Introduction

Confidentiality represents a core principle of research ethics and forms a standard practice in social research (Kaiser 2009). Researchers rarely choose to break confidentiality, for several reasons. Breaching confidentiality may expose informants to harm (Baez 2002). It also reduces the trust that is required when collecting valuable data (Finch 2001), especially in qualitative research and research with vulnerable participants (Liamputtong 2007). Furthermore, it may undermine the researcher’s credibility and ability to conduct research and future studies in given community (Israel 2004). In addition, it demonstrates a lack of respect for the informants by ignoring his or her privacy and right to self-decide (see American Anthropological Association 2012).

In spite of the negative consequences, some sociological and anthropological codes of ethics allow for a breach of confidentiality, for example in the case of child abuse (British Sociological Association 2017). In addition, the law may obligate researchers to disclose any knowledge of specific crimes and, in certain circumstances, to hand over research material to law enforcement authorities. Legal restrictions on confidentiality constitutes a contentious issue among researchers (Lowman and Palys 2014). However, few systematic studies regarding the ways in which researchers manage their knowledge of illegal and/or harmful activities exist. An exception is provided by Wiles et al. (2008), who have studied the opinions and experiences of British researchers working with vulnerable groups concerning the protection of confidentiality and its accidental or intentional breach. The study demonstrated the diversity of researchers’ attitudes towards confidentiality breaches when their research participants were at risk of harm or where information needed to be disclosed pertaining to their research participants’ involvement in crime or other illegal activities. On the one hand, most researchers working with vulnerable groups such as children and people with learning disabilities felt legally bound to report abuse, and so informed their informants of this possibility prior to data collection. The researchers claimed that they would always consider the consequences of a breach of confidentiality in advance, and would ask participants whether they agreed to disclosure. On the other hand, the researchers who conducted studies in criminal environments kept their knowledge confidential and did not feel compelled to disclose information regarding illegal activities. Indeed, they believed that otherwise they would not be able to carry out their research. In some cases, researchers attempted to avoid knowledge of non-convicted offenses, thus freeing them from the dilemma of whether or not to report. They were also aware that their data might be of interest to law enforcement authorities.

Although Wiles et al.’s study offers important insights regarding the ways in which confidentiality is managed in social research, it treats some issues briefly and broadly. For example, Wiles et al. did not distinguish between different types of crimes, and instead wrote quite generally about illegal activities. It is therefore unclear whether researchers working in criminal environments possessed the same attitude towards confidentiality in the context of both serious and minor crimes. Wiles et al. also did not explain how awareness that the research data may be of interest to law enforcement authorities influenced the ways in which researchers managed information about crime or other illegal activities. Moreover the authors did not elucidate the forms of abuse that researchers working with children and people with learning disabilities would be prepared to report. I believe that it is important to consider these issues in detail, because they may have important consequences for research participants, researchers and the research process as a whole, as well as for employers, universities and peer researchers.

Unlike Wiles et al.’s (2008) investigation, this article focuses exclusively on the decisions of researchers to maintain or break confidentiality in situations where they were aware of illegal activities and/or harm. I understand confidentiality broadly, as embracing anonymity as well (see Saunders et al. 2015). Thus confidentiality refers to: 1) not disclosing what participants said or did during research, unless they consent to do so, and only in ways agreed; and 2) concealing the identity of the participants (Surmiak 2018). The article draws on in-depth interviews with Polish sociologists and anthropologists who conduct qualitative research with vulnerable participants. The experiences of Polish social researchers have yet to be discussed in the literature on this subject, which is instead dominated by researchers from the United States of America (USA) and Western Europe (e.g., Clark 2010; Colnerud 2013; Corden and Sainsbury 2006; Dawson and Kass 2005; Wiles et al. 2008). However, it cannot be assumed that the experiences of American and Western European researchers are universal. Rather, it is more plausible to assume that research practices are contingent on context, including the institutional settings. It should be noted here that in Poland, unlike in English-speaking countries (e.g., the United Kingdom and USA), there exists a rather low degree of institutional ethics regulation in social science: there are few ethics review boards and they are usually not mandatory.

In this article I focus on two issues. First, I analyze the perspectives and experiences of the researchers regarding the maintenance or breaking of confidentiality in the context of their knowledge about crime and other illegal activities and harm. I am interested in the specific situations in which researchers maintain confidentiality, the situations where they are ready to break it, and the situations in which they actually do so. Second, I examine how my informants justified their choices to maintain or breach confidentiality. For instance, I am interested in the extent to which the decisions of researchers are influenced by legal norms and the rules of professional ethics. My aim is to discuss the common ways of thinking among researchers and to analyze the factors that influence qualitative research ethics in practice.

Ethical and Law Regulations of Confidentiality

Researchers should make decisions on confidentiality on the basis of, inter alia, ethical standards in the relevant scientific community and law regulations. Thus, I discuss below the confidentiality rules in professional codes and ethical guidelines for sociologists and anthropologists and the provisions of Polish law, which may limit the guarantees of confidentiality in social research. I also discuss two positions of researchers on the legal obligation to breach confidentiality: law-first perspective and ethics-first perspective.

Ethical Regulations and Confidentiality

Research community standards are reflected in the rules concerning professional codes of ethics. In many Anglophone countries (e.g., the USA and UK), researchers must additionally consider the directives of the relevant research ethics committees. However, very few research ethics committees in the social sciences exist in Poland and, like in Germany (Von Unger et al. 2016), they tend not to be compulsory. Ethical standards in sociological research in Poland (a code of ethics for anthropologists has not yet been developed) are defined by the “Code of Ethics” of the Polish Sociological Association (2012). According to this document the highest possible degree of confidentiality shall be guaranteed for the participants (section 10). In addition, researchers should not provide guarantees of confidentiality that they are not able to keep (s.10). However, no mention is made that they can break confidentiality. In contrast, some Anglophone professional codes of ethics in the social sciences allow the researcher (in exceptional cases) to break confidentiality, for example in “health- or life-threatening” situations (ASA 2018) or child abuse (BSA 2017). It is possible that the absence of reference to such a situation in “Code of Ethics” of the Polish Sociological Association is linked to the historical context. During the communist period, the state authorities occasionally demanded scientists to denounce their colleagues or the researched (Kulczycki 2009). Therefore, the authors of the Code may associate a breach of confidentiality with a political denunciation and therefore consider such an action unethical.

Even though no directive to violate confidentiality if required by law is included in professional ethical codes and guidelines, researchers are obligated to do so as citizens. Thus, if the law requires a breach of confidentiality, researchers may be confronted with conflicting obligations: on the one hand, compliance with the law, and on the other hand, maintaining confidentiality (Palys and Lowman 2007). Many ethics committees and ethical codes for social researchers, including the Polish code of ethics for sociologists, recommend that researchers consider the risk of compromising the confidentiality of the researched prior to conducting the research. Consequently, some researchers consciously avoid asking their informants questions about illegal activities (e.g., Sluka 1995) or conducting research with or among people who may break the law (e.g., Van Maanen 1983). In addition, some Anglophone codes of ethics (e.g., ASA 2018) and research ethics committees (Lowman and Palys 2014) recommend that participants be informed about confidentiality as part of the written informed consent process. However, in Poland social researchers rarely used written informed consent in academic qualitative research due to the aforementioned poor institutionalization of ethical control and researchers’ reluctance to formalize ethical aspects of the research process (Surmiak 2018).

Legal Restrictions on Confidentiality

According to the Personal Data Protection Act (2018, 1997c), a social researcher in Poland, like other social researchers from the European Union (EU), has a duty to protect the personal data of his or her research participants. Moreover, the Constitution of the Republic of Poland (article 47, 1997a) obligates the researcher to protect their privacy. Under Polish law, the disclosure or use of information acquired as a result of, inter alia, research activities, is subject to a criminal penalty (266§1, The Criminal Code 1997b). However, the judiciary may release the researcher from the obligation to preserve professional secrecy, for example for the purposes of criminal proceedings (180§1, The Code of Criminal Procedure 1997d). In this instance, the researcher is formally obligated to reveal research data (including personal confidential information) and/or to testify as a witness. Polish researchers, like researchers in the UK, do not enjoy the same freedoms researchers in the USA, who can obtain (under certain conditions) a Confidentiality Certificate from the National Institute of Health. This document protects the research data from forced disclosure at the federal, state or local level (with certain limitations, see more: https://humansubjects.nih.gov). Therefore, if Polish researchers do not want to testify and/or make their data available to the judiciary, they must be prepared to face legal consequences. To my knowledge, researchers in Poland, like researchers in the UK and Canada (Lowman and Palys 2014), are unlikely to receive subpoenas to disclose confidential research data.

Another potential legal constraint on the confidentiality guaranteed by researchers is the obligation to report a crime. Under Polish law, anyone who learns of a crime has a public duty to report it to the prosecutor or the police. However, there is no penalty for failure to notify, which merely constitutes a moral obligation (304§1, The Code of Criminal Procedure 1997d). The most serious crimes are the exception. Thus, under Article 240 of the Criminal Code (1997b), citizens have a legal obligation to notify the public prosecutor or the police of any prohibited acts committed or attempted under the penalty of deprivation of liberty for up to three years (although this does not apply to a crime victim, for example). Such prohibited acts include actions against state security (e.g., assault on the President of the Republic of Poland, terrorism, genocide, murder or imprisonment). Following the amendment of the Criminal Code in 2017, the list of prohibited acts also includes grievous bodily harm and certain sexual crimes (e.g., rape with special cruelty). The researcher has a legal obligation to report these crimes to the prosecution authorities even if they promised confidentiality. In addition, according to Article 12 of the Act on Counteracting Domestic Violence (2010), the researcher should also report to the police or the public prosecutor any witnessing of domestic violence or suspicion of committing a crime prosecuted ex officio with the use of domestic violence.

The Law-First Perspective and the Ethics-First Perspective

Discussion of legal restrictions on confidentiality can be found in the literature. Lowman and Palys (2014) have distinguished two perspectives: “the law-first perspective” and “the ethics-first perspective.” Proponents of the former believe that a researcher has an obligation as citizen to obey the law. As Stone (2002: 24) argues: “scholars and universities are not above the law.” Researchers who adhere to this viewpoint usually inform participants about legal confidentiality restrictions before commencing their research (e.g., Brown 2018; James 2013). In contrast, proponents of the ethics-first approach are of the opinion that a researcher should be able to maintain confidentiality “without any a priori limit” (Palys and Lowman 2010: 280) even against the law (and then bear the legal consequences of his/her choice). Lowman and Palys (2014) argue that the ethics-first approach protects the researched from harm, is on the side of academic freedom, and facilitates research on sensitive topics. Scholars who adopt an ethics-first perspective do not report illegal activities to law enforcement authorities and sometimes actively participate in such activities as an aspect of their research (e.g., Goffman 2015; Pearson 2009; Sluka 1995). Moreover, they may refuse to make their research material available to the judiciary, in spite of negative consequences (e.g., Brajuha and Hallowell 1986; Lowman and Palys 2014; Scarce 1994). It follows that researchers’ approaches to legal confidentiality restrictions are contingent on their response to the question posed by Becker (1967): “Whose side are we on?” However, opting for one of the sides (in this case, confidentiality or law) does not solve the dilemma. The course of qualitative research is somewhat unpredictable, hence researchers may always experience “ethically important moments” (Guillemin and Gillam 2004) that challenge their beliefs. One of the challenges may be knowledge of harm and low violation acquired during the course of research.

Methods

The paper draws on in-depth interviews with 34 Polish qualitative researchers. Each interview explored the maintenance or breaking of confidentiality in a situation in which the researcher was aware of harm, crime or other illegal activities. The interviews generally focused on the researchers’ ethical experiences and opinions. For example, I was interested in their ethical choices and the reasons behind these choices. Research participants were drawn from 12 universities and research centers in Poland. They were at various stages of their academic careers, including nine professors and associate professors, 16 postdoctoral researchers and seven Ph.D. students. All possessed qualitative research experience with vulnerable participants, this is, with people who are particularly susceptible to harm due to their unprivileged position (such as the homeless, sex workers, the poor, criminals, refugees, people with disabilities and children). Participants defined themselves as sociologists (n = 16), anthropologists (n = 15) or both (n = 3). I recruited the interviewees using maximum variation sampling (Patton 1990). Participants were primarily identified through university websites and my research network, and I contacted most of them via email. The researchers received an invitation to participate along with basic information about my research project. I also requested permission to record the interviews. I conducted all interviews in Polish from February 2017 to June 2018. I introduced myself to research participants as a postdoctoral researcher with an academic background in sociology and cultural anthropology. Some were aware of my experience in conducting research with sex workers. The interviews lasted an average of 2.5 h. All interviews were digitally recorded and fully transcribed.

The analysis of the interviews consisted of several steps. First, I read all the transcriptions to get a general impression of data, looking for preliminary themes associated with my research goals. I listed such themes for each interview with a short note containing information about the research tools used by the interviewed researcher, research objectives and context, research paradigm and ethical assumptions. For example, I distinguished the theme: the behaviour of researchers when they hear of illegal activities and/or harm. Second, I read each transcription again to apply meaning categorisation (Kvale 1996), i.e. reducing longer statements to simple categories (one or more). I created the categories on an ad hoc basis during the analysis, labelling them to reflect the sense of the statement—for example, ‘knowledge of harm’, ‘knowledge of crime’, ‘violation of confidentiality’. At the same time, I performed meaning condensation (Kvale 1996), i.e. I summarised the meaning of statements marked with a given category in a few words or sentences. The third step of the analysis was to re-read the coded and uncoded transcriptions in order to verify the validity of the themes (and sometimes to add sub-themes), as well as the categorisation and meaning condensation. Next, I assigned to each theme fragments of the interview statements along with the aforementioned notes from the interviews. In the last stage of the analysis I read the coded research material several times and interpreted it.

It should be stress that I did not assume deliberate disclosures in any case. However, I was aware of the fact that the participants in my study could be recognizable by research peers, raising the issue of guaranteeing them “internal confidentiality” (Tolich 2004). I discussed this issue with the researchers, generally before collecting the data. The participants were able to decide the extent to which they wanted to remain anonymous. They were also offered the opportunity to receive and edit the interview transcripts (Wiles et al. 2006). During some of the interviews, the researchers indicated the fragments they did not want me to use or those that I should use with caution. Some also highlighted so in the comments on the interview transcript. I respected all such requests (Surmiak 2018). Therefore, when I quote or refer to researchers’ statements in the article, I disclose identifying information to various degrees (e.g., the area of research). I also change their names and sometimes their gender (see Kaiser 2009; Saunders et al. 2015).

Results

Confidentiality and Knowledge of Law Violation

The question of knowledge of law violation appears in my interviews in the context of two issues: 1) the choice between maintaining or breaking confidentiality; and 2) the awareness that the data may be of interest to law enforcement authorities or the judiciary. In the first case, my informants stated that they had not passed knowledge of minor offences acquired during the research to law enforcement authorities. Researchers learned about theft, handling stolen goods, drug sales, extortion, unregistered economic activity and poaching, among other issues. Some researchers (those who undertook ethnographic research) also played an active part in illegal activities such as squatting.Footnote 1 At the same time they believed that they acted ethically. The question thus arises of where, in researchers’ opinions, is the limit beyond which confidentiality should be broken and law enforcement authorities should be notified of a crime?

Most of my interviewees declared that they would break confidentiality if it concerned a serious crime such as murder, sexual violence, and physical violence that threatens health or life, especially that of a child. Some believed that in such cases, even circumstantial evidence would be sufficient to break confidentiality. Others claimed that they would only react if they had witnessed such events themselves. Moreover, my interviewees were more likely to break confidentiality if the serious harm of which they had become aware occurred in the present and not in the past. Perhaps this was due to their conviction that it is possible to remedy harm in the present, but not from the past. Interestingly, unlike the participants of Wiles et al.’s (2008) study, my interviewees generally did not inform their participants that in certain circumstances they would be prepared to break confidentiality. I believe that there are several reasons for this. First, they did not expect to attain knowledge of serious crimes so they did not suppose that they would feel obligated to break confidentiality. Second, they did not seek information about situations of serious harm so that they would not have to break confidentiality. Third, some researchers felt that communicating confidentiality restrictions precluded the development of mutual trust, especially in highly suspicious social environments. As Feenan (2002: 775) has noted in the context of paramilitary violence in Northern Ireland, informing about confidentiality restrictions may result in “inhibiting or scaring off interviewees with what might appear to be reading them their rights.” Instead, at the beginning of the interviews, Freenan informed his interviewees that he was not interested in any real names or identifying details. Some of the participants in my research claimed to do likewise.

I have thus far discussed the declarations of researchers concerning a breach of confidentiality. The question then arises as to what they do when they learn about a serious crime (or harm). Indeed, some of my interviewees found out about crimes involving physical or sexual violence. I have distinguished three types of reaction of researchers when they know of a crime involving physical or sexual violence: breach of confidentiality, maintaining confidentiality and the uncertainty of whether to maintain or breach confidentiality. Only one researcher decided to breach confidentiality in a case of knowledge of serious violence. Jarek was a prisoner and at the same time a researcher who conducted covert participant observation for his thesis (and so was bound by research ethics principles). During this research, Jarek witnessed the physical harassment of a new prisoner by his peers, as a result of which the victim was seriously injured. He initially tried to stop the perpetrators, but they threatened to beat him up. Therefore, he decided that he would inform the ward about the violence. As he put it: ‘snitching, but I said, well, I’ll take the risk because it’s for a good cause’. However, before he could do so another prisoner had already informed the prison staff about the violence, exposing himself to revenge. This story may serve as a warning to researchers to be careful about their own safety when reporting a crime. I also consider this an example of a decision to break confidentiality in spite of the fact that Jarek did not explicitly promise this to anyone (although it can be argued that confidentiality was tacitly assumed). Under Polish law, he had a public duty to report this crime. However, when making such a choice, Jarek did not follow either the law or the sociological research ethics, but his own values, which I will write about later. In other words, he did not adopt either the law-first perspective or the ethics-first perspective.

Unlike Jarek, most of my interviewees maintained their research participants’ confidentiality in spite of knowledge of physical violence. For example, Edek studied the way of life of a certain internally conflicted subculture. In the course of his ethnographic research, he often heard of specific cases of violence in this community, especially of men against women. Sometimes he was asked to intervene: ‘If you are so wise, tell (...) the prosecutor’s office, that (...) he beats his wife and put her head in the pot. Should I file a report to the prosecutor’s office or not? This is an ethical dilemma. Maybe he really put her head in the pot, and I suspect that he did. Because it happens in this community, violence against women is really there’. Edek believed that he should not pass such information on to anyone. Nor did his attitude change when he witnessed the corporal punishment of a child: ‘a daughter [of the hosts] comes and sits there with us for a while and I think she looked at me too much. She was so young, I don’t know, 14 years old or something. She looked at me too much, I don’t know why, but in a moment a mother comes up and slaps her face, slaps her face, but it’s so [hard]’. When I asked him what he had done in this situation, he answered: ‘Nothing, nothing, just nothing. And then there was the need to deal with it somehow. And I think that when you enter someone’s house and see their way of life, it’s just different. It’s the system of values you have there, and some rules <not to beat a child> and so on, you can maintain them in your own home’. As mentioned earlier, according to article 12 of the Act on Counteracting Domestic Violence (2010) researchers should report the witnessing of domestic violence or the suspicion that a crime of domestic violence is being committed to the prosecutor or police. However, in the example given above, Edek pointed to the values of the community examined and not to the legal regulations applicable to the researcher and the researched (because they lived in Poland, they were bound by Polish law). Edek adopted an ethics-first perspective, i.e. he believed that both in the case of suspicion of a crime and the corporal punishment of a child, as a researcher he should maintain his research participants’ confidentiality.

In the aforementioned cases, Edek took it for granted that he should maintain confidentiality. However, some researchers did experience ethical dilemmas in similar situations. For example, Bartek found out that staff at a homeless centre physically abused some of its residents. Despite his willingness to help the victims, following some reflection he ultimately decided not to pass this knowledge to the proper authorities. Indeed, he believed that maintaining confidentiality would be more ethical. Like the interviewees above, Bartek followed the ethics-first perspective, i.e., he kept confidentiality, in spite of his knowledge of the crime. However, it is worth noting that his dilemma was not a choice between law and confidentiality, but rather was related to his uncertainty as to what would cause less harm to the participants: maintaining or breaking confidentiality. Certainly, he was afraid that if he informed about the wrongdoings, the centre could be closed down and the homeless would have nowhere to go.

Interestingly, most of the participants in my research seemed not to take the law into account in their decisions to maintain or breach confidentiality (although there were some exceptions). Rather, they often considered it the personal decision of the researcher. Moreover, the researchers often used the word ‘snitching’ to describe a breach of confidentiality in the case of knowledge of a law violation. Thus, they saw such a breach as a violation of certain informal social principles (according to some researchers it also goes against professional ethics) rather than as acting in the public good. It may be that this reluctance to violate confidentiality and inform authorities of a violation of the law stems from the low trust in public institutions that is characteristic of post-communist societies (Jaskulowski and Surmiak 2016; Sztompka 1999).

As to the second context of guilty knowledge (awareness that research data may be of interest to law enforcement agencies), some participants in my research, as in Wiles et al.’s (2008) study, were aware of this issue and took steps to protect their data and informants. Thus, some of them resigned from publishing information about the illegal activities of their research participants, while others anonymized such information prior to publication. Two researchers dealing with criminal groups consulted with their informants about the anonymization of data before publishing their books (see Sluka 1995). They did not want to expose them or themselves to negative consequences. In addition, several of my research participants claimed that the law should exempt researchers from the obligation to provide law enforcement authorities with any confidential information from their research. For example, Klaudia stated: “If someone were interested in it, so what, should we give recordings, transcripts, should we talk in detail about our participants? I would rather not do it.” In accordance with the literature, my analysis indicates that reluctance to make research material available to law enforcement authorities was typical both of researchers in favour of the law-first and the ethics-first approach (see Stone 2002; c.f. Lowman and Palys 2014). The literature also demonstrates that researchers who adhere to the law-first approach will comply with the law as long as they receive subpoenas. However, it is difficult to state how my researchers would react: only one of my interviewees mentioned that she was asked by a state institution to transfer research data on the activities of a criminal group, and refused (without negative consequences for herself). She did not want to talk about the details. However, it was evident from her story that she had adopted an ethics-first perspective: her refusal was based on a sense of obligation to maintain confidentiality. The question arises as to whether it can be assumed that the majority of my interviewees took such an approach. In other words, was it more important for most researchers to maintain confidentiality when they learned about crimes and other illegal activities than the law requirements? I will return to this question in the conclusions.

Confidentiality and Knowledge about Harm

My interviewees learned about harm that was not related to illegal activities, for example, betrayal of a person’s partner or falling seriously ill. Hammersley and Traianou (2012: 61) have argued that “in identifying harms we are engaging in evaluation: judging some actual or potential outcome to be bad for someone.” Thus, judgments of harm always depend on a person’s values and are to some extent subjective. Of course, the fact that the researchers identified harm did not in itself pose a challenge to the guarantee of confidentiality. For example, if they found out that someone was living in very difficult social conditions, but both social care workers and the local community knew of this situation, they did not usually question the promised confidentiality. A similar case comprised if they found out about harm that had occurred in the past. The difficulty of guaranteeing confidentiality only appeared when they believed that they could remedy the harm (e.g., protect the informants from a deterioration in their situation) by providing certain information (for example) to relevant institutions.

Having analyzed my interviews, I can distinguish four approaches towards breaking confidentiality in a case where the researcher learns of some harm. I refer to these as: passive, collaborative, caring (although it may also be described as paternalistic) and balancing. The researchers adopted these attitudes both in situations where the research participant was a harmed person and the perpetrator of someone else’s harm.

My interviewees occasionally adopted a passive attitude, that is, not only did they maintain confidentiality, but they also did not take any action to prevent harm from occurring. The researchers usually acted in this way when the harm concerned family matters, as in the case of harmful (according to the researchers) way of raising children (e.g., verbal violence), financial abuse of a research participant by family members, or infidelity in a relationship. Many researchers perceived these matters to be private and so did not wish to interfere. However, some thought that this was not right. For example, one of my interviewees said: ‘You would say to someone (...) to someone who gave you so much and suddenly you haven’t thanked them for the interview yet, and you say: but you know what, I think you could treat the children a little better’. This quote shows, I think, how important the commitment to reciprocity was for some researchers, which I will discuss in more detail later in this article.

Sometimes my interviewees adopted a collaborative approach, i.e., they sought permission from research participants to intervene. Wiles et al. (2008) have argued that researchers face a problem when participants do not want an issue to be disclosed. Dominika experienced such a problem in a study involving children with attention-deficit/hyperactivity disorder (ADHD). The researcher has a rule that she will ask children if she can talk to other people (e.g. parents) about what they are doing together. During the research she learned from one of the children that a teacher had insulted him in the classroom. Dominika explained to the boy that what the teacher had done was wrong. She also suggested that he talk to his parents about the event. However, the boy did not want to tell his mother about it, nor did he allow the researcher to do so. Dominika explained: ‘So I started looking with the boy for another adult whom he trusted and could talk to and felt would somehow support him’. In this example, Dominika defined the teacher’s behaviour as harmful because it reduced the boy’s self-esteem and his willingness to communicate with teachers at school. For this reason, she believed that the case required adult intervention. Because the boy did not want to tell his mother about the incident, Dominika decided not to breach confidentiality. Instead, she and the boy looked for a solution with which both of them would be comfortable. Dominika’s intervention was successful and did not require her to break confidentiality. However, other researchers did not always have the opportunity or willingness to negotiate with research participants regarding whether and to whom to discuss harm. Sometimes they took a caring (or in other words, paternalistic) approach, i.e., they decided to break confidentiality in the name of protecting a research participant’s wellbeing. Researchers acted in this way when the threat of harm concerned participants (with whom they made friends) or when these participant friends posed a threat to themselves. For example, one of the interviewees warned his research participant friends not to trust the other participants because, in his opinion, they had a tendency to lie and gossip which could lead to harm. The risk of harm in this case consisted of a threat to the intimacy, privacy and trust among the participant’s friends. However, the researcher himself decided what would be potentially harmful to the research participants and how to counteract it. His decision resulted from his close relationship with the participants, which gave him a sense of responsibility for their well-being and at the same time a ‘right’ to interfere. One might wonder, however, whether the researcher’s behaviour did not harm the other research participants about whom he warned his friends. And if it did harm them, was it ethically justified by wanting to protect his participants’ friends? According to the researcher, in such a situation one is always unfair to someone, so one has to choose something.

Another approach could be described as a balancing approach. My interviewees rarely decided to openly and directly break confidentiality. They were more likely to balance on the verge of maintaining and breaking confidentiality, especially when they encountered child harm. In other words, researchers tried to subtly communicate or suggest to responsible persons that someone had been harmed or that someone was the perpetrator of harm. For example, Jola was researching relations between the inhabitants of a community home. Initially, it was difficult for her to gain their trust, despite the promise of confidentiality, because they saw her as someone who was co-operating with the staff. It was only over time that she managed to deepen her knowledge of relations within the institution. During her conversations with the residents, she learned of cases of harassment and hostility. She said: “What I allowed myself to do was to ask the educators in a very general way whether they recognize such people [victims] and what actions they take. But I couldn’t indicate that this is the person who has been injured by this person”. Thus, Jola did not want to break confidentiality or interfere in the environment under study, and so attempted to subtly indicate malevolent actions. She hoped that by asking appropriate questions she would draw educators’ attention to the issue of violence in the community home.

Researchers’ approaches to harm depended on the circumstances (i.e., they may have been different in various situations) and occasionally changed over time. However, in general (as in cases where the researcher learned about crimes and illegal activities) the researchers decided to maintain confidentiality. Interestingly, researchers were more likely to break confidentiality in the context of relatively low-key harm than in the case of a serious crime. I will return to this topic when discussing the reasons for breaking confidentiality. First, however, I will address the factors behind maintaining confidentiality.

Maintaining Confidentiality and the Role of the Researcher

Researchers tended to justify maintaining confidentiality in the context of knowledge about illegal activity, crime and/or harm by referring to their understandings of the role of the researcher. Thus, my interviewees believed that the role of the researcher is primarily about producing knowledge, not judging the activities of research participants. The conviction that the main goal of the researcher is to produce knowledge had two consequences pertaining to confidentiality. First, many of my interviewees perceived the disclosure of information about illegal activities to the prosecuting authorities as an action that goes beyond the researcher’s duties. As Edek briefly stated in the context of knowledge about the domestic violence: “This is research, not an investigation”. Second, most researchers felt that protecting confidentiality enabled them to collect valuable data and therefore was not only ethically but also methodologically justified. Researchers who shared such a perspective were only willing to break confidentiality in exceptional situations, such as a threat to someone’s health or life.

Many of my interviewees perceived the role of the researcher in neutral terms: a researcher should not judge his or her research participants. Such a belief appears to be related to the positivist epistemology that assumes the axiological neutrality of the researcher. My informants who shared such an assumption tried to approach research without personal emotions: they separated their research from their personal experiences. As Lucyna noted: “We are [researchers] schizophrenics because of this sneaking through these worlds, entering, exiting these roles.” This suggests that Lucyna saw herself as becoming “somebody else” when undertaking research, pushing aside her personal values. According to this logic, a researcher who does not have the right to judge what is good and what is bad has no grounds for breaking confidentiality in the case of knowledge about an illegal activity, crime or harm. It is worth noting, however, that not all of the participants who believed that the researcher should not judge the participants justified this belief in terms of axiological neutrality. Some researchers referred to the “rights-sensitive autonomy” of their participants: “the liberty people have to choose their own way of life” (Meyer 1987: 267). Researchers who took such a point of view did not feel that they had the right to interfere in the autonomous decisions of their informants. For example, Dominika felt that a boy who was publicly offended by a teacher should be able to decide whether, to whom and in what form he would relate the incident. Although she expressed her opinion, she did not impose her view on him or make him do anything.

Moreover, some researchers (especially anthropologists) justified their refrainment from judging the behavior of participants in terms of moral relativism (Levy 2002). In other words, they believed that the researcher’s role is to avoid evaluating the behavior of people from different cultures (including subcultures) through the prism of the researcher’s culture. Furthermore, they claimed that the researcher should not seek to evaluate the participants but rather understand them. For example, the aforementioned Edek, who saw a mother slapping her daughter’s face, did not react and did not breach confidentiality because he knew that such behaviour was the norm in that community. He felt that he should accept this different norm. In addition, at times some researchers accepted the perspectives of their research participants on moral issues. Thus, they “suspended” (at least to some extent and rather temporarily) their own values. For example, Krystian, who discovered that his informants were going to beat a former member of their radical group, kept it confidential, inter alia, because (as he said metaphorically), he “turned on their morality.” Thus, he saw such an action as being consistent with the values of his participants. An extreme example of such an approach has been presented by Goffman (2015), who conducted ethnographic research with a group of black men in an urban ghetto. Goffman helped her research participants (who were also her friends) to search for the murderer of one of the group’s members. She claimed that she wanted to take revenge: “I got into the car because, like Mike and Reggie [other participants in her study – author’s note], I wanted Chuck’s killer to die” (Goffman 2015: 262). In response to criticism and accusations of complicity in the attempted murder, Goffman defended herself, stating that she did not believe that they would find the murderer (Lubet 2015). Nevertheless, Goffman’s story and the experiences of the participants in my research show that the consequences of adopting a moral perspective during research can be quite serious. This provokes the reflection: Does conducting ethnographic research legitimize such an approach ethically? There are no easy answers to this question.

My interviewees also attributed guest status to the researcher, with certain consequences for confidentiality in the context of knowledge about illegal activity, crime and/or harm. First, in their opinion, the researcher-guest stays in a research field only conditionally. At any time, his or her access to the informants’ world may be denied if he or she fails to obey the rules of hospitality. In my interviewees’ opinions, a researcher who breaks confidentiality is not behaving in accordance with these rules. For example, had Edek decided to breach confidentiality and communicate his observations and/or suspicions of domestic violence to the police or social workers, then he would probably not have been able to continue his research in this community. Second, many interviewees perceived the hospitality they received and the opportunity to obtain knowledge from the informants as a kind of gift of trust that should be reciprocated. This reciprocity included confidentiality. The researchers emphasised this especially when they conducted research with people who did not fully understand what the research was about, and when they experienced difficulty gaining the trust of the group—as in the case of the aforementioned Jola, who conducted research in a community home. Third, my interviewees were aware that hospitality was temporary and that at some point they would return to their “reality.” According to this line of thinking, if they broke confidentiality, it would not be them but their informants who would bear the consequences of such a decision. Moreover, the feeling of being a guest (i.e., a stranger from the outside) caused some researchers, and especially those conducting short-term research (e.g., based on interviews) to feel uncertain about how to interpret what they saw or heard during research.

Maintaining Confidentiality and the Consequentialist Point of View

Many of my informants decided to keep confidentiality in the context of knowledge about illegal activities and/or harm because they adopted (at least to some extent) a consequentialist point of view, based on assessment of the impacts of a particular action. According to consequentialism, one should choose the action that leads to the best and/or least negative effects (Hammersley and Traianou 2012). My interviewees attached the greatest importance to the consequences that a breach of confidentiality would bring to the participants of their research. They decided to maintain confidentiality because they believed that, on balance, it would have a better effect on the informants than breaking confidentiality. For example, aforementioned Bartek, who learnt that staff at an aid institution were using serious violence against its residents, was afraid that passing this information to the police would result in the closure of the institution. According to him, the victims of violence could be worse off than in the current situation, as they would not have any support or place to live. This example raises the question of whether an action taken with a positive intention is morally justified, even if it is expected to have a negative effect. On the grounds of deontological ethics, that is, ethics based on principles and values to be observed even if they have negative effects, this question can be answered in the affirmative (see Saja 2015). However, Bartek thought in a consequentialist manner: he believed that the potential negative effects of a breach of confidentiality outweighed the positives, and so determined that confidentiality should be maintained in this case.

Some researchers explained that they had kept confidentiality because they were unsure of the reliability of the information regarding the crime (or harm). They also believed that it would be difficult to prove that a crime of which they had been informed had been committed owing to a lack of evidence. In such cases, the researchers recognized that the negative consequences of a breach of confidentiality outweighed the positives. For example, some ethnographers researching highly internally conflicted groups, like Edek, suspected that information shared with them about violence (e.g., that a husband was beating his wife) or involvement in a crime (e.g., financial malpractice) might have been invented or exaggerated. As they suggested, the informants could talk about alleged crimes in order to harm their opponents. Thus, the negative effects of accusing innocent people, and for some interviewees also violating the researcher’s own impartiality and neutrality, outweighed the potential benefits of violating confidentiality. Moreover, the researchers sometimes tacitly believed that since they are not the only people who have knowledge of the crime/ harm, they do not have to react.

Several participants in my research indicated that a breach of confidentiality could have negative consequences for themselves and for peer researchers as well. Breaking confidentiality could result in a requirement to interrupt their research. It might in cases also endanger the safety of the researcher. For example, in studies of criminals the reporting of a crime to the police could expose the researcher to the risk of retaliation. Breaking confidentiality might also have negative connotations for peer researchers, as it could render it more difficult or impossible to conduct future research in a given environment. As Jarek, who conducted research in prison, claimed: “If you were to go to the prosecutor with everything, such research would not be possible at all.” Such an argument is frequently raised by criminal researchers (e.g., Lowman and Palys 2014; Wolfgang 1981), based on a conviction that the value of the knowledge gained from research is greater than the positive effects of a potential breach of confidentiality. However, the evaluation of some actions based on the calculation of positive and negative effects has its drawbacks. For example, it is not possible to predict all of the consequences of potential decisions (direct and indirect), and they are difficult to weigh and compare (see Saja 2015). Some researchers did not find this problematic, perhaps because they did not believe in the positive effects of their possible intervention. However, aside from their views regarding the role of the researcher and the calculation of profits and losses, an equally important reason why researchers maintained confidentiality was their personal values. I discuss this issue in the following section.

Keeping Confidentiality and Researchers’ Personal Values

Some researchers justified the maintenance of confidentiality with reference to personal values. Among them were researchers who referred to the Weberian post-positivist paradigm, as well as those who identified themselves with various critical approaches. Thus, some of my interlocutors drew a Weberian distinction between “value-judgments” (werturteil) and “relevance to values” (wertbeziehung). Following Weber (1949), they believed that the researcher should avoid value-judgments. At the same time, however, and still in accordance with Weber, they assumed that it was impossible to conduct research without “relevance to values” because it influences the choice of research topic and the formulation of the research problem. My interviewees who conducted research in criminal environments treated their choice of research topic as a kind of moral declaration; they were able to accept certain illegal and extra-legal activities and maintain confidentiality. As Wolfgang (1981: 352) has claimed in the context of criminological research: “if any researcher doubts the moral obligation to maintain confidentiality, he [sic] should abstain from this kind of research.”

A relatively small number of my interviewed researchers, especially anthropologists, rejected Weber’s distinction between value-judgments and relevance to values. They also did not believe in the worldview of researcher neutrality. Paradoxically, although they referred to a different paradigm, they also kept confidentiality in the context of knowledge about an illegal activity or crime (similarly to Weber’s followers). For example, Jaga did not inform the police that the participants of her research possessed marijuana, because in her personal opinion this should be legalized. Moreover, due to their left-wing views, some researchers criticized the economic and political system as a well as an institutional way of treating refugees, undocumented immigrants or the unemployed. As a result, they perceived the participants in their research as being victims of structural violence and they tolerated their actions even if the system defined them as illegal, e.g., undeclared work or minor thefts. Furthermore, as noted, some anthropologists not only kept confidentiality in such situations, but also personally engaged in illegal activities with their participants, such as afromentioned squatting. They were driven by a desire to get to know the world of the informants and/or a desire to help. However, personal values were not only the factor behind maintaining confidentiality, but also the reason (as I write later) to decide to break confidentiality in the context of knowledge about serious crimes and/or harm to the research participants.

Justifications for Breaking Confidentiality

Researchers have often argued that they would break confidentiality in the context of knowledge of a crime that threatens someone’s health or life, because their system of values dictates them to do so. As Ania claimed, “We have to decide what is more valuable at the moment. I believe that people’s health and life should always be that value.” Researchers believed that people’s life and health should be protected at the expense of violating confidentiality or even the possibility of conducting research. Like participants in Wiles et al.’s (2008) study, many of the researchers in my study were particularly sensitive to child harm. However, this sense of moral duty, as I have already noted, did not always result in the breaking of confidentiality. Therefore, the question arises: Why did some researchers decide to break confidentiality?

From the aforementioned statements of Jarek, who wanted to inform prison staff about the harassment of a fellow prisoner, it appears that he made this decision because he considered human health and life to be of paramount importance. His value system did not seem to differ from the researchers who did not violate confidentiality even where they learned of a serious crime. It is worth noting, however, that Jarek witnessed violence, the context and meaning of which he understood very well, because he himself was a member of the prison community. For this reason, I think that it would be difficult for him to distance himself from this situation by referring to the categories used by other researchers in the case of knowledge of violence, such as the notion that the researcher has guest status. Moreover, it is possible that the fact that Jarek carried out covert observation was of some importance. Consequently, unlike in the case of overt research, his violation of confidentiality did not directly jeopardize the success of his research or future studies in the prison environment.

Researchers who took a caring approach, that is, they broke confidentiality because they believed that their informants were threatened with harm, were also convinced that the welfare of their informants was more important than maintaining confidentiality. When making their decisions, they were guided by the obligations arising from their relationships with their research participants, and not by the principles of code ethics. Perhaps also, although it did not appear in their narratives, they perceived the possible negative consequences of breaking confidentiality – both for the informants, the research, as well as for themselves – as relatively insignificant. I think that researchers were more likely to break confidentiality and to take the balancing approach in the case of knowledge of minor harm than in the case of serious crime, because they considered the possible negative consequences of such actions for various individuals (e.g., participants, themselves, the other researchers) as being relatively smaller.

Conclusion

Thanks to the reflexivity and self-awareness of my interviewees, their statements provide a clear insight into the complexities of their choices regarding confidentiality in the context of knowledge of violations of law or harm. The results of my research, as well as that of Wiles et al. (2008), indicate that researchers took their obligation to protect confidentiality very seriously. In both studies, interviewees declared that they were only willing to break confidentiality if they learned about physical or sexual harm, especially when it affected children. However, my analysis shows that in research practice, even in such cases, researchers tended to opt to protect confidentiality. In this article I have distinguished three (albeit often interconnected) reasons for maintaining confidentiality in the context of knowledge about illegal activity, crime and/or harm. They revolve around three issues: 1) perceiving the role of the researcher as an axiologically neutral person, whose main purpose is to collect data thanks to the “hospitality” of the research participants; 2) predicting the negative consequences of breaking confidentiality, especially for the informants; and 3) researchers’ personal values, including their political views. My analyses show that the researchers justified their breach of confidentiality primarily in terms of their personal values and their willingness to protect their research participants.

Contrary to the participants in Wiles et al.’s study, my interviewees generally did not inform their research participants about legal restrictions on confidentiality, nor did they ask them about their consent to disclose information about harm (although there were exceptions). These differences in approach to law may reflect the distinctive social research culture of Poland relative to Anglophone countries. In addition, most of the participants in my research did not comply with the legal regulations on breach of confidentiality, unlike some of the participants in Wiles et al., who felt obligated to take legal regulations into account when making decisions. Thus, the question returns of whether my research participants were in favour of the ethics-first approach, i.e., was the guarantee of confidentiality more important to them than the legal regulations ordering them to violate it? Some researchers explicitly objected to the sharing of confidential research material with law enforcement authorities, and one interviewee refused to disclose it. Moreover, the majority of the researchers maintained confidentiality, despite the fact that according to the law they should have notified law enforcement authorities of a crime. In my opinion, however, the aforementioned question is flawed, because researchers generally did not choose between the confidentiality principle and the law that ordered the breaking of confidentiality. Indeed, their reasoning was more complex. In deliberating over maintaining or breaking confidentiality, my interviewees took various ethical norms (mainly related to not harming the informants), personal values and/or methodological presumptions (e.g., related to conducting research such as conceptions of the role of the researcher) into account. In other words, when researchers attained knowledge of illegal activities, crime and/or harm, they balanced different ethical principles, personal values and methodological ideas, rather than simply selecting between confidentiality and legal requirements.

The results of my research indicate that researchers did not provide a single, universal answer to the question of whether to keep or break confidentiality when aware of illegal activity, crime and/or harm. They did not comply with the directive of absolute respect for the guarantee of confidentiality contained, for example, in the code of ethics of the Polish Sociological Association or the requirement of unconditional assurances of confidentiality imposed by some research ethics committees (Downes et al. 2014). Nor did they comply with the directive requiring researchers to assure confidentiality to the extent permitted by the law (Palys and Lowman 2010). My research reveals a disjuncture between procedural or codified ethics and ethics in practice (Guillemin and Gillam 2004). Any guidance on how to deal with knowledge of illegal activity, crime and/or harm has the disadvantage that it does not take into account the research context or the personal values of the researchers themselves. On the one hand, any ethical or legal codification does not and should not exempt researchers from making their own decisions (Taylor 1987). On the other hand, relying only on researchers’ own judgments also has its limitations. Researchers may make mistakes, for example by overestimating the importance of the negative consequences of a breach of confidentiality in certain circumstances, and underestimating the positive consequences related to the protection of an informant’s wellbeing. Alternatively, they may too hastily decide to break confidentiality, which can also bring negative results, for example, it may worsen the situation of the people it was supposed to protect. Therefore, a balance is required between professional ethical obligations, law regulations and researchers’ own morality. Certainly, any serious unethical behaviour on the part of the researcher (e.g. towards the participants) would negate the validity of research. In my opinion, however, this does not mean that in Poland there is a need for the formal control of social researchers’ conduct (e.g. through ethical committees). Although such control is a standard in many English-speaking countries, it receives a lot of criticism from qualitative researchers, among others, for its ineffectiveness and restriction of academic freedom (e.g. Dingwall 2008; Hammersley 2009). While a certain amount of institutional control is unavoidable, we should also develop and deepen the ethical imagination of researchers. In other words, ethical decisions are a complex sphere, full of nuance and ambiguous situations that cannot be predicted and regulated once and for all. We need series of loose guidelines and sensitizing frameworks about how to deal with complexities and moral shading of the dilemmas between keeping and breaking confidentiality in research practice. This may stimulate researchers’ imagination and enhance reflexivity. My paper discuss the experience of researchers who have learned about illegal activities and/or harm during research with vulnerable participants, but it may has a wider significance, because researchers are apt to encounter crime or harm no matter what groups they study – although the profile of low violation and harm might be quite different.