Transitional Justice, Education, and Sexual Violence Stigma: The Results of a Schools‐Based Study in Bosnia‐Herzegovina

Within the context of armed conflict, the problem of sexual violence&#8208;related stigma is routinely acknowledged. Sustained efforts to tackle it, however, have often been lacking. This article argues that transitional justice processes have an important role to play in fighting stigma, and in creating new attitudinal spaces that enable those who have suffered sexual violence to tell their stories without fear of being mocked or judged. Underscoring crucial linkages between education and transitional justice, the article introduces a novel schools&#8208;based project in Bosnia&#8208;Herzegovina (BiH). Based on analysis of over 800 questionnaires, it demonstrates how a series of talks delivered in BiH high schools on the issue of conflict&#8208;related sexual violence led to some positive and important attitudinal changes vis&#8208;&#224;&#8208;vis common rape myths. These educational efforts to tackle stigma are theorized in the article as highlighting a crucial attitudinal dimension of transitional justice.


INTRODUCTION
Speaking in June 2016 at an event to mark the International Day for the Elimination of Sexual Violence in Conflict, the then United Nations (UN) Deputy Secretary-General, Jan Eliasson, underlined that`Sexual violence is unique in often stigmatizing the victim, rather than the perpetrator of the 509 for transitional justice in tackling sexual violence stigma. Part of this role involves fostering the development of facilitative ecologies ± a term used in this article to denote social attitudinal spaces wherein victims/survivors feel supported and able to tell their stories without fear of being judged, criticized or marginalized. If the concept of facilitative ecologies thus foregrounds a major attitudinal dimension of transitional justice work, it also highlights crucial linkages ± which to date remain underexplored ± between education and transitional justice. The empirical part of the article presents the results of the author's ongoing impact research in high schools in Bosnia-Herzegovina (BiH). By engaging young people on the issue of sexual violence, the aim of this project ± which can be viewed as an example of attitudinal transitional justice ± is to challenge common rape myths that feed social stigma. To focus on schools alone is not sufficient to create facilitative ecologies, but it is a significant first step.
The article is divided into four sections. The first section focuses on one of the central goals of transitional justice, namely, giving victims 12 a voice. It argues that notwithstanding important developments ± particularly in the international prosecution of sexual violence ± aimed at giving victims a voice, these developments will remain limited as long as the wider social attitudinal environment is left unaltered. The second section builds the argument that transitional justice processes should give more attention to the wider social context, in order to transform social ecologies and enhance thè voice' that victims/survivors have. Schools represent one part of these social ecologies, and the third section introduces the author's continuing work in BiH high schools aimed at fighting sexual violence stigma. The final section explores the impact of this work. Analysing the data from over 800 questionnaires, it demonstrates that the interactive school talks on sexual violence did bring about some positive attitudinal changes. These results underscore that transitional justice processes should not only be about dealing with the past, but also about giving more attention to negative social attitudes that can exacerbate the suffering of victims/survivors in the present.

GIVING VICTIMS A VOICE
As the burgeoning field of transitional justice continues to grow and expand, scholars increasingly disagree on a number of core issues. For example, where do the boundaries of transitional justice lie; what should transitional justice encompass; are we expecting too much from it; what is the 512 12 Although this article refers to individuals who have suffered sexual violence as victims/survivors', it uses the term`victims' in the context of transitional justice. This is consistent with standard transitional justice terminology, which primarily speaks about`victims' rather than`survivors'. relationship between transitional justice and peacebuilding? 13 Regardless of how broadly or narrowly the parameters of transitional justice are drawn, a commonly-stated goal of transitional justice ± and in particular of truth and reconciliation commissions (TRCs) and criminal prosecutions ± is to give victims a voice. 14 Citing from the final report of South Africa's TRC during the 1990s, for example, Verdoolaege notes that`The emphasis of the HRV [human rights violations] hearings was on``the validation of the individual subjective experiences of people who had previously been silenced or voiceless''.' 15 The website of the International Criminal Tribunal for the former Yugoslavia (ICTY), which recently completed its mandate, underlines that the Tribunal's work`provided thousands of victims the opportunity to be heard and to speak about their suffering'. 16 At an event to commemorate the twentieth anniversary of the adoption of the Rome Statute, Herman von Hebel, the then Registrar of the International Criminal Court (ICC), emphasized the`need to ensure that the voice of victims will continue to be heard and to be increasingly heard.' 17 Within the judicial framework of the ICC and the Extraordinary Chambers in the Courts of Cambodia (ECCC), there is provision for victims not only to testify but also to participate in trial proceedings. 18   risk that within the courtroom victim participation is a token effort, rather than a genuine representation of voice and legal agency.' 22 In general, and not only in the particular context of criminal trials, the extent to which victims' voices are actually heard ± and the potential impact of these voices ± will depend on multiple factors, from the prevailing political context and the level of regime change to simple practicalities and the country-specific operationalization of individual or combined transitional justice mechanisms. Discussing Canada's Indian Residential Schools TRC (2008±2015), for example, James expresses`concerns about how a victimcentred approach functions in a sociopolitical context dominated by the institutional perpetrators and individual beneficiaries of the injustices.' 23 Particular voices, moreover, may be marginalized, a point which feminist scholars have frequently stressed. 24 Reflecting on the work of the South African TRC (1994±1998), for example, Ross notes that: . . . the emphasis on spectacular events of violence ± abduction, torture, killing, and severe ill-treatment ± forecloses an understanding of the subtleties of oppression and the diverse ways in which apartheid affected different categories of people. 25 Women's experiences were frequently silenced. 26 Even when victims are given a platform, the emphasis on public remembering and recounting may conflict with local preferences and ways of dealing with the past, as Shaw found during her work with war victims in Sierra Leone. In her words,`The friction of a globalized universal traveling through a particular place and time may enable wheels to turn, but sometimes those wheels may get stuck.' 27 The above examples highlight that the aim of giving victims a voice is far from straightforward. When one takes a step back and observes the wider picture, however, it is unquestionable that important progress has been made. Emphasizing this point, Karstedt  rights regime, in international courts, tribunals or in Truth and Reconciliation Commissions, makes us aware of the distance travelled and of achievements gained. 28 Of particular significance are the developments that have occurred in relation to rape and other forms of sexual violence ± crimes that have historically been marginalized. Regarding the post-war Nuremberg trial of 22 leading Nazi war criminals, for example, McDonald notes that`the French prosecutor specifically asked forgiveness from the tribunal for his decision to avoid citing atrocious details of crimes of sexual violence, although he had no problem reciting atrocious details of other war crimes.' 29 Moreover, although the voluminous trial transcripts contained extensive evidence of sexual violence, this was not reflected in the 732-page index that accompanied them:`Neither``rape'' nor``women'' is included in any heading or subheading in this index.' 30 The word rape was similarly absent from the Tribunal's judgment. 31 The International Military Tribunal for the Far East fared considerably better in this regard. For example,`the Tokyo Indictment did include allegations of gender-related crimes', 32 and its judgment recorded the mass rapes of approximately 20,000 Chinese women in Nanking in 1937. 33 However, the Tokyo Tribunal did not give any of these women the opportunity to testify, and it critically ignored the thousands of`Comfort Women' who were forced into sexual slavery to serve the needs of the Japanese imperial army. 34 If women's bodies were traditionally framed as war booty, 35 over the last two decades there have been major strides forward in the international prosecution of sexual violence. At the International Criminal Tribunal for Rwanda (ICTR), for example, the trial of the former mayor of Taba commune, Jean-Paul Akayesu, was not only the first case to define rape in 516 international law, 36 but also`the first case in history that rendered a conviction for either genocide or crimes against humanity through sexual violence.' 37 At the ICTY, the Trial Chamber in Delalic Â et al. established that rape can constitute an act of torture; 38 the Furundz Ïija trial was the first to focus solely on sexual violence; 39 and in the Kunarac et al. trial, the definition of enslavement as a crime against humanity was broadened to include sexual enslavement. 40 Several cases have drawn much-needed attention to the use of conflict-related sexual violence against men, 41 and successful convictions in cases such as the Charles Taylor trial at the Special Court for Sierra Leone (SCSL) 42  general agreement with the factors put forward by the Prosecutor, to be taken into consideration in determining whether enslavement was committed. These are the control of someone's movement, control of physical environment, psychological control, measures taken to prevent or deter escape, force, threat of force or coercion, duration, assertion of exclusivity, subjection to cruel treatment and abuse, control of sexuality and forced labour. . This was a highly significant judgment, not least because Bemba was the first defendant to be These important developments in the prosecution of sexual violence have been accompanied by various procedural efforts aimed at facilitating the testimony process. Article 68(2) of the Rome Statute, for example, states that: As an exception to the principle of public hearings provided for in Article 67, the Chambers of the Court may, to protect victims and witnesses or an accused, conduct any part of the proceedings in camera or allow the presentation of evidence by electronic or other special means. 45 It adds that these protective measures`shall be implemented' in cases involving victims of sexual violence or child victims/witnesses, unless the Court ± having regard to all the circumstances and, in particular, the views of the victim or witness ± orders otherwise. 46 While acknowledging these advances, some scholars underline that criminal trials, by their very nature, accord only a limited voice to those who have suffered sexual violence. Drawing on their experiences at the ICTY, for example, Gopalan et al. recall that`It was frustrating for some when their only experience of the courtroom was a harsh cross-examination by the defence without any opportunity to tell their story in their own words first.' 47 Making a similar point, Henry distinguishes between two types of testimony: one that focuses on the facts of what happened (the description of the rape) and the other that centres on the emotional aftermath of rape and its impact. While criminal trial processes call for the first type of narrative, this is probably not the story that victims would want, or feel comfortable, to 518 convicted of rape at the ICC: J.N. Clark,`The First Rape Conviction at the ICC: An Analysis of the Bemba Judgment' (2016) 14 J. of International Criminal Justice 667. He was subsequently sentenced to 18 years' imprisonment. On 18 June 2018, however, the Appeals Chamber reversed the judgment and acquitted Bemba. It found that the Trial Chamber had made a number of`serious errors' in assessing whether Bemba, as a commander, had taken all necessary and reasonable measures to prevent or repress the commission of crimes by his subordinates. Among these errors, the Appeals Chamber found that`the Trial Chamber erred by failing to properly appreciate the limitations that Mr Bemba would have faced in investigating and prosecuting crimes as a remote commander sending troops to a foreign country': tell'. 48 It is important to stress, however, that although the rigours of the trial process undoubtedly constrain the types of narrative that are permissible in a court of law, 49 this is not the whole picture. Whether they occur in the formal setting of a courtroom or in a more community-based forum, testimony and story telling take place within a wider context. 50 To date, transitional justice practice has critically neglected these social ecologies ± and the way in which they impact on the narratives and stories articulated by victims/ survivors of sexual violence.

THE MISSING LINK: SOCIAL ECOLOGIES AND TRANSITIONAL JUSTICE
Social ecology refers to the interaction and interlocking of different micro, meso, exo, and macro systems. Embodying the fundamental idea that we are part of the environment that surrounds us, 51 it entails`investigations that go beyond the immediate setting containing the person to examine the larger contexts, both formal and informal, that affect events within the immediate setting'. 52 From a social-ecological perspective, therefore, it is artificial to look at individuals and individual behaviour in isolation. They are necessarily part of a wider whole and of`[n]esting interacting spheres of social relationships . . .'. 53 The concept of social ecology has been utilized and discussed in many different contexts, including health, 54 bullying, 55 resilience, 56 and crime. 57

519
To date, however, it remains strikingly underexplored in relation to transitional justice. This is surprising because these social ecologies can significantly affect how individuals deal with experiences of trauma. Highlighting this point, Herman notes that:`Because traumatic life events invariably cause damage to relationships, people in the survivor's world have the power to influence the eventual outcome of the trauma.' 58 By extension, and relatedly, social ecologies necessarily influence processes of story telling and narrative, shaping what can and cannot be said ± and how it is said. As Porter underlines,`Narratives are made up of a series of stories that are told within social, cultural, and political contexts.' 59 These contexts, which necessarily frame any transitional justice process, are particularly important in relation to story telling involving sexual violence. According to Caruth: trauma . . . is always the story of a wound that cries out, that addresses us in the attempt to tell us of a reality or truth that is not otherwise available. This truth, in its delayed appearance and its belated address, cannot be linked only to what is known, but also to what remains unknown in our very actions and our language. 60 Yet, if victims/survivors live in an environment where they are judged, blamed, and stigmatized, the scope for their wounds to`cry out' may be restricted. 61  (and particularly other women) who readily cast blame. 63 In one case, an interviewee's neighbour had accused her of having willingly`slept' with enemy soldiers. 64 Another interviewee explained that her closest neighbour, and former friend, had told her that because she chose to stay in the area when the war broke out in 1992, it was her own fault that she was raped. 65 These were not isolated cases, 66 and interviewees frequently expressed feeling unsupported. Indeed, one of the major challenges of working with victims/ survivors of conflict-related sexual violence in BiH is precisely the fact that some of them are afraid to speak out, fearing negative reactions from others. To cite Hackett and Rolston,`At the individual level, trauma and fear may lead to silence rather than speech, while at the social level there may not be spaces in which stories can be told and listened to sympathetically.' 67 Transitional justice has a crucial role to play in fostering these spaces through the development of facilitative ecologies. The concept of paratranslation, used in translation studies, is useful for illustrating this point. Paratranslation stresses`the essential role performed by paratextual elements in translation, that is, their participation, together with text, in the construction of meaning of the published work.' 68 In an interview, the words spoken constitute the main text. Everything else surrounding the interview forms the paratext.
Extending the analogy, societal attitudes about sexual violence, and the existence of stigma, constitute a crucial paratext that interlocks with and influences the main text ± that is, the narratives of victims/survivors. Transitional justice is overwhelmingly concerned with the main text. It focuses on victims'/survivors' individual stories; what happened to them, when, how, who hurt them? What it neglects is the broader paratext ± and the fact that individuals`construct narratives from culturally availablè`n arrative templates'',``public narratives'' or``meta-narratives'' . . .'. 69 521 63 According to Hayes et al., part of the explanation for victim blaming lies in just world belief (JWB) and, more specifically,`just world-other' ± the conviction that people only get what they deserve. When this belief is shaken, people look for ways to re-stabilize it.`To maintain one's JWB, it may be natural for people to neutralize or deny the occurrence of injustice, or even hold the victim accountable for their misfortune in order to restore one's belief in a just world-other': R.M. Hayes, K. The central point is that if transitional justice processes overlook the wider social context in which victims/survivors narrate and tell their stories, core transitional justice goals ± including truth telling and giving victims a voice ± may be compromised. The focus of transitional justice thus needs to extend beyond victims and perpetrators. In addition to looking at the suffering and actions of individuals, it also needs to address the fact that`Traumatic events destroy the sustaining bonds between individual and community'. 70 More specifically, it needs to help to restore what Laub has called an`empathic dyad' 71 between victims/survivors and those around them. For this to happen, it is essential that transitional justice processes tackle wider societal attitudes. The ICTY, for example, which completed its mandate at the end of 2017, insisted that its judgments`contributed to . . . combatting denial and preventing attempts at revisionism'. 72 In the absence of concerted efforts to tackle the social-attitudinal context that sustains denial and revisionism, however, such claims remain unconvincing. As one illustration, reactions in Croatia and BiH to the courtroom suicide of Slobodan Praljak 73 ± in November 2017 ± showed that this former Bosnian Croat general and convicted war criminal was highly respected and admired. To his supporters, he was a war hero and a victim of a huge injustice. 74 His crimes and wrongdoings were conveniently overlooked. 75 Similarly, many convicted war criminals continue to enjoy widespread backing and adulation from within their own ethnic group. 76 This spotlights a critical gap within transitional justice work ± both in BiH and more generally ± that stems precisely from a failure to target social attitudes. 522 This imperative`attitudinal dimension', in turn, foregrounds an important role for education in transitional justice practice. Ramõ Ârez-Barat and Duthie point out that`. . . to date neither education reform nor the teaching of the recent past has been treated with the seriousness it deserves within the scope of transitional justice.' 77 Seeking to redress this, in October 2016, I commenced a research project focused on high schools in BiH. Through interactive discussions with young people on the issue of sexual violence, the aim was to directly confront social attitudes and prejudices that foster stigma, and to thereby contribute to developing new narrative spaces within a wider framework of facilitative ecologies.

ADDRESSING SEXUAL VIOLENCE STIGMA IN BIH
During my work with male and female victims/survivors in BiH in 2014 and 2015, the issue of social stigma emerged as a recurring theme. Despite this, the problem has largely been left unaddressed, within both transitional justice processes and the civil society sector. Non-governmental organizations (NGOs) that work with victims/survivors have repeatedly called for justice'. According to Bakira Hasec Ïic Â, for example, the president of the organization Z Ï ena ± Z Ï rtva Rata who was herself raped during the Bosnian war,`The most important thing is the war crimes verdict; we want to fight for justice, and the only justice is that war criminals are arrested and sentenced.' 78 More recently, there has been a growing emphasis on the need for reparations and, more specifically, compensation. A lawyer from TRIAL International, for example, has underlined that`The symbolic message of a compensation award is strong: it is a societal recognition of what she 79 [the victim/survivor] went through. In that sense, it also holds strong preventive potential.' 80 Supporting victims/survivors, however, also requires more systemic change, and specifically attitudinal change. Perceiving a critical gap that needed to be addressed, I sought to explore ways of harnessing education to tackle sexual violence stigma, by piloting a series of interactive discussions in high schools across BiH. If schools form part of a larger social-ecological system, 81 they are also ± and relatedly ±`a place for prevention and changing social norms'. 82 Young people in BiH high schools did not directly experience the war. However, a myriad of factors ± from political rhetoric and media stories to the ongoing discovery of mass graves and commemorations of massacres ± all contribute to keeping the past alive. The war, moreover, has indirectly touched BiH youth in different ways. They might have lost members of their family; 83 their parents may have subconsciously`passed on' their own trauma through their words, fears, and beliefs; 84 and many post-war stressors still exist. According to research by Al-Sabah et al., for example: Student focus group members described the persisting influences of the war on their lives in a variety of ways . . . including war-related trauma reminders, loss reminders, and secondary adversities . . . that impaired their concentration at home and school. 85 Another legacy of the war that directly affects young people is the education system, which is deeply fragmented. Multiple curricula, the persistence of`two schools under one roof', 86 and the fact that students learn different versions of history and religion, depending on their ethnicity, all contribute to fostering ethnic divides. 87 In this climate, speaking to young people about the use of sexual violence during the Bosnian war ± which remains a highly-charged topic ± could potentially create new tensions. To avoid this, and to shift the focus away from BiH, the project was purposefully designed around the global issue of conflict-related sexual violence. The aim was to create a`neutral' space in which students could reflect on and discuss a range of questions pertaining to sexual violence, 524 thereby gaining new knowledge and understanding that they could ultimately apply to the local context. Previous initiatives, in particular, Education for Peace, 88 have highlighted the transformative potential of education in BiH. 89 This project, which like Education for Peace was`designed to be both universal and specific', 90 can itself be viewed as a form of peace education.
Having previously worked closely with Snaga Z Ï ene, 91 a women's NGO in Tuzla, I trained the organization to deliver a series of interactive talks on conflict-related sexual violence in high schools across BiH. 92 At the start of each talk, students were shown two short video clips ± the first of a female survivor and activist, Jineth Bedoya Lima, in Colombia, 93 and the second of a teenage boy in the Democratic Republic of Congo who suffered anal rape. As an ice-breaker, students were asked to describe how they felt as they watched the clips.`Sad',`angry',`powerless' were the most commonly-used words. As the discussion progressed, they were asked to reflect on a range of questions, including: what is sexual violence, why are rape and other forms of sexual violence used as weapons of war, how do these crimes affect human lives, and are those who directly experience sexual violence themselves at fault. Challenging what Sjoberg has termed`gender normativity', 94 the talks underlined that fact that sexual violence affects men as well as women, and that the perpetrators of these crimes are not exclusively male. 95 The students who participated in the talks were requested to complete a pre-and post-talk questionnaire. They were asked some general questions about the use of sexual violence during the Bosnian war (for example, do they feel comfortable speaking about the topic, and is it an issue that BiH society should openly speak about) and in conflict situations more generally. The main part of the pre-talk questionnaire focused on gender and society, and students were asked to indicate their level of agreement or disagreement with the following six statements:`I think that if a man is insulted, he should be prepared to physically fight to defend his honour'; 96`I think that if a wife does something wrong, her husband has the right to punish her';`I think there are situations when a woman deserves to be beaten';`I think that if a woman dresses provocatively or gets drunk, she is inviting men to rape her'; I think that when a woman is raped, she is usually to blame for putting herself in that situation'; and`A real man would not allow himself to be raped'. Consistent with the project's social-ecological emphasis, students were asked not only what they themselves think in relation to these statements, but also what their friends and family think. This was a way of gaining insights into the provenance of their views. In the post-talk questionnaire, students were asked the same set of statements. They were also asked to reflect on the talks; were they useful, would they like more talks on sexual violence in school, did they rethink any of their views? For the above-mentioned statements, a four-point Likert scale was used and students were given the following response options:`Strongly agree', Agree',`Disagree', and`Strongly disagree'. It is important to note that scholars continue to debate what constitutes an`optimal' Likert scale. As Adelson and McCoach point out: Some have argued that including a middle category allows respondents to indicate a neutral response and be more discriminating in their response, making the scale scores more reliable and the scale more preferred by subjects . . . 97 Conversely: . . . others have expressed concern that with a middle category respondents will be less discriminating and declare themselves neutral more often, while omitting the neutral point forces respondents to be more thoughtful, resulting in more precise ratings. 98 Leaning towards the second view, I opted to use a four-point Likert scale rather than a five-point or seven-point version. For the purpose of analysing the data, a four-point scale was also used, running from 1 = strongly agree to 4 = strongly disagree.
In the first stage of the project, interactive talks were given in seven high schools in Tuzla Canton (five in Tuzla, one in Srebrenik and one in Z Ï ivinice) in November and December 2016. From these seven schools, 278 students (177 female students and 101 male students) completed a pre-and post-talk questionnaire. The data from these schools will be referred to as the Tuzla 526 96 It should be noted that because the talks did not address issues around male honour, this particular statement was only included in the pre-talk questionnaire but not in the post-talk questionnaire. Canton dataset. In the second stage of the project, between January and May 2017, interactive talks took place in a further 14 schools throughout BiH. Six of these schools were in the BiH Federation (Busovac Ïa, 99 Mostar, Odz Ïak, Vares Ï, and Vitez), four were in Republika Srpska (East Sarajevo, Foc Ïa, Srebrenica, and Vis Ïegrad) and four were in Brc Ïko District. From across these 14 schools, 529 students (267 female students and 262 male students) completed questionnaires. These 529 questionnaires will be referred to as the main dataset.
All of the 807 students (444 female and 363 male) who took part in the talks were in their final year of high school and the majority of them were 18 years old. A small number of them were 17 years old. They were randomly selected to participate in the project and all of them were given a project information sheet and an informed consent form to complete. Each school talk lasted approximately one hour and the talks were given to two groups of students in each school. On average, there were 20 students in each group. All of the groups were mixed gender, with the exception of two of the groups in Tuzla where all of the students ± due to the gender composition of the school ± were male. The majority of the talks were delivered by the head of Snaga Z Ï ene, a practising doctor. Two psychologists from the NGO were also on standby, to speak to any students who became upset or distressed. Additionally, Snaga Z Ï ene's contact details were provided on the participant information sheet.
As anticipated, there were some important gender differences in the students' answers to the questionnaire. 100 In the main dataset (529 students), for example, the statement`If a man is insulted, he should be prepared to physically fight to defend his honour' elicited significantly different 527 99 In Busovac Ïa, talks took place in two schools that were officially`two schools under one roof'. The Bosniak and Bosnian Croat students in these two schools were following different curricula. 100 In contrast, students generally gave very similar answers regardless of their ethnicity (Bosniak, Bosnian Serb or Bosnian Croat) and there were few regional or interentity differences (that is, differences between students in the BiH Federation and those in Republika Srpska, or in Brc Ïko District). Three exceptions can be noted. First, regarding the statement`If a wife does something wrong, her husband has the right to punish her', 42 per cent of all students in the main data set strongly disagreed. However, 10 per cent more students from Republika Srpska (RS) strongly disagreed compared to students from the Federation and Brc Ïko District. Secondly, in response to the question`Should Bosnian society as a whole talk more about the rapes and sexual violence committed during the Bosnian war', the majority of students (47 per cent) selected yes, 13 per cent said no and 40 per cent did not know. The majority of students from RS answered`I don't know' and a higher proportion said no compared to students in the Federation and Brc Ïko District. Thirdly, asked whether their views had changed as a result of the project, 35 per cent of students in total answered in the affirmative. A slightly higher proportion of students from RS (38 per cent) answered yes compared to students from the Federation (34 per cent) or Brc Ïko District (32 per cent). responses from male students (M 2.89, SD 0.85) and female students (M 3.45, SD 0.72), t(526) À8.1, p `0X001). Only 2 per cent of female students and 5 per cent of male students strongly agreed with the statement. A higher percentage of female and male students ± 7 per cent and 27 per cent respectively ± agreed with it. Interestingly, a lower percentage of female students (33 per cent) disagreed with the statement than male students (42 per cent). However, while 57 per cent of female students strongly disagreed with the statement, only 26 per cent of male students did so. In the Tuzla Canton dataset, there was also a significant ± and more pronounced ± difference between the responses of male students (M 2.75, SD 0.93) and female students (M 3.46, SD 0.64), t(267) À7.41, p `0X001.
In some cases, however, the gender differences were less marked than anticipated. Vis-a Á-vis the statement`When a woman is raped, she is usually to blame for putting herself in that situation', the average response in the pretalk questionnaires was significantly different between male students (M 3.43, SD 0.69) and female students (M 3.67, SD 0.6), t(523) À4.3, p `0X001. In the post-talk questionnaires, similarly, the average responses of male students (M 3.44, SD 0.68) and female students significantly differed (M 3.72, SD 0.52), t(502) À5.1, p `0X001. Yet, while there was a skew among female students towards strongly disagree in both the pre-and post-talk questionnaires, the difference between male and female responses was less prominent than for the previous statement about male honour. In the Tuzla Canton dataset, there were also significant differences between the average responses of male and female students in relation to this particular statement. These differences were observed in the pre-test measure (male: M 3.4, SD 0.66; female: M 3.76, SD 0.48), t(269) À5, p `0X001, as well as in the post-test measure (male: M 3.3, SD 0.57; female: M 3.79, SD 0.44), t(264) À7, p `0X001. Although these gender differences were more significant than those in the main dataset, they were smaller than those seen (in both datasets) vis-a Á-vis the statement about male honour.
Nevertheless, a small proportion of female respondents did still express views supportive of victim blaming. In the main data set, for example, 11 per cent of female students (compared with 13 per cent of male students) agreed with the statement`If a woman dresses provocatively or gets drunk, she is inviting men to rape her'. A further 3 per cent strongly agreed (5 per cent of male students did so). In the Tuzla Canton dataset, the percentage of female students agreeing with this statement was slightly higher at 15 per cent (only 2 per cent strongly agreed), although more male students (24 per cent) also agreed (and 9 per cent strongly disagreed). According to Cowan,`The women who have a tendency to victim blame in situations in which women are harassed and raped appear to have a predisposition to dislike and distrust women.' 101 While this article does not embrace such a sweeping statement, the fundamental point is that efforts to tackle sexual violence stigma in BiH need to be gender inclusive and to avoid essentializing`men' as the ultimate problem. CARE International's work with young men in the Western Balkans is a case in point. While its`Young Men Initiative' is both important and necessary, the strong association that it makes between men and violence 102 means that young women ± and their own negative social attitudes ± are overlooked.
Looking at wider social attitudes, the analysis revealed that responses tò My family think . . .' were generally aligned to responses for`I think . . .'. In the main dataset, this was the case for all the statements with the exception of`There are times where a woman deserves to be beaten' (`My family thinks . . .', M 3.7;`I think . . .', M 3.6). In contrast, respondents often perceived their friends as less likely to disagree strongly with each of the gender and society statements. The biggest difference in mean scores was in relation to the statement`If a man is insulted, he should be prepared to physically fight to defend his honour'. This difference could be seen in both the main dataset (`I think . . .', M 3.2;`My friends think . . .', M 2.9) and in the Tuzla Canton dataset (`I think . . .', M 3.4;`My friends think . . .', M 2.9). This may indicate that participants simply felt inclined to provide more progressive responses to the`I think . . .' statements. As Randall and Fernandes note,`. . . individuals may present themselves in a favorable light, regardless of their true feelings or actual behavior'. 103 However, the data might also be read as highlighting social pressures ± or perceived social pressures ± on young people to adopt certain viewpoints. While the questionnaire data are limited in this regard, it would be interesting as a follow-up to explore the perceptions that young people have about each other's opinions. More broadly, the data underscore the need for transitional justice processes to give more attention to social ecologies ± which include families and peer pressures ± that inform and sustain individual viewpoints. As Herman accentuates,`. . . social context is created by relationships with friends, lovers and family.' 104 An essential part of addressing negative attitudes and prejudices relating to sexual violence is to create a space for discussion and questions. In BiH, such a space currently does not exist. It is therefore unsurprising that in the pre-talk questionnaire, 90 per cent of students in the main dataset revealed that they felt very uncomfortable speaking about the issue of sexual violence during the Bosnian war. The fact that 95 per cent of female students (as compared to 85 per cent of male students) expressed feeling very uncomfortable further underscores the critical importance of developing gender-inclusive approaches. 105 Interestingly, 55 per cent of female students and 38 per cent of male students said that BiH society as a whole should talk more about the sexual violence committed during the war, for example, through media discussions and community events. Only 11 per cent of female students and 16 per cent of male students disagreed (and 35 per cent of female students and 46 per cent of male students were unsure). In other words, the data reveal strong cross-gender support for the creation of new discursive spaces around the issue. The students' receptiveness, in turn, helps to explain the fact that there were some important attitudinal shifts reflected in the post-talk questionnaire.

EVIDENCE OF ATTITUDINAL CHANGE
Immediately following the conclusion of each interactive school talk, the students were asked to complete a second questionnaire. This was designed as a tool for gauging the impact of the talks and for assessing whether any positive attitudinal shifts in the students' views had occurred. As previously noted, students were requested to express their levels of agreement or disagreement with the same set of gender and society statements that formed a key part of the pre-talk questionnaire. They were not asked a second time about the views of their families and friends.
It was highlighted in the previous section that in both the main dataset and the Tuzla Canton dataset, there was a statistically significant difference 530 104 Herman, op. cit., n. 48, p. 9. 105 In the Tuzla Canton dataset, in contrast, only 46 per cent of female students (and 26 per cent of male students) said that they felt very uncomfortable speaking about sexual violence during the Bosnian war (while 26 per cent of female students ± and 25 per cent of male students ± said that they felt uncomfortable). It is unclear why substantially fewer students in the Tuzla Canton dataset compared to the main dataset expressed feeling very uncomfortable.
between the average responses of female and male participants to the pretalk and post-talk measures. However, in the main dataset, the difference in average total attitudinal change between female (M 0.47, SD 2.1) and male (M 0.46, SD 2.4) participants was not significant: t(476) À0.04, p b 0X05. In other words, notwithstanding the fact that male students initially expressed more negative attitudes than female students, the data indicate that the school talks ultimately had a similarly positive effect on both genders. In the Tuzla Canton dataset, although there was a greater average total attitudinal change between female students (M 0.35, SD 1.57) and male students (M 0.06, SD 1.69), t(262) À1.43, p b 0X05, it was not statistically significant.
Returning to the main dataset, analysis of the data reveals that there was a significant positive difference between the overall scores in the pre-talk (M 16.7, SD 2.8) and post-talk (M 17.2, SD 3) attitudinal measures: t(485) À5.1, p `0X001. Separate paired t-tests on the individual pre-and post-talk statements indicate that statistically significant attitudinal changes occurred in three of the five statements (excluding the statement`I think that if a man is insulted, he should be prepared to physically fight to defend his honour', which as previously noted was only included in the pre-talk questionnaire but not in the post-talk questionnaire). These changes, two of which caused the observed positive effects in the main dataset, can be seen in Table 1.
Clear evidence of positive attitudinal change occurred vis-a Á-vis two particular statements:`If a woman dresses provocatively or gets drunk, she is inviting men to rape her' and`A real man would not allow himself to be raped'. Issues surrounding victim blaming and the use of sexual violence 531 Table 1. Attitudinal changes in the main dataset.

Main dataset
Pre-test Post-test T, sig against men were directly addressed in the talks. In relation to the statement I think there are times when a woman deserves to be beaten', there was a significant reduction in the mean score after the talks. While this was not the desired result, it was caused by a slight shift from students selecting strongly disagree' to`disagree'. However, when the students' responses are divided into two groups ± a strongly agree/agree group and a strongly disagree/disagree group ± the mean responses across the pre-and post-talk questionnaires are identical. In short, students varied in the extent to which they disagreed with the statement, but the overall numbers who disagreed with it in some way did not change. This is unsurprising; the issue of domestic violence was only indirectly touched upon during the talks. The data thereby suggest that this is a topic that should be given more attention within schools, and indeed this is one of the themes that this project will address as it moves into its next phase (discussed below).
In the Tuzla Canton dataset, the results were more modest. There was a significant difference between the overall scores in the pre-talk (M 14.13, SD 1.81) and post-talk measures (M 14.39, SD 1.93), t(264) À2.56, p `0X05. As can be seen in Table 2, however, significant attitudinal change occurred only in relation to the statement:`If a woman dresses provocatively or gets drunk, she is inviting men to rape her'. It is important to note that the pre-talk version of the questionnaire (which was initially envisaged as a pilot) used in this part of BiH did not include the statement`A real man would not allow himself to be raped'. This was only included in the post-talk questionnaire. This means that when assessing the overall impact of the talks, the mean scores for the Tuzla Canton dataset necessarily look lower, although the two datasets are not comparable due to the different number of pre-and post-talk statements. The post-talk mean scores for thè real man' statement, however, were similar in both the Tuzla Canton dataset 532 2) and the main dataset (M 3). As this statement generated the principal attitudinal change in the main dataset, it is possible that its absence from the pre-talk questionnaire used in Tuzla Canton concealed attitudinal changes around this particular item. It is interesting to note that in the posttalk data from Tuzla Canton, the average response to this statement was significantly different between male students (M 2.8, SD 0.92) and female students (M 3.6, SD 0.66), t(261) À8.1, p `0X001. This is similar to the post-talk difference between male participants (M 2.8, SD 1) and female participants (M 3.33, SD 0.86), t(497) À6.22, p `0X001 in the main dataset. Analysis of the pre-and post-talk questionnaires thus indicates that the interactive school talks did have some positive impact on students' ideas and thinking about sexual violence. That the distinctive positive post-talk attitudinal shifts occurred in relation to only two statements (and one statement in the Tuzla Canton dataset) is consistent with the findings of Black et al.'s study. 106 This suggested that`perhaps specific subscales within sexual assault attitudes may be more susceptible to attitude change and attitude change maintenance.' 107 Moreover, the positive attitudinal shifts occurred in relation to two statements which can be seen as examples of`rape myths'. Rape myths have been defined as`prejudicial, stereotyped, or false beliefs about rape, rape victims, and rapists' 108 and`a combination of stereotypical attitudes about rape with the cultural functioning of a myth'. 109 In this regard, it is important to acknowledge that the questionnaire data tell us little about the social prevalence of the beliefs reflected in the two statements ± or about their functionality and social impact. According to Newcombe et al., however, the most common clusters of attitudes that constitute rape myths are: victim responsibility (e.g.,`She asked for it'), disbelief of rape claims (e.g.,`It wasn't really rape'), and the belief that rape only happens to certain kinds of women (e.g.,`Only women who dress suggestively are raped') . . . 110 Both the notion that a woman can invite men to rape her, based on how she dresses or how much alcohol she has consumed, and the idea that onlỳ lesser' men are raped clearly contain elements of victim blaming and 533 imputed responsibility. From this perspective, they can be viewed as rape myths.
In the post-talk questionnaire, students were explicitly asked whether their opinions had changed as a result of the project. In the main data set, 35 per cent of students answered in the affirmative, of whom 57 per cent were female and 43 per cent were male. The majority of students (55 per cent) maintained that their views had not changed, while 10 per cent were unsure. In the Tuzla Canton dataset, similarly, 35 per cent of students (40 per cent of female students and 28 per cent of male students) said that their views had changed following the talks. A further 40 per cent (an equal percentage of female and male students) answered that their views remained unaltered, and 25 per cent (21 per cent of female students and 33 per cent of male students) did not know whether their views had changed. While these results present a somewhat mixed picture, the fact that 35 per cent of students in both datasets claimed that their views had changed as a result of the talks is encouraging.
It is also encouraging that a high proportion of students in both datasets seemingly appreciated the value of the talks. When asked whether their participation in the project had been useful, 33 per cent of students in the main dataset said that it was`Very useful' and 45 per cent said that it was Useful'. More female students (43 per cent) selected`Very useful' than male students (22 per cent), but more male students (48 per cent) selected Useful' than female students (41 per cent). Only 14 per cent of students selected`Not useful' as their answer, of whom 11 per cent were female and 18 per cent were male; and only 8 per cent felt that the talks were a waste of time (5 per cent of female students and 11 per cent of male students). Overall, the average response of female students was skewed slightly towards an assessment of the talks as`Very useful' (M 1.8, SD 0.8), while the responses of male students were skewed slightly towards`Not useful' (M 2.2, SD 0.9), t(466) 5.12, p `0X001.
In the Tuzla Canton dataset, students were given the option of five possible answers. A near majority (49 per cent) said that the talks were very useful and 28 per cent found the talks useful. A further 18 per cent found them somewhat useful. Only 3 per cent of students said that the talks were not very useful and 2 per cent maintained that the talks were a waste of time. In contrast to the main dataset, a slightly higher percentage of male students (50 per cent) than female students (48 per cent) found the talks very useful, whereas more female students (31 per cent) than male students (24 per cent) found the talks useful. An equal percentage of female and male students (3 per cent) assessed the talks as not very useful, and more male students (4 per cent) than female students (1 per cent) viewed the talks as a waste of time.
Consistent with these results, the majority of students overall ± 58 per cent in the main dataset and 86 per cent in the Tuzla Canton dataset ± said that they would like more school talks on the topic of conflict-related sexual violence. Although a considerably higher percentage of female students (69 per cent) than male students (46 per cent) in the main dataset expressed support for further talks, in the Tuzla Canton dataset support was extremely high among both genders (89 per cent and 80 per cent respectively).
The key aim of this project is to utilize education as a way of creating new facilitative ecologies that enable victims/survivors of sexual violence to speak about their experiences without fear of being judged or stigmatized. 111 Yet, the creation of facilitative ecologies necessarily requires long-term (attitudinal) change; and what the questionnaire data cannot tell us is whether those students who appeared to alter their views as a result of the school talks have since reverted to their pre-talk views. It is, moreover, impossible to assess this through additional measures as the students have now left school. In other words, there is no way of gauging the long-term impact of the talks. According to Suarez and Gadalla, however,`Most rapeprevention programs have been found to have only a short-term impact on the participants . . .'. 112 It is also important to acknowledge that there are multiple factors that shape and inform young people's views about sexual violence, from social media and popular culture to families and peer pressure. One of the obvious shortcomings of this project is that it has sought to create new discursive spaces within BiH high schools without also addressing the wider social ecologies in which these schools exist. While this necessarily raises further questions about the long-term impact and success of the talks, it also supports the argument that transitional justice processes need to give more attention to the wider environments in which victims/survivors live. Facilitative ecologies cannot be created in isolation, or in a single sector. Beyond schools, they also need to be created, inter alia, within communities, within institutions and within the media. Situating the questionnaire data within this wider context highlights not only its limitations but also its potential significance. In particular, the data lend support to the argument that there is a crucial attitudinal dimension of transitional justice that needs to be developed and emphasized. 535

CONCLUSION
Lambourne has developed a transformative model of transitional justice that accentuates transformation rather than simply transition. As she theorizes it: Transformative justice not only deals with the past but also establishes conditions and structures in order to ensure justice in the present and the future, creating a longer-term vision and commitment than suggested by the term`transitional justice'. 113 Transformative justice, which should be viewed as an integral component of transitional justice, also extends to social attitudes and social milieu. The very concept of`dealing with the past' becomes vacuous if social attitudes and prejudices that keep the past alive, or that prolong the suffering of waraffected groups, are left unaddressed. Focused on the issue of sexual violence stigma, this article has argued that the extent to which transitional justice processes aid and benefit victims/survivors of conflict-related violence will be limited as long as little or no effort is made to alter the wider`contextual configuration' 114 and to foster facilitative ecologies in which men and women are able to narrate their stories without fear of the consequences. These facilitative ecologies can potentially enable new, more diverse and richer forms of narrative and story telling within transitional justice. They can also contribute to actively engaging victims/survivors in the fight against sexual violence stigma. Based on their work in northern Uganda, for example, Baines and Stewart underline how, through story telling,`survivors might renegotiate their social marginalization and insist on their innocence and social worth'. 115 The article has emphasized the importance of education in this regard, underlining the latter's`transformative potential'. 116 Although it has focused on BiH, the schools project which it has presented is an example of a more attitudinal-focused approach to transitional justice that can open up new discursive and narrative spaces ± and thus new textual/paratextual relationships ± in which victims/survivors are supported rather than stigmatized. The importance and relevance of such an approach extends far beyond BiH and the former Yugoslavia.
Moving forward, this project was never simply about one-off talks. The aim was always to bring about wider curricular change in BiH, so that young people in BiH high schools could have the opportunity to talk about sex, 536 sexual violence and how to support victims/survivors. Snaga Z Ï ene and I are now working closely with various Ministries of Education in BiH, with a view to introducing sex and relationship education classes into their C Ï OZ (C Ï as Odeljenske Zajednice) curricula. 117 These classes will cover, among other things, issues surrounding domestic abuse, sex and consent, conflictrelated sexual violence, rape myths, and stigma, and will commence with immediate effect in Oras Ïje.