Risky business: voluntary sex workers as suitable victims of work-related crimes in a legalised prostitution environment

Many studies have analysed the violence to which sex workers (SWs) are exposed in countries where prostitution is criminalised, but violence in sex work when it is a legal and freelance activity has seldom been studied. This study is based on non-systematic participant observation and 14 interviews conducted with cisgender and transgender freelance SWs exercising their profession in Switzerland. Findings show that all the participants had been victims of work-related offences and that their victimisation can be interpreted through the updated version of Routine Activities Theory. Few reported the victimisation to the police although in general, they said that they appreciated the Swiss police. Most common perpetrators were customers, passers-by, and colleagues. The places where offences occurred can be defined as risky settings. The police, the SWs themselves and other actors such as social workers are identified as controllers, while the legal framework and NGOs themselves can be labelled as super-controllers. Our study suggests that public policies should not be limited simply to regulating prostitution, but also continuing to ensure SW’s safety by developing evidence-based prevention strategies.


Introduction
Sex work or prostitution can be defined as the exchange of sex for money, goods, or other types of services (Jenkins 2020). Attitudes towards prostitution have varied throughout history as they are shaped by the dominant culture of each period, from tolerance in the Roman Empire to punishment with fines, detention and even death penalty in the Puritan period that followed the rise of Protestantism in the 16th century (Jenkins 2020). In Western Societies, attitudes towards sex and, indirectly, prostitution has been undergoing major, continuous and heterogeneous changes since the 1960s, when the invention of the combined oral contraceptive pill sparked the sexual revolution and the normalisation of contraception. Meanwhile, the rise of feminism further impacted attitudes towards this practice, resulting in differing views ranging from considering it a profession to those to considering it a form of contemporary slavery. This debate led to the development of the term sex worker (SW) in the 1980s (Jenkins 2020), a decade also marked by the rise of the HIV/ AIDS epidemic which and the development of the internet in the 1990s and widespread access pornographic websites (Weiss 2018).
Today, sex work is a heterogenous phenomenon that includes different forms of services and is exerted by people with diverse socioeconomic backgrounds and motivations (e.g. see Harcourt and Donovan 2005). Its regulation across Europe also varies widely (Danna 2014). Sex work is illegal (although in practice it seems to be alegal) in Eastern European countries, whilst the Scandinavian nations and France tolerate SWs but punish their clients. Quite oppositely, in some Central European countries-Austria, Belgium, Germany, The Netherlands, and Switzerland-sex work is considered a legal freelance employment.
Our study takes place in Switzerland, where sex work is a legal activity subject to taxation and social insurance, whose specific regulation varies across the cantons (states) that compose the country, organised in the form of a confederation. In the canton where this research was conducted, sex work is permitted both indoors-in so-called erotic massage salons-and outdoors, where it can only take place at night and in a specific area. Customers can be solicited during the night in the streets of that area, but the sexual services can only be performed in the erotic massage salons. The latter can be bars, clubs or private rooms, and they are placed under the responsibility of a manager who rent rooms to the SW. The Swiss criminal code (art. 199) punishes with a fine those engaging in sex work outside these places and times as well as pimping and sexual exploitation with fines and/or-depending on the seriousness of the case-imprisonment (arts. 182 and 195).
A pioneer European study on SWs victimisation that inspired our research was conducted in Spain-whose law does not forbid nor regulates sex work-by Rosemary Barberet (2000), almost a quarter of a century ago. Barberet interviewed 50 SWs and found that they had been victims of physical assault, robbery, kidnapping, and rape (defined as non-consensual sexual intercourse) while doing their work. The perpetrators were clients and passers-by who always took away the SWs' cash. The offences were not reported to the police because the SWs feared the consequences of a criminal procedure, distrusted the authorities of the criminal justice system, or were unable to identify the perpetrator. Several SWs mentioned to Barberet (2000) that the police had told them that victimisation was a "side effect" of their job.
Research suggests that the risk of victimisation for a SW is lower in countries where prostitution is legal than in those where it is criminalised or has been "abolished" (Brents and Hausbeck 2005;Sanders and Campbell 2007;Abel et al. 2010;Benoit et al. 2019). In that regard, former studies suggest that legalisation reassures SWs and their rights, hence increasing their willingness to report offences, suffered or witnessed, to the police. However, comparisons are problematic because most of the studies on SWs' victimisation have been conducted in countries where prostitution is illegal or alegal (Weitzer 2009).

Theoretical framework: Routine activity theory
Routine activity theory (RAT, Cohen and Felson 1979;Felson 1986;Sampson et al. 2010;Sherman et al. 1989) is one of the most tested and corroborated theories in criminology (Eck and Madensen 2015). Scholars such as Sanders and Campbell (2007) and Miin Chai et al. (2023) have applied opportunity-based theories to the study of sex work. These theories focus on proximal causes of crime-instead of distal causes such as structural inequalities-and posit that that certain conditions and actors must converge in time and space in order for the crime to occur.
In its original version, RAT argues that a predatory crime requires a motivated offender, a suitable target, and the absence of a capable guardian (Cohen and Felson 1979). In its latest version ( Fig. 1), proposed by Sampson et al. (2010), RAT states that, for a crime to occur, a potential offender and a suitable target must converge in the same place (Cohen and Felson 1979;Felson 1986) in the absence of controllers-which can be handlers, guardians, or managers (Felson 1986;Sherman et al. 1989)-or super-controllers that encourage the controllers to prevent crime (Sampson et al. 2010).
The controllers of a potential offender can be handlers-a significant other, family members, or friends-that discourage him or her from offending. A location can have a manager-a concierge or shopkeeper-whose task is to keep the premises safe. Lastly, a suitable target can be controlled by a guardian-friends, a neighbour, or the police-that provides protection. (Felson and Eckert 2019) remind that the police are unlikely to be present when a crime occurs; consequently, the best guardians are not police officers but common citizens whose mere presence serves as a reminder that the target is not alone.
Similarly, other handlers, managers, and guardians may not always comply with their task, and it is in that context that super-controllers become essential. These can be regulatory agencies, the media, family relationships, courts or markets that can motivate controllers or supervise whether they are exercising their function appropriately. They can also motivate other super-controllers to do the same.

Aims of the study
This article's aim is twofold: First, it studies the extent and characteristics of SWs' victimisation; second, it places it within the framework of the latest version of Routine activity theory (Cohen and Felson 1979;Sampson et al. 2010), an opportunity-based theory that seems useful to analyse the proximal causes of this kind of victimisation.
The research was guided by the following questions: (1) What is the prevalence 1 and variety 2 of sex work-related victimisations ?
(2) At what rate are sex work-related victimisations reported to the police ?
(3) What is the SWs' perception of the Swiss police ? (4) Who are the perpetrators of the SWs' victimisation ? (5) In which places the victimisation of the SW takes place ? (6) Who act as controllers (i.e. guardians, managers and handlers) in preventing the SWs' victimisation? (7) Who act as super-controllers in motivating the controllers to prevent the SWs' victimisation ?

Data collection
This study is based upon approximately 1100 h of non-systematic participant observation distributed over 18 months, which set a trustworthy context for conducting 14 semi-structured interviews with indoor and outdoor SWs (N = 14) exercising their profession in the canton of Vaud (Switzerland). The participants are freelancers, in the sense that they do not work under coercion from pimps or sexual exploitation networks. The first author is a research-practitioner who, in collaboration with the person in charge of an NGO, developed a semi-structured research protocol that was applied in this study. During the first year, we followed an ethno-criminological approach that allowed the first author to assess the situation in the field and to take informed decisions about the profiles of the SWs to be interviewed, the questions to be asked, and the appropriate times and places to conduct the interviews. With theoretical saturation in mind, we attempted to interview cisgender and transgender SW-the latter are underrepresented in the available research (Weitzer 2005)-both with legal and illegal working status. It is to note that we were unable to target male SWs due to their absence from the field. The individual interviews were conducted in 2017 at times and places chosen by the SWs, typically in tea rooms and bars during the day, but also in the streets and in erotic massage salons during the night. We preferred the former because they offered the opportunity of discussing with the SWs outside or their working hours and out of sight of their colleagues or acquaintances. The field researcher paid for soft drinks and pastries, and conducted the interviews in Romanian, French, or Spanish, depending on the native tongue of the SWs or the language in which they were more fluent. Interviews began with general questions about the SWs' situation, focused then on their experience in sex work in Switzerland, and concluded with the sensitive issue of their victimisation. All interviews were recorded but one, for which we took notes manually. The recorded interviews were transcribed in the same language in which they took place, but quotations are presented here in English.

Data analysis
Following the principles of content analysis (Castleberry and Nolen 2018), we started with a horizontal analysis that consisted in browsing each transcript to identify the recurrent topics. This allowed a subsequent transversal analysis of each of those topics across the transcripts.
The main variables that emerged from the interviews were introduced in a data analysis software that allowed estimating the prevalence and variety of the victimisations self-reported by the SWs. Prevalence refers to the proportion of SWs who were victims of a work-related offence since they started working as such in Switzerland. The eight offences identified through the interviews are: (1) threats, (2) sexual assault, (3) physical assault, (4) fraud, (5) robbery, (6) wrongful restraint, (7) theft, and (8) insults. Variety refers to the range of different types of offences suffered by the SWs. It was computed as a lambda index by adding the number of different victimisations reported by each participant (one point for each type of offence, except insults which were excluded to keep only the most serious offences) and dividing the total by the number of members of the sample that suffered at least one victimisation. 3 Table 1 illustrates the sociodemographic characteristics and experience in sex work of the 14 participants in our study. The sample includes eight cisgender women and

Variety of victimisation
Number of different types of victimisation reported by each participant Number of participants who reported atleast one victimisation Indoors n = 4 Combined n = 5 Administrative situation Legal n = 7 Illegal n = 7 six transgender women. On average, at the time of the interview they were 35 years old and had started doing sex works five years earlier. Eleven of them are non-Swiss citizens coming primarily from Eastern Europe (n = 7), among whom two transgender participants belong to the Romanian Roma ethnic group. Most of the members of the sample are single (n = 9) and have no children (n = 10). Four of them have uncompleted university studies and the rest have finished primary or secondary school (n = 9) or an apprenticeship (n = 1). Seven participants have an illegal resident or work status. One of them is a non-European citizen that cannot apply for a work permit, and the rest are European citizens who have remained in Switzerland for more than 90 days (the maximum number of days allowed without a work permit) or do not fulfil the criteria to obtain a work permit (i.e. earn at least 2000 Swiss francs per month, have a reasonable business plan, and have an address in Switzerland).
Being in an illegal situation, combined with the fact that the police check frequently the erotic massage salons to control that the law on sex work is applied properly, has an impact on the working location-indoor or outdoor-chosen by the SWs to exercise their profession. Five of the members of our sample work in the prostitution area from 10 PM to 5 AM, soliciting customers in the street and performing the sexual service in the car or the house of their client, or in their own house. Four SWs exercise their profession indoors, primarily in erotic massage salons, although some work also in their house or, even if it is illegal, in hotels. The other five SWs work both indoor and outdoor.

Settings
As mentioned in the Introduction, in the city where our study was conducted, street sex work is permitted in a specific area (three streets) of a neighbourhood that cannot be considered degraded or particularly dangerous and is located at approximately ten blocks (one kilometre) from the city centre. The neighbourhood combines a zone of apartment buildings with a commercial zone composed of large streets with relatively high buildings-some under construction-that host small light industries, commercial companies, and offices, two education centres, a theatre, a bar, a concert house, and a skate park. Soliciting in the sex work area is allowed from 10 PM to 5 AM.
The sex work area-where soliciting is allowed from 10 PM to 5 AM-is not isolated from the rest of the neighbourhood, which means that, apart from the SWs and their potential clients, there are many passers-by (neighbours, visitors, partygoers, etc.) going through it by car or on foot. Police also patrols the neighbourhood by car and on foot. In addition, the social workers from the NGO that assists the SWs also start their shift at 10 PM and do their work inside a caravan placed in one of the streets of the area, where SWs can have a cup of coffee, tea or other soft drink and receive condoms as well as informative brochures.
Customers usually arrive by car and drive nearby the SWs, who are waiting for them on the sidewalk, to ask about pricing and types of services. Upon reaching a mutual agreement on the terms of service between the SW and the customer, the SW will proceed to provide the service. The service may be performed inside the customer's vehicle, or at a designated location, e.g. a hotel room, the client's home, a parking lot, or the countryside.
As mentioned above, sex work is allowed indoors in the premises of erotic massage salons that have obtained an administrative licence for erotic tariffed activities. The managers of these salons rent their rooms to SWs-usually for the entire day-after checking that they are legally entitled to work. Managers cannot impose any obligation to the SW-such as working a certain number of hours, receiving a minimum number of clients, performing specific sexual practices, or accepting clients the SW would prefer to reject-nor can they take a percentage of their earnings. The salons are usually opened from morning to midnight, unless they are placed in clubs or cabarets, which also have a night shift. Table 2 illustrates the prevalence of the victimisations self-reported by the participants (N = 14). They have all been victims of at least one offence during their working experience in Switzerland. However, only three (two with legal status and one without) have reported the offence to the police. The main reasons given for not doing so were related to their fear of being fined or deported from Switzerland because they worked in illegal places, at illegal times, or did not have a work permit. In addition, one SW did not speak French, a Swiss SW explained that she feared that, if she complained to the police, her entourage would know about her job, and one considered that the police are useless.  Table 3 illustrates the variety of the offences against the participants. As a reminder, insults were excluded from the calculation to study only the most serious offences. On average, the lambda variety of the sample's victimisation is 2.1, which means that each victimised SW endured 2.1 different work-related victimisations during her working experience (5 years on average). When broken down into different categories, transgender SWs reported a lambda variety of 3.3, whilst for cisgender SWs it was only of 1.3. Moreover, transgender SWs reported a maximum of six different types of victimisations compared to only 3 for cisgender SWs. Similarly, SWs exercising their profession both outdoors and indoors report a higher variety than those working in a single place, be it outdoors or indoors (3.4 vs. 1.3-1.5). On the contrary, there are no differences in victimisation according to the legal status of residence (variety of 2 in both cases).

Threats
Eight SWs reported having been threatened with physical aggression by their customers (four victims), colleagues (two) or the pimps of other colleagues (two) if they did not comply with the perpetrator's demands. Frustrated customers threatened the SW during the sexual service when they could not reach their climax, or when something did not go as planned: The condom broke, and he threatened that he would kill me if I were sick. (Interviewee No. 9, cisgender woman with legal status).
Threats from other colleagues or other colleagues' pimps arise in the context of "territorial" disputes, typically at night and in the street. Senior SWs consider they have priority to choose a specific spot to work-for example, a street corner-and do not hesitate to chase potential competitors, intimidating them if necessary. The law on prostitution does not go into such detail, suggesting that the tacit rule in the street should be "first come, first served", but in practice SW prefer to work systematically in the same spot for strategic reasons. For example, because there are more passersby or to help their usual clients find them easily.
She's jealous: a blonde over there, she's a shit (…), she didn't let me go to the other side (…). It's not her place, here there is a place for everyone. (Interviewee No. 2, transgender woman with legal status).

Property crime
Six SW reported having been victims of property crime while working. The offences include larceny thefts (three victims)-being deprived of their property without force or violence-, robberies (one victim)-being deprived of their property by force or threat of force-and fraud (four victims), usually in the form of clients promising to pay for their services but not keeping the engagement.
In two of the larcenies, the thieves were clients that discreetly took the money from under the bed or the drawer where the SW had put them when the SW went to another room, like the toilet. The robbery occurred in the street at night, and the perpetrator pretended to be a client until he reached an unguarded place where he took the SW money by force or threatening her with a knife.
Three out of the four cases of fraud coincide with the stereotype of the client promising the SW to pay after the intercourse, but never complying. The victims can be indoor or outdoor SW. Nevertheless, some fraud schemes can be quite sophisticated and entail substantial and legal consequences. That was the case of one SW in an erotic massage salon whose owner took part of her money with the promise of registering her officially (work permit, taxes, etc.), but discovered years later, when she was summoned by the Tax Administration, that she was still in an illegal situation and had debts towards the State.

Insults
Seven SW reported having been victims of verbal aggression while working. The perpetrators were mainly pedestrians walking by the prostitution area who called them by names such as "bitch", "slut", or "faggot". These offences' perpetrators were largely pedestrians walking through the prostitution neighbourhood at night. During the field observation, the first author also witnessed some SWs insulting others. Finally, in the context of escalating physical or sexual assaults, the SWs were also insulted by frustrated customers.
He took out his condom and he just cum inside of me, you know? And then I moved myself, but I had it [the sperm] running down (…). And he says to me: 'Shitty bitch'. (Interviewee No. 7, transgender woman without legal status).

Sexual assaults
Six SW reported having been victims of sexual assaults committed by their clients. The latter forced them to have sexual intercourse without their consent, using force, cuffs, or even drugs.
… it happened in his secondary residence. I had been there already several times […] and we have tried several BDSM [Bondage-Domination-Sadism-Masochism] stuff, and once he attached me to the bed. The rules were very clear (…) and we repeated them several times. He insisted. I said 'no' to him the first times, but then I accepted to be handcuffed to the bed. And he did everything which I told him not to do. It clearly was a rape. (Interviewed No. 13, woman without legal status).
Interviewee No. 7 also mentioned cases of customers taking off their condom during the intercourse, ignoring the risk of spreading sexually transmitted infections, such as HIV, hepatitis, syphilis, chlamydia, or gonorrhoea, or the possibility of an unwanted pregnancy.

Physical assaults
Four SW reported having been victims of physical assaults while working. The perpetrators were pedestrians or clients. For the latter, the setting was similar to that explained in the case of threats: the aggression was the consequence of frustrated expectations about the sexual service. Typically, a customer that does not reach climax during the time stipulated for the service wants the SW to continue working and becomes violent when the SW refuses to do so.
He wanted to force me to stay there until he cum because he had paid me. And I told him: 'Yes, you paid me for an amount of time, and that time is over'. And he got very angry, and, fortunately, he grabbed only the pack of humid tissues (…): He threw them in my face, and he shook me. (Interviewee No. 9, cisgender woman with legal status).
SWs' descriptions of physical assaults by pedestrians match cases of unprovoked violence as the perpetrators carried out the assaults without prior interaction and for no apparent reason.
He came from the corner and hit me with the fist in my forehead. I fall with my head on the floor, and I lost consciousness. I don't know how much time… maybe five or six seconds, I don't know exactly… (Interviewee No. 7, transgender woman without legal status).

Wrongful restraint
Four SW reported having been obliged to remain in a room or a car with clients who, dissatisfied with the service, wanted them to continue working. These wrongful restraints lasted from some minutes to a maximum of one hour.

Perceptions of the Swiss police
Five out of the nine SW that discussed police perceptions with the interviewer expressed their appreciation for the Swiss police. Comparing them with their native country police, they agreed that they were much more polite.
Here [in Switzerland, in comparison to Italy] you feel safe; here the police pass, and they say hello to you, they don't bother you. (Interviewee No. 4, woman with legal status).
Five of the 10 SWs asked about police efficiency, considered that police failed to prevent crime because they do not patrol the streets more often and that it took them a long time to intervene when called upon. Conversely, four SW stated that the police are supportive and intervene when needed.

A RAT application to SWs' victimisation
In Fig. 2, we have applied the latest version of the crime triangle developed within RAT to the different kinds of victimisation reported by the SW interviewed for this research. The purpose is not to test RAT, but to use it as an analytical framework to identify the various actors involved in SWs' victimisation and crime prevention.
The potential offenders that our interviewees reported are customers, passersby, colleagues, and managers of the erotic massage salons. Sexual assaults (triangle d in Fig. 2) as well as wrongful restraints (triangle f) are typically committed by customers, while the range of potential offenders is large for the rest of offences. For instance, passers-by are often involved in property crimes, physical assaults, and insults. The latter, together with threats, can also come from colleague SWs.
The places where these crimes occur can be considered as "risky settings" (in the sense of Felson and Eckert 2019), especially at night: empty streets, large parking lots, the customer's car or house, or recreational settings such as the erotic massage salons. Even though they are not fully isolated due to the presence of blocks of flats and events nearby, at night they are deserted and to count on efficient guardians is more difficult than during the day.
As far as controllers are concerned, whenever they are present, the police or the NGO's social workers act as guardians and can play a role in preventing SWs' victimisation. They are 'motivated' by the legal framework provided by the law on prostitution and the Swiss criminal code, who could metaphorically act as supercontrollers. In this regard, the NGO serves a dual role as both a controller and super-controller, where the board and management motivate the field social workers to fulfil this responsibility. In addition, the managers of the erotic massage salons play literally that role in the sense of RAT, and our field observations suggest that private security agents-who patrol open garages and buildings at night-also play the role of managers of places. Both are motivated to comply with the task of managing the place because it represents their source of income. Nevertheless, our data does not allow any general conclusion regarding handlers. It can be noted, however, that the strategies applied by SWs to calm down their clients when tension arises (for examples, see Molnar and Aebi 2022) show that, somehow, they are playing the triple role of handlers, guardians, and potential victims.

Discussion
This paper highlights the risk of victimisation faced by cisgender and transgender female SWs even when they operate in a context in which prostitution is legal. Sex work is performed in risky settings: it usually takes place at night, in rather isolated places and in constant contact with strangers. Consequently, SWs are susceptible to endure threats, insults, property offences, and physical or sexual assault. It is to note that although our article focuses on the concrete settings and proximal causes of SWs' victimisation, structural causes such as poverty, lack of access to education, gender-based violence, etc., should not be forgotten.
The levels of victimisation observed in Switzerland-where the most frequent offences are threats and insults-are lower and much less serious than the ones observed in Russia (Karandikar and Próspero 2010) or the USA (Chan and Beauregard 2019). The profiles of some of the perpetrators and the places where victimisation occurred have some points in common with previous research conducted in countries where prostitution is illegal (Barberet 2000;Földhazi 2010;Karandikar and Próspero 2010;Mitjans Núñez and Molnar 2016;Bungay and Guta 2018;Peitzmeier et al. 2019) but also some interesting differences. For example, former studies have highlighted that SW are largely the victims of their customers and passers-by, but also of their pimps and the police. In our study, we found that cisgender and transgender female SWs were victims of male customers and passers-by, as well as of other cisgender and transgender female SWs, but none had been the victim of the Swiss police. This suggests that legalisation acts as a protective factor. In addition, we have also identified threats by colleague SWs and their pimps that had not been identified by previous research, perhaps because the question was not raised by the researchers. Similarly, fraud has seldom been studied in this context, but we identified frauds perpetrated by the managers of the salons. In particular, the SW interviewed in our study affirmed that they had been threatened by their colleagues largely because of the strong competition for clientele. In our view, this is a salient element that should be taken into consideration when prevention programs are designed. As in other workplaces, SW have problems with colleagues that may escalate into violent offences if not prevented. In addition, colleague SW can have the role of guardians and therefore the social cohesion among SW seems highly relevant for crime prevention programmes.
Our findings reveal associations between categories of perpetrators and types of victimisations. Thefts and physical assaults are mainly committed by pedestrians unknown to the SW, while sexual assault and fraud are more likely to be committed by their customers. These patterns are consistent with opportunity-based theories (Hindelang et al. 1978;Cohen and Felson 1979) and can be explained by a differential exposure to the risk of victimisation. The latter is higher when the SW works in various settings (indoors and outdoors), when the victim is isolated with the offender (i.e., when there are no controllers), like cisgender and transgender SW with their male customers, and is influenced by the type of activity performed, which explains the kind of crimes committed in that context. There seems to be, nevertheless, some specialised offenders that do not hesitate to impersonate a client to rob a SW. In addition, sexual arousal plays a major role in the victimisations committed by the customers. Threats, as well as physical and sexual assaults are often triggered by the clients' inability to climax. This is consistent with the literature showing that sexual arousal decreases self-control and limits the personal strategies available to manage problematic situations (Ariely and Loewenstein 2006).
The transgender SWs in our sample are overrepresented among victims, a finding which is exactly the opposite of Mitjans-Núñez and Molnar's (2016). In that study, conducted in Spain, transgender women were the less victimised group, but the authors clarified that the transgender SW themselves attributed their low level of victimisation to their tall and fit body shape; on the contrary, the transgender women in our sample are largely short and slim, and it is well known that physical characteristics affect the risk of victimisation (Felson 1996). Nonetheless, the available research on victimisation of transgender SW is scarce (Weitzer 2005) and our sample too small to draw any conclusion. Having worked on this topic in two different countries, the first author considers that future research should try to measure the community's general degree of tolerance or hostility towards transgender individuals as it may have an influence on their levels of victimisation.
As in other countries that regulate sex work (Abel et al. 2010;Benoit et al. 2019), SW have a positive perception of the Swiss police, but this does not mean that they consider them as competent nor that they report the victimisations suffered. This is a result consistent to other posterior studies conducted in Switzerland (Molnar and Ros 2022). However, offence reporting is also influenced by the legal situation of the SW in the country, as we have seen that legalisation does not mean that all SW are respecting the applicable rules. In addition, police patrolling may discourage potential customers, in such a way that SW face a conflict between their need for protection and their need for privacy. Somehow similarly, police officers face a conflict of roles: on the one hand, they are SW's guardians; on the other, they must enforce the law, which sometimes means punishing SW working without a permit or working at forbidden places or times. This is a classic conflict faced also by community police officers (Armenta 2016). Nevertheless, our study could not shed sufficient light on super-controllers such as the media, the politicians, or police management. This is therefore an area that deserves further investigation. This study includes diverse SWs' voices-cisgender and transgender, documented and undocumented-in one of the few countries where sex work is regulated. Data collection was preceded by 18 months of social work with the participants through an NGO that has been supporting SW for more than 20 years, something that should have reduced the bias due to the social desirability effect (Paulhus 1984). Similarly, 12 of the 14 SW could use their mother tongue during the interviews (French, Romanian, or Spanish), which increased the fluidity of the interviews and the rapport of confidence between interviewer and interviewee. As a means of increasing reliability, we created a semi-quantitative database with the information collected. This allowed a relative quantification of the phenomena studied, avoiding thus the anecdotal use of quotations and suspicions of cherry-picking. Instead, the quotations presented in the paper were used to contextualise the findings.
Limitations of our study should also be mentioned. First, we cannot overlook that we are facing here a potential conflict of roles, because the first author was engaged both as a social worker and as a researcher, and that could have led to unconscious biases during the data collection and analysis. For that reason, we interviewed not only SW with whom there was some sort of close relationship, but all SW who agreed to be interviewed, while maintaining a balance with respect to gender and legal status. Second, our research could not reach men who exercise sex work, a population that presents different characteristics and exercise in different manners compared to the cisgender and transgender female SW (see for e.g. Minichiello et al. 2002Minichiello et al. , 2013. Third, other the limitations of this study are mainly related to its small sample, which indirectly affects the crime index used. Prevalence, for example, is a sort of period prevalence starting when each interviewee began SW, which means that its length differs from person to person, as there are not enough cases to concentrate on the last year only. Similarly, the responses of the interviewees were too generic to allow measuring the incidence or frequency of their victimisations, a measure that would have allowed a more detailed analysis of the phenomena under study.

Conclusion
This study was guided by several research questions whose answers can be summarised as follows: The 14 SWs that participated in our study had been victims of work-related offences, particularly threats, insults, and sexual assaults. Few SWs reported the offences suffered to the police although in general, they said that they appreciated the Swiss police. General variety of victimisation in on average of two different offences, although transgender SWs are overrepresented among the victims of multiple offences: having been on average victims of three different offences. The perpetrators of the offences were customers, passers-by, colleagues, other colleagues' pimps, and managers of erotic massage salons. There was a certain degree of specialisation among the perpetrators, in the sense that clients were more likely to commit sexual assault and fraud, whilst passers-by were more likely to commit theft and physical assault. The places where offences occurred can be defined as risky settings. The SWs themselves, the police, the NGO's social workers, the managers of the erotic massage salons, and employees of private security enterprises fulfil roles of controllers, whilst the Swiss criminal code, the law on prostitution and the NGO's management fulfil roles of super-controllers.
Our results highlight that prostitution remains a hazardous job even when it is legal. This suggests that public policies should not be limited to simply regulating prostitution, but also to ensuring SWs' safety through the development of evidence-based prevention strategies. In that perspective, future research could focus in the specific context in which sex work takes place to propose tailored situational crime prevention strategies (Cornish and Clarke 2003). In the city where this study was conducted, the context is given by the prostitution area and the erotic massage salons. Our findings also suggest that, in the immediate future, interventions could focus on increasing offence reporting by the SW, increasing their knowledge of the procedure they should follow and of their rights as crime victims, as well as increasing the social cohesion among the SW.