrast with fighting for IS (3%), suggesting that more severe transgressions are associated with more culpability.3TablePerceptions of Collaboration as Voluntary (Percentage of Respondents)Type of CollaborationVoluntaryInvoluntaryAn IS fighter973Cook for IS fighters8812Married to an IS fighter8317Janitor who worked for the IS municipality7429Resident of Mosul who paid taxes to IS991Multivariate OLS regression analysis finds that volition and culpability have separate effects on punishment and forgiveness. On average, if an act is perceived as voluntary, the respondent is 0.53 points (SE = 0.08, 11% of the full scale) harsher in punishment and 20 percentage points (SE = 0.03) less likely to forgive the collaborator. Figure 3 shows the marginal effects of interactions between these two factors. The most striking result is that taxpayers receive a punishment that is 1.42 points harsher (28% of the 5‐point scale) when they are perceived as having voluntarily paid taxes compared to having done so involuntarily; voluntary tax payment is treated as harshly as involuntary participation in acts of collaboration that directly support fighters (cooks and wives). Voluntary collaborators are also 38 percentage points less likely to be forgiven than those who were perceived to be coerced. This outcome demonstrates the importance of considering perceived volition of collaboration in addition to the type of collaboration when determining preferences for retribution and reconciliation. Notably, the effect of perceived volition may vary depending on the type of collaboration.3FigurePerceptions of Voluntariness Interacted with Enemy ActNote: The figure depicts point estimates (circles and squares) with 95% confidence intervals (horizontal lines) and robust standard errors clustered at the individual. The act of collaboration is displayed on the y‐axis.DiscussionOur study develops an empirical framework for theorizing and testing the microfoundations of attitudes toward punishment and forgiveness of enemy collaborators. The evidence suggests the importance of variation in individual collaborator culpability and agency for post‐conflict transitional justice and reconciliation.The experimental design we employ is of immediate relevance to policy makers working to reestablish the rule of law and lasting peace in post‐conflict settings. Our results reveal that public opinion is on average more forgiving of IS collaborators than the harsh, one‐punishment‐fits‐all approach taken by the Iraqi government.27We find no effect of reported trust in Iraqi courts on our results. The policy implication here is that lighter punishments—including restorative, noncarceral sanctions such as community service or amnesty—should be considered for more cases than the law currently allows. On the other end of the spectrum, much of the sample still refuses to forgive the most condemnable acts of collaboration (e.g., fighters), indicating that reconciliation—even after the implementation of what respondents believed to be appropriate punishment—could be very difficult to achieve in these harder cases. Much more work is needed to understand the conditions under which, if any, violent collaborators can be accepted as rehabilitated.An important limitation of our study is that we were only able to vary a small number of identity characteristics given limited time and resources. It is possible that other social identity characteristics (e.g., religious identity) would have stronger effects on the likelihood of forgiveness and reconciliation. There is a rich debate over the role of ingroup versus outgroup dynamics in post‐conflict reconciliation that is beyond the scope of our study given the lack of diversity within our sample. That we did not find strong effects of collaborator identity characteristics on punishment demonstrates why public opinion should not be the sole basis for the design of transitional justice mechanisms; it is inhumane to hold juvenile offenders to the same standard as adults.Engaging with an active scholarly debate as to whether personal victimization at the hands of an enemy group leads to either an increased desire for revenge and retribution (Hall et al. 2018; Hirsch‐Hoefler et al. 2016) or pro‐sociality and reconciliation through a posttraumatic growth mechanism (e.g., Blattman 2009), we find some support for the “revenge hypothesis” in terms of punishment, but not forgiveness. Our study is not the first to question the substantive effects of victimization on post‐conflict attitudes (e.g., Dyrstad and Binningsbø 2019).Importantly, perceptions of volition behind collaboration are strongly associated with harshness of punishment and have an interactive effect with different types of collaboration. Multivariate regression analysis finds that volition is more strongly associated with retribution and reconciliation than victimization (SI, pp. 14–15). In‐depth, qualitative work in post‐conflict areas underscores that the perceived voluntariness of collaboration is an important factor in willingness to forgive and allow the reintegration of former enemy collaborators into their home communities (McKay and Mazurana 2004). More broadly, experimental work in social psychology suggests that intentionality intensifies the perceived damage caused by a transgression (Darley and Huff 1990). Our research fills a gap in the existing quantitative empirical literature on post‐conflict justice and reconciliation, which does not consider variation in agency behind enemy collaboration.ConclusionUnderstanding the microfoundations of preferences for punishment and forgiveness of enemy collaborators in post‐conflict societies is a necessary first step toward reconciliation. Most of the people who support and enable insurgencies, coups, and occupations are civilians, not fighters (Arjona 2016; Petersen 2001; Weinstein 2006). Thus, this research has important theoretical and substantive policy implications for the case of post‐IS Iraq and beyond.Overall, our results challenge scholars and practitioners in the field of transitional justice—where policies are often designed by elites working at the macrolevel—to turn their attention to the microlevel processes through which individuals affected by conflict form attitudes toward different types of collaborators. Including the voices of victims in the process of transitional justice policymaking may bolster the legitimacy of state institutions in contexts where historical experiences with violence, repression, or political exclusion have undermined trust in governments. Our experimental design provides a framework for testing the determinants of justice and willingness to forgive, as well as other important outcomes, including support for peace processes and reintegration of former collaborators. We hope it will spark a broader research agenda on the determinants of justice for, reconciliation with, and reintegration of enemy fighters and civilian collaborators across varied contexts.We suggest several directions for future research. First, additional factors and mechanisms that could affect our outcomes should be further explored. Petersen et al. (2012) suggest that people assess the prospective utility of reintegrating criminals to their society when determining severity of punishment and whether to forgive. Apologies or humility from transgressors has also been shown to encourage forgiveness (e.g., McCullough, Fincham, and Tsang 2003). Other work shows that prospective assessments of threat drive preferences over justice (Hirsch‐Hoefler et al. 2016). Revkin and Kao (Forthcoming) find that traditional leaders (i.e., tribal or religious) can increase acceptance of offenders in the community.Second, much remains to be learned about subnational variation in popular views of transitional justice processes. The generalizability of our findings should be tested in other areas of Iraq that differ in their demography as well as exposure to IS. In a cross‐national comparative perspective, Mosul could be thought of as a hard case for post‐conflict reconciliation because of the extreme violence that IS engaged in. Yet, within Iraq, our Sunni‐only sample of Moslawis is likely an easier case for reconciliation than one drawn from areas with significant Shia, Christian, and Yazidi populations who were severely persecuted by IS. Since our sample includes people who collaborated with IS, it may be more empathetic with collaborators. Additionally, since IS had a high degree of control over the city, there is more room for collaborators to claim they were acting under coercion. Replicating our experiment in other areas of Iraq could advance our understanding of how IS collaborators are perceived among non‐Sunnis and those who experienced differing levels of IS rule (from none to contested to complete control).Third, our design could be employed to make contributions to understanding the microdynamics of enemy collaboration and the potential for reconciliation beyond Iraq. Future studies could run our design in other settings that differ from ours in important ways: variation in regime type (e.g., democracies vs. autocracies), cultural or religious norms and legal traditions, levels of development, duration of conflict and enemy rule, and different patterns of violence and intensities of violence. Another important question is how the passage of time since the cessation of conflict affects prospects for reconciliation, which can be assessed through longitudinal data collection. Our finding that variations in perceptions of enemy culpability and agency shape attitudes toward transitional justice and reconciliation is likely to hold across all of these varying contexts. Replication of our experimental design in other settings would contribute to the development of a more comprehensive, systematic model of attitudes toward post‐conflict retribution and reconciliation.ReferencesAbdul‐Zahra, Qassim, and Susannah George. 2018. “Iraq Holding More than 19,000 Because of IS, Militant Ties.” Associated Press. https://apnews.com/article/after-the-caliphate-islamic-state-group-ap-top-news-international-news-iraq-aeece6571de54f5dba3543d91deed381Aharoni, Eyal, and Alan J. Fridlund. 2011. “Punishment without Reason: Isolating Retribution in Lay Punishment of Criminal Offenders.” Psychology, Public Policy, and Law 18(4): 599–625.Alicke, Mark D. 2000. “Culpable Control and the Psychology of Blame.” Psychological Bulletin 126(4): 556–74.Arjona, Ana. 2016. Rebelocracy: Social Order in the Colombian Civil War. Ithaca, NY: Cornell University Press.Arjona, Ana, Nelson Kasfir, and Zachariah Mampilly. 2015. Rebel Governance in Civil War. New York: Cambridge University Press.Bakke, Kristin M., John O'Loughlin, and Michael D. Ward. 2009. “Reconciliation in Conflict‐Affected Societies: Multilevel Modeling of Individual and Contextual Factors in the North Caucasus of Russia.” Annals of the Association of American Geographers 99(5): 1012–21.Blair, Graeme, Kosuke Imai, and Jason Lyall. 2014. “Comparing and Combining List and Endorsement Experiments: Evidence from Afghanistan.” American Journal of Political Science 58(4): 1043–63.Blattman, Christopher. 2009. “From Violence to Voting: War and Political Participation in Uganda.” American Political Science Review 103(2): 231–47.Boon, Susan D., and Lorne M. Sulsky. 1997. “Attributions of Blame and Forgiveness in Romantic Relationships: A Policy‐Capturing Study.” Journal of Social Behavior and Personality 12(1): 19–26.Daly, Erin. 2001. “Between Punitive and Reconstructive Justice: The Gacaca Courts in Rwanda.” NYU Journal of International Law and Politics 34(2): 355–96.Darley, John M., and Charles W. Huff. 1990. “Heightened Damage Assessment as a Result of the Intentionality of the Damage‐Causing Act.” British Journal of Social Psychology 29(2): 181–88.Davenport, Christian, and Molly Inman. 2012. “The State of State Repression Research since the 1990s.” Terrorism and Political Violence 24(4): 619–34.David, Roman. 2014. “International Criminal Tribunals and the Perception of Justice: The Effect of the ICTY in Croatia.” International Journal of Transitional Justice 8(3): 476–95.David, Roman, and Susanne Choi Yuk‐Ping. 2005. “Victims on Transitional Justice: Lessons from the Reparation of Human Rights Abuses in the Czech Republic.” Human Rights Quarterly 27(2): 392–435.de Greiff, Pablo. 2014. “Report of the Special Rapporteur on the Promotion of Truth, Justice, Reparation and Guarantees of Non‐Recurrence.” United Nations General Assembly, Human Rights Council, A/HRC/27/56.Dyrstad, Karin, and Helga Malmin Binningsbø. 2019. “Between Punishment and Impunity: Public Support for Reactions Against Perpetrators in Guatemala, Nepal and Northern Ireland.” International Journal of Transitional Justice 13(1): 155–84.Enright, Robert D. 1991. “The Moral Development of Forgiveness.” Handbook of Moral Behavior and Development, Vol. 1, eds. William M. Kurtines and Jacob L. Gewirtz. Lawrence Erlbaum Associates, Inc.: New Jersey, US. 123–52.Fabbe, Kristin, Chad Hazlett, and Tolga Smmazdemir. 2019. “A Persuasive Peace: Syrian Refugees’ Attitudes towards Compromise and Civil War Termination.” Journal of Peace Research 56(1): 103–17.Fearon, James, and David Laitin. 1996. “Explaining Interethnic Cooperation.” American Political Science Review 90(4): 715–35.Gibson, James L., and Amanda Gouws. 1999. “Truth and Reconciliation in South Africa: Attributions of Blame and the Struggle over Apartheid.” American Political Science Review 93(3): 501–17.Hainmueller, Jens, Daniel Hopkins, and Teppei Yamamoto. 2013. “Causal Inference in Conjoint Analysis: Understanding Multidimensional Choices via Stated Preference Experiments.” Political Analysis 22(1): 1–30.Hall, Jonathan, Iosif Kovras, Djordje Stefanovic, and Neophytos Loizides. 2018. “Exposure to Violence and Attitudes towards Transitional Justice.” Political Psychology 39(2): 345–63.Hewstone, Miles, Mark Rubin, and Hazel Willis. 2002. “Intergroup Bias.” Annual Review of Psychology 53(1): 575–604.Hill, E. Wayne. 2001. “Understanding Forgiveness as Discovery: Implications for Marital and Family Therapy.” Contemporary Family Therapy 23(4): 369–84.Hirsch‐Hoefler, Sivan, Daphna Canetti, Carmit Rapaport, and Stevan E. Hobfoll. 2016. “Conflict Will Harden Your Heart: Exposure to Violence, Psychological Distress, and Peace Barriers in Israel and Palestine.” British Journal of Political Science 46(4): 845–59.Honey, P. Lynne. 2017. The Oxford Handbook of Women and Competition. Oxford: Oxford University Press.Human Rights Watch. 2017. “Flawed Justice: Accountability for ISIS Crimes in Iraq.” https://www.hrw.org/sites/default/files/report_pdf/iraq1217web.pdfHumphreys, Macartan, and Jeremy Weinstein. 2007. “Demobilization and Reintegration.” Journal of Conflict Resolution 51(4): 531–67.Jones, Nicholas A., Stephan Parmentier, and Elmar G. M. Weitekamp. 2012. “Dealing with International Crimes in Post‐War Bosnia: A Look through the Lens of the Affected Population.” European Journal of Criminology 9(5): 553–64.Kalyvas, Stathis. 2006. The Logic of Violence in Civil War. New York: Cambridge University Press.Kalyvas, Stathis N., and Laia Balcells. 2010. “International System and Technologies of Rebellion: How the End of the Cold War Shaped Internal Conflict.” American Political Science Review 104(3): 415–29.Kasfir, Nelson. 2005. “Guerrillas and Civilian Participation: The National Resistance Army in Uganda, 1981–86.” Journal of Modern African Studies 43(2): 271–96.Latimer, Jeff, Craig Dowden, and Danielle Muise. 2005. “The Effectiveness of Restorative Justice Practices: A Meta‐Analysis.” The Prison Journal 85(2): 127–44.Lyall, Jason, Graeme Blair, and Kosuke Imai. 2013. “Explaining Support for Combatants during Wartime: A Survey Experiment in Afghanistan.” American Political Science Review 107(4): 679–705.Mampilly, Zachariah. 2011. Rebel Rulers: Insurgent Governance and Civilian Life during War. Ithaca, NY: Cornell University Press.McCullough, Michael E., Frank D. Fincham, and Jo‐Ann Tsang. 2003. “Forgiveness, Forbearance, and Time: The Temporal Unfolding of Transgression‐Related Interpersonal Motivations.” Journal of Personality and Social Psychology 84(3): 540–57.McKay, Susan, and Dyan E. Mazurana. 2004. Where Are the Girls? Girls in Fighting Forces in Northern Uganda, Sierra Leone and Mozambique: Their Lives during and after War. International Center for Human Rights and Democratic Development, Montreal, Canada.Olson, Mancur. 1993. “Dictatorship, Democracy, and Development.” American Political Science Review 87(3), 567–76.Petersen, Michael Bang, Aaron Sell, John Tooby, and Leda Cosmides. 2012. “To Punish or Repair? Evolutionary Psychology and Lay Intuitions about Modern Criminal Justice.” Evolution and Human Behavior 33(6): 682–95.Petersen, Roger D. 2001. Resistance and Rebellion: Lessons from Eastern Europe. Cambridge: Cambridge University Press.Revkin, Mara. 2018. “After the Islamic State: Balancing Accountability and Reconciliation in Iraq.” New York: United Nations University Centre for Policy Research.Revkin, Mara Redlich. 2020a. “Competitive Governance and Displacement Decisions under Rebel Rule: Evidence from the Islamic State in Iraq.” Journal of Conflict Resolution 65(1): 46–80.Revkin, Mara Redlich. 2020b. “What Explains Taxation by Resource‐Rich Rebels? Evidence from the Islamic State in Syria.” Journal of Politics 82(2): 757–64.Revkin, Mara Redlich, and Elisabeth Wood. 2020. “The Islamic State's Pattern of Sexual Violence: Ideology and Institutions, Policies and Practices.” Journal of Global Security Studies 6(2): 1–20.Revkin, Mara, and Kristen Kao. In Press. “No Peace Without Punishment? Reintegrating Islamic State Collaborators in Iraq.” The American Journal of Comparative Law.Samii, Cyrus. 2013. “Who Wants to Forgive and Forget? Transitional Justice Preferences in Postwar Burundi.” Journal of Peace Research 50(2): 219–33.Scott, Elizabeth S., N. Dickon Reppucci, Jill Antonishak, and Jennifer T. DeGennaro. 2006. “Public Attitudes about the Culpability and Punishment of Young Offenders.” Behavioral Sciences & the Law 24(6): 815–32.Shaw, Rosalind, Lars Waldorf, and Pierre Hazan. 2010. Localizing Transitional Justice: Interventions and Priorities after Mass Violence. Stanford, CA: Stanford University Press.Sly, Liz. 2015. “The Hidden Hand behind the Islamic State Militants? Saddam Hussein's.” Washington Post.Stewart, Megan. 2018. “Civil War as State‐Making: Strategic Governance in Civil War.” International Organization 72(1): 205–26.Stylianou, Stelios. 2003. “Measuring Crime Seriousness Perceptions: What Have We Learned and What Else Do We Want to Know?” Journal of Criminal Justice 31(1): 37–56.Tabak, Shana. 2011. “False Dichotomies of Transitional Justice: Gender, Conflict and Combatants in Colombia.” New York University Journal of International Law and Politics 44(1): 103–63.Tellez, Juan Fernando. 2019. “Worlds Apart: Conflict Exposure and Preferences for Peace.” Journal of Conflict Resolution 63(4): 1053–76.Terpstra, Niels. 2020. “Rebel Governance, Rebel Legitimacy, and External Intervention: Assessing Three Phases of Taliban Rule in Afghanistan.” Small Wars & Insurgencies 31(6): 1143–73.Tyler, Tom R. 2003. “Procedural Justice, Legitimacy, and the Effective Rule of Law.” Crime and Justice 30: 283–357.United Nations. 2006. Handbook on Restorative Justice Programmes. New York: United Nations Office on Drugs and Crime.Verini, James. 2016. “Surviving the Fall of ISIS.” National Geographic. https://www.nationalgeographic.com/culture/article/islamic-state-isis-iraq-mosul-syria-offensiveWeinstein, Jeremy. 2006. Inside Rebellion: The Politics of Insurgent Violence. Cambridge: Cambridge University Press.Wistrich, Robert S. 2013. Who's Who in Nazi Germany. New York: Routledge.Wood, Elisabeth. 2006. “The Ethical Challenges of Field Research in Conflict Zones.” Qualitative Sociology 29(3): 373–86.Wood, Elisabeth Jean. 2003. Insurgent Collective Action and Civil War in El Salvador. Cambridge: Cambridge University Press.Worthington, Everett L., Jr. 2006. Forgiveness and Reconciliation: Theory and Application. New York: Routledge. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png American Journal of Political Science Wiley

Retribution or Reconciliation? Post‐Conflict Attitudes toward Enemy Collaborators

Download PDF
16 pages

Loading next page...
 
/lp/wiley/retribution-or-reconciliation-post-conflict-attitudes-toward-enemy-LLWceqVOGv

References (62)

Publisher
Wiley
Copyright
© 2023 by the Midwest Political Science Association.
ISSN
0092-5853
eISSN
1540-5907
DOI
10.1111/ajps.12673
Publisher site
See Article on Publisher Site

Abstract

Verification MaterialsThe data and materials required to verify the computational reproducibility of the results, procedures, and analyses in this article are available on the American Journal of Political Science Dataverse within the Harvard Dataverse Network, at: https://doi.org/10.7910/DVN/0ONBF6.After violent conflicts over territory end, there is often a widespread belief that those who lived under the rule of an armed group were “collaborators”1Collaboration is a widely used term in conflict research, but it is not often defined. We use collaborator to mean a person who supports an armed group, whether voluntarily or under coercion. and are therefore complicit in crimes perpetrated by the enemy. As an Iraqi interviewee attests: “People assume that everyone who stayed in Mosul is an Islamic State supporter or member, but many of us were victims.”2Author interview with “Khaled” (38, accountant) in Mosul, Iraq (April 2017). To ensure anonymity, interviewees are identified by pseudonyms. Generally, policy makers, scholars, and civilians living outside of enemy‐held territory take for granted “unlimited and unwavering support of the population for the political actor who claims to represent it,” treating individuals involved in conflict as constituting an undifferentiated and monolithic “entity that must be ‘won’ by political actors” (Kalyvas 2006, 6–7). In reality, an enemy collaborator can be both a victim and a perpetrator, or lie somewhere on a continuum between the two. Many residents of territory captured by an armed enemy group are victims of its violence and only comply with its policies in order to stay alive. Others do so willingly, or even enthusiastically.The territorial defeat of the Islamic State (IS, also known by its Arabic acronym, Daesh), a Sunni armed group that captured large swaths of territory in Iraq and Syria, provides a unique opportunity to collect data on public attitudes toward enemy collaborators at a time when peace processes and accountability mechanisms are still at an early stage in development. Our fieldwork in Iraq indicates that war‐torn communities perceive significant variation in the culpability of different types of accused enemy collaborators. At the time of our survey in Mosul, IS's de facto capital from 2014 to 2017, many former civilian employees of IS's administrative and service‐providing institutions were still living in the city without fear of prosecution or reprisals. One municipal worker (“Zyad”) explained, “No one blames us for keeping our jobs when IS arrived because we needed to feed our families, and we continued doing the same work we had done before, just with new bosses. Besides, quitting was not an option because it would have been an act of rebellion, which would have put me and my family at risk for severe punishment.”3Author interview with “Zyad” (35, municipal services) in Mosul (April 2017). Yet, Iraqis affiliated with IS in other ways were not being forgiven so easily. Widows of IS fighters reported that they would rather remain indefinitely in camps for internally displaced persons because they fear for their safety and that of their children if they return to their former hometowns. For example, “Laila,” whose brother's house was attacked with grenades as a result of the family's ties to IS, said, “I am afraid that if I return, my neighbors would kill me in my sleep.”4Author interview with “Laila” (40) in Hajj Ali IDP Camp (December 2017). What explains the stark difference between the cases of these two collaborators? One (“Zyad”) is perceived as innocent while the other (“Laila”) is facing death threats.To assess variation in attitudes toward the punishment and forgiveness of enemy collaborators, we employ an experiment embedded in an original survey of 1,458 Sunni Arab residents of Mosul (Moslawis). We evaluate the effects of randomly varied identity attributes of hypothetical IS collaborators and a range of collaborative acts on preferences for punishment and willingness to forgive. Although attitudes toward reconciliation are necessarily context‐specific, our research design can be applied in a wide variety of post‐conflict settings—including civil wars, foreign occupations, and post‐authoritarian transitions—to further our understanding of the prospects for legitimate justice and reconciliation at the microlevel. By widening our analytical lens to consider a broad spectrum of enemy collaboration, this study challenges a false dichotomy between victims and perpetrators found in much of the existing research on transitional justice (Tabak 2011). Our results demonstrate that variation in the type of enemy collaboration is an important determinant of preferences for post‐conflict punishment and forgiveness. This finding is very strong, remaining robust even in interactions with the identity attributes (e.g., age, gender, tribe) manipulated in our experiment as well as with respondent identity characteristics (e.g., age, gender, educational background). We argue that the predominant focus of existing transitional justice work on ingroup versus outgroup dynamics has distracted scholars from considering a potentially stronger determinant of preferences for punishment and forgiveness: variation in individual culpability.Observationally, we explore several mechanisms that may mediate the effects of collaborator culpability. Our data provide some support for the “revenge hypothesis”—that victimization at the hands of an enemy group leads to an increased desire for retribution—although the effects are substantively small by comparison. Instead, a key finding is that attitudes regarding punishment and forgiveness are strongly associated with perceived volition behind the act of collaboration, a mechanism that is understudied in the empirical literature and should receive greater attention in models of attitudes toward transitional justice. In emphasizing this mechanism, we contribute to a growing literature on the dynamics of civilian agency during conflict (e.g., Arjona 2016; Wood 2003) by empirically evaluating how it shapes prospects for reconciliation after conflict ends. This study provides uniquely fine‐grained data on the factors that shape perceptions of individual enemy collaborators, which have not been systematically tested or theorized by scholars thus far. Moreover, it offers a replicable research design and an expandable theoretical framework for furthering research on perceived collaborators in other post‐conflict settings.Why Study Public Perceptions of Former Enemy Collaborators?When conflicts end, national and international elites determine transitional justice policies, often paying scant attention to the opinions of local populations (Jones, Parmentier, and Weitekamp 2012). Similarly, studies of post‐conflict public opinion focus on attitudes toward different state‐imposed justice mechanisms, including lustration laws, truth commissions, or peace processes (Fabbe, Hazlett, and Smmazdemir 2019; Hall et al. 2018; Hirsch‐Hoefler et al. 2016; Tellez 2019). Although these studies make important contributions, the microfoundations of reconciliation with individual enemy collaborators remain poorly understood. As a result, there is increasing concern that transitional justice processes do not adequately address the concerns and needs of victims on the ground, whose support for and belief in the legitimacy of these processes is necessary for the reintegration of individual perpetrators into their local communities (e.g., Shaw, Waldorf, and Hazan 2010), without which sustainable peace cannot be achieved. To fill this gap, our study takes a bottom‐up, microlevel, and victim‐centered approach that gives voice to the people whose attitudes matter most for conflict resolution and sustainable peace.Rich descriptive work provides vivid accounts of the wide range of engagement options available to civilians, from defiance to full support of armed groups (e.g., Arjona 2016; Kalyvas 2006). Petersen (2001, chap. 1) develops a scaled spectrum of roles for defectors versus collaborators, placing participation in violence at the extremes, to theorize about when and why individuals decide to take on these different roles. The various types of collaboration that civilians engage in have been shown to shape the internal organization of enemy groups, in addition to the establishment or collapse of their rule (Arjona 2016; Arjona, Kasfir, and Mampilly 2015; Weinstein 2006). We posit that variation in the type of collaboration also matters for the design of post‐conflict transitional justice processes. Individuals develop attitudes toward collaborators based on their actions, which inform perceptions of their culpability and preferences for punishment.Previous quantitative research has largely overlooked this important variation in the different roles and varying culpability of enemy collaborators, gauging attitudes toward armed groups as a whole (e.g., Blair, Imai, and Lyall 2014; Dyrstad and Binningsbø 2019) or the most egregious type of enemy collaboration—participation in violence (e.g., David 2014; Lyall, Blair, and Imai 2013; Samii 2013). Empirically, we know much less about public opinion toward civilian collaborators despite their importance. Parties to conflicts over territorial control rely heavily on civilians to obtain food, water, shelter, labor, and information (Arjona 2016; Kalyvas 2006; Weinstein 2006; Wood 2003).A handful of studies randomize some features of enemy combatants and ask about justice mechanisms to examine the effects of ingroup versus outgroup identities. David (2014) finds that social identity matters for popular perceptions of justice among Serbs and Croats, with outgroup members receiving harsher scrutiny (David 2014, 489). In South Africa, ex‐combatant leaders are more likely to be blamed for their actions than their subordinates, and those who were pro‐apartheid in particular, although these outcomes are mediated by respondent race (Gibson and Gouws 1999). Another study asks respondents about forgiveness of “people of other nationalities for the violence they have committed in the last fifteen years,” finding that personal experience with violence and its effects, not ethnic hostility toward outgroup members, negatively correlates with willingness to forgive (Bakke, O'Loughlin, and Ward 2009, 1017). Notably, this type of experimental design does not allow us to distinguish whether it is the reminder of past violence, the various events that have taken place within the last 15 years, the many different outgroups invoked by the prompt, or combinations of some or all of these factors that drive attitudes.A common thread linking these studies—and informing ours—is the core idea that characteristics of enemy collaborators shape attitudes toward justice and reconciliation. Yet, in line with other work on attitudes toward enemy combatants (e.g., Lyall, Blair, and Imai 2013; Samii 2013), these studies do not examine collaboration by civilians, nor do they consider variation in individual identity characteristics of collaborators such as gender or age. In sum, the existing literature fails to explore more nuanced understandings of guilt and commensurate justice for individuals engaged in lesser crimes. Our research design experimentally manipulates the identities and actions of hypothetical enemy collaborators in order to make causal inferences about microlevel determinants of attitudes toward justice and forgiveness.Our work also has important and timely policy implications. Although public opinion should never be the sole basis for the design of transitional justice policies because of the risk of “mob justice” (Daly 2001, 383), it should be taken into consideration by policy makers to facilitate durable peace after conflict. Social psychologists have found that sensitivity to popular concerns about the fairness of legal institutions is an important determinant of trust in the police and courts, warning that failure to consider public opinion may increase the likelihood of noncompliance with laws and state authorities (e.g., Tyler 2003). Work in criminology finds that justice processes are more likely to lead to reconciliation between adversarial groups and a reduction of criminal recidivism when they take into account the voices of victims (Latimer, Dowden, and Muise 2005). In communities where victims of violence are deeply dissatisfied with the state's official response, they may take matters into their own hands, resulting in extrajudicial revenge (Human Rights Watch 2017). Thus, there is a real need for safeguards to ensure that transitional justice processes protect the fundamental rights of perpetrators—including the right to due process in trials (United Nations 2006, 36)—and that these processes are acceptable to the society at‐large.Conducting Multimethod Research in Mosul, IraqWe conducted an original survey of 1,458 Mosul residents from March to April 2018 with an Iraqi research firm.5The company was the Independent Institute for Administration and Civil Society Studies. A team of enumerators recruited from Mosul conducted face‐to‐face surveys with tablets. In addition to the quantitative data, our research draws on qualitative evidence from fieldwork in Mosul and other areas of northern Iraq in 2017 (supporting information [SI], pp. 17–20). This fieldwork enabled us to ensure the appropriateness of our survey questions for the context and to validate the realism of the experiment. It included visits to public institutions that were previously administered by IS, as well as observations of trials of alleged IS members. We carried out interviews with 61 individuals from areas previously controlled by IS, as well as 17 lawyers, judges, and experts involved in prosecuting and defending suspected IS collaborators.Case SelectionCivilian collaboration with enemy rule and governance is a widespread phenomenon. Focusing solely on civil wars, Stewart (2018) estimates that one‐third of all rebel groups active from 1945 to 2003 engaged in governance of civilians. Scholarly accounts are replete with descriptions of how collaboration is the lifeblood of any armed campaign to control territory (e.g., Kalyvas 2006; Petersen 2001).6For a long list of armed groups that provided public services, see Arjona (2016, chaps. 1–3). Throughout history and around the world, armed groups such as the Revolutionary Armed Forces (FARC) in Colombia (Arjona 2016), the Taliban in Afghanistan (Terpstra 2020), the Liberation Tigers of Tamil Eelam (LTTE) in Sri Lanka (Mampilly 2011), the Bakonozos of Northern Uganda (Kasfir 2005), and the Nazis in occupied territories during the Second World War (Wistrich 2013) have relied heavily on a range of collaborators to capture and hold territory. The spectrum of collaboration in these cases ranged from horrific acts—such as murder, kidnapping, torture, rape, and even genocide—to much less severe transgressions that did not involve violence; civilian collaborators in these cases performed a wide range of nonmilitary functions and services, including tax collection, sanitation, healthcare, education, and supplying or otherwise supporting combatants.IS rule in Mosul is just one example of the broader phenomenon of enemy rule and governance. Across a wide variety of settings and cases, rebel governance, defined as “the set of actions rebels engage in to regulate the social, political, and economic life of non‐combatants during war” (Arjona, Kasfir, and Mampilly 2015, 3), has become more common with the increase in civil wars since the end of the Cold War (Kalyvas and Balcells 2010). Likewise, the broader phenomenon of enemy governance is characterized by significant civilian participation—whether voluntary or coerced—in the enemy's governing institutions. IS set up a government, established its own rules of conduct, and maintained control of substantial territory for 3 years. The group operated a variety of institutions that provided public goods and basic services, which necessitated a civilian bureaucracy staffed by employees who generally did not engage in violence. Moslawis witnessed a wide range of collaboration, including (1) compliance with taxation (Revkin 2020b); (2) social integration with the group through marriage to its members and the enrollment of children in IS‐controlled schools; (3) employment in IS's civilian workforce as teachers, doctors, or cooks; and (4) recruitment as fighters (Revkin 2020a). Since Moslawis confronted these different types of collaboration on a daily basis, they can recognize and imagine the scenarios described in our experiment.Residents of Mosul are now grappling with the question of what to do with former collaborators and under what conditions they might be forgiven for their transgressions. Surveying people in the place where these transgressions occurred provides important insights into the microfoundations of attitudes toward reconciliation in a context where there is an urgent need to reduce the risk of conflict recurrence. Thus, Mosul is a particularly relevant setting in which to collect data on attitudes toward former enemy collaborators.Our sample includes only Sunni Arabs living in Mosul in June 2014—when IS arrived—and therefore having had some exposure to IS. Sunnis made up more than 97% of the city's population at the time of the survey. The sample is not representative of Iraq as a whole. Including representative samples of other identity groups in Iraq (e.g., Shias) not only would have been prohibitively costly, but it also would have shifted the focus of this study to the very salient ethno‐sectarian divides in Iraq. We expect that Shias would be much less forgiving of former IS collaborators than the Sunni respondents in our sample given IS's particularly harsh treatment of Shias. As our goal was instead to focus on the effects of varying degrees of enemy collaboration on the prospects for post‐conflict reintegration and reconciliation, we intentionally held this very salient social identity constant in our study. It is also important to note here that the conflict with IS was not fought purely along ingroup versus outgroup lines: IS killed more Sunni Muslims than any other religious or ethnic group (Verini 2016). In sum, our sample includes both IS collaborators and victims, identities that are not mutually exclusive. It is representative of populations living in territories that have experienced enemy rule and governance and are now grappling with the question of how to assess the culpability of the enemy collaborators among them. Our findings would not generalize to cases where a defeated enemy did not seek to govern, such as armed groups whose sole purpose is economic predation of a territory's resources or where chaos ensued. Such cases are rare, however, as even “roving bandits” have incentives to establish a social contract with civilian populations to profit over the long term (Arjona 2016; Olson 1993).We were also motivated by the immediate policy implications that such a study in Mosul could have for post‐transitional justice in Iraq. At the height of its expansion in 2014, IS governed millions of people. When IS was militarily defeated in 2017, it left behind a population that is now widely perceived as collectively complicit in the group's crimes. The government is currently facing the monumental challenge of reintegrating this population back into their local communities, but authorities have taken a heavy‐handed approach that fails to differentiate between voluntary and involuntary collaboration, and more serious crimes and lesser offenses. Iraq's Anti‐Terrorism Law criminalizes membership in any terrorist group without requiring proof of a specific criminal act; anyone with a plausible connection to the group can easily be sentenced to life in prison, the minimum punishment allowed by the law.7Law No. 13 (2005). More than 8,000 accused IS collaborators have been convicted in trials that are often decided in under 30 minutes, with a conviction rate of around 98%,8Author observations of trials in Tel Kaif (December 2017). and more than 3,000 have been sentenced to death (Abdul‐Zahra and George 2018).This one‐punishment‐fits‐all approach—which is widely perceived as collective punishment of Sunnis—is generating new grievances that could fuel the emergence of an “IS 2.0” (Revkin 2018). A correlation between repression and radicalization has been documented in many contexts (Davenport and Inman 2012). And it has been argued that the rise of IS—which emerged from the remnants of al Qaeda in Iraq—was fueled by resentment over the collective punishment of Sunnis through de‐Baathification and incarceration (Sly 2015).In addition to its direct and immediate implications for the case of Iraq, this work speaks to a broader set of ongoing as well as future cases of post‐conflict transitional justice processes. In all cases of conflict involving territorial sovereignty, victors face the challenge of walking a fine line between under‐ and overpunishment of former enemy collaborators. When punishment of a perpetrator falls short of what the victims believe is commensurate to the crime committed, the resulting perception of an “injustice gap” increases the likelihood of victims being dissatisfied with the outcomes (Worthington 2006). Yet, excessive punishment may be perceived as victors’ justice and delegitimize transitional justice efforts (de Greiff 2014, 18). Popular opinion data can help policy makers find the middle ground.Theorizing Post‐Conflict Justice PreferencesWe develop an original theoretical framework for analyzing how individuals in conflict‐affected societies form preferences for punishment and forgiveness of other community members who collaborated to varying extents with an enemy. As noted, previous research has established that social identity is an important determinant of reconciliation in intergroup conflicts (e.g., David and Yuk‐Ping 2005; Gibson and Gouws 1999; Samii 2013). Our theoretical contribution is to highlight the importance of an additional factor that has been understudied in the conflict literature: culpability. We posit that perceived culpability varies depending on an individual collaborator's physical proximity to and social intimacy with the enemy. In this section, we develop and test several hypotheses about the expected effects of variation in individual collaborator attributes and acts of enemy collaboration during conflict on preferences for accountability and prospects for reconciliation.9We preregistered these hypotheses and secondary expectations prior to data collection. See Kao, Kristen and Mara Revkin. “Pre‐Analysis Plan: ‘Reintegration of Enemy Collaborators: Survey Experiments in Mosul,’” Evidence in Governance and Politics (March 29, 2018), https://egap.org/registration/4395.Variation in Types of Collaboration and Perceived CulpabilityWe argue that information about the type of collaboration, which is closely linked with the perceived culpability of the collaborator, is an important determinant of preferences for punishment. Studies in social psychology posit that more severe transgressions in personal relationships are more difficult to forgive (e.g., Boon and Sulsky 1997). Work in political science indicates that enemy fighters from more abusive units are less likely to be reintegrated back into society (Humphreys and Weinstein 2007) and that commanders are held more responsible than subordinates (Gibson and Gouws 1999). Considerable evidence across a variety of contexts suggests that individuals see violent behaviors as the most serious offenses (Stylianou 2003).To explore variation in perceptions of culpability behind collaboration, we chose collaborator roles that vary in their proximity to violence. In addition to the role of combatants, we examine a spectrum of nonmilitary roles that collaborators—both men and women—may perform. The specific roles included in the experiment are based on our fieldwork in Iraq as well as common patterns of enemy collaboration in other conflicts. In particular, we hypothesized that fighters would be most harshly punished and least likely to be forgiven due to the violent nature of their collaboration. Additionally, we expected that respondents would prefer harsher punishments for collaborators who were physically closest to enemy combatants (cooks for fighters) and those who were the most intimate with them (wives of fighters) in contrast with collaborators in civilian roles for institutions that provided services to other civilians (janitors working in IS's department of municipal services) or those who financially supported the insurgency (taxpayers).10We validated our assumption that different acts of collaboration are associated with varying levels of severity. We found that 98% of the sample agrees that being a fighter is the most condemnable transgression, followed by civilians directly involved with fighters. Those not directly involved with fighters are ranked as the least condemnable (SI, p. 5).Identity TraitsAlthough not the main focus of this study, we believed that the individual identity characteristics of collaborators would have either direct or moderating effects on attitudes toward punishment and forgiveness. Some identities are seen as less agentic and therefore less culpable than others. Shared identities between the respondent and a collaborator may trigger empathy and forgiveness.Based on previous research in criminology, we expected respondents to prefer more lenient punishments for and be more forgiving of younger collaborators. Juveniles are generally assumed to be less agentic than adults because they are easily influenced by those around them and they may not be able to distinguish between right and wrong (Scott et al. 2006). Similar to most penal codes around the world, Iraqi law requires reduced punishments for children, taking into consideration their age and the stage of their mental development at the time of the offense.11Iraq's Penal Code (Act No. 111 of 1969), Articles 67–78.We also expected respondents to prefer more lenient punishments for women than for men. Previous studies find that women in other contexts are perceived as being less responsible when they cause harm, and they receive lighter punishments compared to men, ceteris paribus (Honey 2017). Interviews in Iraq support this expectation. “Fadila” explained that when her husband decided to join IS and she expressed misgivings, he replied, “You can leave and I will keep the kids.”12Author interview with “Fadila” (35, wife of an IS fighter) in Ninewa (December 2017). Anecdotal evidence suggests that female collaborators should be perceived as less culpable than men because, as “Fadila” put it, “We did not have a choice.”Research demonstrates that members of the same group tend to favor one another and punish outsiders (e.g., David and Yuk‐Ping 2005; Samii 2013). In conflict settings in particular, ingroup biases become more pronounced (Hewstone, Rubin, and Willis 2002). In social psychology, McCullough, Fincham, and Tsang (2003) also suggest that shared social identity encourages empathy with and increased benevolence toward a transgressor. In the context of the tribal society of Iraq, these previous findings concerning ingroup biases led us to expect that people should be more empathetic, lenient, and forgiving of transgressions by members of their own tribe (ingroup) in comparison with members of other tribes (outgroups).VictimizationPreferences for punishment of enemy collaborators may be affected by whether an individual was victimized by the enemy group. Many studies find that exposure to violence decreases willingness to forgive, reconcile, and cooperate with transgressors (e.g., Bakke, O'Loughlin, and Ward 2009; Hall et al. 2018; Hirsch‐Hoefler et al. 2016). Based on this previous research, we expected that individuals who have been victimized by an enemy group—as measured by the death or injury of family members or property destruction—would prefer harsher punishments for enemy collaborators than those who did not experience such a personal loss.Volition of CollaborationWe expected the perceived volition behind the acts of collaboration to be associated with preferences for punishment of collaboration. Assessing volition behind acts of transgression, while difficult, is a key component of most psychological models of blame and responsibility (e.g., Alicke 2000, 57). Experimental research finds that belief in intention is a major driver of preferences for punishment of criminals (Aharoni and Fridlund 2011), and that more intentional transgressions are more difficult to forgive (e.g., Boon and Sulsky 1997). We test whether these findings generalize to a post‐conflict setting.Perceived volition in engagement might be expected to have stronger effects for lesser collaborative acts; as the act increases in condemnability, intent may not be as impactful on attitudes toward retribution or reconciliation. Volition is a particularly important factor in contexts where the collaborator may be acting under considerable economic or physical duress, as is the case in wartime environments where enemy groups exercise coercive control over territory and people. For instance, IS required all residents of its territory—except for the extremely poor—to pay taxes. Yet, some IS supporters may have voluntarily paid taxes, seeing their payment as an appropriate payment for services that the group was providing. Ten percent of our sample openly agreed that the fees collected by IS were fair in exchange for the services that IS was providing.13Another 74% disagreed and 16% refused to answer. Altogether, 54% of the sample admitted paying taxes to IS in some form. Analyses of engagement in this act of collaboration on our outcomes are shown in the SI (p. 13).There was also considerable variation in the voluntariness of marriage to IS fighters. Some women were already married when their husbands decided to join the group. Others married fighters for the purpose of gaining social status. As one interviewee explains: “In poor neighborhoods, some women believed that they could become princesses by marrying IS emirs.”14Author interview with “Salem” in Mosul (April 2017). Emir means “prince,” which IS used to refer to high‐ranking officials. Still others were coerced into marrying fighters through social pressure, economic duress, or physical threats.Punishment, Forgiveness, and ReconciliationThis work measures two interrelated outcomes: forgiveness and punishment. We adopted a widely cited definition of forgiveness: the lessening of negative feelings, thoughts, and behaviors toward transgressors (McCullough, Fincham, and Tsang 2003). Forgiveness is “one of the most important processes in the restoration of interpersonal relationships after conflict” (Hill 2001). Based on previous research, we also assumed that benevolence toward transgressors is generally conditional upon the belief that justice has been served (Enright 1991, 128). Yet, although punishment may facilitate forgiveness in some cases, it is not established that punishment is necessary—which we allow for in our design.Our experimental prompt is designed to trigger attitudes toward reintegration and reconciliation by priming respondents with the information that the collaborators want to move back into their neighborhood (and could therefore interact with them in the future). We ask respondents whether—after selecting the punishment that they feel is appropriate for a given hypothetical collaborator—they would be willing to forgive these persons. Although we cannot identify the causal effect of punishment on forgiveness, this second dependent variable nonetheless enables us to examine the potential for reconciliation through correlational analysis.Experimental DesignTo evaluate respondents’ beliefs about the type of justice deserved by former collaborators, we implement a ratings‐based conjoint experiment. The design significantly reduces the number of participants needed while maintaining sufficient power to test multiple hypotheses by randomizing each potential driver of outcomes independently of others. We included a follow‐up question on willingness to forgive after punishment, which we use to assess prospects for reconciliation.The experiment randomizes enemy collaborators’ identities and the nature of their collaboration with IS. As identity characteristics, we included gender,15Though the majority of IS fighters were male, IS had some female combatants. age, and whether or not the collaborator is a member of the respondent's tribe to serve as identity cues.16We did not vary ethno‐religious identities because IS was predominantly a Sunni Arab organization. Almost all Moslawis identify with a tribe. We also randomize acts of collaboration. The five collaboration roles specified in the experiment are (1) fighting for IS, (2) working as a cook for IS fighters, (3) being married to an IS fighter, (4) working as a janitor for the IS municipality, and (5) paying taxes to IS.Every respondent evaluated three separate profiles that were generated by randomizing the attributes listed in Table 1. The total sample of evaluated profiles was 4,275. Before the enumerator reads the descriptions of the hypothetical collaborators, the respondent is told, “I am going to read you some hypothetical scenarios about people from Mosul who are being prosecuted for their past cooperation with Daesh (IS). These people now want to move back into your neighborhood. I would like you to choose the type of punishment that you view as appropriate for this person. The person is a [insert profile].”1TableRandomized Enemy AttributesDimensionAttributesGender– Man– WomanAge– 15– 35Tribal member– Respondent's tribal group– Member of another tribeType of collaboration– IS fighter– Cook for IS fighters– Married to an IS fighter (limited to female candidates)– Janitor at the municipality employed by IS's government– Resident of Mosul who paid taxes to ISAfter the respondent is read a collaborator profile, she is told: “A thorough investigation concluded that this is the only act of collaboration that the person committed. I have ordered the following punishments from least harsh to most harsh. I would like you to choose the type of punishment you deem appropriate for this former Daesh collaborator, who now wants to move back into your neighborhood.” The responses to this question make up our dependent variables and include the following options: no punishment necessary (least harsh), mandatory community service (e.g., picking up trash, rebuilding homes) for six months, imprisonment for 3 years, imprisonment for 15 years, and capital punishment (most harsh).This first sentence in the prompt serves two purposes. First, it helps address the concern that respondents might impute other types of collaboration to the profile by specifying that the stated act is the only one committed. Second, it encourages respondents to view the five punishment types as an ordered scale from least to most harsh. To validate our ranked scale, we ran a pilot study in which we asked 100 Moslawis to rank the punishments from least to most harsh. We also ran a post‐experiment validation check of this question (SI, p. 11).17Over 90% of those asked agreed with our ranking in the pilot and full sample; the 10% that do not simply add noise to our findings. Only 4% of the sample preferred a different type of punishment, suggesting the scale of punishments we offered was well‐aligned with actual preferences.Research EthicsResearch in conflict areas raises unique ethical challenges (Wood 2006), as well as security concerns for researchers and their subjects. Asking respondents in post‐conflict settings to share their experiences and attitudes about a recent conflict in which they and their families may have experienced several personal harms has the potential to trigger retraumatization (Revkin and Wood 2020). Another concern is that such questions could increase polarization, particularly in multiethnic communities. We took these concerns very seriously and conducted semi‐structured interviews with Iraqis, as well as our enumerators from Mosul, to assess whether any of the questions would be offensive or disturbing to respondents.The overwhelming feedback was that the questions would not be upsetting to respondents because the aftermath of the very recent conflict was still, tragically, a constant presence in the lives of residents of Mosul in 2018. Conversations about IS collaborators, their culpability, and justice mechanisms were commonplace in 2018 and therefore less sensitive than might be expected. For example, one Moslawi expressed frustration about the lack of accountability: “Family members of IS fighters, who were beneficiaries of IS and its crimes, are living among us, and no one is holding them accountable.”18Author interview with “Walid” (33, store clerk) in Mosul (April 2017). Many of our interlocutors were eager to discuss their experiences with IS and preferences for justice, apparently viewing this study as an opportunity to tell their stories to a broader audience than they could otherwise reach.We found similar enthusiasm for participation in the household survey, with only 15% of potential respondents declining. Within the survey, the response rate for arguably the most sensitive question—“During the first six months of Daesh rule, did you believe that Daesh was doing a better job of governing Mosul than the Iraqi government did previously?”—was even higher: Only six respondents (4% of the sample) said they did not know or refused to answer.Given the extent to which IS collaboration was already being publicly debated by Iraqis at the time of the survey, as well as an informed consent procedure that allowed all respondents to opt out of the survey at any time, we do not believe that the survey exposed respondents to significant risk.19We also designed a list experiment to assess sensitivity about preference for IS governance. The results suggest that our results were not significantly affected by social desirability bias (SI, pp. 9–10). The SI (pp. 15–20) discusses other potential risks to survey respondents and the steps taken to minimize those risks.20The Yale University Institutional Review Board approved the survey instrument (Protocol #2000022022), observations of trials (Protocol #2000021840), and interviews with Iraqis (Protocol #1506016040).Analyses and ResultsExamining the distribution of the dependent variable across all types of collaboration, the two most frequently selected options were no punishment (28%) and capital punishment (33%), indicating that there is considerable variation in the preferences of Moslawis concerning justice. Table 2 shows that IS fighters and those who were most closely associated with fighters (cooks for and wives of fighters) receive consistently harsher punishments than those less closely associated with fighters (janitors who worked for the IS municipality and taxpayers). More than three‐quarters of the sample sought capital punishment for IS fighters, whereas a similar proportion did not think any punishment was necessary for taxpayers. About one‐third of the sample considered death to be an appropriate punishment for cooks and wives of fighters, although for married women half of the sample sought less than 15 years in prison. The largest gaps in perceptions of appropriate punishment are between those who are intimately involved with violence in contrast to those who are more distant from it. Importantly, more than two‐thirds of the sample sought a restorative punishment (community service) or no punishment for janitors at the IS municipality and for taxpayers.2TablePunishments Preferred for Types of Collaboration (Percentage of Cases)ActNo PunishmentCommunity Service3 Years in Prison15 Years in PrisonCapital PunishmentIS fighter1251378Cook for fighters314222536Married a fighter1717161831Janitor at municipality412711615Paid taxes747549Total2814121433We employ ordinary least squares regression (OLS) to assess the average marginal component effect (AMCE) of each of the profile attributes, pooling across all respondents and tasks.21The results are robust to ordinal logistical analysis. Following Hainmueller, Hopkins, and Yamamoto (2013), we expect OLS regression to be a consistent estimator of the AMCE. This allows us to estimate the effects of profile attributes on degree of punishment through the following equation:Punishmentik=θ0+θ1Genderik+θ2Ageik+θ3Tribeik+θ4Collaborationik+εik,$$\begin{eqnarray*} Punishmen{t_{ik}} &=& {\theta _0} + {\theta _1}Gende{r_{ik}} + {\theta _2}Ag{e_{ik}}\\ && + \ {\theta _3}Trib{e_{ik}} + {\theta _4}\textit{Collaboration}_{ik} + {\varepsilon _{ik}}, \end{eqnarray*}$$where i denotes the respondent and k denotes which round of three rounds each respondent completes. Punishmentik is the outcome on the scale of least to most severe punishment. The analysis is run with robust standard errors clustered at the level of the respondent to account for within‐respondent correlation across the rounds. The error term εik refers to any random variation and, importantly, the effects of any additional determinants of preferences for punishment not accounted for in our model. The point estimates from the OLS regression are displayed in Figure 1. The dependent variable is the 5‐point scale of punishment, in which 1 is no punishment, 2 is six months of community service, 3 is three years in prison, 4 is 15 years in prison, and 5 is the death penalty.1FigureEffects of Collaborator Identity and Type of Act on PunishmentNote: The figure depicts point estimates (circles) with 95% confidence intervals (horizontal lines) and robust standard errors clustered at the individual. The circles on the vertical line at 0 denote the reference category for each attribute.Former IS fighters receive punishments that are higher than all other acts of collaboration to a statistically significant level (see Figure 1). On average, former IS taxpayers receive punishments that are 2.97 points lower than IS fighters (shown on the vertical dashed line at 0), accounting for approximately 59% of the entire 5‐point scale, with a standard error (SE) of 0.06. In other words, punishments for taxpayers were nearly three levels less harsh than for fighters, which on our 5‐point scale is the difference between 6 months of community service and capital punishment.22Additional analyses reveal that taxpayers receive significantly less harsh punishments than former collaborators who worked as cooks or janitors for IS or were married to IS fighters.We also find that civilian collaborators who were directly involved with fighters (e.g., women married to fighters and cooks for fighters) receive harsher punishments than those who did not work directly with fighters (e.g., janitors working for the IS municipality). On average, cooks receive punishments that are 0.87 points (SE = 0.05) lower than fighters—a difference of 17% of the 5‐point scale. Women married to fighters and janitors receive, respectively, punishments that are on average 1.36 (SE = 0.07, 27% of the scale) and 2.37 (SE = 0.07, 47% of the scale) points less harsh than former fighters.Contrary to expectations, respondents prefer harsher punishments for members of their own tribe (0.09 points equivalent to 2% of the scale, SE = 0.04). This finding suggests that respondents may hold members of their own tribe to a higher moral standard than members of other tribes, consistent with a theory of “ingroup policing” (Fearon and Laitin 1996). However, this effect is substantively quite small. Additionally, more lenient punishments are selected for younger collaborators (15 years old) than for older ones (35 years old) by 0.07 points (SE = 0.04), although this outcome is only significant at the p < .10 level and is substantively small. We lack support for the expectation that female collaborators would receive more lenient punishments than male collaborators, finding instead that women and men are seen as equally culpable for their actions.Identity characteristics of collaborators included in our experiment do not seem to have a substantial effect on perceptions of culpability, not even in interaction with the different acts.23There are a few exceptions, but the effects are substantively small. See the SI (pp. 6–7). These results underscore the notion that variation in the type of collaboration is an important, overlooked determinant of preferences for justice.ForgivenessWe ask a post‐treatment question for each profile to better understand the implications of variation in collaborator acts and identity characteristics on forgiveness: “Given the punishment you have selected, would you forgive this person?” This question sheds light on whether respondents are able to forgive and are potentially open to the related objectives of reconciliation and reintegration. Overall, a high proportion of respondents (59%) who did not choose the death penalty24Those who chose the death penalty, who make up 34% of the sample, were not asked this question. were willing to forgive collaboration with IS after punishment. Within this sample, 29% who were presented with profiles of IS fighters were willing to forgive them, about one‐third of those who were presented with profiles of cooks were willing to forgive them, while 42%, 72%, and 85% were willing to forgive women married to fighters, janitors for the IS municipality, and taxpayers, respectively.We employ OLS regression to analyze this outcome, coding those who chose the death penalty as not being willing to forgive collaborators.25The results are robust to logit regression. The findings are also largely robust to dropping out respondents who chose the death penalty (SI, p. 13). Compared to fighters, cooks are 15 percentage points more likely to be forgiven on a 0–1 scale (SE = 0.02). Likewise, respondents are significantly more likely to forgive wives of fighters (by 24 percentage points, SE = 0.02) and janitors working for the IS municipality (by 55 percentage points, SE = 0.02) than they are to forgive fighters. Taxpayers are 71 percentage points (SE = 0.02) more likely to be forgiven than fighters.Notably, women are less likely to be forgiven by about 3 percentage points (SE = 0.014). Although this finding is statistically significant, the effect is small in magnitude. Yet, it is in line with qualitative research documenting the intense stigma surrounding female enemy collaborators (e.g., McKay and Mazurana 2004). Age and shared tribal membership are insignificant. Figure 2 shows that the actions of former collaborators matter more than their identities for reconciliation.2FigureEffects of Identity and Act on Forgiveness of Former Enemy CollaboratorsNote: The figure depicts point estimates (circles) with 95% confidence intervals (horizontal lines) and robust standard errors clustered at the individual. The circles on the vertical line at 0 denote the reference category for each attribute.These results mirror those for the punishment outcome presented above, which suggests that severity of the act may drive both punitive preferences and willingness to forgive. If we add punishment to the forgiveness analysis, harsher punishments are negatively correlated with forgiveness, even when holding the act constant in subsamples. It is also notable that 90% of those who chose no punishment for a collaborator were willing to forgive them. These outcomes indicate that those who are already more punitive are less forgiving, and by contrast those who are less retributive are also more open to reconciliation (SI Table B4, p. 12). Further research should examine how punishment relates to forgiveness.Correlates of Punishment and ForgivenessWe employed observational analyses to explore the effects of respondent characteristics on forgiveness along with two additional preregistered hypotheses. The first hypothesis predicts that personal victimization should lead to a hardening of hearts and a desire for retribution against collaborators, which we refer to as the “revenge hypothesis.” The second hypothesis examines perceived agency behind the act of collaboration, an empirically unexplored potential mechanism explaining why different types of collaboration are met with different levels of punishment or forgiveness, which we refer to as the “volition hypothesis.” Our analyses did not find substantively significant differences among respondents of different ages, genders, and tribal identities.The Revenge HypothesisFollowing other studies on attitudes toward transitional justice mechanisms (Hall et al. 2018; Hirsch‐Hoefler et al. 2016), we examine whether personal victimization by an enemy group is related to our results. We compare respondents who had their residence seriously damaged or confiscated or had a member of their household injured or killed and hold IS responsible for at least one of these grievances (60% of the sample) with those who did not experience such harms. Moslawis with these grievances against IS are about 0.26 points (SE = 0.05, about 5% of the full 5‐point scale) harsher in their assessment of appropriate punishment; they are also 6 percentage points (SE = 0.017) less likely to forgive. Thus, we find some support for the revenge hypothesis, but the effect is substantively small.In interaction with type of act committed, our measure of personal victimization loses statistical significance; only the type of act remains a significant driver of punishment. Average marginal effects reveal that victimization is significantly associated with harsher punishments for janitors (by 0.33 points, SE = 0.13, or 9% of the scale) and less forgiveness for taxpayers by 9 percentage points (SE = 0.38; SI, p. 15).26Measuring victimization as those who experienced the death or injury of a family member reveals a 0.19‐point (SE = 0.06, 5% of the full scale) increase in harshness of punishment; this specification is not significantly correlated with forgiveness.The Volition HypothesisTable 3 explores the relationship between perceptions of different types of collaboration as voluntary and preferences for punishment and forgiveness. Respondents overwhelmingly perceived paying taxes to IS as an involuntary act (92%) in contrast with fighting for IS (3%), suggesting that more severe transgressions are associated with more culpability.3TablePerceptions of Collaboration as Voluntary (Percentage of Respondents)Type of CollaborationVoluntaryInvoluntaryAn IS fighter973Cook for IS fighters8812Married to an IS fighter8317Janitor who worked for the IS municipality7429Resident of Mosul who paid taxes to IS991Multivariate OLS regression analysis finds that volition and culpability have separate effects on punishment and forgiveness. On average, if an act is perceived as voluntary, the respondent is 0.53 points (SE = 0.08, 11% of the full scale) harsher in punishment and 20 percentage points (SE = 0.03) less likely to forgive the collaborator. Figure 3 shows the marginal effects of interactions between these two factors. The most striking result is that taxpayers receive a punishment that is 1.42 points harsher (28% of the 5‐point scale) when they are perceived as having voluntarily paid taxes compared to having done so involuntarily; voluntary tax payment is treated as harshly as involuntary participation in acts of collaboration that directly support fighters (cooks and wives). Voluntary collaborators are also 38 percentage points less likely to be forgiven than those who were perceived to be coerced. This outcome demonstrates the importance of considering perceived volition of collaboration in addition to the type of collaboration when determining preferences for retribution and reconciliation. Notably, the effect of perceived volition may vary depending on the type of collaboration.3FigurePerceptions of Voluntariness Interacted with Enemy ActNote: The figure depicts point estimates (circles and squares) with 95% confidence intervals (horizontal lines) and robust standard errors clustered at the individual. The act of collaboration is displayed on the y‐axis.DiscussionOur study develops an empirical framework for theorizing and testing the microfoundations of attitudes toward punishment and forgiveness of enemy collaborators. The evidence suggests the importance of variation in individual collaborator culpability and agency for post‐conflict transitional justice and reconciliation.The experimental design we employ is of immediate relevance to policy makers working to reestablish the rule of law and lasting peace in post‐conflict settings. Our results reveal that public opinion is on average more forgiving of IS collaborators than the harsh, one‐punishment‐fits‐all approach taken by the Iraqi government.27We find no effect of reported trust in Iraqi courts on our results. The policy implication here is that lighter punishments—including restorative, noncarceral sanctions such as community service or amnesty—should be considered for more cases than the law currently allows. On the other end of the spectrum, much of the sample still refuses to forgive the most condemnable acts of collaboration (e.g., fighters), indicating that reconciliation—even after the implementation of what respondents believed to be appropriate punishment—could be very difficult to achieve in these harder cases. Much more work is needed to understand the conditions under which, if any, violent collaborators can be accepted as rehabilitated.An important limitation of our study is that we were only able to vary a small number of identity characteristics given limited time and resources. It is possible that other social identity characteristics (e.g., religious identity) would have stronger effects on the likelihood of forgiveness and reconciliation. There is a rich debate over the role of ingroup versus outgroup dynamics in post‐conflict reconciliation that is beyond the scope of our study given the lack of diversity within our sample. That we did not find strong effects of collaborator identity characteristics on punishment demonstrates why public opinion should not be the sole basis for the design of transitional justice mechanisms; it is inhumane to hold juvenile offenders to the same standard as adults.Engaging with an active scholarly debate as to whether personal victimization at the hands of an enemy group leads to either an increased desire for revenge and retribution (Hall et al. 2018; Hirsch‐Hoefler et al. 2016) or pro‐sociality and reconciliation through a posttraumatic growth mechanism (e.g., Blattman 2009), we find some support for the “revenge hypothesis” in terms of punishment, but not forgiveness. Our study is not the first to question the substantive effects of victimization on post‐conflict attitudes (e.g., Dyrstad and Binningsbø 2019).Importantly, perceptions of volition behind collaboration are strongly associated with harshness of punishment and have an interactive effect with different types of collaboration. Multivariate regression analysis finds that volition is more strongly associated with retribution and reconciliation than victimization (SI, pp. 14–15). In‐depth, qualitative work in post‐conflict areas underscores that the perceived voluntariness of collaboration is an important factor in willingness to forgive and allow the reintegration of former enemy collaborators into their home communities (McKay and Mazurana 2004). More broadly, experimental work in social psychology suggests that intentionality intensifies the perceived damage caused by a transgression (Darley and Huff 1990). Our research fills a gap in the existing quantitative empirical literature on post‐conflict justice and reconciliation, which does not consider variation in agency behind enemy collaboration.ConclusionUnderstanding the microfoundations of preferences for punishment and forgiveness of enemy collaborators in post‐conflict societies is a necessary first step toward reconciliation. Most of the people who support and enable insurgencies, coups, and occupations are civilians, not fighters (Arjona 2016; Petersen 2001; Weinstein 2006). Thus, this research has important theoretical and substantive policy implications for the case of post‐IS Iraq and beyond.Overall, our results challenge scholars and practitioners in the field of transitional justice—where policies are often designed by elites working at the macrolevel—to turn their attention to the microlevel processes through which individuals affected by conflict form attitudes toward different types of collaborators. Including the voices of victims in the process of transitional justice policymaking may bolster the legitimacy of state institutions in contexts where historical experiences with violence, repression, or political exclusion have undermined trust in governments. Our experimental design provides a framework for testing the determinants of justice and willingness to forgive, as well as other important outcomes, including support for peace processes and reintegration of former collaborators. We hope it will spark a broader research agenda on the determinants of justice for, reconciliation with, and reintegration of enemy fighters and civilian collaborators across varied contexts.We suggest several directions for future research. First, additional factors and mechanisms that could affect our outcomes should be further explored. Petersen et al. (2012) suggest that people assess the prospective utility of reintegrating criminals to their society when determining severity of punishment and whether to forgive. Apologies or humility from transgressors has also been shown to encourage forgiveness (e.g., McCullough, Fincham, and Tsang 2003). Other work shows that prospective assessments of threat drive preferences over justice (Hirsch‐Hoefler et al. 2016). Revkin and Kao (Forthcoming) find that traditional leaders (i.e., tribal or religious) can increase acceptance of offenders in the community.Second, much remains to be learned about subnational variation in popular views of transitional justice processes. The generalizability of our findings should be tested in other areas of Iraq that differ in their demography as well as exposure to IS. In a cross‐national comparative perspective, Mosul could be thought of as a hard case for post‐conflict reconciliation because of the extreme violence that IS engaged in. Yet, within Iraq, our Sunni‐only sample of Moslawis is likely an easier case for reconciliation than one drawn from areas with significant Shia, Christian, and Yazidi populations who were severely persecuted by IS. Since our sample includes people who collaborated with IS, it may be more empathetic with collaborators. Additionally, since IS had a high degree of control over the city, there is more room for collaborators to claim they were acting under coercion. Replicating our experiment in other areas of Iraq could advance our understanding of how IS collaborators are perceived among non‐Sunnis and those who experienced differing levels of IS rule (from none to contested to complete control).Third, our design could be employed to make contributions to understanding the microdynamics of enemy collaboration and the potential for reconciliation beyond Iraq. Future studies could run our design in other settings that differ from ours in important ways: variation in regime type (e.g., democracies vs. autocracies), cultural or religious norms and legal traditions, levels of development, duration of conflict and enemy rule, and different patterns of violence and intensities of violence. Another important question is how the passage of time since the cessation of conflict affects prospects for reconciliation, which can be assessed through longitudinal data collection. Our finding that variations in perceptions of enemy culpability and agency shape attitudes toward transitional justice and reconciliation is likely to hold across all of these varying contexts. Replication of our experimental design in other settings would contribute to the development of a more comprehensive, systematic model of attitudes toward post‐conflict retribution and reconciliation.ReferencesAbdul‐Zahra, Qassim, and Susannah George. 2018. “Iraq Holding More than 19,000 Because of IS, Militant Ties.” Associated Press. https://apnews.com/article/after-the-caliphate-islamic-state-group-ap-top-news-international-news-iraq-aeece6571de54f5dba3543d91deed381Aharoni, Eyal, and Alan J. Fridlund. 2011. “Punishment without Reason: Isolating Retribution in Lay Punishment of Criminal Offenders.” Psychology, Public Policy, and Law 18(4): 599–625.Alicke, Mark D. 2000. “Culpable Control and the Psychology of Blame.” Psychological Bulletin 126(4): 556–74.Arjona, Ana. 2016. Rebelocracy: Social Order in the Colombian Civil War. Ithaca, NY: Cornell University Press.Arjona, Ana, Nelson Kasfir, and Zachariah Mampilly. 2015. Rebel Governance in Civil War. New York: Cambridge University Press.Bakke, Kristin M., John O'Loughlin, and Michael D. Ward. 2009. “Reconciliation in Conflict‐Affected Societies: Multilevel Modeling of Individual and Contextual Factors in the North Caucasus of Russia.” Annals of the Association of American Geographers 99(5): 1012–21.Blair, Graeme, Kosuke Imai, and Jason Lyall. 2014. “Comparing and Combining List and Endorsement Experiments: Evidence from Afghanistan.” American Journal of Political Science 58(4): 1043–63.Blattman, Christopher. 2009. “From Violence to Voting: War and Political Participation in Uganda.” American Political Science Review 103(2): 231–47.Boon, Susan D., and Lorne M. Sulsky. 1997. “Attributions of Blame and Forgiveness in Romantic Relationships: A Policy‐Capturing Study.” Journal of Social Behavior and Personality 12(1): 19–26.Daly, Erin. 2001. “Between Punitive and Reconstructive Justice: The Gacaca Courts in Rwanda.” NYU Journal of International Law and Politics 34(2): 355–96.Darley, John M., and Charles W. Huff. 1990. “Heightened Damage Assessment as a Result of the Intentionality of the Damage‐Causing Act.” British Journal of Social Psychology 29(2): 181–88.Davenport, Christian, and Molly Inman. 2012. “The State of State Repression Research since the 1990s.” Terrorism and Political Violence 24(4): 619–34.David, Roman. 2014. “International Criminal Tribunals and the Perception of Justice: The Effect of the ICTY in Croatia.” International Journal of Transitional Justice 8(3): 476–95.David, Roman, and Susanne Choi Yuk‐Ping. 2005. “Victims on Transitional Justice: Lessons from the Reparation of Human Rights Abuses in the Czech Republic.” Human Rights Quarterly 27(2): 392–435.de Greiff, Pablo. 2014. “Report of the Special Rapporteur on the Promotion of Truth, Justice, Reparation and Guarantees of Non‐Recurrence.” United Nations General Assembly, Human Rights Council, A/HRC/27/56.Dyrstad, Karin, and Helga Malmin Binningsbø. 2019. “Between Punishment and Impunity: Public Support for Reactions Against Perpetrators in Guatemala, Nepal and Northern Ireland.” International Journal of Transitional Justice 13(1): 155–84.Enright, Robert D. 1991. “The Moral Development of Forgiveness.” Handbook of Moral Behavior and Development, Vol. 1, eds. William M. Kurtines and Jacob L. Gewirtz. Lawrence Erlbaum Associates, Inc.: New Jersey, US. 123–52.Fabbe, Kristin, Chad Hazlett, and Tolga Smmazdemir. 2019. “A Persuasive Peace: Syrian Refugees’ Attitudes towards Compromise and Civil War Termination.” Journal of Peace Research 56(1): 103–17.Fearon, James, and David Laitin. 1996. “Explaining Interethnic Cooperation.” American Political Science Review 90(4): 715–35.Gibson, James L., and Amanda Gouws. 1999. “Truth and Reconciliation in South Africa: Attributions of Blame and the Struggle over Apartheid.” American Political Science Review 93(3): 501–17.Hainmueller, Jens, Daniel Hopkins, and Teppei Yamamoto. 2013. “Causal Inference in Conjoint Analysis: Understanding Multidimensional Choices via Stated Preference Experiments.” Political Analysis 22(1): 1–30.Hall, Jonathan, Iosif Kovras, Djordje Stefanovic, and Neophytos Loizides. 2018. “Exposure to Violence and Attitudes towards Transitional Justice.” Political Psychology 39(2): 345–63.Hewstone, Miles, Mark Rubin, and Hazel Willis. 2002. “Intergroup Bias.” Annual Review of Psychology 53(1): 575–604.Hill, E. Wayne. 2001. “Understanding Forgiveness as Discovery: Implications for Marital and Family Therapy.” Contemporary Family Therapy 23(4): 369–84.Hirsch‐Hoefler, Sivan, Daphna Canetti, Carmit Rapaport, and Stevan E. Hobfoll. 2016. “Conflict Will Harden Your Heart: Exposure to Violence, Psychological Distress, and Peace Barriers in Israel and Palestine.” British Journal of Political Science 46(4): 845–59.Honey, P. Lynne. 2017. The Oxford Handbook of Women and Competition. Oxford: Oxford University Press.Human Rights Watch. 2017. “Flawed Justice: Accountability for ISIS Crimes in Iraq.” https://www.hrw.org/sites/default/files/report_pdf/iraq1217web.pdfHumphreys, Macartan, and Jeremy Weinstein. 2007. “Demobilization and Reintegration.” Journal of Conflict Resolution 51(4): 531–67.Jones, Nicholas A., Stephan Parmentier, and Elmar G. M. Weitekamp. 2012. “Dealing with International Crimes in Post‐War Bosnia: A Look through the Lens of the Affected Population.” European Journal of Criminology 9(5): 553–64.Kalyvas, Stathis. 2006. The Logic of Violence in Civil War. New York: Cambridge University Press.Kalyvas, Stathis N., and Laia Balcells. 2010. “International System and Technologies of Rebellion: How the End of the Cold War Shaped Internal Conflict.” American Political Science Review 104(3): 415–29.Kasfir, Nelson. 2005. “Guerrillas and Civilian Participation: The National Resistance Army in Uganda, 1981–86.” Journal of Modern African Studies 43(2): 271–96.Latimer, Jeff, Craig Dowden, and Danielle Muise. 2005. “The Effectiveness of Restorative Justice Practices: A Meta‐Analysis.” The Prison Journal 85(2): 127–44.Lyall, Jason, Graeme Blair, and Kosuke Imai. 2013. “Explaining Support for Combatants during Wartime: A Survey Experiment in Afghanistan.” American Political Science Review 107(4): 679–705.Mampilly, Zachariah. 2011. Rebel Rulers: Insurgent Governance and Civilian Life during War. Ithaca, NY: Cornell University Press.McCullough, Michael E., Frank D. Fincham, and Jo‐Ann Tsang. 2003. “Forgiveness, Forbearance, and Time: The Temporal Unfolding of Transgression‐Related Interpersonal Motivations.” Journal of Personality and Social Psychology 84(3): 540–57.McKay, Susan, and Dyan E. Mazurana. 2004. Where Are the Girls? Girls in Fighting Forces in Northern Uganda, Sierra Leone and Mozambique: Their Lives during and after War. International Center for Human Rights and Democratic Development, Montreal, Canada.Olson, Mancur. 1993. “Dictatorship, Democracy, and Development.” American Political Science Review 87(3), 567–76.Petersen, Michael Bang, Aaron Sell, John Tooby, and Leda Cosmides. 2012. “To Punish or Repair? Evolutionary Psychology and Lay Intuitions about Modern Criminal Justice.” Evolution and Human Behavior 33(6): 682–95.Petersen, Roger D. 2001. Resistance and Rebellion: Lessons from Eastern Europe. Cambridge: Cambridge University Press.Revkin, Mara. 2018. “After the Islamic State: Balancing Accountability and Reconciliation in Iraq.” New York: United Nations University Centre for Policy Research.Revkin, Mara Redlich. 2020a. “Competitive Governance and Displacement Decisions under Rebel Rule: Evidence from the Islamic State in Iraq.” Journal of Conflict Resolution 65(1): 46–80.Revkin, Mara Redlich. 2020b. “What Explains Taxation by Resource‐Rich Rebels? Evidence from the Islamic State in Syria.” Journal of Politics 82(2): 757–64.Revkin, Mara Redlich, and Elisabeth Wood. 2020. “The Islamic State's Pattern of Sexual Violence: Ideology and Institutions, Policies and Practices.” Journal of Global Security Studies 6(2): 1–20.Revkin, Mara, and Kristen Kao. In Press. “No Peace Without Punishment? Reintegrating Islamic State Collaborators in Iraq.” The American Journal of Comparative Law.Samii, Cyrus. 2013. “Who Wants to Forgive and Forget? Transitional Justice Preferences in Postwar Burundi.” Journal of Peace Research 50(2): 219–33.Scott, Elizabeth S., N. Dickon Reppucci, Jill Antonishak, and Jennifer T. DeGennaro. 2006. “Public Attitudes about the Culpability and Punishment of Young Offenders.” Behavioral Sciences & the Law 24(6): 815–32.Shaw, Rosalind, Lars Waldorf, and Pierre Hazan. 2010. Localizing Transitional Justice: Interventions and Priorities after Mass Violence. Stanford, CA: Stanford University Press.Sly, Liz. 2015. “The Hidden Hand behind the Islamic State Militants? Saddam Hussein's.” Washington Post.Stewart, Megan. 2018. “Civil War as State‐Making: Strategic Governance in Civil War.” International Organization 72(1): 205–26.Stylianou, Stelios. 2003. “Measuring Crime Seriousness Perceptions: What Have We Learned and What Else Do We Want to Know?” Journal of Criminal Justice 31(1): 37–56.Tabak, Shana. 2011. “False Dichotomies of Transitional Justice: Gender, Conflict and Combatants in Colombia.” New York University Journal of International Law and Politics 44(1): 103–63.Tellez, Juan Fernando. 2019. “Worlds Apart: Conflict Exposure and Preferences for Peace.” Journal of Conflict Resolution 63(4): 1053–76.Terpstra, Niels. 2020. “Rebel Governance, Rebel Legitimacy, and External Intervention: Assessing Three Phases of Taliban Rule in Afghanistan.” Small Wars & Insurgencies 31(6): 1143–73.Tyler, Tom R. 2003. “Procedural Justice, Legitimacy, and the Effective Rule of Law.” Crime and Justice 30: 283–357.United Nations. 2006. Handbook on Restorative Justice Programmes. New York: United Nations Office on Drugs and Crime.Verini, James. 2016. “Surviving the Fall of ISIS.” National Geographic. https://www.nationalgeographic.com/culture/article/islamic-state-isis-iraq-mosul-syria-offensiveWeinstein, Jeremy. 2006. Inside Rebellion: The Politics of Insurgent Violence. Cambridge: Cambridge University Press.Wistrich, Robert S. 2013. Who's Who in Nazi Germany. New York: Routledge.Wood, Elisabeth. 2006. “The Ethical Challenges of Field Research in Conflict Zones.” Qualitative Sociology 29(3): 373–86.Wood, Elisabeth Jean. 2003. Insurgent Collective Action and Civil War in El Salvador. Cambridge: Cambridge University Press.Worthington, Everett L., Jr. 2006. Forgiveness and Reconciliation: Theory and Application. New York: Routledge.

Journal

American Journal of Political Science – Wiley

Published: Apr 1, 2023

There are no references for this article.