Inspired by individual-level research on direct and indirect as well as reactive and proactive aggression, this article proposes to differentiate direct and indirect types of hate crime. We use the largest hate crime database in Poland (N = 3,153 incidents) to analyze: (1) temporal trends in the relative prevalence of two types of hate crime; (2) the involvement of hate group-affiliated and non-hate group-affiliated perpetrators; and (3) the targeting of victims that are perceived to pose more of a symbolic (vs. more of a realistic threat) to the majority group. Results indicate that direct hate crime was more likely than indirect hate crime to be perpetrated by members and affiliates of hate groups, was more likely to target outgroups seen as posing symbolic rather than realistic threat to the majority group, and was also positively related to societal levels of negative intergroup attitudes and negatively related to unemployment. The findings also show that the two types of hate crime are differently predicted by factors indicative of the social and political climate of the country (e.g., unemployment, political preferences, xenophobia). Although the results were only obtained in one cultural context and will benefit from further validation, they provide very promising initial evidence for the predictive utility of distinguishing direct and indirect hate-crime.
Bias-motivated intergroup violence has a long history ( Throughout the article we use the terms “hate crime” and “bias-motivated crime” interchangeably as they are synonyms (i.e., the “hate” in “hate crime” pertains to intergroup bias that motivates the perpetrator to select their victim, see
To better understand the causes and consequences of hate crime, most research to date has focused on three areas of inquiry (
The current paper aims to fill this significant knowledge gap. Specifically, drawing from social psychological literature on individual and collective aggression (
Hate crimes are crimes in which the victim has been selected based on their (perceived) membership in a social group rather than due to their individual characteristics ( The Polish Criminal Code does not define “hate crime” as a separate legal category, however, there are provisions in the Polish Constitution that guarantee equality for all and ban discrimination on any ground (Article 32), ban the existence of organizations propagating racial and ethnic hatred (Article 13); protect human dignity (Article 30), and guarantee national and ethnic minorities the right to preserve their own language and culture. Additionally, the Criminal Code enhances penalties for some crimes (battery, murder, aggravated assault, and threats) if they occurred as a result of bias motivation due to the victim’s ethnicity, race, nationality, political affiliation, religious creed or worldview (Articles 118 and 119). Articles 256 and 257 of the Criminal Code also penalize propagating totalitarian regimes and hatred as well as insults and assaults due to the victim’s national, ethnic, racial, or religious identity.
Lack of legislation is just one of the limitations that burden investigations of hate crime. Depending on the country of interest, the definition of hate crime and the number and types of social categories (e.g., race, gender, sexual orientation) explicitly protected by the law differ, which makes cross-country comparisons difficult (
Aggression is defined as behavior that has the intention of harming another person, who is motivated to avoid the harm (
Most psychological research on aggression focuses on its interpersonal aspects. This research has established that aggression is situation-specific, functional, and that it is not a unitary phenomenon (i.e., that different types of aggression exist;
Aggression fulfils two main functions: hostile and instrumental (
Similarly to the study of hostile and instrumental aggression, researchers of intimate partner violence (e.g.,
Unlike studies on individual-level aggression and intimate partner violence, research on intergroup aggression typically treats it as a unitary phenomenon (i.e., it examines predictors of its occurrence, without analyzing its forms or components). Most of the extant research on intergroup aggression focuses on the influence of group identification (
Addressing this shortcoming, we introduce a more differentiated approach to hate crime. Inspired by research on individual-level aggression (
In contrast, indirect hate crime does not inflict physical harm on the target and its target is less concrete (e.g., the outgroup as a whole or its parts). Indirect hate crime is very frequently verbal and aims to communicate that the targeted outgroup (or its members) is unwelcome in a given place and/or not accepted in a given community; it may also communicate threats to the outgroup. Overall, indirect hate crime assumes a more instrumental character (see
Below we propose and test a set of predictors of direct and indirect hate crime. The idea that differences in types of hate crime and their unique predictors should be investigated for a better understanding of this phenomenon has been articulated in earlier work (
Given the novelty of our approach, our main goals were to provide evidence that the two hypothesized types of hate crime can be differentiated and that they are differentially predicted by a set of variables established in the literature as predictors of hate crime and intergroup hostility. To this effect we decided to focus on three sets of predictors: perceptions of threat from the outgroup, perpetrators’ associations with hate groups, and the political and economic conditions of the society.
An extensive body of research demonstrates that perceptions of outgroups shape intergroup emotions and behavior (
Difficult life conditions, e.g., political or economic turmoil, have been shown to increase competition over scarce resources (i.e., increase perceptions of realistic threat) and relate to increased intergroup violence (
While considerable diversity exists among hate groups, all are characterized by bigoted ideology and some form of organization. Most (though not all) also espouse white supremacist views, right-wing ideology, and affiliate with a religion (
In light of the above, it seems important to differentiate hate crime perpetrators who belong to or sympathize with hate groups and those who do not. Since hate groups tend to engage in more severe forms of bias-motivated crime (
The investigation of the effects of economic and political context on the dynamics of hate-motivated intergroup aggression has a tradition much older than the definition of hate crime. In their seminal study, inspired by the frustration-aggression hypothesis (
Besides indicators of economic prosperity, like unemployment or the GDP, socially shared attitudes towards members of various minority groups (e.g., ethnic, national, religious) might also constitute an important hate crime predictor. Prejudice is, after all, what puts “bias” in bias-motivated crime. Already in the 1950s,
There is some evidence that intergroup attitudes shared by the general public relate to bias-motivated crime. In a study of right-wing violence in Germany,
Although the relations between societal-level economic indicators and attitudes, and bias-motivated violence are not always straightforward, for exploratory purposes we included three indicators that may be potentially important to differentiating direct and indirect hate crime. In line with the realistic group conflict theory, we expected (H3) that negative economic conditions (operationalized as decreasing GDP and growing unemployment) predict greater prevalence of direct hate crime. We also expected that higher levels of negative intergroup attitudes (H4) and greater prevalence of right-wing political preferences in a society (H5) are positive predictors of direct forms of hate crime as both of these indicators are associated with hate crime (e.g.,
The present study was conducted in Poland, a country characterized by relatively low levels of ethnic and religious diversity (
Our analyses were conducted on the database of incidents that occurred between 1991 and 2013 The year 1991 was chosen as the first year for our analyses, as it was the first year since the cataloguing began that had data for more than half of the months in the year. The catalogue for the years prior to 1991 was very sparse. The year 2013 was the last available fully reported year at the time when this project began.
Character of the event |
|
---|---|
Indirect hate crime | Direct hate crime |
Desecration of religious/sacred objects/buildings (e.g., cemeteries, mosques, synagogues) Public statements, hateful banners at concerts or mass gatherings Publications and media statements (on the radio, TV etc.); internet publications Verbal aggression: name calling, threats, making “monkey sounds” towards Black people Propagation of extreme ideologies; nationalist demonstrations; displays of fascist/Nazi symbols and gestures Dissemination of materials (books, movies, etc.) that propagate extreme right-wing ideologies (fascism, Nazism) Hateful graffiti in public space; dissemination of posters and flyers with hateful contents Discrimination by the institutions of the State Other events that do not fit the above categories but are not direct, physical violence |
Destruction of property of an outgroup and/or outgroup organization Vandalism: destruction of public property; riots; hooligan activities Use of physical violence against members of an outgroup; battery Theft, appropriation of outgroup member(s)’ possessions Sexual violence: rape, attempted rape, sexual harassment Homicide; also death as a result of battery or other actions (e.g., arson) Disturbance of a demonstration, rally, public event; intrusion in the victim’s place of residence, workplace, or other location |
Categories of perpetrators | |
---|---|
Hate group members | Others |
Members of particular hate groups/organizations Affiliates of hate groups/organizations People associated with sports: hooligans, players, team members, referees etc.a Members of right-wing subculture (e.g., skinheads) |
Unknown/other perpetrators Members of political parties; people in public/political office (national or local) Representatives of the Catholic Church Neighbors, fellow city/village dwellers Local administration members Members of the Parliament Owners/staff of restaurants, bars, hotels, etc.; employees of centers for asylum seekers Public persons (but not politicians), e.g., celebrities, people working in the media, etc. |
Categories of victims | |
Perceived as symbolically threatening | Perceived as realistically threatening |
LGBTQIA persons Socially excluded groups (e.g., homeless people) Alternative/counter culture groups Members of the sport community Ideological opponents Religious groups Other kinds of difference |
National or ethnic minorities (e.g., Jews, Roma, Germans, Ukrainians, etc.) |
aMost of the people in this category were racist soccer hooligans attacking or shouting racist slurs at people of color (including players).
Characteristics of the perpetrators and victims Please note that rather than relying on absolute numbers of perpetrators who did or did not belong to hate groups, on raw numbers of victims belonging to groups perceived as realistically or symbolically threatening, and on raw numbers of direct or indirect hate crime elements in a given incident, we decided to rely on proportions. This was motivated by the uneven numbers of reported incidents of hate crime in different years and by the fact that the main focus of the present work was to analyze the differences between direct and indirect hate crime, rather than their absolute numbers.
The “Brown Book” distinguishes between nine types of hate crime victims (e.g., ethnic and national minorities [such as Jewish people, Roma people, Ukrainians, Belarussians]; political adversaries). Based on the integrated threat theory (
The “Brown Book” distinguishes between eight types of perpetrators. In order to simplify our analyses and to verify H2, we categorized them as either members of hate groups (e.g., skinheads) or those who did not belong to such groups. For a more conservative test of our hypothesis, perpetrators categorized as “other” and “unknown” in the original source were included in the non-hate group category. As noted above, some events were perpetrated by both hate group members and non-members. The former were involved in 68.4% of the catalogued events (see
In order to account for the role of the economic, political, and intergroup context at the country level in shaping the nature of hate crimes, a number of country-level variables were included in the analyses.
The economic situation in the country was approximated using the unemployment rate (
The political climate was operationalized as the average declared political preference (left- versus right-wing). The data came from the Centre for Public Opinion Research (CBOS), a Polish polling agency which has conducted large, annual, representative sample surveys of political attitudes among Poles since 1989 (
We also created an index of negative intergroup attitudes. The index was based on CBOS’ annual representative sample survey of attitudes towards other nations and towards ethnic and national minorities in Poland (
The main dependent variable was an index of relative prevalence of indirect vs. direct type of hate crime elements. As indicated above, each event was categorized as exemplifying at least one of 16 hate crime types. In order to create the index, four competent judges (scholars specializing in the study of intergroup relations with at least a master’s degree in social science) received definitions of direct and indirect hate crime and were asked to categorize the 16 types of events differentiated in the “Brown Book” as representing either one or the other. The inter-rater reliability was high: Krippendorff's Alpha for nominal variables calculated using bootstrap with 1000 resamples was .87 with 95% CI [.77, .96] (
Because the numbers of recorded hate crime incidents differed vastly between different years (see
In order to analyze the relation between the independent (event-level and contextual-level) variables and the relative prevalence of indirect (vs. direct) type of hate crimes we tested several multilevel models, using robust maximum likelihood estimators. We used the stepwise strategy recommended by
In the first step we computed a null model (see “Null model” in
Variable / Model | Null Model | Fixed |
Fixed |
Random model: |
Random model: |
Random model with interactions |
---|---|---|---|---|---|---|
Intercept | 0.53 (0.03)*** | 0.62 (0.02)*** | 0.62 (0.01)*** | 0.63 (0.01) | 0.62 (0.01)*** | 0.64 (0.01)*** |
Level 1 predictors (event-related) | ||||||
Victims (0 = only symbolic threat; 1 = only realistic threat) | 0.43 (0.04)*** | 0.43 (0.04)*** | 0.43 (0.04)*** | 0.42 (0.04)*** | 0.42 (0.03)*** | |
Perpetrators (0 = other; 1 = hate groups) | -0.22 (0.03)*** | -0.21 (0.03)*** | -0.20 (0.02)*** | -0.26 (0.03)*** | -0.23 (0.02)*** | |
Level 2 predictors (contextual variables) | ||||||
Neg. int. attitudes | -0.37 (0.09)*** | -0.49 (0.18)** | -0.29 (0.09)** | -0.39 (0.12)*** | ||
Political attitudes | -0.10 (0.04)* | -0.12 (0.05)* | -0.04 (0.07) | -0.09 (0.05)† | ||
Unemployment | 0.01 (0.002)*** | 0.01 (0.003)*** | 0.01 (0.002)*** | 0.01 (0.002)*** | ||
GDP growth | 0.003 (0.005) | 0.002 (0.006) | 0.003 (0.006) | 0.004 (0.006) | ||
Cross level interactions | ||||||
Victims x Neg. int. attitudes | 0.93 (0.20)*** | |||||
Perpetrators x Political attitudes | -0.29 (0.07)*** | |||||
Random part | ||||||
Within year variability |
0.20 (0.005) | 0.138 (0.006) | 0.14 (0.006) | 0.13 (0.006) | 0.13 (0.006) | 0.13 (0.006) |
Between year variability |
0.03 (0.009) | 0.006 (0.003) | > 0.001 (0.001) | 0.001 (0.001) | 0.002 (0.002) | 0.001 (0.001) |
Variance of slopes Victims |
0.02 (0.005) | 0.003 (0.003) | ||||
Variance of slopes Perpetrators |
0.01 (0.006) | 0.005 (0.003) | ||||
Deviance | 3817.64 | 2018.01 | 1991.73 | 1943.86 | 1974.37 | 1903.45 |
†
In the second model (“Fixed Level 1 predictors” in
In the third model (“Fixed Level 2 predictors” in
In the next steps we added a random component for the type of victims and perpetrators (in two independent models; see “Random model: Victims” and “Random model: Perpetrators” in
In order to interpret the interaction effects two separate plots were created. Results showed that when right-wing political preferences were stronger in society, hate group perpetrators were more involved in direct hate crimes. No effect was observed when right-wing political preferences were relatively less prevalent in society (see
An analysis of the second interaction term showed that when negative intergroup attitudes were more widespread in society, outgroups perceived as more realistically threatening were targeted with hate crime of a more indirect type, however, when negative intergroup attitudes were relatively low the effect was opposite and those same outgroups were more likely to be victims of more direct types of hate crime (see
The main purpose of the current study was to introduce a more fine-grained approach to analyzing and understanding hate crime. Drawing insights from the psychology of violence (
The results show that indirect hate crime was slightly more prevalent than direct hate crime (50.4% vs. 40.7%). Nearly 70% of recorded hate crime perpetrators were members or associates of hate crime groups and about 60% of hate crime victims belonged to groups perceived as realistically threatening the ethnic majority. Moreover, groups seen as posing more of a realistic threat (
Changes in the GDP did not relate to the type of hate crime but greater unemployment in a given year was related to greater prevalence of indirect hate crime, which contradicted our predictions (H3). In line with our hypothesis (H4) greater prevalence of negative intergroup attitudes in society was associated with more direct forms of hate crime. Societal-level political preferences were only marginally related to a greater prevalence of direct hate crime in the final model, providing somewhat weak support for H5.
Results of our analyses support the utility of differentiating direct and indirect hate crime and contribute to the literature in several ways. To our best knowledge, the present research is the first to analyze different types of hate crime, relations among characteristics of the victims, the perpetrators, and the social context, and the characteristics of hate crime events. We did so by using a series of multilevel models and provide evidence that the two hypothesized types of hate crime may be reliably differentiated and that they are predicted by a set of explanatory variables.
We also demonstrated that hate group members were more likely to engage in direct hate crime (when compared to general population perpetrators who are more likely to commit indirect hate crimes). This is in line with research showing that hate groups openly proclaim violence against groups they deem undesirable (
The results of our analyses also showed that groups perceived as realistically threatening were more likely to be targeted with indirect types of hate crime. While this was not in line with our predictions, a possible explanation for this effect is that Poland is a very ethnically homogeneous country–over 96% of the population is ethnically Polish (
At the societal level, negative intergroup attitudes and right-wing political preferences were linked to greater prevalence of direct hate crimes. Additionally, a cross-level interaction showed that when right-wing political preferences were widespread, hate group members were more likely to be involved in direct hate crimes. We believe that this could be interpreted as an example of societal level radicalization. Specifically, when social norms become more hostile (i.e., when the societal-level intergroup attitudes become more negative and right-wing political preferences more prevalent), the extreme part of the political scene may radicalize even further. Psychologically this mechanism might be understood as an attempt to protect optimal distinctiveness: when the mainstream society moves to the right, hate groups (that are predominantly right-wing) go even further with their actions in order to remain distinct (
A significant cross-level interaction showed that more negative intergroup attitudes in society relate to groups perceived as realistically threatening becoming likely targets of indirect hate crime. At the same time, when the level of negative intergroup attitudes decreases, realistically threatening groups become targets of more direct hate crime. We suggest that these results might be related to shifting social norms and to the limited “availability” of groups seen as realistically threatening in Polish society. Overall, indirect hate crime is more likely to be perpetrated by people who do not belong to hate groups. When social norms shift towards greater intergroup hostility (stronger negative intergroup attitudes) it seems to result in an overall increase in hate crime. Specifically, when controlling for direct hate crime perpetrated by hate groups members, it can be seen that indirect violence toward groups seen as realistically threatening perpetrated by people who do not belong to hate groups is also on the rise. This finding is in line with results showing a positive relation between negative intergroup attitudes (islamophobia) and presence of hate speech (
Contrary to our predictions, a less prosperous economy, approximated by higher rates of unemployment (but not GDP which was not a significant predictor) was related to a greater prevalence of indirect hate crime. This effect could be explained by the relative deprivation theory (
The presented study is not without limitations. First, the “Brown Book” is not an exhaustive catalogue of all hate crime events. It was created by a non-governmental organization whose monitoring procedures rely on a network of voluntary informants. This means that we cannot fully rule out that there exists a bias in terms of some crimes being more likely to be recorded than others. Having said that, the system of data collection that the “Brown Book” relies on is standardized (i.e., volunteers systematically monitor local and national press and cross-check reports) and provides a more reliable picture than official hate crime statistics. Second, the numbers of crimes recorded in the early 1990s (i.e., at the beginning of the cataloguing effort) were lower than those recorded in the late 1990s and in the 2000s, making analyses using raw numbers challenging. In order to circumvent the problem of unequal frequencies we focused on the relative prevalence of different types of hate crime rather than absolute numbers. With this approach, only one basic assumption needs to be met: the procedure for collecting and processing data needs to be constant. We have no reason to doubt that this assumption was met, given the descriptions of the data collection provided in the “Brown Book” itself (
In conclusion, research presented in this article constitutes the first analysis of direct and indirect hate crime. We provide evidence that the relative prevalence of the two types of hate crime is significantly shaped by the type of target group, type of perpetrator (hate group member versus not), and a host of societal-level variables. We hope that following this approach will contribute to a better understanding of hate crime as a social phenomenon, as well as to hate crime prevention and more adequate victim support.
The reported research and preparation of the manuscript have been supported by the National Science Center of Poland Harmonia grant (2017/26/M/HS6/00114).
The authors have declared that no competing interests exist.
The authors have no support to report.
Hate crime database in Poland, the “Brown Book” compiled by the Never Again Foundation; All resources available freely on: Center for Public Opinion Research (CBOS), Political attitudes and attitudes toward minorities: Data on unemployment Data on GDP for Poland