Keyword

1 Sex Work: Definitions, Delimitations, and State of the Art

Criminological research on sex work is not a hot topic, but it is not scarce either, even though most studies have been conducted in countries that forbid this activity. In this chapter, I address the methodological and ethical aspects of criminological research on sex workers that my colleagues and I conducted in Switzerland, a country where sex work is a legal activity. Through the sharing of my own research experience, inspired naturally by past studies, I develop a reflection about conducting criminological research when wearing two “hats”: one as a social worker and the other as a researcher.

1.1 Sex Work in Switzerland

It is imperative to define what we mean by sex work as this is a varied phenomenon (Azhar et al., 2020; Harcourt & Donovan, 2005) that includes many sexual practices under remuneration in a diverse context. Sex work, also called prostitution in the past, has been regarded differently depending on the epoch and culture (for a review, see Jenkins, 2020). Currently, its regulation or criminalization depends on each country (Danna, 2014; Jahnsen, 2019). For instance, in countries such as Romania or Croatia, prostitution is a misdemeanor, while in the Netherlands, Germany, and Switzerland, it is a freelance economic activity. Other countries, such as France and Sweden, criminalize customers of prostitution but not the sex workers per se, who are considered as victims of the society. Moreover, it is important to note that sex workers are a heterogenous “group” composed of persons of highly different socioeconomic and cultural backgrounds, and therefore, making generalizations should be avoided if heterogeneous samples are not sufficient (Shaver, 2005).

Switzerland is a federal country that is composed of smaller administrative units, the cantons, which possess a great autonomy in ruling themselves. Areas such as the management of prisons, police, and justice are canton-dependent, for instance. The same is true for sex work. Specifically, the Canton of Vaud, situated in Western Switzerland, where my research was conducted, considers sex work as “the activity of a person who habitually engages in sexual acts or acts of a sexual nature, with a fixed or undetermined number of clients, for remuneration” (art. 1, LOI 943.05 sur l’exercice de la prostitution [LPros], 2004). Street sex work is permitted only in a specific area in Lausanne, the capital of the canton (Ville de Lausanne, 2016). At night, sex work through street soliciting is permitted from 10 p.m. until 5 a.m. Sex workers are therefore allowed to solicit their customers in the street, but the sexual service must be conducted elsewhere.

On the other hand, indoor sex work is permitted in the whole canton in premises that have obtained an administrative license for erotic tariffed activities. These are named by the law as “erotic massage salons” (LOI 943.05 sur l’exercice de la prostitution [LPros], 2004; Ville de Lausanne, 2016) and can be an apartment, which is the most prevalent type of erotic massage salon, but also bars or clubs with private rooms. In the latter, the sex workers gather in the main bar area and initiate interactions with customers of the bar. If the sex workers and the customer agree upon the conditions of the sexual service (price, sexual practices, etc.), they go to the private area where they can conduct the service in a private room. However, advertising sex services at a hotel or at one’s domicile is not allowed and is punishable with a penal fine (art 199, Swiss penal code).

All erotic massage salons have a manager who rents the rooms to the sex workers. In some instances, sex workers can work alone in a studio, or together with more sex workers in a multiple-room apartment. The managers must ensure that the premises are in a proper state and that the sex workers are legally entitled to work and have a valid work permit or authorization. However, the managers cannot exploit profit from the sex work per se, which means that all gains, in principle, go to the sex workers.

1.2 Criminological Focus: Victimization in Sex Work

Victimization in sex work is a relevant topic because of the various risks that sex workers face during their job (Rekart, 2005). In this regard, victimization has been studied in many countries, as mentioned above, but almost all of these countries prohibit sex work (Barberet, 2000; Berger et al., 2018; Brents & Hausbeck, 2005; Bungay & Guta, 2018; Campbell et al., 2019; Chan & Beauregard, 2019; Cunningham et al., 2018; Karandikar & Próspero, 2010; Peitzmeier et al., 2019; Sanders & Campbell, 2007) with only a few exceptions (Földhazi, 2010; Ros, 2021). These studies described the occupational hazards faced by sex workers, who have an even higher likelihood of becoming victims of murder compared to the general population. However, sex workers’ risks also seem dependent on the legal framework of each country (Benoit et al., 2019). In that regard, it has been demonstrated that sex workers’ victimization is less in countries where sex work is legal (Benoit et al., 2019), even though more extensive research is needed on these legalities.

To a lesser extent, sex workers’ drug use has also caught the attention of academia, specifically of those sex workers who work to finance their drug addiction (Surratt et al., 2004; Young et al., 2000). These studies focused specifically on drug-addicted sex workers, while the drug consumption among a general sample of sex workers has only scarcely been addressed by scholarship. One of the only exceptions is the Swiss study of Lociciro et al. (2017) who found that the rates of consumption of illegal drugs are higher than those of the general population, but they were still low and did not exceed 10% among sex workers.

Academia has endorsed the use of qualitative methods to study the victimization and drug use of sex workers. Semi-structured interviews have been conducted in different countries with different laws for sex work (Barberet, 2000; Brents & Hausbeck, 2005; Bungay & Guta, 2018; Campbell et al., 2019; Földhazi, 2010; Karandikar & Próspero, 2010; Sanders, 2001; Sanders & Campbell, 2007). Quantitative methods such as structured questionnaires have been used to a lesser manner (Berger et al., 2018; Campbell et al., 2019; Lociciro et al., 2017; Peitzmeier et al., 2019; Surratt et al., 2004; Young et al., 2000). Another type of data that proved useful for conducting research was police data (Chan & Beauregard, 2019; Cunningham et al., 2018; Potterat, 2004) and, to a lesser extent, focus groups (Surratt et al., 2004). It is worth mentioning that Campbell et al. (2019) applied participatory action research in which they incorporated their gatekeeper (an aid organization) in the design of their study.

2 Fieldwork: Being in the Right Place at the Right Moment with the Right Colleagues

Between 2016 and 2021, together with other colleagues, we interviewed or surveyed around 140 sex workers in three main studies whose aims were to understand sex workers’ victimization, trust in reporting to the police, and their drug use (Molnar et al., 2021; Molnar & Aebi, 2022; Molnar & Ros, 2022). We used interviews, questionnaires, and participant non-systematic observations (see Fig. 9.1). sex workers are considered a hard-to-reach population: One cannot just go into the field without knowing anyone or anything about the topic. Former researchers mobilized gatekeepers (e.g., nongovernmental organizations [NGOs] for accessing sex workers (see Barberet, 2000; Berger et al., 2018; Campbell et al., 2019; Karandikar & Próspero, 2010; Lociciro et al., 2017; Surratt et al., 2004). Other actors who proved useful were the police (Barberet, 2000), owners of sex work clubs (Barberet, 2000; Brents & Hausbeck, 2005), attorneys (Brents & Hausbeck, 2005), and the press (Barberet, 2000; Brents & Hausbeck, 2005). The details of these studies are too vast to be explained in this manuscript, which is the reason why we invite the readers to delve into the references for further details.

Fig. 9.1
A diagram represents the timeline of the surveys follows. 2017, victimization and routine precaution, qualitative interviews N = 14. 2018, drug use, survey N = 91. 2020 to 2021, victimization and drug use, survey N = 40.

Studies conducted on the domain of sex work

Full size image

Our studies were possible because we were practitioners working for an NGO supporting sex workers and assisting them in various domains, such as distributing prophylactic materials, giving advice about the prevention of sexually transmitted infections, assisting them with their administrative duties (e.g., working permits and tax declarations), and accompanying them to the hospital, the police, and elsewhere. We therefore did not need a gatekeeper because we were in the field on a regular basis, in both the prostitution areas and the erotic massage salons. In the former, the social workers of the NGO were onsite from 10 p.m. to 1:30 a.m., driving a caravan to three different emplacements. Sex workers could enter the caravan and stay for a while, have a beverage and a snack, obtain prophylactic materials (condoms, gel, and tissues), and receive orientation and counseling. On a weekly basis, the social workers visited the erotic massage salons of the whole canton of Vaud, usually during the afternoon, and offered sex workers the same kind of information and materials. I thereby approached my field via my professional practice, and most of the participants knew me long before I started the data collection. My ethnic background was an asset in this case because I was born in Romania and grew up in Spain, and I am fluent in Romanian, Spanish, French, and English. My linguistic skills greatly facilitated my research because the sex workers in Switzerland are frequently foreigners. The rapport was constructed during months in which, due to my work, confidentiality and anonymity were constantly addressed.

The first study (Molnar & Aebi, 2022) referred to the work-related victimization that sex workers (N = 14) endured during their professional experiences, their perceptions of the Swiss police, their reporting to the police, and the prevention techniques that they used to prevent victimization. The second study (Molnar et al., 2021) addressed the drug consumption of sex workers (N = 91), which was previously scarcely studied.Footnote 1 The third study (Molnar & Ros, 2022; Ros & Molnar, 2022) studied the impacts of the COVID-19 pandemic on the sex workers.

3 Getting Started: Pragmatism as a Guiding Light

The first study that we conducted in 2017 (Molnar & Aebi, 2022) was the foundation for the other two, and therefore, it is the most important of the three. I started working for the NGO in 2016 as a social worker, since the position of criminologist did not exist. At that time, I had already been volunteering for 4 years in Spain for Doctors of the World (Médicos del Mundo) in assisting sex workers, and I was especially interested in the work-related violence endured by sex workers. The director of the NGO, Silvia Pongelli, also quickly manifested her interest in this topic, because the association lacked systematic knowledge on this area, and they wished to do more to tackle violence. We therefore decided to initiate a study on the violence endured by sex workers.

Our research, far away from academic purposes, had a purely pragmatic goal: to understand sex workers’ victimization in order to propose prevention strategies to the sex workers that we encounter on a daily basis. This was later reached via the Previst project (Molnar, 2019, 2021) that we designed and that was financed by the Swiss Federal Police. The data collection would take place approximately 1 year after I was hired by the NGO as a social worker. At that time, my field experience was about 1100 hours. These observations were not done with any goal since they were part of “the job,” but, of course, they allowed me to gather considerable insider knowledge prior to conducting the interviews.

Nowadays, it is typical to start the research by submitting a protocol to the ethics committee (see Berger et al., 2018), but this was not our case. At that time, no one belonged to any university and the board of the NGO was positive about the project and its specifics. The director and I had some background in social science research, and so we planned the research and brainstormed on the potential risks, being logically inspired by articles that had already problematized these issues (especially Shaver, 2005).

We prepared a research protocol in which we planned the methodology that was to be used. In this case, it was a qualitative method using interviews. We also brainstormed and decided on many of the ethical and methodological aspects. Similar to Barberet (2000) and Sanders (2001), we planned an interview guide containing the main topics we wanted to address in our interviews.

This was also the case for the study of the sex workers’ drug use and the impacts of the pandemic on the sex workers. For both cases, we prepared questionnaires guided by the feedback from fellow colleagues (social psychologists, doctors, and social workers) that were as well-pretested with several people (depending on the case with sex workers and non-sex workers). We also planned to keep a diary during the fieldwork in which we reported our own impressions about the questionnaires. Some quotations from the diary are used in the next sections to illustrate the challenges we faced.

Following the recommendation of Shaver (2005) regarding the need to include heterogeneity among the sex worker samples, our first study’s goal was to reach around 12 to 16 sex workers with four different profiles: (1) cisgender women, (2) transgender women, (3) sex workers of legal status, and (4) sex workers of illegal status. We decided that we would not include those sex workers whom we knew were victims of trafficking in human beings or pimping because of the distinctive characteristics and types of violence endured, which could not be generalized to the other sex workers. Nevertheless, this was also a way to protect the potential participants by avoiding situations such as those mentioned by Sanders (2001): “On one occasion a sex worker was physically assaulted by her pimp as punishment for taking part in the research” (Sanders, 2001, p. 8).

Basic ethical rules regarding the anonymous and confidential characteristics of the study (Barberet, 2000; Karandikar & Próspero, 2010) were also applied in our study, but we emphasized the fact that no one from the NGO would know about their answers. Our interviews would follow the same approach used for our interventions, which was to show the maximum respect for the self-determination of the sex workers. This is slightly different from other studies, such as Shaver’s study (2005, see quotation below). That research, even if ethically correct, was conducted solely by scholars, and therefore, it is understandable that they were “insistent” when recruiting participants. We decided to avoid this in order to follow the line of action of the NGO.

We also made it clear during these conversations that participation was voluntary and that we would take no for an answer and move on politely. In doing so, however, we also pointed out that, although it was their right to say no to an interview, it was our job to keep on trying. (Shaver, 2005, p. 301)

Together with the NGO’s director, we decided another important aspect was my safety and that of my interviewees. We also decided that I would recruit the sex workers during working hours when I worked in tandem with a colleague. At all times, I would inform her about where I was going and at what time to conduct the interviews. We decided that we would obtain consent orally so that the sex workers could avoid providing too much personal data. This is not different from other studies (Bungay & Guta, 2018; Karandikar & Próspero, 2010; Sanders, 2001; Sanders & Campbell, 2007). It is highly probable that today this might be criticized by the ethical committees at most universities, but hard-to-reach populations do not share the same type of cultural views as people belonging to Western, Educated, Industrialized, Religious, and Democratic (WEIRD) societies.

4 Techniques Employed: Much Diversity, But Never Enough Data!

4.1 Sampling Challenges and Participant Recruitment

The first challenge that one faces when conducting research with sex workers is the sample size. In this regard, we kept in mind the following:

The size and boundaries of the [sex worker] population are unknown, so it is extremely difficult to get a representative sample. The traditional methods of sampling such populations—snowball sampling, key informant sampling, and targeted sampling—do not solve this problem. (Shaver, 2005, p. 296).

In each of our studies, we aimed to cover all profiles and heterogeneity of sex work. For that, we approached sex workers who worked (1) in the streets, (2) in the erotic massage salons, (3) those who came to the office of the NGO, and (4) those working on the Internet. The latter were included only for the last study. In terms of sample distribution, we obtained a majority of cisgender women and some transgender sex workers but no men. In terms of age, we reached approximately a representative sample of sex workers. The same is true for people in legal and illegal situations.

Having in mind the need for heterogeneity in the samples, we approached the sex workers most of the time during our working hours and let them know about our research. It might not be obvious from the beginning, but the naming and explaining of a study are cardinal issues. For instance, Barberet (2000) avoided using the word “victim” and instead used “safety.” In our case, we decided to use a broad description and present the study as “a study on the experiences in sex work” as well as “experiences of sex workers during COVID-19 pandemic.” However, with regard to the drug use study, we had no choice but to call it, “Use of psychoactive substances and medication among sex workers active in the streets of Lausanne (prostitution area) and in the salons of the canton of Vaud.” We had no major issues about the naming, except for one sex worker who inferred that we assumed that sex workers were a vector for COVID-19 and a few who commented that “they do not take drugs,” to which I replied that we are interested in knowing about those who take drugs and those who do not.

Details count when recruiting sex workers participants: being very conscious of the space we occupied in relation to those approached; never corner a person in a doorway, or if with a partner, never approach a sex worker from two sides. Leaving them room to move away is a courteous way to demonstrate that you are not the police. More important, it provides a clear indication that the choice to participate is theirs. (Shaver, 2005, p. 302).

We applied the same principles when recruiting the sex workers. Interestingly, sex workers were interested in being interviewed but not in the places that we planned in our first protocol (Molnar & Aebi, 2022). We hoped that interviews were carried out offsite and outside of working hours in a neutral environment that facilitates discussion out of the sight of indiscreet third parties. However, the sex workers proposed that the interviews happen in the prostitution area or the massage parlors. Even though at the beginning I was quite tenacious to enforce this “rule,” from the 14 sex workers that I interviewed, this “ideal scenario” happened only around eight times. Actually, then I realized that some sex workers—especially those who worked indoors or who combined indoor with outdoor work—were tremendously busy and had no time for drinking coffee with a researcher. We therefore conducted the interviews in the caravan of the NGO when it was transited by other sex workers who were not speakers of that language; in the street, next to the caravan but far away from other people; in one of the remotest streets of the prostitution neighborhood; and even in the erotic massage salons in empty rooms. Surprisingly, this seemed to have worked since the sex workers showed signs of feeling comfortable despite the environment.

Therefore, for the other two studies (Molnar et al., 2021; Molnar & Ros, 2022), we were more relaxed and confident about this manner of interviewing, and most of our data collection took place “in the field.” This was not unique to our study. For instance, Barberet (2000) faced the same problem: Even though her team’s ambition was to interview the sex workers in “private places,” those working at rural clubs did not having any possibility of going outside. Földhazi (2010) and Sanders (2001) stated that they chose sex workers’ place of work, without going into too much detail.

When the COVID-19 pandemic emerged, we decided to go hybrid. Besides face-to-face surveys, we also conducted online surveys, taking, for example, the innovative study of Campbell et al. (2019) who interviewed the sex workers remotely. Equipped with face masks and hand sanitizers, we surveyed sex workers at the NGO’s office, erotic massage salons in the afternoons, and the street at night. We also rendered our questionnaire online in all four languages mentioned previously—French, English, Romanian, and Spanish—and disseminated it either via the erotic announces formularies or via WhatsApp by sending it to the public telephone number that was announced in the advertisement. Only seven people answered the online questionnaire, but among the 40 members of the sample, this seems like an adequate proportion. Nevertheless, several sex workers misunderstood our study and wrote us back asking for more information.

[11:16, 17.02.2021] SW2: “Excuse me who are you and where did you get my phone number? I don’t give out personal information.”

[11:17, 17.02.2021] Lorena: “Hi, I’m Lorena Molnar. I work for the University of Lausanne, and we are doing a study on the impact of the pandemic on the lives of sex workers. I found the number on [website].”

[11:17, 17.02.2021] SW2: “I don’t know who you are.”

4.2 Procedure During the Data Collection

Most studies conducted their data collection face-to-face and audiotaped and transcribed the interviews (Barberet, 2000; Brents & Hausbeck, 2005; Bungay & Guta, 2018; Földhazi, 2010; Sanders & Campbell, 2007). This was our procedure: For all interviews, except one, I could audio record. For the remaining one, I took manuscript notes. This was due to the fact that the sex worker did not feel comfortable about being recorded.

For the surveys, we did not have this need, as the data were recorded on an online survey platform called LimeSurvey. I admit that we did not really put much thought into our choice of the platform. It was a secure platform and the one for which the subscription was paid by the NGO and later by the University of Lausanne. It is still important to consider this aspect beforehand in order to ensure that the answers are protected and confidentiality is assured. I collected the data on a tablet or on my phone via an online questionnaire in Romanian, French, Spanish, and English. There were also a couple of sex workers who were not fluent in any of the languages proposed, and for reaching out them, we counted on the assistance of a social worker from the NGO who was fluent in the specific languages.

Already in the 1990s, scholars discussed “validity” and “reliability” in the domain of sex work. An instrument is valid if it measures in an efficient manner a phenomenon and is reliable if, when used in repeated occasions, one always obtains the same outcome no matter who uses it (Aebi, 2006). Vis-à-vis the former, the impossibility of reaching out to high-level sex workers such as escorts or male sex workers is a challenge. Regarding the latter, the social desirability, the labeling of the questions, or the characteristics of the interviewer can pose serious problems in research. For instance, Barberet (2000) discussed the sex workers who claim “that everything is great” and do not elaborate much; this is a sign that the sex workers are not being honest with the researcher.

In our studies, we faced similar validity challenges. Even though we included several escorts, we could seldom find any men who did sex work. Another measuring problem was related to the estimation of the number of times that they had been victims of a crime. Sex workers did not remember these details and just answered “many times,” “sometimes,” “a couple of times,” etc. The same happened with their estimation of their monthly income, which prevented us from calculating their economic status. Regarding reliability, it is also worth mentioning that the labeling of our questions was challenging in some instances. For example, in the drug use study (Molnar et al., 2021), we realized that even though a sex worker is a regular drug user, she would be acquainted with a maximum of three substances. However, our questionnaire addressed the consumption of more than 10 drugs. As well, when asking the sex workers if they were at risk for COVID-19 and enumerated to them the list of conditions that put one at risk (advanced age, high blood pressure, diabetes, etc.), a sex worker responded negatively, but afterward she let us know by coincidence that she actually has high blood pressure.

The interview went very well, seemed wary at first, then very happy to talk (thanked me several times). Hard to get her to tell me how many clients insisted that she lowers prices. She didn’t say how much she was making...I don’t feel like she didn’t want to say, but couldn’t figure out monthly average (even when I insisted, “Can you give me an approximate figure?”). (Fieldwork journal, January 2021).

The data storage of the participants’ responses is seldom mentioned by the scholars that studied the sex workers. The exception is Campbell et al. (2019) who mention that the interviews “were recorded on a digital recorder, under strict data management procedures due to the sensitive nature of the data and the need for high levels of security around identity of the participants.”

In our case, we stored the first interviews on an external hardware and then on a safe server at our university. Surveys were stored on the university’s LimeSurvey platform. In this regard, anonymization is a highly relevant aspect: Samples with sex workers tend to be small and local, rendering them easily recognizable. To protect the sex workers’ anonymity, I avoided disclosing in our papers too many of the socio-demographic characteristics of specific interviewees, and I only provided general information about the sample. Also, we “baptized” our participants with pseudonyms, in one case, chosen randomly from the novel, The Girl with the Dragon Tattoo (Larsson, 2017).

4.3 Data Analysis

The section data analysis seems to be the least extensive in the papers the address sex workers, and I believe that this should be different. In our studies, we used both quantitative and qualitative techniques for analyzing the data. First, content analysis, used by many scholars (Barberet, 2000; Sanders & Campbell, 2007), was especially enlightening to understand the victimization endured by the sex workers and its contexts, as well as the feelings and perceptions of the sex workers during and after their victimization. We analyzed the interviews horizontally, going through each transcript, and after identifying the recurrent topics, we also analyzed them transversally, going across transcripts through each topic identified.

We also created a small database with relevant variables emerging from the interviews that could not be grasped in a qualitative manner; for instance, the number of different victimizations endured by the same person is known as the so-called variety (Aebi, 2006). In this way, we could analyze that data by groups (transgender versus cisgender), status (legal versus illegal), the number of victimizations reported, the multiple victimizations, etc. We gained a better picture of the victimizations of the sample group, while not neglecting the context details.

The two surveys that we conducted with 91 sex workers and 40 sex workers (Molnar et al., 2021; Molnar & Ros, 2022) allowed us to do descriptive analyses of frequencies and cross-tabulation. In the first survey, we also conducted a logistical regression. Since the sample was not large (N = 91) and the sex workers’ drug consumption was low, we could not compute a “predictive model” of the risk factors that increased drug use in sex work. Instead, we only analyzed two risk factors—age and place of work—in relation to using alcohol and tobacco. The reason is that these were the substances consumed at a higher rate and therefore the only ones that could be statistically analyzed. Although this is unfortunate, it is the reality of many surveys that explore rare phenomena among hard-to-reach populations.

5 Ethical, Emotional, and Safety-Related Considerations

In my view, the first ethical challenge about studying sex work is its socially controverted character (Benoit et al., 2019). In this regard, almost everyone has an “philosophic” position about the phenomenon and several scholars perceive sex workers as perennial victims. This can influence the relationship with the participants and the manner in which we treat them. Furthermore, “Prostitution is commonly treated as an identity category rather than a revenue-generating activity” (Shaver, 2005, p. 297). This can increase the risk that researchers do not treat their participants as adults with agency but as “inherent victims.” Our position vis-à-vis sex work coincided with the one adopted in the legislation of Switzerland. In other words, we considered the sex workers as adults who, for different reasons, worked in the sex industry. It must be stressed that in our studies we also adopted a post-positivist philosophy, which consists of describing and analyzing the data as objectively as possible, while keeping a reflexive orientation (Phillips & Burbules, 2000). This did not prevent us from being misunderstood and having our manuscripts rejected by anonymous referees who had different philosophical positions regarding sex workers.

Other ethical dilemmas appeared when conducting research in sex work, namely the psychological and safety-related risks, participants’ recompense, and sex workers’ stigmatization. Barberet (2000) refers to the participants’ sadness and tears after remembering some traumatic event, and Sanders (2001, p. 8) mentions the risks for sex workers’ safety:

On one occasion, a sex worker was physically assaulted by her pimp as punishment for taking part in the research. This highlighted the risks that individuals took to be part of the research. These ethical dilemmas highlight the difficulties of doing research in a community that is saturated with violence and suspicion of officials.

Having this in mind, we aimed to minimize this risk by being transparent about the research with the whole group of sex workers. We approached a group and explained in public our study in generic terms (“experiences in sex work,” “challenges of sex work,” and “substance consumption”) and tell them that the NGO is interested in these topics and would like to discuss with as many sex workers as possible. We also stated that we wanted to discover the general trends and not the details about individuals in order to understand in which areas we can assist better.

Physical threats also affected the researcher. For instance, they might come simply from being outdoors at night, from the sex workers, or from criminals such as pimps or criminal networks. Although I did not directly face most of them, I was, on a few occasions, confronted by strong reactions from the sex workers. The sex workers, especially if they are going through a difficult time, might easily feel judged, such as a sex worker who thought that we had inferred that she was a vector of the coronavirus. Passersby or customers can also pose a risk. For instance, as a woman in her 20s, I was faced with passersby who assumed that I was a sex worker and requested a sexual service.

Regarding the psychological risks that sex workers encounter when participating in these kinds of studies, we decided to pose our questions in a broad manner, e.g., “Did a customer impose sexual practices that you were not comfortable with?” After affirmative responses, we tried to clarify the event by asking for more details. Conversely, if the participant decided not to elaborate, we respected her wish without further insistence.

The latter situation is related to the need to connect and sympathize with the sex workers (Barberet, 2000). However, an almost unavoidable asymmetrical situation emerges at the end of an interview. In principle, a criminologist would be satisfied if they found out the relatively important prevalence of sex workers’ victimization and interesting details that can enlighten further research and public policies. Conversely, a participant who has been the victim of a crime, or of several, does not necessarily feel the same enthusiasm. On the contrary, she might feel especially vulnerable after the interview.

In our research, we referred the participants to the NGO, letting them know that social workers are also available to discuss violence and rights in case of victimization and accompany the sex workers to victim assistance services. Listening to the sex workers’ experiences of violence can affect the researchers’ mental health. Mirror neurons activate automatically in humans, and naturally, we were emotionally affected when we met people suffering from crude violence, poverty, or, a few times, illnesses such as HIV. Writing in a journal or debriefing colleagues or my closest acquaintances has always been beneficial. Personally, I prefer debriefing about my emotions with my inner circle and letting fellow colleagues offer better methodological advice.

In the first two studies, we offered a sort of recompense either paying for the beverage of the sex workers who agreed to be interviewed or offering her something such as a body lotion. However, the COVID-19 pandemic highly limited our possibilities to offer recompense since the study was conducted without any funding. Sanders (2001) develops a very interesting reflection by illustrating her own dilemmas regarding this issue. The advantage highlighted was the acknowledgment of sex workers’ participation. The disadvantages of giving a recompense were the feeling of exploiting the sex workers or the fear that they would use the money to buy drugs. Most of the time, researchers cited a material recompense for the sex workers that consisted of the payment of some pocket money of around $20 (Barberet, 2000; Bungay & Guta, 2018; Campbell et al., 2019; Karandikar & Próspero, 2010; Sanders, 2001).

The last ethical consideration is the risk of stigmatizing the sex workers. In my view, scholars should be selective about their presence in the field and shift weight onto the relevance of our studies and the inconvenience that our presence is going to cause to the sex workers’ participants. They might feel like research “objects” who are often studied but seldom assisted in tackling their daily problems. We should also reflect in advance on the manner in which our research might affect sex workers’ public image. For me, this poses a great dilemma because, in my view, one must be faithful to the data and avoid activism-led science. At the same time, one also needs to be aware of the limitations of the studies and report and communicate them in the clearest manner.

6 Plan, But Be Flexible: Lessons Learned and Methodological Perspectives

This section summarizes the lessons that I, an early career researcher in criminology, learned during my fieldwork. First, understanding the context before collecting data from participants is highly useful. I recommend taking advantage of all opportunities to talk with the actors involved in sex work: sex workers, social workers, neighbors, police, etc. Once one has the whole picture, the data collection is likely to be richer. Understanding all parties could also be positive for the researchers’ and participants’ safety. This was, in my case, happening naturally because of my role as a social worker. However, this should be part of the research protocol if one does not count as an “insider” or one who is not an insider themselves.

Second, when planning a study, consider that in central European countries, there are many foreign sex workers, hence having a multilingual team can be an invaluable asset and increase the validity and the reliability of the research. During the research protocol, it is vital to be rigorous but also open-minded and creative. Most research protocols in sex work might need to be revisited and adjusted. As well, there might be threats or difficulties that have not been anticipated. Therefore, it is imperative to imagine a series of scenarios that could degenerate, such as if a sex worker has a panic attack when conducting the interview, a participant insists on being part of an individual interview, a participant expects more recompense, or a participant who does not show up.

As recommended by Barberet (2000), research with sex workers needs empathetic researchers; thus, “emotion” is something to anticipate when designing the research protocol in the sense that the interviewer might react to crude events with a certain degree of emotion. For instance, during the COVID-19 pandemic, a sex worker disclosed to me that she had been homeless and could not eat much during that time. I could not help my reaction of telling her that I was sorry that she was facing such a difficult situation. I also paused the interview to let her know that the NGO could assist her in finding temporary housing and give her vouchers for the supermarket. In that regard, it might be desirable that early career researchers receive emotional training so they can debrief others and prepare themselves for the anticipated fieldwork. Alone in front of a mirror, practice approaching participants, asking questions, and dealing with refusals. In relation to refusals, I recommend to never take the refusals of no-shows personally or visibly become irritated with the participant. This is wrong both from an ethical and strategic point of view. Some potential participants lose interest in participating in the study and it is their right, but others might recover their interest after some time if the researcher is still in the field.

During the data collection, be prepared to conduct imperfect research. One prepares with much care the interview plan and the topics and subtopics one would like to address in order to answer satisfactorily to the research questions. In actual practice, it is possible that all topics will not be covered during the interview because the participant is willing to address the theme that is most important for them. It is recommended to redirect the conversation, but just as important is to give space to the interviewee and avoid over-insisting. In case of conducting surveys in the streets or in erotic massage salons, collecting the data on online platforms is a safer option. If one is the victim of a theft or simply loses the device, the data will not be lost or accessible. Because of sex workers’ lack of time and lack of accustomedness, I recommend avoiding long questionnaires, such as the International Crime Victims Survey (van Dijk et al., 2007). If applied to sex workers, this questionnaire should be adapted. The length of the questionnaire is not a straightforward issue to decide. On one hand, one needs nuance in order to understand satisfactorily a phenomenon, but on the other, too much nuance can challenge the reliability of the study by the response burden.

Take care of your participants and their privacy. Consider beforehand your storage and anonymization strategy. Be faithful to your commitments, and honor your promises toward your interviewees and be grateful for the time and intimacy they shared with you. However, if you feel overwhelmed, I suggest leaving the field at least temporarily. Too much emotion can blur your perceptions and the validity of your research.

7 Conclusion

Research on the sex work industry is an interesting but challenging area. The collaboration with NGOs is highly valuable since they already have a strong rapport with the sex workers. Observations, interviews, and surveys are possible and complementary, but they need careful consideration in order to be adapted to the sex workers’ characteristics and to the sex work settings, i.e., legality or illegality. Safety of both the researchers and the participants should be at the core of the research protocol, which should be flexible enough to handle unexpected events. Emotional intelligence is vital in order to interact with interviewees and to gain perspectives when gathering and analyzing the data.