Hate Crime Victims in Serbia: A Case Study of Context and Social Perceptions

This article provides an overview of the Republic of Serbia’s legal framework that incorporates strong guarantees for protection from discrimination, national minorities’ rights, and prosecution of (ethnic) hate crimes, but also describes a social context loaded with strong prejudices. To illustrate the above, I present a case study of two similar incidents of alleged hate crimes reported in a local Serbian newspaper. In both cases, the victims were young men belonging to ethnic minorities. In 2015, within a period of two months, a Serb was attacked in the Croatian capital, Zagreb, and an Albanian-speaking man in the Serbian town, Novi Sad. The articles attracted online comments, 205 and 134 respectively, mostly from readers from Serbia. These comments elicited what are likely to be honest responses because of the relative anonymity provided to authors. By analyzing commentaries on these newspaper items, this article compares social responses to hate crime cases where victims belonged to different ethnic groups and where the incidents occurred in different geographic and social contexts. representative sample of 1,200 Serbian citizens over the age of 15 years responded to 45 questions. Of the sample, 52 per cent were women and 48 per cent were men. The research into ethnic distance was conducted using the Bogardus social distance scale which measures attitudes toward different social groups. In addition to ethnic groups (Albanians, Croats, Hungarians, Bosniaks/Muslims, Roma and Serbs), this survey included persons with physical disabilities; persons with mental disabilities; elderly persons; women; poor persons; lesbian, gay, bisexual and transgender persons (LGBT); religious minorities; refugees; migrants; foreigners; and persons with HIV/AIDS. With regard to the ethnic minorities, as expected within Serbia, the ethnic distance expressed by ethnic Serbs was closest with other members of that population group. Serbs expressed greatest reluctance to accept Albanians and, to lessening degrees, Croats and Roma and Muslims/Bosniaks. Among the national groups offered for selection by respondents to the survey, Hungarians were the least unfavorable. Results indicated that 45 per cent of the interviewees would not like to have an Albanian in their family, and 23 per cent would not like a Croat as a family member. Regarding less intimate forms of relationship, 35 per cent did not want to have an Albanian as a teacher to their children compared to 14 per cent disliking Croats as teachers. Lastly, 24 per cent did not want Albanians as citizens of Serbia compared to 9 per cent not accepting Croats. Interestingly, in several measurements of attitudes towards ethnic groups, the distance toward Albanians decreased but overall ‘general’ unacceptance increased. Nevertheless, the sequence of unpopularity remained the same: Albanians, Croats/Roma, Bosniaks and, lastly, Hungarians.


Introduction
Hate crimes are criminal acts motivated by bias or prejudice towards particular groups of people. This definition is offered by the Organization for Security and Co-operation in Europe (OSCE) through its Office of Democratic Institutions and Human Rights (ODIHR). Historical circumstances, social context and national legislation will determine which of the many individual/identity characteristics shall be further protected through a nation's hate crime legislation.
The OSCE/ODIHR advocates that, when developing hate crimes legislation, protected characteristics should be limited to those which function as a marker of group identity and those that have been the basis for past and recent incidents. Consideration should also be given to the practical implications of implementation for the investigators and the prosecutors (ODIHR 2009: 37-39). Within the OSCE region, race, national origin and ethnicity are the most commonly protected characteristics, closely followed by religion (ODIHR 2009). Internationally, these characteristics were recognized during the early period of hate crime law making.
Based on the OSCE/ODIHR hate crime definition, we can establish that ethnic hate crimes are acts of hostility motivated by the victim's real or perceived ethnicity or nationality. In countries with persistent inter-ethnic tensions, particularly those which have experienced inter-ethnic violence in recent times, these crimes are of particular significance. Ethnic hate crimes may destabilize inter-state relations in cases where the victim belongs to a minority group in one state and represents a majority in their 'kin state'.
Moreover, a link has been established between ethnic hate crimes on the one hand and genocide and terrorism on the other, such that genocide and terrorism can be viewed as a continuation of hate crimes (Perry 2014: 11). Hate crimes do not just affect marginalized groups but are also a threat to broader national security (Perry 2014: 11). Regardless of whether hate crimes precede or follow large-scale acts of violence, they may be viewed as elements of the same pattern of violence where it is difficult to distinguish where one ends and the other begins. Ethnic hate crimes enacted in fragile post-conflict countries may tip the situation back into conflict. These crimes usually go hand-in-hand with hate speech which may circulate in public discourse for decades after the conflict ends or appeases (Chapman 2014: 35).
Although ethnic hate crimes do not exclusively target minorities as the victims may belong to the majority community (OSCE MC Decision No. 9/09), ethnic hate crimes are usually the product of the everyday stigmatization, marginalization and stereotyping of minorities. Prejudices help justify immoral acts committed toward the 'Other' (Aronson 1999), and the perpetrators of hate crimes might act because of their beliefs based on the prejudices or negative emotions towards persons whom they perceive as 'different' (Levin and McDevitt 2008). Prejudices and negative attitudes towards others result in violence and tensions and, in this environment, hate tends to accumulate. As noted in the OSCE decision on hate crimes (OSCE MC Decision No. 9/09), manifestations of intolerance may give rise to conflict and violence on a wider scale.
In the Balkans, this trend has been present over the past century and inter-ethnic violence has peaked twice during that time span. The first peak was evident during World War II (WWII) and the second, after 50 years of relative inter-ethnic harmony, during the break-up of the former Socialist Federal Republic of Yugoslavia (SFR Yugoslavia) in the 1990s. The latter rise in hate incidents began, however, in the late 1980s. Due to the economic and political crises of the time, these incidents, which were previously severely punished by law, went unpunished (Levin and Rabrenovic 2001). All of this contributed to the weakening of the interethnic cohesion and resulted in its opposites: ethnic cleansing and acts of genocide. The roots of the ethnic conflict lay in the inability of the government to legitimate and then support national (ethnic) minorities (Levin and Rabrenovic 2001 In 2009, the Republic of Serbia strengthened its overall non-discrimination policies by adopting the Law on the Prohibition of Discrimination. The 2009 law has defined discrimination as … any unwarranted discrimination or unequal treatment, that is to say, omission (exclusion, limitation or preferential treatment) in relation to individuals or groups, as well as members of their families or persons close to them, be it overt or covert, on the grounds of race, skin colour, ancestors, citizenship, national affiliation or ethnic origin, language, religious or political beliefs, gender, gender identity, sexual orientation, financial position, birth, genetic characteristics, health, disability, marital and family status, previous convictions, age, appearance, membership in political, trade union and other organisations and other real or presumed personal characteristics.
The law also recognizes severe forms of discrimination, such as: Causing and inciting inequality, hatred and enmity on the grounds of national, racial or religious affiliation, language, political opinions, gender, gender identity, sexual orientation or disability … Non-discrimination provisions are included in relevant sectoral legislation governing employment, education and media. For example, the 2015 Law on Foundation of Education provides for protection from discrimination in schools.
The Serbian Criminal Code introduced several articles which provide protection from violence targeting national minority identities. This includes, inter alia, incitement of national, racial and religious hatred and intolerance (Article 317) as well as violating the right to expression of national or ethnic affiliation (Article 130) and violating the right to use a language or alphabet (Article 129). Specific hate crime legislation was enacted by amendments to the criminal code in 2012 which were accepted in 2013 through the adoption of Article 54a of the criminal code. This article of the code establishes the obligation for the court to consider hatred based on race, religion, national or ethnic affiliation, sex, sexual orientation or gender identity as an obligatory aggravating circumstance unless it represents an element of a criminal act as such. Finally, as part of the EU accession process, Serbia has adopted several policy documents which outline activities aimed at combatting hate crimes including those based on national or ethnic origin. 1 Nonetheless, the situation in practice does not reflect the progressive legislation described above. This has been observed by, for example, the European Commission against Racism and Intolerance (ECRI). The ECRI Report on Serbia (2011) has noted that the court practice regarding racist crimes is problematic since there are no data and only a few cases have been processed. 2 Even when there are final court decisions, the penalties remain minimal. The lack of adequate judicial response in hate crime cases is also criticized by the representatives of Serbian civil society organizations such as the Lawyers' Committee for Human Rights (YUKOM) (Nenadović 2017). 3 Inter-ethnic relations in the Republic of Serbia are an extension of the recent and historical, explicit and silent conflicts with essentially all bordering territories, as well as stereotypes that Balkan nations hold regarding each other. The WWII hostilities included conflicts with neighboring Hungary, Bulgaria and Albania, and the more recent 1990s armed conflicts involved-in addition to Serbia-Bosnia and Herzegovina, Croatia and Kosovo. The conflicts in the 1990s culminated with NATO (North Atlantic Treaty Organization) intervention in Serbia and Montenegro, which resulted in the UN Interim Administration Mission in Kosovo in 1999, and the self-proclaimed independence of Kosovo in 2008.
As with the rest of the region, the ethnic stereotypes in Serbia are strong and persistent. Surveys from 1999 (Centar za slobodne izbore i demokratiju (CeSiD) et al. 2012) and 2016 (Commissioner for Protection of Equality 2016) into citizens' attitudes towards discrimination in Serbia indicate insignificant changes during the intervening period in the perception of other ethnicities. The research on auto-stereotypes and hetero-stereotypes of Serbs in Serbia (Biro and Popadic 1999) has examined views that Serbs have of themselves and nine ethnic groups that reside in Serbia (Croats, Romanians, Hungarians, Bulgarians, Macedonians, Muslims/Bosniaks, Montenegrins, Albanians, and Slovenians). The sample included 400 individuals and the survey was conducted in three Serbian regions: the Autonomous Province of Vojvodina; the Belgrade area; and Central Serbia. The results of the 1999 survey indicated that the majority of Serbs held predominantly positive stereotypes about themselves, seeing Serbs as hospitable (90%), proud (80%), sensitive (82%) and courageous (78%). Subsequently, only a few Serbs saw themselves as inhospitable (2%), cold (6%) and cowardly (8%). The strongest negative stereotypes were about Albanians: they were considered uncivilized (82%), dirty (81%), unsophisticated (77%) and unfriendly (77%). Croats were considered by the majority as dishonest (70%), selfish (70%), confrontational (66%) and clean (73%) (Biro and Popadic 1999 representative sample of 1,200 Serbian citizens over the age of 15 years responded to 45 questions. Of the sample, 52 per cent were women and 48 per cent were men. The research into ethnic distance was conducted using the Bogardus social distance scale which measures attitudes toward different social groups. In addition to ethnic groups (Albanians, Croats, Hungarians, Bosniaks/Muslims, Roma and Serbs), this survey included persons with physical disabilities; persons with mental disabilities; elderly persons; women; poor persons; lesbian, gay, bisexual and transgender persons (LGBT); religious minorities; refugees; migrants; foreigners; and persons with HIV/AIDS. With regard to the ethnic minorities, as expected within Serbia, the ethnic distance expressed by ethnic Serbs was closest with other members of that population group. Serbs expressed greatest reluctance to accept Albanians and, to lessening degrees, Croats and Roma and Muslims/Bosniaks. Among the national groups offered for selection by respondents to the survey, Hungarians were the least unfavorable. Results indicated that 45 per cent of the interviewees would not like to have an Albanian in their family, and 23 per cent would not like a Croat as a family member. Regarding less intimate forms of relationship, 35 per cent did not want to have an Albanian as a teacher to their children compared to 14 per cent disliking Croats as teachers. Lastly, 24 per cent did not want Albanians as citizens of Serbia compared to 9 per cent not accepting Croats. Interestingly, in several measurements of attitudes towards ethnic groups, the distance toward Albanians decreased but overall 'general' unacceptance increased. Nevertheless, the sequence of unpopularity remained the same: Albanians, Croats/Roma, Bosniaks and, lastly, Hungarians.
The above-mentioned attitudes are partially reflected in hate crimes statistics which the Serbian authorities and civil society organizations (CSOs) have been collecting regularly since 2006 for the OSCE annual Hate Crimes Report (2009-2016 reports for Serbia may be accessed from the OSCE/ODIHR website at http://hatecrime.osce.org/serbia). In 2014, CSOs reported 44 hate crimes of which 28 were motivated by the victim's racial or ethnic background. Of these 28 incidents, 12 were committed against persons belonging to the Roma community and 16 against other minorities. The records include a description of incidents and note that, in the majority of the 16 'other' cases, most of the victims were Albanians (13) whose property was damaged after the football match between Serbia and Albania in Belgrade in October 2014. In 2015, the total number of recorded incidents had decreased from 44 to 21. Yet, 15 out of 21 incidents were motivated by the victims' religious, racial or ethnic background. The majority of these 15 represented an attack against property. In 2016, CSOs reported 17 hate incidents, with the majority of these again motivated by the victims' religious, racial or ethnic background.

Public perception of hate crime victims in Serbia: A case study
Contrary to the increasing attention given to hate crimes victims, less effort has been invested in uncovering people's perceptions and attitudes about these crimes (Kraig and Waldo 1996). Yet, there is evidence that demographic characteristics determine how disruptive an individual may regard the crime and how likely it is that a person would report personal knowledge of a victim of hate crime (Kraig and Waldo 1996). In Kraig and Waldo's 1996 study carried out among the students at a large midwestern university in the United States on the perceptions of hate crimes, it was noted that the participants' perceptions of what hate crimes involve, why they occur, and who the victims are, suggest that victims encounter a range of reactions when they inform others that they have experienced a hate crime. For example, the participants acknowledged the disruptiveness of the incident to the victim, but did not necessarily attribute the incident to the victim's real or perceived membership of a group (Kraig and Waldo 1996).
My principle hypothesis is that a hate crime victim is perceived by the public depending on the ethnic affiliation of the victim and that of the majority population. If victims share the common characteristic (in this case, the ethnic background) with the majority, the public shall display more empathy whereas if the victim belongs to a minority group, denial, mistrust and sentiments of hate might be experienced in addition to the initial attack. These responses shall become more evident if the victim and the public belong to groups with a record of inter-ethnic hostilities.

Two media articles on inter-ethnic hate incidents
Based on the above assumption that hate crimes are perceived differently by different social groups and as an illustration of social perceptions of hate crimes, I have analyzed online comments responding to two articles about alleged hate crime incidents which were published in the Serbian daily newspaper, Blic, in 2015. The articles and the comments are part of the online edition of Blic.
These two articles were selected because they described the same type of hate crime incident: a physical assault by a group of perpetrators, driven by a single (hate) motive. In both cases, the victims were men from the same age group (late teens/early twenties). The main difference in these two cases was the ethnic background of the victims and the perpetrators. Both articles appear to be serious and not written in the sensationalistic manner. The author in both instances (Anicic) had demonstrated a neutral if not negative attitude toward the incidents without manifesting any bias. There was evidently different treatment, however, of the titles of the articles and this may be attributed to the editor and not the journalist herself. The first article included 'naming and shaming' ('Croats beat up a young man only because he is a Serb') whereas the second title was neutral in referring to the perpetrators ('Novi Sad's bloody face-beaten because he spoke Albanian with his mother'). Both articles included photos of the victims, with visible injuries and stitches following the attacks.
The first article, 'Croats beat up …', describes an incident that happened to a 21-year-old Serb who was traveling by train from Serbia to Slovenia, through Croatia (Anicic 2015a). The train stopped at the main railway station in Zagreb, Croatia's capital, and several 'hooligans' wearing clothes with football club insignias boarded the train with baseball bats. They used racial slurs against Serbs and attacked the victim on the train. The article further reported that a victim was interviewed by Croatian police officers but not provided with adequate assistance. He continued his trip to Slovenia where he was hospitalized with severe injuries for which he later received treatment in a hospital in Serbia. The article also mentioned the assertion from the Croatian Police Directorate that the injured man refused medical assistance and continued his trip. Furthermore, the article reported that the Serbian Ministry of Foreign Affairs requested urgent delivery of all information relevant to this case including which investigative measures had been taken. Finally, the Serbian Embassy asked for clarification as to why the victim did not receive adequate medical assistance in Croatia.
The second article, 'Novi Sad's bloody face …' (Anicic 2015b), tells the story of a young Albanianspeaking man-whether he was an ethnic Albanian was not stated-who was beaten by two young men (the author did not use the term hooligans this time) after he finished a phone conversation in Albanian with his mother. The victim claimed that he was approached by two men who asked him for a cigarette after which they pushed him to the ground and severely beat him. A passer-by saw him and helped him, and the attackers consequently ran away. The victim noted that the attackers were young adults wearing casual clothes and no visible football insignia. He added that he saw no particular reason for the attack apart from the fact that he was speaking Albanian. He stated that he had never experienced anything similar in this town of Novi Sad where he was born and raised. The article noted that the case had been reported to the police and that the police began an investigation.

Methodology
The expansion of online news media allows for more commentary by readers than when online access was not available. The media is experiencing a shift from print to online, and people are using it not only for information but also for interaction and communication where they respond to articles but also to other commenters and for space for interaction (McMillen 2013). Social media, which is becoming part of many people's lives, allows for a level of anonymity, thus enabling an environment where negative sentiments including hate speech can proliferate.
For this study, I analyzed the online reactions of people commenting on both articles. This method has, however, serious limitations. The first is due to the practice of the media to introduce premoderating systems which do not allow content that directly incites hate. For example, current Blic website rules for posting comments (available at https://www.blic.rs/pre-slanja-komentaramolimo-vas-da-procitate-sledeca-pravila/gsv5wj2) require that 'comments which include racist or chauvinistic messages shall not be posted' and that 'Blic reserves a right not to allow offensive comments or those which incite racial or ethnic hatred …'. This may result for the majoritiy of the offensive language or comments inciting hate not to be posted.
The second limitation to this study is the respondent sample may not be representative of Serbian society or the Blic tabloid or online readers, particularly in regards to education level. According to the 2015 survey titled 'End to delusions on media circulation' undertaken by the Research Solutions Partner, Blic is the most popular daily newspaper in Serbia with a majority of the readers aged 35-44 years. Additionally, the majority (62%) of Blic readers has attained secondary education as their highest level of education, 22 per cent has attained college education as their highest level, and 14 per cent has attained a university degree. According to the 2011 Census, 49 per cent of the Serbian population has attained secondary education as their highest level of education, six per cent a college education, and just 10 per cent a university degree. Accordingly, the respondent sample might be better educated than the Serbian population in general.
The data for this analysis include online comments for the articles reporting on the two alleged inter-ethnic hate incidents. This analysis provides an insight into the perceptions and attitudes of those who chose to comment on the hate crimes incidents. I posed three principal questions in an attempt to reveal how correspondents perceived the conflict in terms of motive, how they related to the victim, and who they blamed for the incidents.
In order to group the data and illustrate the differences in social perception of these two cases by persons from a predominantly Serbian readership, the coding list contained the following categories: relevancy of responses; the sequence of responses; the attitude towards the victim; perception of the motive; and responsibility for the attack. The coding sheet also included data on commentators, such as presumed ethnicity and gender.
Regarding the presumed ethnicity of the author of the comment, four categories were coded: Serbian; Croatian; other; and those for whom the identities could not be presumed. The manner by which I determined the ethnicity of commentators in the sample is speculative and contestable yet is based on the presumption that most readers of the Serbian newspaper are Serbs. I have only recorded commentators as Croatian or other ethnicities when signature nicknames suggested their ethnicity, or where ethnicity was explicitly stated.
Regarding the gender of the commentators, three categories were applied: male; female; and comments where gender could not be presumed. I was able to determine gender with more certainty by looking at the signature nicknames. Linguistic rules of the local languages (Serbian and Croatian) also allow such distinction to be interpreted due to, for example, different endings to adjectives used for the names of men and women.
Comments posted in response to each of the two articles were coded according to relevance, sequence, attitude, perception and responsibility as follows: • The relevance of the responses included three categories: responses which directly responded to the attacks described in the article; responses which related to the article IJCJ&SD 28 Online version via www.crimejusticejournal.com © 2018 7(2) but at the same time introduced a new topic (for example, 'this is why I don't go to the Croatian coast' or '… and what did they do to us?!'); and, finally, off-topic responses (trolling). • In sequence of responding, the distinction was made between the initial response, reactionary supportive response, and reactionary conflicting responses. • The attitude towards the victim was graded according to apparent strong empathy, empathy, neutrality, animosity or cynicism, and strong animosity/hate speech. It was also noted if the comment did not refer to the victim. • The perception of attack was described as ethnically motivated, not ethnically motivated, and 'displays no attitude toward the attackers' motive'. • The responsibility for the attacks in the comments was coded as: those who blamed the perpetrators; those who put blame on the victim; those who blamed a third element, such as poverty, former government, and discontent and poverty among young people; and those who did not indicate who was to be blamed for the attack.

Relevance of the responses
Applying the above criteria, I analyzed a total of 339 online comments and present results for those I gauged to be the most relevant/partially relevant and illustrative findings. I coded as relevant those which directly referred to the incidents described in the newspaper articles. Those coded as partially relevant generally referred to some form of inter-ethnic intolerance, personal ethnic-related experiences, or a related topic but from a different non-ethnic angle. Most comments posted (205) referred to the first article, with 134 referring to the second article For the first article, there were 181 (88%) relevant and partially relevant comments, compared to 123 (92%) for the second article. Comments that could be categorized as most relevant amounted to 79 (38%) for the first article and 78 (58%) comments for the second (Figure 1).

Figure 1: Relevance of responses to Articles 1 and 2
Examples of partially relevant comments include: I don't know what to say when I read some of your comments. It seems that some of our nationalists are still stuck in the fourteenth century and the Kosovo battle. It seems that Turks also had complexes from us when they beat us so badly.

Sequence of responses
Regarding the sequence (Figure 2), 103 comments (50%) represented initial responses for the first and 65 (49%) for the second article. For the first article, 25 of the remaining comments (12%) were in agreement with the initial comments compared to 35 (26%) for the second article. The overall number of comments indicates that 'ethnic' and 'inter-ethnic' themes attract attention, spark dynamic public debate, and suggest that personal 'ethnicity related' experiences are not uncommon. The content of the comments also indicates that inter-ethnic conflicts which happened over 25 years ago remain alive in contemporary discourse, shaping current attitudes and prejudices.

Attitude towards victims: From empathy to hate speech
The attitude towards the victim represents the first of three principle questions examined in this analysis. Of all respondents, 120 (58%) in the first article compared to 56 (42%) in the second made no reference to the victim.
Of the remaining 85 persons who responded to the first article, 6 (7%) demonstrated apparent strong empathy with the victim, 47 (55%) empathy, 30 (35%) had a neutral stance and two (2%) displayed animosity or cynicism towards the victim. No evidence of 'strong animosity/hate speech' was present in comments for this article (Figure 3).

Figure 3: Attitude towards victim (when expressed) in Article 1
In the second article, of the 78 persons (58%) who did make reference to the victim, 2 (3%) showed high empathy, 30 (38%) empathy, 15 (19%) a neutral stance, 25 (32%) animosity or cynicism and, in six cases (8%) strong animosity/hate speech was present. This includes those who exhibited hate speech in their initial comments and those who agreed with such statements (Figure 4).

Figure 4: Attitude towards victim (when expressed) in Article 2
Examples of strong animosity/hate speech included the following statements: They should speak Serbian. Both them and the Hungarians. Enough with tolerance! You live in Serbia. We cannot say anything when we come to you. In any language.

Perceptions of the attack
A majority of commentators who displayed an attitude about how they perceived the attacks regarded them as ethnically motivated ( Figure 5). In the first article, out of 117 respondents who displayed an attitude about the attack, 99 (85%) thought it was ethnically motivated. For the second article, out of 80 respondents who displayed an attitude about the attack 61 (76%) thought it ethnically motivated. Similarities in the responses indicate that the victim's origin in these cases does not play a major role. An example of commentary which viewed the incident as an ethnically motivated is: Thugs ran away again. They do not have a courage to accept responsibility for expressing hostility in this manner, only because a person communicated in a different language. (Vrbas in response to Anicic 2015b)

Responsibility for the attacks
Regarding the question on the responsibility for the attacks for the incidents (refer to Figure 6), 152 commentators did not make any reference to who should be blamed. For the first article, this accounted for 92 (45%) of the comments and, in the second, for 60 (also 45%). Among the 187 commentators who did make reference to who was responsible for the attacks (113 and 74 for, respectively, Articles 1 and 2), the Albanian-speaking victim in Article 2 was blamed by 36 persons (49%), substantially more so than the Serbian victim in Article 1 who was blamed by just four (4%) respondents.
The majority (100 or 49%) of the commentators for the first article blamed the perpetrator. The victim was blamed for the attack by only four (2%) commentators, with three of those four identified as ethnic Croats. For the second article, commentators blamed the attack on perpetrators in 29 comments (21%), the victim in 36 (27%), and a third element (including a poor economic situation, wars in 1990s, and an overall social context that promotes violent communication) in nine (7%) comments. Further analysis of the comments showed that, in comments for the first article where the victim was a Serb, among those 53 who displayed a level of empathy towards the victim, the majority 41 (77%) viewed the incident as ethnically motivated. The majority of 44 (83%) also indicated that they thought that the perpetrators were responsible for the incident.
In comments for the second article regarding the Albanian-speaking victim, 26 (81%) of the 32 who displayed a level of empathy towards the victim also viewed the incident as being ethnically motivated; additionally, 25 (78%) of them indicated that the perpetrators were responsible for the incident, a lesser proportion than for the first article. In contrast, the majority of the 31 (23%) who displayed a level of animosity or hate towards the Albanian-speaking victim saw the incident as ethnically motivated (26 commenters or 84%) but viewed the victim as responsible for the attack (23 commenters or 89%).
The data as analyzed support my principle hypothesis about the perception of the victim. Specifically, a victim who shares the protected characteristic of the majority of the population is considered by that population as an 'authentic' ethnic hate crime victim. Moreover, that person bears no responsibility for the attack and was attacked for no reason other than ethnicity. The level of empathy of the majority population toward the victim is high and the perpetrator's bias motives are assumed. Conversely, in the case of a victim from a minority ethnic group, the bias motivation of the perpetrator is denied, relativized or, in the worst case, justified. This is evidenced even if the overall attitude toward the victim is sympathetic.
As if they know how does the Albanian language sound … This more looks like arrogant kids, and he found himself in the wrong place at the wrong time. (Avg responses in Anicic 2015b) The real motive behind the attack is unknown. Personally, I don't believe that the reason is Albanian language. There were so many attacks on people of different, age, looks, and behaviors. The reason is in their crazy heads … I feel sorry for this young man! (Novosadjanka response in Anicic 2015b) Hate crime legislation per se does not protect any particular group in society; it 'protects' certain characteristics which represent a basis for groups' identities (such as ethnicity, skin color, sexual orientation). Thus, in theory, a hate crime victim may belong to both a minority and a majority group within a population. In practice, however, the bulk of victims are members of various minority and usually disadvantaged social groups (LGBT persons, persons of Roma ethnicity, persons belonging to different ethnic minorities in ethnically heterogeneous states). The impact of a negative attitude toward hate crime victims within these groups might further affect their underprivileged position in the society.

Gender variance
The majority of the commentators in both articles could be identified as male. For the first article, 169 were coded as male (83%) and 27 as females (13%). This compared to 127 men (95%) and only 6 women (4%) for the second article (plus one unknown).
In the first article, a slim majority of the women (14 or 52%) made no reference to the victim. Of the 13 (48%) who did, all but one displayed empathy or high empathy. The comment which was arguably the most favorable towards the Serbian victim was made by a woman who identified as a Croat, which contradicts a common belief on Serbian-Croatian antagonism: Unfortunately these idiots spread hate among the normal people. It is true that most of them hate Serbs and their own life and always seek trouble. I am a Croat, and I am sorry because of this. I hope that this young man recovers soon and that he doesn't feel hate towards everybody because we are not all the same. (Uaha response in Anicic 2015a) The attitudes of women towards the Albanian-speaking victim differ from the majority attitude. Five of the six women (83%) expressed empathy; the other one expressed strong animosity by siding with the previous hate speech comment.
You are so full of hate, this is the reason why you talk like that. Leave the newcomers alone, so that you don't become a newcomer yourself. (MIRKA-NS response in Anicic 2015b) Official statistics for female perpetrators of hate crime in Serbia are unknown but available data and media reports indicate that this area of crime contains a higher ratio of male perpetrators than other crimes. For example, a 2017 survey (Statistical Office of the Republic of Serbia 2017) measured the prevalence of female perpetrators in all crime at about 10 per cent. For criminal acts, that usually constitutes hate crime (bodily injuries, security threats) and the percentage is between three and seven.
Accordingly, the all-male participation in the hate incidents discussed in this article supports the male-dominated nature of perpetrators of crime. The 'hate crime audience' responding to the two incidents studied here was also predominantly male. Furthermore, the data indicate that men more so than women tend to display a hostile attitude and discredit a victim with whom they do not share protected characteristic. This is explicable if we consider present day ethnic hate crime as a reflection or a 'micro segment' of former SFR Yugoslavia conflicts which, like other wars, remain a masculine pursuit. The comments of the majority of women for these two articles also reflect, on a micro level, a peacekeeping position that women have historically held in the conflicts, and Balkan conflicts in particular. The first anti-war demonstrations were organized by an alliance of women's organizations and those women became the engine behind the anti-war movement in Serbia which started spontaneously after the outbreak of war in the 1990s (Licht and Drakulic 1996)

Final notes
It is important to note that Blic is a Serbian daily newspaper and that the majority of the readers and commentators are Serbian. The ethnic composition of the sample in the first article could be 183 (89%) Serbs, 15 (7%) Croats and two (1%) other, with the ethnic origin undetermined for five (3%). In the second article, 128 (96%) appeared to be Serbs compared to five (4%) of 'others' and one not determined. Again, the manner, in which I determined the ethnicity of commentators in the sample is speculative and contestable, yet is based on presumption that the most readers of the Serbian newspaper are Serbs. I have only recorded commentators as other ethnicities when their signature nicknames suggested this, or when they clearly indicated their ethnicity. Where there was doubt, I coded them as 'not be able to determine'.
Despite identified limitations to the sample and methodology, this analysis provides an example of how two hate crimes and their victims were perceived by some online readers and commenters of the most popular newspaper website in Serbia. According to the Gemius Audience monitoring system (available at https://rating.gemius.com/rs/tree/domains), Blic online had over 2.6 million readers in March 2018, 300,000 more that the second leading newspaper website. It provides an insight into public discourse and may contribute to a better understanding of the perceptions of hate crimes, particularly those based on ethnic background.
The results as analyzed are in line with my hypothesis that people perceive hate crimes differently if they share the 'protected characteristic' with the victim. In this analysis, almost 62 per cent of commentators displayed a level of empathy and two per cent a degree of hostility towards the Serbian victim whereas 32 per cent appeared hostile towards the Albanian-speaking victim.
The findings also suggest that post-conflict hate crimes in Balkan countries represent a part or a continuation of the intolerant behavior that culminated in war in the late twentieth century. This suggests that, 20 years after the conflict, these societies did not engage in successful tolerance promotion, reconciliation or 'facing the past' processes. In 2005, Serbia's government of the Autonomous Province of Vojvodina launched a permanent tolerance campaign 'Affirmation of multiculturalism and tolerance in Vojvodina' which was aimed at young people between 14 and 19 years of age. If campaigns such as this were used elsewhere within Serbia, it appears that these actions had only a limited effect.

Conclusion
In the last 15 years, the Republic of Serbia has taken important steps to improve its human rights record including the protection of minority communities. This has been by utilizing human rights instruments, adopting a national legislative framework, and establishing institutions and human rights protection mechanisms. Regrettably, this has not resulted in significant changes in interethnic relations, and several social surveys (as referenced by Gavrilovic and Petrusic 2014) have noted that the system of values and attitude towards others shows very few changes.
As noted by Perry (2009: 13), hate crimes are a product of particular contexts that marginalize communities. Ethnic hate crimes in fragile post-conflict societies such as Serbia are cause for additional concern and attention. Such crimes can be a manifestation of stereotyping, as suggested by this analysis of the two articles, and a consequence of on-going political processes.
In the cases described above, with an inter-ethnic Croatian-Serbian hate incident and a Serbian-Albanian incident, the current political developments cannot be overlooked. Namely, unresolved issues between Serbia and Croatia, which include war crimes cases, missing persons, property issues, protection of the Croatian minority in Serbia and vice versa, and issues related to the undefined border between two countries affect inter-ethnic relations in practice. This contributes to how inter-ethnic hate crimes which involve persons of Serbian or Croatian ethnicity are perceived and prosecuted. With regard to Serbia and Kosovo (the latter with a predominantly Albanian population), the situation is even more complex: in addition to the same unresolved issues as with Croatia, their relations are further hampered by Kosovo's ambiguous status, despite the Belgrade-Pristina political dialog and the subsequent Brussels Agreement aimed at normalizing relations, which was concluded in 2013.
In such an environment, hate crime consequences extend beyond one particular attack. The victims belonging to minority population groups may not only experience discrimination leading to violence but also post-attack denial, belittlement and further victimization by the majority. The level of hostility may depend on the current political climate.
Serbia needs to undertake further steps to protect not only the most vulnerable parts of society but also society itself. The OSCE Ministerial Council Decision on Combatting Hate Crimes (OSCE MC Decision No. 9/09) calls for the participating states to apply several different measures to address hate crimes. Four of these I view as the most relevant. Firstly, the firm, persistent and coherent implementation of anti-hate legislative policies is essential. Secondly, application of existing hate crime legislation should be supported by further training of the police and judiciary. Next, offering support to victims should include counselling and legal and consular assistance as well as effective access to justice. Finally, tolerance, mediation and conflict prevention measures through work with the communities should be applied in co-operation with state stakeholders.