Swedish Women’s Experiences of Misogynistic Hate Crimes: The Impact of Victimization on Fear of Crime

The overall aim of this study is to fill a knowledge gap regarding misogynistic hate crimes, since only one previous study has focused on victims’ experiences. Drawing from a sample of 1,767 female students, the results show that women with experiences of misogynistic hate crimes are more likely to be subjected to sexual harassment, repeat victimization, and to have been targeted by strangers. They consistently report higher levels of fear of crime by comparison with both non-bias victims and non-victims. Finally, the results support the thesis that misogynistic hate crime, like other forms of hate crime, has a message effect.


Introduction
It is well established that women fear crime more than men (Collins, 2016;Hale, 1996;Henson & Reyns, 2015;Madriz, 1997;Stanko, 1995;Warr, 1984). Hate crime scholars have argued that this heightened fear is in part the result of misogynistic hate crimes, defined in this study as crimes that target women due to their gender, either based on prejudice or hatred against women as a group or based on the perception of women as easy and defenseless targets (Bespinar & Canel-Cinarbas, 2017;Mason-Bish, 2015;McPhail, 2003;Perry, 2001).
Hate crimes are known to have a more pronounced impact on victims, with victims reporting higher rates of anxiety, stress, depression and fear by comparison with nonbias victims (Herek et al., 1999;McDevitt et al., 2001;Pezzella & Fetzer, 2017). Further, hate crimes have a negative effect that extends far beyond the direct victim, since the offenses spread fear and insecurity within entire minority communities and contribute to the marginalisation of particularly vulnerable groups (Bell & Perry, 2014;Iganski, 2001;Noelle, 2002;Paterson et al., 2019;Perry & Alvi, 2011). Hate crimes are therefore often referred to as message crimes.
The issue of including sex or gender as categories in hate crime legislation has been controversial, largely due to a lack of knowledge about misogynistic hate crime. Only one previous study has focused on the victims' experiences (Mason-Bish & Duggan, 2019). The results show that sexual offences are common and that there is a heightened risk among young women and among those with an androgynous or transgender identity.
Drawing on the data from a mixed method study of Swedish university students, the overall aim of this study is to fill a knowledge gap regarding victimization patterns, the victims' fear of crime, and the message effect of misogynistic hate crime. The study provides more in-depth knowledge about victimization among the risk groups identified by Mason-Bish & Duggan (2019), since this study includes both cis-and trans women at a sensitive stage of life characterized by high victimization rates. The study contributes to the literature by shedding light on the question of whether the impact on victims is in line with research on hate crime based on other motives, that is, whether the impact is more pronounced by comparison with non-bias crime and whether there is a message effect. Since uninformed conceptions about whether misogynistic hate crime fits into the hate crime paradigm have an impact on hate crime legislation in Sweden and other countries, the study makes an important contribution to the field by examining these questions more closely.

Women's Fear of Crime and Victimization
In their research review on fear of crime, Henson and Reyns describe gender as "Quite possibly the strongest and most widely-accepted predictor of fear of crime" (Henson & Reyns, 2015, p. 95). Women's fear has been found to be relatively stable over time, and consistently higher in comparison to men (Haynie, 1998;NTU, 2019). A recent meta study based on 114 studies identified gender as the strongest predictor of fear of crime, followed by own experiences of victimization (Collins, 2016). Furthermore, researchers have shown that women's fear of crime increases with threatening and intimidating interactions with male partners and male strangers. This means that fear is also affected by threatening acts that might not necessarily be illegal (Scott, 2003;Stanko, 1995).
It has been argued that it is women's fear of sexual assault and rape that increases their fear of other forms of interpersonal crime, since they are perceived to be contemporaneous offenses (Warr, 1984). Although the "shadow of sexual assault" has found empirical support (Ferraro, 1996;Fisher & Sloan, 2003;May, 2001;Özascilar, 2013;Wilcox et al., 2006), others have raised criticisms against the manner in which the theory is usually tested (Hirtenlehner & Farrall, 2014). It has been further argued that the theoretical framework is unsatisfactory, since it does not explain the process by which women develop such an extensive fear of sexual offenses (Rader & Haynes, 2011;Reid & Konrad, 2010).
Feminist scholars in the field have instead pointed to the process by which fear of crime is learned as a more fruitful source of explanation. These authors argue that women learn to fear crime by receiving messages about their alleged vulnerability from family members, friends and the media (Madriz, 1997;Rader & Haynes, 2011;Reid & Konrad, 2010;Stanko, 1995). Cultivation theory further describes how the media's focus on violent crime results in a process by which individuals adopt a skewed world view, over-estimating the risk of being victimized. Research on cultivation theory has shown that fear increases with consumption of non-fiction and local news, the extent of media consumption, and the degree of attention paid to the media (Grabe & Drew, 2007;Weitzer & Kubrin, 2004).
Consequently, through social interactions and media-consumption, women learn to think about the risk of victimization as an inherent part of the definition of what it means to be a woman. These sentiments are reinforced as women are expected to engage in protective strategies, such as avoiding being outside alone at night. Women who do not conform to these expectations might be met with blame or mistrust when victimized (Madriz, 1997;Rader & Haynes, 2011;Reid & Konrad, 2010;Stanko, 1995). For example, in their coverage of victimization, the media tend to portray women of color as being more risk-taking and as having a "bad character" by contrast with white women (Slakoff & Brennan, 2017). Such narrative practices ascribe guilt to the victim rather than the offender. It has also been established that experiences of indirect victimization -having friends with experiences of victimization -have a fundamental impact on women's fear of crime as part of this socialization process (Hale, 1996;Russo & Roccato, 2010).
The desire to explain women's fear of crime is partially based on the notion that they are less frequently victimized in comparison to men, a statement often confirmed by data from large national victimization surveys (Ringel, 1997;Stanko, 1995;Warr, 1984). In turn, this has led some to describe women's fear of crime as exaggerated, irrational and unfounded (Reid & Konrad, 2010;Skogan, 1987;Stanko, 1995).
This pattern of victimization does not hold true in the Swedish context, as women have reported higher victimization rates than men in the Swedish National Crime Survey since 2012. The most recent survey showed that the largest differences between men and women were in the lowest age group, 16 to 24 years. Half of the women reported experiences of interpersonal crime during the past year, compared to one-third of the men. Women in the same age group were also more likely to have been subjected to repeat victimization, with one-third having been victimized more than once during the past year, in comparison to one-fifth of the men. Moreover, women's fear of being outside in their own neighborhood and their worry regarding criminality in society both increased between 2012 and 2016, to plateau in the years since (NTU, 2019).

Gender and Hate Crime Legislation
The inclusion of gender in hate crime legislation has been heavily debated. The most frequent counterargument is often referred to as the "intimacy problem" exemplified in cases of domestic violence. These cases are often interpreted as being based on interpersonal conflict rather than misogyny (Bespinar & Canel-Cinarbas, 2017;Gelber, 2000;Gerstenfeld, 2018;Hall, 2013;Hodge, 2011;Iganski & Levin, 2015;Mason-Bish, 2015;McPhail, 2003). This position is also linked to a strict interpretation of hate crime legislation as being based on animus, that is, direct prejudice, hostility or hatred for the victim's group (Gerstenfeld, 2018;Hodge, 2011;Iganski & Levin, 2015;McPhail, 2003). Another argument repeatedly raised in the literature is the "overflow argument," whose proponents argue that gendered hate crime cases would overburden the justice system, since violence against women is so frequent in comparison to other forms of hate crime (Gerstenfeld, 2018;Hall, 2013;McPhail, 2003). Other arguments that have been raised include a lack of guidelines for working with misogynistic hate crime cases, and that the group "women" is too large to be given protection under hate crime law (Hodge, 2011;Mason-Bish, 2015;McPhail, 2003).
Those arguing for the inclusion of gender in hate crime legislation often position themselves in favor of the discriminatory selection model, holding that women are often targeted because they are women (Gerstenfeld, 2018;McPhail, 2003;Perry, 2001). Some also argue that the discriminatory selection of female victims, such as a partner, daughter or female friend, pre-supposes the presence of animus (Campo-Engelstein, 2016), and that male entitlement is an expression of animus (Bespinar & Canel-Cinarbas, 2017). Others hold that hate crime is a misnomer and point to the everyday nature of most hate crime cases. Consequently, they argue that the term should not be interpreted literally, since hate as we think of it is rarely present in hate crime incidents. Seen from this perspective, the inclusion of gender is unproblematic and well in line with hate crimes based on other forms of prejudice or bias (Iganski & Levin, 2015). Moreover, many argue that violence against women instills fear beyond the initial victim, leading women to change their routines and activity patterns in order to avoid victimization, much like other groups targeted by hate crime (Bespinar & Canel-Cinarbas, 2017;Mason-Bish, 2015;McPhail, 2003;Perry, 2001).

Gender and Hate Crime Legislation in Sweden
Hate crime legislation in Sweden covers acts that target victims due to their race, color, ethnicity, religion, sexual orientation, non-conforming gender identity or expression, and other similar circumstances. The offenses covered are agitation against a population group, unlawful discrimination, and any crime in the penal code that targets the victim due to them belonging to any of the groups mentioned above. The Swedish legislation thereby prohibits some acts, such as agitation against a population group, and insults based on membership of the protected categories, which are considered protected under the principle of freedom of speech in some countries. The wording at the end of the list of categories, "other similar circumstances," has led some to argue that the law could be used to enhance the penalty for a crime motivated by misogyny. There are however no cases in which the wording has been tested in court (Granström, 2019).
The question of explicitly including gender or sex in Swedish hate crime legislation has been raised on two different occasions. The first time was in 1995, when the Commission on the Integrity of Women suggested that crimes motivated by hostility against the sex of a woman should be eligible for a penalty enhancement (Granström, 2019;SOU, 1995, p. 60). The suggestion was turned down in favor of the introduction of a new offense entitled violation of a woman's integrity, defined as cases of repeat and systematic domestic violence against women (Granström, 2019).
The second time was in 2015, when legislation was proposed to include gender identity and gender expression in all sections of law under the hate crime umbrella (SOU, 2015, p. 103). The aim of this proposal was primarily to offer transgender individuals explicit protection. While the Swedish government supported the proposal (Lagrådsremiss, 2017), it was dismissed by the Swedish Council on Legislation. The Council on Legislation held that the labels "gender identity" and "gender expression" were too close to each other, and held that there was a lack of evidence supporting the need for protection that would also cover cis-women and cis-men (Lagrådet, 2017). As a result, "non-conforming gender identity and expression" was included instead, even though these labels had been subject to substantial criticism from national LGBTQ-organizations for reflecting transphobic and transhostile attitudes. It is worth noting that the other categories in Swedish hate crime legislation protect minority and majority populations alike, and that there was no demand to prove a general need for protection beyond vulnerable minority groups when race, color, ethnicity, religion and sexual orientation were included.
There is, however, disagreement regarding the lack of evidence for the prevalence of misogynistic hate crime. The Swedish Defence Research Agency and the Swedish Security Police have raised concerns regarding the rise of an incel network in Sweden. The term incel is short for involuntary celibacy, and the networks consist of men who fail to establish intimate relationships with women. These networks have become a breeding ground for deeply misogynistic ideologies. Worldwide, approximately ten terrorist attacks against women have been tracked to incel networks in the few years since their first appearance (Fernquist et al., 2020). One example is the case of Elliot Rodger who killed six and wounded 13 before committing suicide. In a text written prior to the murders and his suicide, Rodger writes extensively about violent fantasies about torturing and killing women for not wanting to have sex with him, as seen in the following citation: "Women's rejection of me is a declaration of war, and if it's war they want, then war they shall have" (cited in Langman, 2016, p. 2). His violent fantasies further extended to men with sexual experiences, since he did not want other men to have sexual access to women when he himself was rejected. Consequently, he targeted both men and women whom he perceived to be sexually attractive, something he equated to being sexually active, as a form of revenge for being unable to form romantic and sexual relationships with women (Langman, 2016;Vito et al., 2018). Mass murderers such as Rodgers often have an online presence on incel forums prior to their crimes. Consequently, a concern raised by the Swedish Defence Research Agency is that the proportion of Swedish visitors on these sites, 240 per million citizens, is larger than, for example, the United States with 43 visitors per million citizens (Fernquist et al., 2020).

Victims' Experiences of Misogynistic Hate Crime
That hate crime targeting women often has a gendered dynamic is well established in previous research. A recent example is found in Mason-Bish & Zempi's (2019) study about Muslim women's experiences of street harassment. In a combined interview and focus group study with 60 participants, they found that Muslim women perceived themselves to be targeted because they did not conform to expectations placed on women to make their bodies available for the male gaze. Another example is McCarthy's (2017) interview study about experiences of intimate partner violence among 17 women with intellectual disabilities. For this group of women, disability hate crime, "mate" crime and domestic violence all merge, as their gender, disability and relationship status with the offender all contributed to the forms of violence to which they were subjected.
Such studies make a very important contribution by clarifying and describing how various systems of oppression are expressed when the targets are minority women. There is, however, a general lack of empirical studies that focus on misogynistic hate crime per se. For the purposes of this study, misogyny is defined as "the system that operates within a patriarchal social order to police and enforce women's subordination and to uphold male dominance" (Manne, 2017, p. 33). As such, misogyny is to be understood as the practical realization of a sexist value system (Manne, 2017). Misogynistic hate crimes occur when this enforcement is expressed in criminalized acts, such as unlawful threats and harassment, physical violence, sexual harassment and rape. Misogynistic hate crimes are often corrective acts in that they are intended to put women who transgress the rigid norms of patriarchy back in their place, simultaneously enforcing the subordinate status of the victim and the dominant status of the aggressor (Perry, 2001).
McPhail (2003) raised the need for empirical studies on misogynistic hate crimes nearly 20 years ago, arguing that the literature on the subject is of a conceptual nature. Her observation remains true today. To the author's knowledge, only one study, conducted by Mason-Bish & Duggan (2019), has examined victims' experiences of misogynistic hate crime. The study consists of a survey with open-ended questions regarding the 85 participants' interpretation of hatred motivated by gender hostility. The victims' emotional responses included anger, distrust and anxiety. Sexual offenses were very common, and the victims often drew the conclusion that the incident was misogynistic since the offenders used derogatory sexual epithets. Like previous studies, Mason-Bish & Duggan (2019) also found an interplay between sexual orientation and gender, as well as a heightened risk for those with transgender or androgynous appearance. The study also indicates that age is a risk factor, as participants described a reduction in victimization frequency as they aged.
The lack of research into women's experiences of misogynistic hate crime has resulted in a policy debate without empirical roots. Though heavily debated, it remains unknown whether misogynistic hate crimes are a "stranger danger" or primarily a domestic violence problem. It is unknown whether women with experience of misogynistic hate crime suffer a greater detrimental impact compared to non-bias victims, as has been found in studies on hate crimes with other motives. Lastly, it is well established that women have a higher fear of crime than men, but it remains unknown whether this is partly the result of the message effect of misogynistic hate crime. The present study will attempt to address this lack of knowledge.

Aim
The overall aim of the study is to fill a knowledge gap regarding misogynistic hate crime, more specifically regarding victimization patterns, its impact on fear of crime, and the potential message effect. The questions examined are: 1) Is there a distinct victimization pattern among women with experiences of misogynistic hate crime? 2) Do women with experiences of misogynistic hate crime report heightened levels of fear of crime? 3) Does misogynistic hate crime have a message effect, increasing fear of crime among women in general?

Method
The present study uses data collected as part of Experiences and Exposure to Hate Crime (EEHC), a mixed-method study that examines hate crime among Swedish students at Malmö University. Data was collected in two waves. The first wave consisted of a survey study about hate crime victimization directed at all students at Malmö University who took courses of at least 15 credits during the fall of 2013. Questionnaires were distributed to 264 out of 277 classes and reached 4,649 students; 2,853 participated, which gives a response rate of 60%. The second wave consisted of an interview study with victims of hate crime and took place during the spring semester of 2014. Only the quantitative survey material will be used in the present study. For a more detailed description of the data collection process, see Andersson (2018).

Sample
For the purpose of the present study, students who identified as women were selected (N = 1,940). Women with experiences of hate crimes with other motives (targeting their religion, ethnic origin, cultural background, sexual orientation or disability), were excluded so as not to dilute or strengthen the results, leaving a final sample of 1,767 women. The mean age of the participants in the study was 25, and a demographic summary can be seen in Table 1. Most of the women are Christian or Atheist, heterosexual, middle class and have a Swedish background.
There are limitations to the present sample that the reader should bear in mind. Firstly, the sample consists of a student population and does not mirror the general population of Swedish women in general with regard to age, educational level and class background. Secondly, it relies on the victims' interpretation of the offender's motive, and might therefore be subject to classification errors. Third, the hate crime legislation in Sweden has a broad scope, and includes acts that are legal in many other countries, such as agitation against a population group. Each of these aspects involve limitations regarding the generalizability and transferability of the results to other contexts.

Operationalization
Fear of crime was measured by asking "How often, if at all, have you felt worried about being subjected to any of the following acts during the past 12 months?" The acts encompassed by the study were 1) Verbal threats and harassment, 2) Threats and/ or harassment via Facebook, Twitter, Instagram or similar, 3) Threats and/or harassment by phone or texting, 4) Being followed or chased by someone, 5) Being pushed, having your hair pulled or being hit with an open hand, 6) Being kicked, punched with a clenched fist, or with something else that could have hurt you, 7) Being threatened or attacked with a weapon, such as a knife or a gun, 8) Being exposed to unwanted physical contact carried out in an offensive and or/unpleasant way, such as touching, kissing, grabbing or fondling, 9) Being forced or threatened to have sex against your will, raped, sexually abused or subjected to similar acts. Answers were reported on a four-grade scale, from regularly (3), occasionally (2), once or twice (1), to never (0). Non-bias victimization was measured by asking about experiences of the following acts (yes/no): 1) Have you been subjected to bullying, threats, or harassment? 2) Have you been subjected to any bullying, threats or harassment on the internet, like Facebook, Instagram, Twitter or similar? 3) Have you been followed, chased, pushed, had your hair pulled, or been hit with an open hand? 4) Have you been kicked, punched with a clenched fist, or something else, like a weapon, that could harm you? 5) Have you been exposed to unwanted physical contact that was carried out in an offensive and or/ unpleasant way, for example touching, kissing, grabbing or fondling? 6) Have you been forced to have sex against your will, raped, sexually abused or subjected to similar acts? 7) Have you been exposed to robbery or burglary? Respondents were asked about experiences of these acts both during the past 12 months, and more than 12 months ago. For the purpose of the present study, these experiences were added together and thus the study uses life-time prevalence of victimization experiences coded as either having experiences of victimization or having no experiences of victimization.
Misogynistic hate crime was measured by asking "Do you believe that the incident/ any of the incidents may have been motivated by a prejudice or hostility against your sex, religion, gender identity, sexual orientation, origin/cultural background or disability? If yes, please tick all that apply to your experience during the incident/incidents." Respondents who categorized the incidents as being motivated by prejudice or hostility against their sex and/or gender identity were categorized as having experiences of misogynistic hate crime (Manne, 2017). Respondents who categorized incidents as being motivated by their religion, sexual orientation, origin/cultural background and/or disability were categorized as having experiences of hate crimes with other motives and were, as mentioned previously, excluded from the sample for the present study.
Relationship to the offender was measured by asking "What kind of relation did you and the offender have?" followed by a list of alternatives and the instruction to tick any that applied (yes/no). These included friends, acquaintances, strangers, partner, ex-partner, family member, student, neighbors, colleagues, and someone working at the university.
Repeat victimization was measured by asking how many times the respondent had been subjected to the acts mentioned previously. The relationship with the offender and repeat victimization was only measured for victimization that had taken place during the past 12 months.
Indirect non-bias victimization was measured by asking (yes/no) "Have any of your friends or acquaintances been subjected to any of these incidents? Please tick all that apply", followed by a list of the same offenses described above.
Indirect misogynistic victimization was measured by asking (yes/no) if their friends' experiences of victimization were based on prejudice or hostility towards their sex and/or gender.

Analytical Strategy
The victimization patterns of women with experiences of misogynistic hate crimes are compared to the victimization patterns of women with experiences of non-bias crime using chi-square tests (χ 2 ). Yates' Continuity Correction is presented for χ 2 -tests for two-by-two tables with Phi (Φ) as the effect size (.1 for small, .3 for medium, .5 for large). For three-by-two tables, Pearson's χ 2 is presented along with Cramer's V for the effect size (.7 for small, .21 for medium, .35 for large) (Cohen, 1988).
Comparisons regarding fear of crime are made between victims of misogynistic crime, victims of non-bias crime, and non-victims using MANOVA. Effect sizes are presented using Eta Squared (η 2 ), with the thresholds .01 for small, .06 for medium, and .14 for large.
Since demographic factors can influence both victimization risk and fear of crime, the results from the MANOVA are further tested in a stepwise regression model. The regression model tests whether victimization experiences and indirect victimization increase fear of crime, while controlling for possible confounders. The dependent variable consists of an index of the variables measuring fear of crime (alpha = .86). Initial analyzes showed the presence of outliers, for which reason cases above the 99th percentile, 14 cases, were excluded. Independent variables and possible confounders are presented in Supplemental Appendix 1.

Victimization Patterns
Of the 1,767 women in the sample, 17% reported having experienced misogynistic hate crime at some point during their life, and 37% reported experiences of non-bias crime (see Table 2). Victimization rates for the past year are slightly lower, with 13% reporting misogynistic hate crime and 24% reporting non-bias crime. The total past-year victimization rate of 37% in this study falls within the normal span for women within the same age group in Sweden, when similar measures of victimization have been used (NTU, 2019). In a previous publication from the EEHC project, a qualitative analysis was conducted focusing on how the study's interview participants came to categorize experiences as targeting their gender. The results showed that four primary factors were central: objectification, humiliation, sadism and discriminatory selection. In the case of objectification, the offender would treat the victims much like a sex toy and act as though they were entitled to their bodies. Humiliation referred to acts in which the offender used denigrating language or gestures to belittle the victim. Sadism was referred to in cases during which the offender expressed joy or excitement at, or took pleasure in, humiliating or causing the victim pain. Discriminatory selection refers to incidents in which the victim describes having been selected because they were perceived as being unable to protect themselves; these cases were more likely to involve property crimes rather than sexual offenses (Andersson & Mellgren, 2016). Thus, it was the offender's behavior and modus operandi that made the victims in this sample distinguish misogynistic hate crime from non-bias crime.
In Table 3, comparisons of past-year experiences of offenses are made between those who have experiences of misogynistic hate crime and non-bias crime respectively. Three offense types account for the large majority of victimization experiences among the women regardless of motive: 1) verbal harassment, threats and bullying, 2) being followed, chased or hit with an open hand, and 3) being subjected to sexual harassment. There are significant differences in victimization patterns between the groups despite this common trend.
The results show that women with experiences of non-bias crime are significantly more likely to have been subjected to property crimes and assault. The effect sizes show, however, that these differences are small. Women with experiences of misogynistic hate crime are, on the other hand, significantly more likely to have been subjected to sexual harassment with a medium effect size. This result is in line with previous research showing that misogynistic crimes often include threats of sexual violence or consist of various forms of sexual harassment (Mason-Bish & Duggan, 2019). Table 3 also presents comparisons in repeat victimization during the past year. The results show that women with experiences of misogynistic hate crime are more likely to have been subjected to three or more incidents, and are less likely to report only one incident, by comparison with women with experiences of non-bias crime. The differences are significant with a medium effect size. The results thus show that women with experiences of misogynistic hate crime are particularly vulnerable to victimization.
The results from Table 4 contest the widespread perception of misogynistic hate crime as primarily spouse-related, since they show that women with experiences of misogynistic hate crime are significantly more likely to have been targeted by a stranger by comparison with non-bias victims. They are also significantly less likely to have been targeted by a partner. All effect sizes are small and these results must therefore be interpreted with caution.

Fear of Crime and Message Effects
A MANOVA model was estimated, comparing fear of crime between non-victims, victims of non-bias crime, and victims of misogynistic crime. The full model shows that there are significant differences in fear of crime between the groups, with a medium effect size, with victimization experiences accounting for 12% of the variance in fear of crime: F (18, 3312) = 25.64, p = .000; Wilk's Lambda = 0.77, η 2 = 0.12. There is a linear relationship between victimization experiences and fear of crime at the aggregate level. Women with no experiences of victimization report the lowest levels of fear of crime (mean 2.15). Fear of crime is higher among the women who have experienced non-bias victimization (mean 4.31), and peaks among the women with experiences of misogynistically motivated crime (mean 6.16). These results are in line with previous research showing that victimization experiences have an effect on fear of crime (Collins, 2016). They also provide support for those who have argued that misogynistic hate crime, like other forms of hate crime, has a more detrimental impact by comparison with non-bias crime (Mason-Bish, 2015;Perry, 2001). In order to obtain more detailed information about fear of crime, followup tests on between-subjects effects were conducted to examine the differences between the groups on each variable.
The results presented in Figure 1 show that women with experiences of misogynistically motivated hate crime consistently report the highest level of fear of crime at the group level in relation to all studied offenses. The results show that these differences were significant for all offenses when comparing women with no experiences of crime with women who had experiences of misogynistically motivated hate crime. The differences between non-bias victims and victims of misogynistic hate crime are generally smaller, and significant in relation to fear of verbal and online harassment, threats and bullying, and being followed or chased, sexually harassed or raped.
The effect sizes are small in relation to fear of online harassment and bullying, phone harassment and bullying, minor assault, major assault, or being attacked with a weapon, or raped; the differences in victimization experiences only account for 2% to 5% of the explained variance in fear. The effects sizes for verbal harassment and fear of being followed or chased are medium, with victimization experiences accounting for 10% and 6% of the variance in fear respectively. Finally, the effect size for sexual harassment is large. Here the differences in victimization experiences account for 19% of the variance in fear, a notably large proportion considering the many factors that contribute to forming fear of crime.
It is also notable that the curves in Figure 1 follow one other for each offense. All groups of women report the highest levels of fear in relation to verbal threats and harassment, fear of being followed or chased, and fear of sexual harassment. Similarly, all groups report lower levels of fear with regard to phone or online harassment and bullying, minor and major assault, and the fear of being attacked with a weapon. A stepwise regression model was estimated to examine whether victimization experiences and indirect victimization increase fear of crime, while also controlling for possible confounders. The results are presented in Table 5. The first model includes background variables that might act as confounders. In the second model, the participants' own experiences of victimization were also included, while experiences of indirect victimization were added in the final model. The explained variance for fear of crime increases from 7% in Model 1, to 20% in Model 2, and 22% in Model 3. Consequently, the respondents lived experiences of crime had the strongest impact on their fear of crime.
While the size of the participants' hometown has a significant impact in Models 1 and 2, this variable loses its significance in Model 3. The other three variables -foreign background, sexual minority status, and age -remain significant across all three models. It should be noted that the correlations for foreign background and age are negative, which means that Swedish women are more fearful than women of immigrant background, and that young women are more fearful by comparison with older women.
The results show that victimization experiences, with the exception of robbery and burglary, lead to higher levels of fear of crime in both Models 2 and 3. The positive correlation shows that women with no victimization experiences report low levels of fear of crime, that fear increases among women with experiences of non-bias crime, and that it peaks among women with experiences of misogynistic hate crime. As such, the results show that lived experiences have a profound impact on women's fear of crime.
Finally, the standardized b-coefficients in Model 3 show that the strongest predictor of fear of crime is having friends with experiences of misogynistic hate crimes. The results therefore support the work of previous authors who have argued that misogynistic hate crime has a message effect, leading to a general increase in fear of crime beyond the initial victim (Bespinar & Canel-Cinarbas, 2017;Mason-Bish, 2015;McPhail, 2003;Perry, 2001), while also showing that crimes against women increase levels of fear regardless of motive.
In summary, the results show that women with experiences of misogynistic hate crime are more likely to be subjected to sexual harassment and repeat victimization, and more likely to be targeted by strangers. Women with experiences of misogynistic hate crime consistently report higher levels of fear of crime compared to both non-bias victims and non-victims. Finally, the results support the thesis that misogynistic hate crime, like other forms of hate crime, has a message effect.

Victimization Patterns Among Victims of Misogynistic Hate Crimes
The degree to which misogynistic hate crimes are comprised of domestic violence cases has been heavily debated (Bespinar & Canel-Cinarbas, 2017;Gelber, 2000;Gerstenfeld, 2018;Hall, 2013;Hodge, 2011;Iganski & Levin, 2015;Mason-Bish, 2015;McPhail, 2003). The results of the present study show that while misogynistic hate crimes sometimes occur in intimate relationships, they are largely the result of men subjecting women they do not know to various forms of sexual harassment. In general, the results show that it is overly simplistic to equate violence against women with misogynistic hate crime. For incidents to be perceived as misogynistic hate crime, they need to contain one or more of these elements: the use of derogatory language that targets the victim's gender, objectification, humiliation, sadism and/or discriminatory selection (Andersson & Mellgren, 2016;Mason-Bish & Duggan, 2019). Consequently, misogynistic hate crimes can be motivated by both animus and discriminatory selection, meaning that gender can be included in hate crime legislation regardless of which of these legal traditions are practiced in a given jurisdiction. More than half of the women subjected to misogynistic hate crime during the past year had been targeted three times or more. Previous research indicates that women who transgress norms are at particularly high risk of becoming targets of misogynistic hate crime (Mason-Bish & Duggan, 2019;Mason-Bish & Zempi, 2019;McCarthy, 2017). Research from the Swedish context also shows that women are met with misogyny when they do not conform to conservative gender stereotypes (Andersson & Mellgren, 2016). These are both likely explanations for the pattern found in this study, of a high level of vulnerability being limited to a smaller group of women in the sample. Future research should focus on identifying vulnerability factors among this frequently victimized group.
The results show that the victimization rates among women are generally high: over half of the respondents have experienced crime at some point during their lives, and one-sixth have experiences of misogynistic hate crime. Some opponents to the inclusion of gender in hate crime legislation have argued that these high victimization rates would overwhelm the justice system (Gerstenfeld, 2018;Hall, 2013;McPhail, 2003). At the same time, reporting rates remain low in places where such laws have been implemented, and the application of penalty enhancement provisions is extremely rare (Hodge, 2011;McPhail, 2003).
The problem of misogynistic hate crimes does not disappear when these offenses go under the radar, either as a result of a lack of legislation or low reporting and detection rates. Instead, these facts -high victimization rates combined with low reporting rates and legal application -call for action on the part of policy makers to both increase the detection of such crimes and to expand the competence and capacity of the justice system to handle these cases. While critics of such legislation have pointed to a lack of guidelines (Hodge, 2011;Mason-Bish, 2015;McPhail, 2003), promising guidelines for the prosecution of misogynistic hate crime cases have been presented (Bespinar & Canel-Cinarbas, 2017;McPhail, 2003;Walters & Tumath, 2014).

Misogynistic Hate Crimes and Fear of Crime
The results show that women with experiences of misogynistic hate crimes have heightened levels of fear of crime, and that they are most afraid of sexual harassment and of being followed or chased. These results are in line with research on hate crime based on other motives, which shows that victims report more severe impacts when compared to victims of non-bias crime (Herek et al., 1999;McDevitt et al., 2001;Pezzella & Fetzer, 2017).
Further, the crime that women are most often subjected to is sexual harassment, and many are repeatedly victimized. Therefore, it might not be fear of rape that drives up the fear of different types of crime, as proposed by Ferraro (1996). Instead, it might be the continuous exposure to "minor" incidents of sexual harassment that leads women to rationally conclude that violations of their physical integrity are to be expected (Scott, 2003;Stanko, 1995). As such, the results show that women's fear is based on rational rather than irrational grounds (Skogan, 1987;Stanko, 1995;Reid & Konrad, 2010).
It has been shown that the media, by over-reporting violent and serious crimes, contribute to women's fear of crime by creating a distorted world view among the public (Grabe & Drew, 2007;Weitzer & Kubrin, 2004). Consequently, future research should examine the interaction between media-consumption and fear of crime among victims of misogynistic hate crime. Such studies should also take into account the fact that female victims who belong to racial minorities are more likely to be stigmatized and blamed in media reporting (Slakoff & Brennan, 2017) and that there might be community-specific impacts as a result of these practices. There is a need for critical self-reflection and evaluation within the media sector regarding the racial and gendered narratives that surround women's victimization, as old stereotypes still dominate this reporting.
The victimization rate in the present study is higher by comparison with many other studies, but is in line with larger Swedish victimization surveys for the appropriate age group (NTU, 2019). In part, this might be due to differences in definitions of crime between countries, and the fact that victimization studies sometimes exclude harassment and threats in order to focus on violent crime. Such studies will consistently over-report the victimization rates of men, who are more likely to be victims of physical assault, while under-reporting the victimization rates of women, who tend to be subjected to various kinds of threats and harassment.
Moreover, the fact that women report increased victimization rates in Sweden is also an important issue for future research. It has been argued that the previously lower levels of victimization among women were the result of lifestyle factors, such as women changing their daily routines and activity patterns to avoid victimization (Reid & Konrad, 2010;Skogan, 1987;Stanko, 1995). These changes might therefore paradoxically be the result of increased gender equality in activity patterns (Gracia & Merlo, 2016). An important area of study here is whether equal exposure to risky settings, such as night clubs, poses the same risk for women and men, or whether women are at higher risk for victimization when utilizing their right to freedom of movement in the public sphere (Gracia & Merlo, 2016;Warr, 1984). If participating as a full member in society comes at the price of victimization for women, this is a serious democratic threat. Finally, it should be examined whether the increased victimization rates among Swedish women are linked to the rise of an online incel network in Sweden (Fernquist et al., 2020). There is also a widespread lack of both qualitative and quantitative research on women's experiences of misogynistic hate crime. Studies from different contexts and population groups are sorely needed, as well as in-depth studies on victim impact.

The Message Effect of Misogynistic Hate Crime
The message effect of hate crime is well known and has been presented as one of the core dynamics that distinguishes hate crimes from non-bias crimes (Bell & Perry, 2014;Iganski, 2001;Noelle, 2002;Paterson et al., 2019;Perry & Alvi, 2011). The results of the present study support the claim that misogynistic hate crimes also have a message effect (Bespinar & Canel-Cinarbas, 2017;Mason-Bish, 2015;McPhail, 2003;Perry, 2001), since the study shows that having friends with experiences of misogynistic hate crime results in higher levels of fear of crime than having friends with experiences of non-bias crime. In the final model of the regression analysis, having friends with experiences of misogynistic hate crime stands out as the strongest predictor of fear of crime. The post-hoc tests from the MANOVA further show a clear trend in fear of crime: having experiences of direct victimization has a clear impact on the level of fear, but the overall pattern of peaks and dips in the fear of different offenses is consistent regardless of victimization experiences. As such, the study supports the notion of fear of crime as a social construct learnt both through one's own experiences and through interactions with victims, as proposed by previous authors (Madriz, 1997;Rader & Haynes, 2011;Reid & Konrad, 2010;Stanko, 1995).
Previous research has also shown that the message effect of hate crime leads to an increased sensitivity to external stressors. In Paterson et al. (2018), the participants read a mock news article about a hate crime incident targeting a person belonging to the same minority group as the participant. Those who had friends that were direct victims of hate crime reported higher levels of anxiety and anger. In the light of these results, it is important to further study how the media reporting of hate crime affects fear and the sense of security among minority groups. Studies should also further examine the increased sensitivity that comes with being both a vicarious and a direct victim of hate crime.

Limitations
While young women are the most common victims of interpersonal crime in Sweden, the women who participated in this study come from relatively privileged backgrounds; most are heterosexual, middle class, Christian or atheist, and of Swedish background. Women in the same age span who also belong to more stigmatized groups are more likely to be victimized, and older women are less likely to be victimized. Generalizations regarding victimization patterns should be limited to women of similar backgrounds and cultural contexts. Furthermore, Swedish legislation covers acts that are legal in other countries. Highly relevant examples are acts of hate speech that are sometimes protected under the principle of freedom of speech, and the Swedish legislation on sexual harassment and rape uses broad definitions of these crimes, and thus includes incidents that might not be illegal in other countries. This also poses limitations to the transferability of the results to other contexts.
Moreover, all studies that rely on the victims' interpretations of the offender's motives might be subject to classification errors. On the other hand, research also shows that hate crime offenders tend to think of themselves as unbiased and formulate stories in which they can deny the bigoted aspects of their own actions (Gadd, 2009).
Finally, the reliance on quantitative methods is a limitation of the current study. Qualitative studies into women's fear of crime and experiences of misogynistic hate crime would make an important contribution by offering a context to the quantitative findings of the present study. More specifically, interview studies with a longitudinal design, (Gålnander, 2020), trajectory analysis (Garcia-Hallett, 2019), or a comparative studies (Gueta & Chen, 2019) can shed further light on the processes that form women's fear of crime.

Conclusions
The present study shows that women are at high risk for victimization, and that these experiences contribute to increasing women's fear of crime. Women's fear of crime has been described as "one of the most oppressive and deceitful sources of informal social control of women" (Madriz, 1997, p. 393). Policy makers in Sweden need to recognize women as a high-risk victimization group; there is an urgent need to study the increased victimization rates among women in order to develop preventive measures. Policies to prevent women from being subjected to crime, both with and without hate motives, need to focus on strengthening women's constitutional rights to freedom of expression and movement.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The data collection was co-funded by the Swedish Crime Victim Authority and Malmö University.

Supplemental Material
Supplemental material for this article is available online.