Violent and non­violent crimes against sex  workers : the influence of the sex market  on reporting practices in the United  Kingdom Connelly, LJ, Kamerade, D and Sanders, T http://dx.doi.org/10.1177/0886260518780782 Title Violent and non­violent crimes against sex workers : the influence of the  sex market on reporting practices in the United Kingdom Authors Connelly, LJ, Kamerade, D and Sanders, T Type Article URL This version is available at: http://usir.salford.ac.uk/id/eprint/46973/ Published Date 2018 USIR is a digital collection of the research output of the University of Salford. Where copyright  permits, full text material held in the repository is made freely available online and can be read,  downloaded and copied for non­commercial private study or research purposes. Please check the  manuscript for any further copyright restrictions. For more information, including our policy and submission procedure, please contact the Repository Team at: usir@salford.ac.uk. mailto:usir@salford.ac.uk 1 Violent and non-violent crimes against sex workers: The influence of the sex market on reporting practices in the United Kingdom Abstract Previous research has shown that sex workers experience extremely high rates of victimization but are often reluctant to report their experiences to the police. This paper explores how the markets in which sex workers operate in the United Kingdom impact upon the violent and non-violent crimes they report to a national support organization and their willingness to report victimization to the police. We use a secondary quantitative data analysis of 2,056 crime reports submitted to the UK National Ugly Mugs (NUM) scheme between 2012 and 2016. The findings indicate that although violence is the most common crime type reported to NUM, sex workers operating in different markets report varying relative proportions of different types of victimization. We also argue that there is some variation in the level of willingness to share reports with the police across the different sex markets, even when the type crime, presence of violence, and other variables are taken into account. Our finding that street sex workers are most likely to report victimization directly to the police challenges previously held assumptions that criminalization is the key factor preventing sex workers from engaging with the police. Key words: sex work; violence; policing; reported victimization Facilitated by processes of globalization, the commercial sex industries have in recent years grown in size, diversified in nature, and emerged as a more visible part of the global economy (Anon & Author C, 2010). These changes to the global sex industries have given rise to a growth in the levels of violence sex workers1 experience. Indeed, it has become almost axiomatic for researchers and practitioners alike to lament the enduring levels of violence against sex workers (Aids Alliances, 2017; Deering et al., 1 In this article, we use the term ‘sex worker’ to describe those engaged in commercial sex work since it is the term commonly used by National Ugly Mugs and its membership, including its self-identifying sex worker members. It is important to recognise, however, that the language one uses to describe the global sex industries and those working within them is not politically neutral. Although the term ‘sex worker’ is contentious – with some feminists preferring instead to use the terms ‘prostituted women’ and/or ‘person engaged in commercial sex work’ – we use it here to acknowledge the global movement for the (labor) rights of sex workers. 2 2014; Lyons et al., 2015). Some feminists assert that prostitution is a form of violence against women (Jeffreys, 1997; Raymond, 1998), and buyers of sex demonstrate more sexually aggressive values and behaviours than men who do not purchase sex (Farley et al., 2015). Other scholars – many of whom view sex work as a legitimate form of labour – reject the view that the sex industry is inherently violent. Instead, they consider high levels of victimization to be the product of state regulation, particularly policies which criminalize clients and/or other aspects of the sex industry (Ellison, 2015; Krüsi et al., 2014; Levy & Jakobsson, 2014). They contend that the majority of sex workers are rational agents who exercise choice and display resilience within the structures exerted upon them (Author A et al., 2016; Shdaimah & Leon, 2015), and only a minority of clients are violent (Kinnell, 2008). While these debates foreground much of the research into the lived experiences of sex workers, there exists agreement that sex workers are often perceived by offenders to represent ‘easy targets’ (Canter, Ioannou & Youngs, 2016; van der Meulen & Valverde, 2013). Many offenders operate with the belief that sex workers will not report their victimization to the police or if they do, the police will not take their report seriously (Hubbard, 2006; Krüsi et al., 2014). The criminalization of the street sex market in the UK is understood to hinder further the reporting and investigation of crimes against sex workers (Kinnell, 2004), whilst its spatial and environmental conditions shape experiences of violence (Maher, Pickering & Gerard, 2013). Research has therefore established that street sex workers experience higher rates of violence than those working in indoor markets (Church et al., 2001; Kinnell, 1993; Stalans & Finn, 2016). In this article, we present findings from secondary quantitative bivariate and multivariate data analyses of 2,056 crime reports submitted between 2012 and 2016 to the UK National Ugly Mugs (NUM) scheme. NUM is a non-profit organisation which collates reports of victimization from sex workers and shares information about dangerous clients across its membership and/or with the police. This paper seeks to answer two key questions: i) How do the different sex markets in the UK effect what crimes are reported to NUM? ii) What effect do the different sex markets in the UK have on sex workers’ willingness to report victimization to the police? This paper makes an important empirical contribution. Whilst previous research has focused largely on localised samples of street sex workers to explore experiences of violent crime, this paper is 3 this first to use a large, national sample to explore reporting practices around both violent and non- violent crime, as well as reported victimization across a range of sex markets. In so doing, we make a theoretical contribution by encouraging academic understandings of sex workers to move beyond generalisations about violent trends and instead, to look at the nuanced differences that arise in the reporting behaviours of sex workers operating across different sex markets. We argue that there is a significant relationship between sex market and the type of crime reported, with sex workers in different markets reporting varying relative proportions of different types of victimization. We also argue that street sex workers are more likely to report their victimization to the police than sex workers operating in other sex markets, regardless of the type or severity of the crime they report, whether the crime is violent or non-violent in nature, the number of perpetrators involved, or the gender of the sex worker. It therefore suggests that the police, and practitioners working to support sex workers, need to review their ‘one-size-fits-all’ approach to addressing violence against sex workers. Instead, their responses must recognise that even sex workers working legally in the UK experience significant barriers to reporting victimization. The paper begins by reviewing existing literature, detailing what is already known about reported crimes against sex workers. Next, it describes the NUM scheme and the methodology used in this study. We then examine the results, before presenting a discussion that focuses on the centrality of the sex market as a key variable which can be related to the type of crimes that sex workers report to NUM and their willingness to engage with the police. Current Perspectives on Sex Worker Victimization Research shows that the extent of physical, sexual and emotional violence against sex workers is extremely high, both in the UK and globally. In a systematic review of research evidence from 41 peer- reviewed articles, Deering et al., (2014) reported that workplace violence is experienced globally by 45-75% of sex workers over their lifetime. In the UK, Kinnell (1993) found that of the 110 women involved in sex work who she interviewed in Birmingham, 68% had experienced violence during the commission of their work. Research carried out in another large city in the UK, Liverpool, found that 80% of sex workers had experienced work-based violence (Campbell & Stoops, 2010). A similar 4 finding emerged from a survey of 240 sex workers, this time across three cities in the UK, which found that 63% had experienced violence in their lifetime (Barnard et al., 2002). Mortality rates are also significantly greater amongst sex workers than in the general population (Goodyear & Cusick, 2007), with one UK-based study indicating that female sex workers are twelve times more likely to be murdered than women from the general population (Ward, Day & Weber, 1999). When compared to homicide rates amongst women working in bars or catering – occupations believed to have high mortality ratios – sex working women are approximately five times more likely to be murdered (Author C et al., 2017). Serial killers are also known to have targeted women on the basis of their involvement in the sex industry, such as the ‘Green River Killer’ in the US in the 1980s and 1990s, and the ‘Ipswich Murderer’ in the UK in 2006. Yet levels of violence are not experienced equally across all sectors of the sex industry. The smallest of the UK sex markets, the outdoor/street market is known to be at very high risk of violence. Comparatively, indoor markets – which include those working in brothels, saunas and parlours, as well as sex workers working independently – are understood to be less ‘risky’ (Kinnell, 2004, 2008; O’Neill et al., 2008; Phipps, 2013). In the aforementioned survey conducted across three UK cities, for example, 50% of street sex workers had experienced violence in the past six months compared to 26% of indoor workers (Church et al., 2001). Surratt et al., (2004) contend that street sex workers are engaged in a social milieu wherein subcultures of violence are commonplace: violence is perpetrated not only by clients but also ‘pimps’, pseudo-clients, vigilantes and other sex workers. They also experience higher levels of stigmatization than their indoor counterparts, even in countries such as New Zealand where sex work is decriminalized (Armstrong, 2016). Although rates of violence are generally lower in the indoor markets, there is a greater risk of robbery (Brooks-Gordon, 2006), which may be due to the greater likelihood of monies kept on the property. Author C (2005) notes, however, that a professionalized occupational culture exists within indoor markets, which operates in part to minimize harm. An additional factor that works to minimize violence for indoor sex workers is that they may form stronger bonds with their clients than street sex workers. Although, as Phipps (2013) points out, this increases their risk of stalking, with a small minority of clients developing problematic fixations on 5 indoor sex workers. Of course, sex workers’ experiences of violence are not only mediated by the sex market in which they operate but also by factors such as class, nationality, race and ethnicity, immigration status, gender and sexuality (Bernstein, 2007; Feis-Bryce et al., 2015; Scott et al., 2005). Accurate estimates of the prevalence of violence against sex workers are no doubt difficult to obtain. Many incidents of violent crimes against sex workers are not reported to the police, meaning an unknown proportion comprises the dark figure of crime (Church et al., 2001). Dodd’s (2002 in Toynbee Hall, 2009) survey of 110 female street sex workers, for example, highlighted that 69% report ‘no or hardly any’ incidents of victimization to the police. Some of the same factors that discourage the public from reporting victimization to the police apply to sex workers (Crime Survey for England and Wales, 2017): they assume the police will not care about or believe their claim to victimization; the police will not take their claim seriously; and/or the criminal justice system will be ineffective in prosecuting offenders (Barnard et al., 2002; Dodd, 2002 in Toynbee Hall, 2009). Rape myth attitudes, which assume that sex workers have given up their right to refuse sex, are common towards sex workers (Monto & Hotaling, 2001; Miller & Schwartz, 1995) and may serve as an additional barrier to reporting victimization. In countries where sex work is (partially) criminalized, sex workers may also be deterred from reporting victimization to the police by their fear of arrest, the arrest of their co-workers or clients, the closure of the premises in which they work, and/or public identification (Boff, 2012; Brooks- Gordon, 2006; Campbell, 2015). Enduring societal attitudes of intolerance and a culture of distaste towards sex workers, particularly those working on the street, may also serve to deter the reporting of victimization (Author C & Anon, 2007; Surrat et al., 2004). Indeed, for Lowman (2000), society constructs sex workers through a ‘discourse of disposability’ as non-citizens. This negative construction of sex workers is promulgated by core institutions such as the police and the media. It appears in official discourses and is expressed by much of the public, and operates to ‘Other’ sex workers by positioning them as immoral, dangerous and dirty. Some studies have employed self-report research methods to circumvent the problems inherent to police recorded crime reports (Barnard et al., 2002; Church et al., 2001; Campbell & Stoops, 2010; Kinnell, 1993). These studies have successfully drawn attention to prevalence of violence against sex 6 workers, differences in levels of violence between sex markets, and the poor relationship between sex workers and the police. Yet they are based on small (convenience) samples and/or have concentrated on particular regional locations. This article is the first to use a large, national data set to explore reported violent and non-violent crimes across different sex markets. While previous research has focused on rates of violence and the nature of the violence sex workers experience, this paper examines reported victimization. Indeed, sex workers’ decisions about whether and what to report – their reporting practices – tell us a great deal about their trust in support organisations and the police. This article makes a theoretical contribution by extending the discussion from its traditional focus on violence in the street sex market to examine reported violence alongside other forms of reported victimization, across a range of sex markets. We therefore contribute to the literature on violence against sex workers (and the broader field of violence against women) by responding to two significant gaps in previous analysis: i) how the sex market may influence violent and non-violent crimes reported to a national support organization; ii) how the sex market may influence sex workers’ willingness to engage with the police. We address these gaps empirically by using national data to highlights nuanced variations in the type of reported victimisation and willingness to engage with the police by sex market. Methods Data: National Ugly Mug Crime Incident Reports This paper uses a secondary quantitative analysis of crime incident reports submitted to the UK National Ugly Mugs (NUM) scheme over a four-year period between 2012 and 2016. With few exceptions (n =10), the crime reports submitted referred to incidents that had taken place within the sampled four- year period. Founded by the UK Network of Sex Work Projects (UKNSWP) in 2012, NUM is a non- profit organization initially funded by the Home Office for a one-year pilot. Members sign up to the scheme to facilitate the sharing of information about dangerous clients across NUM’s membership. They can confidentially report victimization to NUM and this information is collated in one central database, from which members receive legally compliant email or text alerts from NUM about potentially dangerous clients. In December 2016, the membership of NUM was at an all-time high: 3,514 individual members; 539 projects offering frontline support to sex workers around sexual health, 7 wellbeing, housing, the law, and social welfare; and 221 escort sites/forums or indoor sex establishments. One of the key objectives of NUM is to support sex workers to report their victimization to the police, in order to improve historically low rates of engagement in the criminal justice process. If the victim consents, NUM may share anonymous intelligence with the police through their national intelligence systems. The reports members submit to NUM capture information about: where and when an incident occurred, what happened, whether the incident involved violence, the perpetrator(s), and whether the member would like to report it to the police. The data the researchers received from NUM was in the form of a Microsoft Access database, which had been compiled by NUM staff from the crime reports submitted by NUM members. To protect the anonymity of members, any identifying characteristics (including age, race and ethnicity) had been removed from the data by NUM. Informed consent to use the data was obtained by NUM, with all NUM members notified before submitting a crime report that anonymized data may be used for research purposes.2 Sample A total of 2,227 reports were submitted to the NUM database between 16 July 2012 (the date the first report was submitted to NUM) and 21 July 2016. The sample for this research project was 2,056 separate crime reports since 171 reports were excluded: 166 reports (7.5%) in which the sex worker did not specify the sector in which they worked; and five reports (0.2%) about sex workers working across multiple sectors. Because of the lack of other variables describing the sex market, no imputation of missing values for these cases was used. Of these sampled crime reports, the majority were submitted to NUM by sex worker support projects (73.3%). Considerably fewer were submitted by sex workers themselves (21%), the police (4 %) or sex work establishments (0.9%). In a few cases the crime incident was reported jointly by the police and a support project (0.1%), and in some cases (0.6%) the reportee was anonymous or not specified on the crime report. 2 This project received ethical approval from the University of Leeds on 8 July 2016 [Ref: LTSSP-029] 8 Variables3 The ‘sex market’ variable described the type of sex work that the victim identified as engaged in on the crime report submitted to NUM. It was a categorical variable coded as: 0 - street sex market; 1 - traditional indoor sex market (i.e. brothel, sauna or parlour); 2 - escort workers operating through an agency; and 3 - independent or private sex workers. The victim could select more than one category, but only five victims did so. For clarity of analysis, these cases were excluded from the dataset. The ‘type of crime’ variable described the type of reported victimization as identified by the person completing the report. The ‘type of crime’ was recorded as a series of binary variables, each with two categories: 0 - no; 1- yes. The types of crime originally recorded included ‘violence’, ‘rape’, ‘attempted rape’, ‘robbery’, ‘attempted robbery’, ‘fraud’, ‘hate crime’, ‘stalking and harassment’, ‘condom removal’, and ‘refusal to wear a condom’. For the analyses in this paper, ‘rape’ and ‘attempted rape’, ‘robbery’ and ‘attempted robbery’, ‘refusal to use a condom’ and ‘removal of a condom’ were recoded together into three variables: ‘rape/attempted rape’, ‘robbery/attempted robbery’, and ‘condom refusal/removal’ as in the vast majority of cases (97-99%) both types of crime were reported for the same incident. Each crime incident could be reported in more than one of these categories: violent rape, for example, could be categorised by the reportee as both ‘rape’ and ‘violence’ and therefore, a single incident could be classified as more than one type of crime. All reported combinations of the type of crime were explored at the beginning of the analysis (see the results section). As this paper focuses on violence in crimes against sex workers, it also used a variable ‘presence of violence’ which measured whether violence was part of a crime incident reported. This variable had three categories: 0 - no violence involved; 1 - only violence; 2 - violence and another crime. The ‘number of perpetrators’ variable was an ordinal level of measurement, which described the number of perpetrators that were reported as being involved in the crime. Twenty-six cases (1.2% of the sample) were excluded from the analyses as the number of perpetrators was reported either as zero or as ‘unknown’. This was the only control variable with missing values. Pairwise exclusion (available 3 For the descriptive statistics for the variables used in the analyses see Appendix A. 9 case analysis) was used, excluding the cases only if they were missing the data required for the specific analysis. This exclusion method was used instead of imputing missing values because the proportion of missing values was low and our sample was too small for sophisticated imputation methods, such as maximum likelihood and the expectation-maximisation algorithm to produce normal, unbiased and efficient estimates (Jamshidian, 2009). The ‘report to police, anonymous consent’ variable measured whether the sex worker reporting the crime had given their consent for the crime report and their personal details to be shared anonymously by NUM with the police. This was a dichotomous variable: 0 - no anonymous consent given; 1- anonymous consent given. Anonymous consent meant that all identifying details were removed by NUM staff before any information was passed onto the police force local to the area in which the reported crime took place. The ‘report to police, full consent’ variable measured whether the sex worker reporting the crime had given their full consent for the report to be shared with the police. This variable was dichotomous: 0 - no full consent given; 1 - full consent given. Full consent was selected on the crime report if the victim had either already reported or intended to report the crime to the police themselves. Even when full consent was selected, NUM never shared full details about the victim with the police but rather, facilitated engagement between the two parties. ‘Gender of the victim’ measured the self-defined gender of the victim. It had four categories: 0 - female; 1 - male; 2 - transgender (male to female); 3 - not identified. None of the sex workers reporting to NUM identified as transgender (female to male). Method of Data Analysis To examine relationships between two categorical variables – for example, ‘sex market’ and ‘type of crime’, or between ‘sex market’ and ‘report to police, full consent’ – a chi-square test was used. To explore differences in the number of perpetrators a Kruskall-Wallis test was used because the number of perpetrators was an ordinal variable. Finally, logistic regression analysis was used to examine the relationships between the ‘sex market’ and the likelihood of the crime being reported to the police, after 10 taking into account: the type of crime, presence of violence, gender of the sex worker, and number of perpetrators. Results Sex Market: Victim Type and Gender More than half (60.9%, n=1,253) of the sampled crimes reported to NUM were committed against sex workers operating in the street sex market. Independent sex workers were the second most likely group to report a crime incident to NUM (29.7%, n=611), followed by sex workers working in indoor sex establishments such as brothels, saunas and parlours, whom accounted for 5.9% (n=121) of all reports. The numbers reporting victimization to NUM was lowest amongst those escorting through an agency (3.5%, n=71). In terms of gender, the vast majority (93.9%) of the victims defined as ‘female’. Only 4.3% of victims defined as ‘male’ and only 1.3% defined as ‘transgender m-f’, whilst 0.5% did not specify their gender. Types of Crime and the Presence of Violence Violence was present in around half (46.1%) of the all crime incidents reported to NUM. Violence on its own was reported in 15.4% of incidents, while in 30.6% of cases, violence was recorded alongside other crimes. A total of 168 different unique combinations of crime types were recorded in the sample. Table 1 presents the 31 combinations reported in ten or more cases. As can be seen, violence on its own was the most commonly reported type of crime against sex workers. Other frequently reported crimes included stalking and harassment, robbery with or without violence, and crimes categorised to NUM by the reportee as ‘other crime’. Details about these ‘other crimes’ were provided by some reportees as free-text but were removed from the dataset to protect sex workers’ anonymity. Other combinations of crime were less common. This paper did not proceed with the analysis of all possible combinations separately as such analysis would lead to low statistical power and increase the likelihood of Type II error – the likelihood of not detecting differences that actually exist. Sex Market and Type of Crime 11 As can be seen in Table 2, the numbers reporting different types of victimization varied within the overall number of reports. Crimes involving violence were considerably more likely to be reported than other crimes; they constituted 46.1% of all reports in the sample. Other types of crime were reported less frequently; they represented only approximately one-fifth of all reports. Condom removal or refusal was the least reported crime in the sample. Table 2 also shows that the relative proportions of the different types of victimization reported to NUM varied considerably by sex market. Street workers reported the greater number of violent incidents, followed by sex workers working in traditional indoor establishments (brothels, saunas and parlours). Those escorting through an agency and independent workers were considerably less likely to report violence. These differences in reported violence among the different sex markets were statistically significant ((X 2 (3, N= 2,056) =194.7, p<0.001)). Traditional indoor sex workers – those working in brothels, saunas and parlours – reported the greatest number of incidents of robbery or attempted robbery but independent sex workers were least likely to report this type of crime. Again, the differences in reported robbery among the various sex markets were statistically significant ((X 2 (3, N= 2,056) =30.5, p<0.001)). Independent sex workers reported considerably more stalking and harassment than sex workers operating in any other sex market. Approximately a third of all crime reports by independent sex workers were related to stalking and harassment, compared only to 13-15% of the reports from sex workers operating in other sex markets. These differences were statistically significant ((X 2 (3, N= 2,056) =98.3, p<0.001)). Street sex workers and agency escort workers reported the highest number of hate crime incidents; a fifth of all crime reports submitted by these workers were related to hate crime. The fewest reports of hate crime were submitted by sex workers operating in the traditional indoor market. The differences in reported hate crime by sex market were statistically significant ((X 2 (3, N= 2,056) =13.0, p=0.03)). 12 Agency escorts and independent sex workers were significantly more likely to report fraud than sex workers operating in the street sex market or in traditional indoor establishments ((X 2 (3, N= 2,056) =58.5, p<0.01)). Sex workers operating in three sex markets – traditional indoor, agency escort, and independent – were significant more likely to report removal or refusal of a condom than sex workers operating on the street ((X 2 (3, N= 2,056) =19.2, p<0.001)). Escorts and independent sex workers were also significantly more likely than street and traditional indoor sex workers to report other types of crime. There were no significant differences among the different sex markets in reported rape or attempted rape, and sexual assault. Figure 1 shows that sex workers operating in the street sex market were the least likely of all sex workers to report non-violent crimes. In contrast, independent sex workers were most likely to report non-violent crimes. It can also be seen that for crimes involving only violence, reports were considerably more likely to involve sex workers operating in the traditional indoor market and in the street sex market. Finally, street sex workers were most likely to report crimes involving both violence and other types of crime. The relationships between sex market and the presence of violence in reported crime was statistically significant ((X2(6, N= 2,056) =203.01, p<0.001)). Reporting to the Police According to Figure 2, sex workers were considerably more likely to give their consent for the crime report to be shared anonymously with the police than they were to give full consent to report to the police. On average, in nine out of ten cases, the sex worker had given anonymous consent, whilst only in two out of ten cases full consent had been given. Their willingness to give consent also varied significantly by sex market. In general, sex workers operating in the street market were most likely to give both anonymous and full consent for the crime report to be shared with the police. Independent sex workers were the least likely to give both types of consent. The differences between the sex markets 13 in reports to the police were statistically significant: for anonymous consent ((X 2 (3, N= 2,056) =49.9, p<0.001)) and for full consent ((X 2 (3, N= 2,056) =77.9, p<0.001)). In order to examine whether some groups of sex workers are more likely to report their victimization to the police, regardless of the type of the crime they report, or whether the crime is violent in nature, a logistic regression analysis was employed. Four regression models were performed (see Table 3). Models 1 and 3 explored the relationships between type of sex market and reporting a crime to police anonymously and with full consent, taking into account whether violence was present. Models 2 and 4 examined the relationships between type of sex market and reporting a crime to police depending on the crime type reported. All models control for gender and number of perpetrators. As can be seen in Table 3, the type of sex market where the victim works was related to their willingness to report to the police, even after other variables are taken into account. The negative regression (B) coefficients in Table 3, for all sex market categories and in all four models, indicate street sex workers are significantly more likely to give anonymous or full consent than sex workers operating in all other sex markets. Those escorting through an agency are the least likely to give anonymous consent to report to the police, while independent workers are the least likely to give full consent to report to the police. According to the odds ratios Ex(B)(inverted for easier interpretation) in Model 1, street sex workers are on average approximately 2.56, 3.33, and 2.86 times more likely than indoor establishment, escort workers and independent sex workers, respectively, to give anonymous consent to report the crime to police. These effect sizes remain similar in Model 2. In Model 3, before the crime type is controlled for, street sex workers are also 1.79, 1.89 and 2.94 times more likely than indoor establishment, escort workers and independent workers, respectively, to give full consent to report the crime to the police. When the type of crime is taken into account in Model 4, the effect size remains the same for indoor establishment workers: street sex workers are still 1.79 more likely to give their full consent than indoor establishment workers. However, in this model the effect size increased slightly for escort and independent sex workers. After controlling for the crime type, street workers were 2.03 times more likely than escort workers and 3.09 times more likely than independent sex workers to give full consent. 14 The regression results suggest it is not the type of crime reported, the number of perpetrators involved, the gender of the victim, or whether the crime was violent in nature but rather, the sex market in which the sex worker operates that has the greatest influence over willingness to engage with the police. Almost none of the ‘type of crime’ or ‘presence of violence’ variables had a significant relationship to reporting a crime to police. The exception was ‘other crimes’ (B=0.42): sex workers were 1.51 times as likely to give full consent to report a crime to the police if ‘other crime’ was present in the NUM report than if it was not. Another exception was hate crime (B=-0.32): sex workers were 1.38 times as likely to give full consent to report to the police if a hate crime was not present than if a crime was reported to NUM as a hate crime. Discussion This paper aimed to answer two key questions: i) How do the different sex markets in the UK effect what crimes are reported to NUM? ii) What effect do the different sex markets in the UK have on sex workers’ willingness to report victimization to the police? These two important questions have not been explored in previous published work and our findings therefore shed new light on the reporting behaviors of street and indoor sex workers. This study contributes to the understandings of reported victimization, and offers recommendations for improving the criminal justice and practitioner response to violence against sex workers. In relation to the first research question, we found a significant relationship between sex market and the type of crime reported to NUM. Sex workers operating in different sex markets report varying relative proportions of different types of victimization. Although violence is the most frequently reported crime type across all sex markets, street sex workers report the largest number of violent incidents. On the other hand, street sex workers report lower numbers of non-violent crimes, such as stalking and harassment, fraud, and condom refusal or removal than sex workers in other sex markets. Traditional indoor sex market (brothels, saunas and parlours) workers are more likely to report to NUM higher numbers of violent incidents and crimes involving robbery or attempted robbery. They are least likely to report stalking and harassment, hate crime, and fraud. Those escorting through an agency are more likely than other sex workers to report fraud and hate crime. Independent sex workers are the most 15 likely to report non-violent crimes, such as stalking and harassment, fraud, or condom removal and refusal but are least likely to report violent crime and robbery. In relation to the second research question, the findings indicate that the likelihood that the sex worker is willing to give their consent for NUM to share their report with the police is related to the sex market in which they work. This is true for both anonymous and full consent. Street sex workers are significantly more likely to give anonymous or full consent to report to the police than sex workers working in any of the other sex markets. Those escorting through an agency are the least likely to give anonymous consent to share their crime report with the police. Independent sex workers are the least likely to grant full consent. Other variables, such as type of crime, the presence of violence, gender, or the number of perpetrators involved are not significant predictors of whether the sex worker will give their anonymous of full consent to share the crime report with the police. Some of our findings are in keeping with previous research which shows that levels of reported violence are higher in the street market than the indoor sex markets (Church et al., 2001; Kinnell 2008; O’Neill et al., 2008; Surrat et al., 2004). Higher rates of violence against street sex workers may be a consequence of enduring societal stigmatization and marginalization, which enables the violence perpetrated against them to go unchallenged often. Frequently imagined in the public consciousness as immoral and transgressors of feminine norms (O’Neill, 1997), street sex workers have a long history of (state sanctioned) oppression. Indeed, street sex workers may be made vulnerable to violence by laws in the UK and elsewhere that criminalise the street sex market and thus relegate street workers to unsafe spaces (Author C, 2004), encourage risky working practices (Krüssi et al., 2014), and deter the reporting of victimization to the police (Boff, 2012). Lower rates of condom removal/refusal experienced by street sex worker may be explained by their (mis)construction in the public imagination as ‘vectors of disease’ (Day & Ward, 1997), which may deter clients from engaging in this form of victimization. Our findings that sex workers operating in brothels, saunas and parlors report higher numbers of robbery and attempted robbery than those working in the other sex markets may be explained by Brooks-Gordon’s (2006) claim that indoor premises are regarded by perpetrators as ‘soft targets’ for robbery due to the large volumes of cash kept on the premises and the low likelihood that these 16 establishments will report to the police. Levels of engagment with the police by indoor sex establishments may be mediated by their legal status. Indeed, Brents and Hausbeck (2005) report that in legalised brothels in Nevada, the police become the allies of brothel owners in minimizing disorder and keeping sex workers safe. Conversely, when brothels are illegal – such as those criminalised in the UK by the Sexual Offences Act 1957 – perpetrators of robbery may be aware that sex workers are unlikely to engage with the police due to fear of their own arrest or closure of their work places. The analysis we present, however, also makes a threefold contribution that offers new findings that extend empirical and theoretical understandings of reported violent and non-violent crimes in the sex industry. Firstly, we contribute to the literature on the victimization of independent sex workers. To date, little in known about the victimization that independent sex workers experience since research is only beginning to act on the realization that the majority of sex work is now mediated by, or provide through, digital technologies (Author C et al., 2016; Cunningham & Kendall, 2011; Jarvis-King, forthcoming; McClean, 2015;). It is perhaps unsurprising that independent sex workers report higher rates of stalking and harassment than other types of sex workers because they are more likely to have regular clients who may, occasionally, develop unnatural emotional attachments to the sex worker (Phipps, 2013), and strong attachements to their victim are a common charatceristic of stalkers (van de Aa, 2010). Secondly, there has to date been no known research which has explored reported crimes against agency escorts. Yet our research indicates they report more fraud and hate crime incidents to the innovative NUM scheme, suggesting their victimization may manifest in both financial and discriminatory forms. Thirdly, we challenge previous research that has highlighted that sex workers are often reluctant to engage with the police due to fear of arrest, fear of public identification or negative experiences of the police (Boff, 2012; Brooks-Gordon, 2006, Campbell, 2015). Klambauer (2017:13) uses qualitative interview data from both indoor and street workers in the UK to highlight how ‘consciousness of stigmatization’ dominates sex workers’ perceptions of interactions with the police. Our findings, however, suggest some variation in the level of willingness to share reports with the police across the 17 different sex markets, even when the type of crime, presence of violence and other variables are taken into account. That street sex workers are the most likely group to give their full consent for their report to be shared with the police and independent sex workers the least, sits in contrast to previously held assumptions that criminalization is the key factor preventing sex workers from engaging with the police (Dewey & St Germain 2014; Armstrong 2017). Making sense of this in the context of what has been a significant and high profile political attempt to ‘eradicate’ prostution in the past twenty years in the UK (Graham, 2017) requires an understanding of how the heterogenous sex industry works. Although street sex work is illegal in the UK and independent sex working is legal, it is clear from this dataset that it is street sex workers whom are more willing to report victimization to the police. Three key factors may explain this. Firstly, street sex workers are ‘targeted’ by services with a range of remits from providing support to ‘exit’ the sex industry, to welfare, safety and harm reduction. These remits are often steered by government policy and guidance, as well as funding obligations (Carline & Scoular, 2017; Author A, 2016). Therefore, there has been a historically higher level of engagement with support service by street sex workers than their independent counterparts (Anon & Author C, 2017; Penfold et al., 2004). These support agencies may act as an intermediary to facilitate street sex workers to engage with the police directly, and/or practitioners may liaise with NUM and the local police themselves as part of the reporting process. A second explanation for why street sex workers are the most likely group to report victimization to the police may be their closer proximity to the police. There is evidence that within the plurality of policing approaches adopted in the UK (Feis-Bryce, 2018), some progressive local authorities use multi-agency partnership models to ensure key police personnel have protection remits for street sex workers. Anon & Author C’s (2017) analysis of a managed zone in the UK demonstrates that Sex Work Liaison Officers are pivotal in improving reporting practices amongst street sex workers. As a result of the priority given to some street sex workers through these multi-agency partnerships, it may also be the case that this group have more regular contact with the police than those working independently and therefore have greater opportunity to report victimization. The physical availability 18 of street sex workers, in comparison to their more hidden indoor counterparts, may also mean they have far greater levels of contact with the police, as well as other (non-state) organizations. Thirdly, the lower numbers of reports from independent workers may be understood through both their unique working arrangements and their ambivalence to engaging with the police for fear of bringing attention to their income generating activities. Whilst street sex workers often have relationships with local support projects, independent workers typically do not. From a survey of internet-based sex workers, most of whom were independent workers, Author C, Author A and Anon (2016) found that the vast majority of respondents (n = 168/240; 70%) do not access a formalised sex work support projects and instead, find support from other sources, including online professional working forums. It may also be the case that independent sex workers, particularly those taking bookings via digital technologies, may have more information about violent clients that those working in the street sex market. They may have telephone numbers and/or email addresses, which they can then use personally or collectively – by sharing with NUM – to avoid future interactions with dangerous clients. The removed and hidden nature of indoor sex work, particularly independent escorting, also means the police will be seldom aware of these types of sex work taking place, unless where complaints are made from members of the community and/or intelligence suggests visits are required to address concerns around trafficking, undocumented workers, or other forms crime (Author A, 2016). In this regard independent workers usually operate ‘below the radar’ and are less inclined to contact the police about their experiences. They may fear not being believed by the police or judged by officers, and/or they may fear being publicly-identified as a sex worker (Boff, 2012; Campbell, 2015; Toynbee Hall, 2009). Limitations Although this paper makes an important empirical and theoretical contribution, there are, however, some limitations. Whilst our analysis indicates that street sex workers report higher levels of violence than those working in other sex markets, we cannot know from this dataset whether street sex workers actually experience more violence than their indoor counterparts. It may be the case that street sex workers are simply more likely to report violence to NUM and the police than indoor workers. As such, 19 these reporting practices may be, at least in part, the consequence of sex worker support projects – which are frequently targeted at street workers – encouraging street sex workers to engage with NUM. While our research identifies and offers potential explanations for variations in the relative proportions of different types of victimization and sex workers’ willingness to engage with the police between sex markets, new qualitative research could further elucidate these complex relationships. Indeed, future research in the form of semi-structured interviews could ask sex workers operating in different markets how and why the environment in which they work influences their reporting practices. Conclusion and Recommendations Through an analysis of a large national dataset of 2,056 crime incident report, we have demonstated that sex workers working in different sex markets report varying relative proportions of different types of vicitimization. Sex workers operating in the street market are the most likely to report victimization involving violence, whilst rates of non-violent crimes – such as stalking and harassment and fraud – are more likely to be reported to NUM by independent sex workers. The findings presented here also demonstrate that street sex workers are more willing to report their victimization to the police, either anonymously via NUM or directly themselves. The different ways in which the street and indoor markets operate has a significant bearing on the nature by which, and extent to which, different types of sex workers engage with the reporting process when they have experienced harm. As such, sex workers operating within the legal framework in the UK have no greater access to justice through the criminal justice system than those working outside of the law. Based upon these conclusions, we recommend that the police, and practitioners who work with sex workers in a social and health care context, refrain from adopting a ‘one-size-fits-all’ approach to supporting sex workers to report victimization. Instead, their responses must recognise variations in reporting behaviors across the different sex markets, and redress indoor sex workers’ reluctance to engage with the police. This will require significant and long-term investments in the cultivation of sex worker-police relationships, which have historically been fractious. Specifically, there is a need for more proactive police work that seeks to accentuate the message the sex workers will be treated with respect and dignity if they come forward as a victim of violent or non-violent crime. Legal reform in 20 the UK is required, to facilitate sex workers to engage with the police, without fear of arrest. For academics, this research represents a call to ensure that in discussions around violence, generalisations and blanket statements are avoided: a great deal of heterogeneity exists in the reporting behaviours of sex workers operating across the different sex markets in the UK. References Aids Alliances (2017) Sex work and institutional violence. Retrieved from ttps://www.aidsalliance.org/news/889-sex-work-and-institutional-violence Armstrong, L. (2016). Who’s the slut, who’s the whore?: Street harassment in the workplace among female sex workers in New Zealand. Feminist Criminology, 11(3), 285-303. Armstrong, L. (2017). From law enforcement to protection? Interactions between sex workers and police in a decriminalized street-based sex industry. British Journal of Criminology, 57(3), 570–588. Barnard, M., Hart, G., & Church, S. (2002). Client violence against prostitute women working from street and off-street locations: A three city comparisons. Retrieved from https://www.viceversadundee.org.uk/files/client_violence_research.pdf Bernstein, E. (2007). Sex work for the middle classes. Sexualities, 10(4), 473-788. Boff, A. (2012). Silence on violence. Improving the safety of women. The policing of off-street sex work and sex trafficking in London. London, UK: London Assembly. Brents, B., & Hausbeck, K. (2005). Violence and legalized brothel prostitution in Nevada: Examining safety, risk, and prostitution policy. Journal of Interpersonal Violence, 20(3), 270-295. Brooks-Gordon, B. (2006). The price of sex: Prostitution, policy and society. London, UK: Taylor and Francis. 21 Campbell, R. (2015). Not getting away with it: Linking sex work and hate crime in Merseyside. In: N Chakraborti & J Garland J (Eds.), Responding to hate crime: The case for connecting policy and research (pp. 57-72). Bristol, UK: Policy Press. Campbell, R., & Stoops, S. (2010). Taking sex workers seriously: Treating violence as hate crime in Liverpool. Research for Sex Work, 12, 9-12. Canter, D., Ioannou, M., & Youngs, D. (2016). Sex and violence: The experience and management of street prostitution. In: D Canter, M Ioannou & D Youngs (Eds.), Safer sex in the city: The experience and management of street prostitution (pp. 1-12). Oxon, UK: Routledge. Carline, A., & Scoular, J. (2017). Saving fallen women now? Critical perspectives on engagement and support orders and their policy of forced welfarism. Social Policy and Society, 14(1), 103- 112. Church, S., Henderson, M., Barnard, M., & Hart, G. (2001). Violence by clients towards female prostitutes in different work settings: Questionnaire survey. British Medical Journal, 322, 524-525. Crime Survey for England and Wales (2017). The Crime Survey for England and Wales: Reasons for not reporting to the police. Retrieved from https://www.ons.gov.uk/peoplepopulationandcommunity/crimeandjustice/adhocs/007750. Cunningham, S., & Kendall, T.D. (2011). Prostitution, technology, and the law: New data and directions. In: L.R. Cohen & J.D. Wright. (Eds.), Research handbook on the economics of family law. (pp. 221-270). Cheltenham, UK: Edward Elgar Publishings. Day, S., & Ward, H. (1997). Sex workers and the control of sexually transmitted disease. Genitourinary Medicine, 73(3), 161-168. Deering, K., Amin, A., Shoveller, J., Nesbitt, A., Garcia-Moreno, C., Duff, P., et al. (2014). A systematic review of the correlates of violence against sex workers. American Journal of Public Health, 104(5), 42-54. 22 Dewey, S., & St Germain, T. (2014). ‘It depends on the cop’: Street-based sex workers’ perspectives on police patrol officers. Sexuality Research and Social Policy, 11(3), 256–270. Ellison, G. (2015). Criminalizing the payment for sex in Northern Ireland: Sketching the contours of a moral panic. British Journal of Criminology, 57(1), 194-214. Feis-Bryce, A. (2018). Policing sex work in Britain: A patchwork approach. In: T. Sanders & M. Laing. (Eds.), Policing the sex industry. London, UK: Routledge. Feis Bryce, A., Campbell, R., Pitcher, J., Laing, M., Irving, A., Brandon, J., et al. (2015). Male escorting, safety and National Ugly Mugs: Queering policy and practice on the reporting of crimes against sex workers. In: M. Laing, K Pilcher & N Smith. (Eds.), Queer Sex Work (pp. 245-254). London, UK: Routledge. Farley, M., Golding, J.M., Schuckman Matthews, E., Malamuth, N.M., & Jarrett, L. (2015). Comparing sex buyers with men who do not buy sex: New data on prostitution and trafficking. Journal of Interpersonal Violence, Online August 31. Goodyear, M.D.E., & Cusick, L. (2007). Protection of sex workers. BMJ, 334(7584), 52-53. Graham, L. (2017). Governing sex work through crime: Creating the context for violence and exploitation. The Journal of Criminal Law, 81(3), 201-206. Hubbard, P. (2006). Out of touch and out of time? The contemporary policing of sex work. In: R. Campbell & M. O’Neill (Eds.), Sex work now. (pp. 1-30). Cullompton, UK: Willan. Jamshidian, M. (2009). Strategies for analysis of incomplete data. In: Hardy, M., Bryman, A. (Eds). The Handbook of data analysis, (pp.113-130), London: Sage. Jarvis-King, L. (forthcoming). Time is money: The commodification of the indoor sex industry. Unpublished PhD Thesis, Leeds, UK: University of Leeds. Jeffreys, S. (1997). The idea of prostitution. Victoria: Spinifex. 23 Kinnell, H. (1993). Prostitutes exposure to rape: Implications for HIV prevention and for legal reform. In: The VII Social Aspects of AIDS Conference, South Bank University, London, UK. Kinnell, H. (2004). Violence and sex work in Britain. In: S. Day & H. Ward (Eds.), Sex work, mobility and health in Europe. (pp. 179-194). London, UK: Routledge. Kinnell, H. (2008). Violence and sex work in Britain. Cullompton, UK: Willan. Klambauer, E. (2017). Policing roulette: Sex workers’ perception of encounters with police officers in the indoor and outdoor sector in England. Criminology and Criminal Justice, Online first. Krüsi, A., Pacey, K., Bird, L., Taylor, C., Chettiar, J., Allan, S., et al. (2014). Criminalisation of clients: Reproducing vulnerabilities for violence and poor health among street-based sex workers in Canada – a qualitative study. BMJ Open, 4(6). Levy, J. & Jakobsson, P. (2014). Sweden’s abolitionist discourse and law: Effects on the dynamic of Swedish sex work and on the lives of Sweden’s sex workers. Criminology and Criminal Justice, 14(5), 593-607. Lowman, J. (2000). Violence and the outlaw status of (street) prostitution in Canada. Violence Against Women, 6(9), 987-1011. Lyons, T., Krüsi, A., Pierre, L., Kerr, T., Small, W. & Shannon, K. (2015). Negotiating violence in the context of transphobia and criminalization. Qualitative Health Research, 27(2), 182-190. McClean, A. (2015). ‘You can do it from your sofa’: The increasing popularity of the internet as a working site among male sex workers in Melbourne.’ Journal of Sociology, 51(4), 887-902. Maher, J., Pickering, S., & Gerard, A. (2013). Sex work: Labour, mobility and sexual services. London, UK: Routledge. Miller, J. & Schwartz, M. (1995). Rape myths and violence against street prostitutes. Deviant Behavior, 16(1), 1-23. 24 Monto, M., & Hotaling, N. (2001). Predicators of rape myth acceptance among male clients of female street prostitutes. Violence Against Women 7(3), 275-293. O’Neill, M. (1997). Prostitute women now. In: G. Scrambler & A. Scrambler (Eds.). Rethinking Prostitution (pp. 3-28). London, UK: Routledge. O’Neill, M., Campbell, R., Hubbard, P., Pitcher, J., & Scoular, J. (2008). Living with the other: Street sex work, contingent communities and degrees of tolerance. Crime, Media, Culture 4(1), 73- 93. Penfold, C., Hunter, G., Campbell, R., & Barham, L. (2004). Tackling client violence in female street prostitution: Inter-agency working between outreach agencies and the police. Policing and Society, 14(4), 365-379. Phipps, A. (2013). Violence against sex workers. In L. McMillan & N Lombard (Eds.), Violence against women (pp. 87-102). London, UK: Jessica Kingsley Publishers. Raymond, J. (1998). Prostitution as violence against women: NGO stonewalling in Beijing and elsewhere. Women’s Studies International Forum, 21(1), 1-9. Scott, J., Minichiello, V., Mariño, R., Harvey, G.P., Jamieson, M., & Browne, J. (2005). Understanding the new context of the male sex work industry. Journal of Interpersonal Violence, 20(3), 320-342. Shdaimha, C., & Leon, C. (2015). ‘First and foremost they’re survivors’: Selective manipulation, resilience, and assertion among prostitute women. Feminist Criminology, 10(4), 326-347. Stalans, L., & Finn, M. (2016). Defining and predicting pimps’ coerciveness towards sex workers: Socialization processes. Journal of Interpersonal Violence, online first, 1-24. Surratt, H., Inciardi, J., Kurtz, S., & Kiley, M. (2004). Sex work and drug use in a subculture of violence. Crime & Delinquency, 50(1), 43-59. 25 Toynbee Hall (2009). Statistics on prostitution in London and the UK. Retrieved from http://www.toynbeehall.org.uk/data/files/Statistics_on_prostitution.pdf van de Aa, S. (2010). Stalking in the Netherlands: Nature and prevalence of the problem and the effectiveness of anti-stalking measures. Antwerp, Belgium: Maklu. van der Meulen, E., & Valverde, M. (2013). Beyond the criminal code: Municipal licensing and zoning bylaws’. In E. van der Meulen, E Durisin & V. Love (Eds.), Selling sex: Experience, advocacy, and research on sex work in Canada (pp. 314-322). Toronto, Canada: UBC Press. Ward, H., Day, S., & Weber, J. (1999). Risky business: Health and safety in the sex industry over a nine-year period. Sexually Transmitted Infections, 75(5), 340-343. 26 Table 1 . Types of crime: most common combinations Type of crime N % Violence 317 15.4 Other crime//crime type not specified 294 14.3 Stalking and harassment 166 8.1 Robbery/Attempted robbery and violence 156 7.6 Robbery/attempted robbery 99 4.8 Rape, attempted rape and violence 90 4.4 Hate crime and violence 78 3.8 Fraud 65 3.2 Hate crime 57 2.8 Rape and other unspecified crime 49 2.4 Hate, robbery/attempted robbery and violence 43 2.1 Sexual assault and violence 36 1.8 Rape, stalking and harassment 29 1.4 Hate crime and other crime 27 1.3 Sexual assault 26 1.3 Violence, stalking and harassment 25 1.2 Rape, robbery and violence 25 1.2 Violence, rape, and hate crime 18 .9 Condom removal & refusal to use condom 17 .8 Robbery and rape 16 .8 Violence, stalking and hate crime 14 .7 Hate crime 12 .6 Robbery/attempted robbery and hate crime 10 .5 Stalking, rape and hate crime 10 .5 Other combinations 377 18 Total N 2,227 27 Table 2. Reported crime type by the type of sex market Type of sex market Street sex workers Indoor establishment sex workers Escort workers Independent sex workers Average for all Reported crime type % % % % % Violence *** 57.1 52.9 33.8 23.4 46.1 Robbery/att. Robbery*** 23.7 31.4 19.7 14.2 21.2 Stalking& and harassment*** 12.6 13.2 15.5 31.3 18.3 Hate crime** 20 8.3 19.7 15.5 18 Other crime*** 17.5 16.5 22.5 25.9 20.1 Rape/attempted rape 19.6 14.9 15.5 15.1 17.8 Sexual Assault 7.1 9.9 5.6 7.2 7.2 Fraud*** 2.7 6.6 12.7 11 5.7 Condom (refusal or removal)** 3.9 5.8 7 7.9 5.3 Note: ***p<0.001; **p<0.01, *<0.05 for chi square test 28 Table 3. Logistic regression estimates Anonymous consent to report to the police Full consent to report to the police Model 1: Presence of violence Model 2: types of crime Model 3: Presence of violence Model 4: types of crime B (S.E) Exp(B) B (S.E.) Exp(B) B Exp(B) B (S.E.) Exp(B) Sex market (street sex worker) Indoor establishment sex worker workerworkersworkers -0.94 ** (0.31) 0.39 -0.947** (0.32) .39 - 0.57*(0.24) 0.56 -0.58*(0.24) .561 Escort worker -1.22***(0.36) 0.3 -1.24 ***(0.36) .29 -0.64* (0.32) 0.53 -0.71*(0.32) .493 Independent sex worker -1.05***(0.19) 0.35 -1.06***(0.19) .35 - 1.0***(0.1 5) 0.34 - 1.13***(0.15) .324 Gender of victim (female) Male -0.189(0.33) 0.83 -0.18 (1.07) 1.10 -.400 (0.35) 0.67 -.42 (0.36) .660 Trans (m to f) -0.053 (0.63) 0.95 -0.09 (1.12) .92 -.036 (0.56) 0.96 -.08 (0.57) .919 Not specified -0.124 (1.07) 0.88 -0.09 (1.23) 1.00 1.39* (0.63) 4.00 1.45* (0.63) 4.247 Presence of violence (not present) Only violence 0.004 (0.26) 1 .033 (0.15) 1.03 Violence and other crime 0.217 (0.21) 1.24 .154 (0.12) 1.17 Presence of type of crime (0- no, 1 –yes) Violence -0.03 (0.22) .97 0.06 (0.14) 1.057 Fraud -0.16 (0.37) .85 -.22 (0.25) .800 Other Crime 0.22(0.26) 1.25 0.42*** (0.18) 1.518 Rape/attempted rape 0.03 (0.22) 1.03 -.07 (0.14) .933 Condom refusal/removal -0.15 (0.35) .86 .38 (0.24) 1.465 Robbery/attempted robbery -0.03 (0.24) .97 -.08 (0.13) .927 Sexual assault -0.47 (0.39) .62 .08 (0.21) 1.082 Stalking and/or harassment -0.03 (0.26) .97 .10 (0.17) 1.105 Hate crime -0.20(0.24) .82 -0.32*(0.14) .723 Number of perpetrators 0.058 (0.1) 1.06 0.05 (0.11) 1.05 0.015 (0.05) 1.01 -.01(0.05) .993 Constant 2.00*** (0.36 7.4 2.31 (0.5) 10.04 1.18*** (0.25) 0.31 -1.16**(0.33) .31 N 2030 2030 2030 2030 R squared 0.06 0.06 0.07 0.08 Note: ***p<0.001; **p<0.001, *p<0.05 29 0 10 20 30 40 50 60 70 80 90 No violence Only violence Violence and other crimes Street sex workers Indoor establishment sex workers Escort workers Independent sex workers Average for all 0 10 20 30 40 50 60 70 80 90 100 Street sex workers Indoor establishment sex workers Escort workers Independent sex workers Average for all Anonymous consent Full consent Figure 1. Sex market type and the presence of violence in crime reporting Figure 2. Reporting to the police by the type of sex workers 30 Appendix A: Descriptive statistics Variable % Sex market (Street) 60.9 Brother/Sauna/Parlour 5.9 Escort through Agency 3.5 Private/independent 29.7 Type of crime (if reported) Violence 46.1 Rape/attempted rape 17.8 Sexual assault 21.2 Robbery/attempted robbery 21.2 Condom (refusal or removal) 5.3 Fraud 5.7 Stalking and harassment 18.3 Hate crime 18.3 Other crime (unspecified) 20.1 Presence of violence (No violence involved) 53.9 Only violence reported 15.4 Violence and another crime 30.6 Report to police, anonymous consent’ (yes) 91.6 Report to police, full consent’ (yes) 24.6 Gender of the victim (Female) 93.9 Male 4.3 Trans male to female 1.3 Not specified 0.5 Number of perpetrators (M=1.31, SD=0.98, min 1, max 15