Introduction

Doing research with survivors of gender-based domestic violence using qualitative interview-based methods is an ethically demanding endeavor. The latency and prevalence of gender-based violence, including intimate partner and other forms of domestic violence, remain high (European Agency for Fundamental Rights, 2014; Shreeves & Prpic, 2019) and continue to be further affected by transnational crises (John, Casey, Carino, & McGovern, 2020; Peterman, et al., 2020; Usher, et al., 2021). Finding research participants is not easy because of fear of stigmatization, which survivors often feel as an integral part of their story. Moreover, in many contexts, the feminist roots of researching this topic may be denied or ignored while research explicitly addressing gender may be seen as implicitly controversial (Graff & Korolczuk, 2021). Most importantly, however, such research entails work with vulnerable people who have experienced trauma with its potentially far-reaching psychological, corporeal and developmental consequences, including the development of post-traumatic stress disorder (van der Kolk, 2005; Kolk, 2015).

The only certainty anyone attempting such research has is that their research participants have deep – often repeated – experience of radical power imbalances and (sometimes longitudinal) abuse typically in multiple areas of their lives. This has been documented in searching for effective care designs with the emergence of trauma- and violence-informed care (Elliott, et al., 2005; Varcoe, et al., 2016). The principles of this care were subsequently translated into trauma- and violence-informed research (Baker & Lalonde, 2020; Newman, et al., 2006; Thompson, 1995) with the main objective of including different aspects and contexts of violence, not just its interpersonal character in the case of gender-based domestic violence. At the same time, a growing number of studies point out the necessity of taking account of vicarious trauma and violence targeting those pointing out violence and its causes, including researchers (Nikischer, 2019; Vidu, et al., 2017).

In this text, we compare the merits and drawbacks of three research designs we used for research with gender-based violence survivors. We focus on what steps we considered in each case in terms of ethical concerns associated with the different aims of the respective research and methods selected. As qualitative researchers, we frequently encounter complex ethical dilemmas that are not always easily derivable from formal ethical codes. Such ethical considerations may arise prior to, during, or after fieldwork, and can be difficult to navigate in practice. Our aim is to contribute to the ongoing discussion on research ethics by defining questions that need to be addressed when designing research with survivors of gender-based domestic violence so that the threat of inflicting further harm is mitigated as we aim for changes on different levels, namely survivors’ empowerment, changes to institutional handling of gender-based violence, and the general public’s awareness of the phenomenon. The changes include invoking greater sensitivity and responsibility towards the communities and individuals with whom we work.

The remainder of the text is structured as follows. Firstly, we outline the theoretical and practical ethical inputs in our research mostly derived from feminist approaches to research with vulnerable groups; secondly, we present our respective individual and collaborative research projects and highlight the ethical dilemmas each design poses. In the discussion section, we compare the different designs as for their capacity to justify their implementation with regard to the possible harm they may cause. In the conclusion, we highlight the ethical issues that need to underpin research with gender-based domestic violence survivors if it is to be methodologically sound and yield results justifying the risks any such research poses to the parties involved.

Research Ethics: Whose Voice and Whose Need Is It?

Ethics is understood as judging human actions and experiences as good or bad as compared to sets of values (Preissle & Han, 2012). Guillemin and Gillam define two types of ethics relevant for research considerations: procedural ethics and ethics in practice (Guillemin & Gillam, 2004). Procedural ethics concerns research design as presented to university research ethics committees (RECs) and institutional review boards (IRBs), which may also include reviews of on-going and completed research. As such, it follows from ethical concerns first identified in scientific research in reflection of abuses of science for racism and in military conflicts (Burgess-Proctor, 2015) and is underpinned by positivist notions of knowledge production conditions, such as detachment and “objectivity”. These notions also inform the very urge to have a binding decision on the notoriously muddy research terrain beforehand. While understandable and an improvement when opposed to lack of accountability, such an endeavor has been shown to be possibly counterproductive in the case of experimental, exploratory and social change-oriented research (Pitt, 2013). This includes research with vulnerable groups, such as violence survivors, although multiple studies indicate that the cost-benefit deliberation of RECs and IRBs is not based on actual research, which rather points to benefits for survivors participating in research (Burgess-Proctor, 2015; Clark & Walker, 2011; Dragiewicz, et al., 2023; Yeater, et al., 2012). As such, RECs and IRBs may work to protect academic institutions (Guillemin, Gillam, Rosenthal, & Bolitho, 2012) but against the interests of vulnerable groups, possibly hampering the research needed for social and political change – this effect of IRBs has been labelled “ethics creep” (Haggerty, 2004; Wynn, 2011) or “mission creep” decrying especially the formalization and bureaucratization of the process (Gunsalus, et al., 2006). Moreover, the fact that the presence of IRBs and RECs and their basic guiding principles are taken for granted and required when publishing research has been identified as potentially contributing to existing geopolitically grounded biases and inequalities in doing science, namely false representation of British and US-based standards as universal (Mamotte & Wassenaar, 2009; Sutrop & Lõuk, 2022).

The positivist origins of procedural ethics that can be frequently detected in various ethical codes (Sutrop & Lõuk, 2022) have been challenged by feminist standpoint theory (Haraway, 1988; Harding, 1986; Preissle & Han, 2012; Sprague & Kobrynowicz, 1999), research into cultural and ethnic bias (Milner, 2007) and queer theory (Butler, 1999; Rivkin-Fish & Hartblay, 2014; Withers, 2010), leading to the formulation of “feminist research principles” that include reflexivity of one’s positionality as well as adherence to ethical standards (Leung, Miedema, Warner, Homan, & Fulu, 2019, fol. 431). This approach strives to integrate ethics in practice, i.e. the ethical decisions forced by moral dilemmas as they appear throughout the research process (Guillemin & Gillam, 2004), with procedural ethics through insisting on reflecting on the ever-evolving situations in the research field in the formalized procedural ethics and related institutions. As shown above, these attempts have not gone unchallenged as the epistemological foundations of what constitutes scientific knowledge differ and continue to affect research with people understood as vulnerable.

We primarily focus on ethics in practice, which is nevertheless affected by procedural ethics as well. The three researches presented below were all conducted in Czechia with their researchers based at two institutions: Czech Academy of Sciences and Charles University. Both institutions have codes of ethics and these differ greatly. The Czech Academy of Sciences’ Code of EthicsFootnote 1 clearly appeals to positivist notions of science citing “objectivity” and impartiality, specifically mentioning “neutrality”. The Charles University’s Code of EthicsFootnote 2 also requires objectivity and impartiality from the staff yet the wording is substantially more nuanced and informed by humanities and social sciences. However, it is not common practice in either of the institutions to have research in social sciences and humanities evaluated for possible ethical issues before the research is undertaken. Rather, researchers are expected to adhere to the respective Codes and are only institutionally held accountable if reported to an Ethical Committee. This situation has ambivalent effects on research with vulnerable groups. It leaves ethical deliberations up to the researchers relying on their commitment to ethical standards, which makes room for experimental and action research. At the same time, the references to “objectivity” in combination with understandings of gender studies as “gender ideology” (Graff, 2014; Graff & Korolczuk, 2021) even within local academia raise questions as to the grounds of reasoning of local ethical committees and subsequently their possible decisions in cases of controversy.

The absence of institutionalized procedural ethics translates into a heightened need for reflexive research at every stage. This is all the more the case in research with people whose access to various forms and types of power is limited, whether because of their gender, sexuality, race, ethnicity, social status, bodily abilities, geographical location, or experience of violence and abuse (Lykes & Hershberg, 2012; Sprague & Kobrynowicz, 1999). In such cases, mitigating harm to research participants is the baseline (Ellsberg & Heise, 2005; Preissle & Han, 2012), but is often not seen as enough to justify research as proponents of participatory action research have claimed (Potts, Kolli, & Fattal, 2022; Wrentschur, 2008). Contrarily, the researcher’s position is one of privilege in that it requires institutionally acknowledged education, capacities and financial support. It is an activity that belongs to those who have power and represent the more powerful part of society (Lykes & Hershberg, 2012; Moree, 2018). At the same time, the researchers themselves may be challenged by the research participants’ stances and experience, but also when presenting research findings in contexts they strive to challenge and change. This can lead to secondary/vicarious trauma (Klein, 2012; Nikischer, 2019; Williamson, et al., 2020), which may further impact the relations between researchers and research participants.

It is multiple power imbalances and experience of violence and abuse that need to dominate responsible research into gender-based domestic violence. Feminist participatory research aims at empowerment of both the individual research participants and the larger disadvantaged group and community (Lykes & Hershberg, 2012). Its ethics center on relationships and their situatedness, replacing mainstream principle-based ethics with ethics of care. This tradition pays special attention to the politics and ethics of representation, as marginalized groups are the ones that have the most at stake when it comes to mis/representation of their lived reality (Preissle & Han, 2012). Academic principles such as avoiding dubious publishers couple with considerations of what information can/should be published and in what format. Representation of findings is impacted by the purpose of the research, such as improving the individual situation of research participants/co-researchers, improving the overall social standing of the group, and achieving social justice/political change. In many instances, and when focusing on the micro-level of social interactions, violence can be individualized, losing sight of its conditioning by and consequences for social structures and hierarchies (Towers, et al., 2017; Walby, 2013). It is the researchers’ responsibility not to lose sight of contextual factors even when these are not directly mentioned, notably when social change is sought (Baker & Lalonde, 2020).

Researchers are however also human beings with their own attitudes, biases, life experience, social backgrounds, political convictions and pre-conceptions related to gender issues (Hofstede, 2001; Preissle & Han, 2012) who may have experienced violence and inequality themselves. All of these aspects should be included in research planning reflectively, because research among vulnerable groups is more ethically sensitive than other kinds of research as comparatively more is at stake for the research participants both individually (possible re-traumatization) and socially.

Given the number of factors listed above, the need to make ethically sound choices becomes evident. Many of the factors may not be readily visible, calling for procedures and methodological choices that minimize such risk. However, there are strategies that help in this process and that might be of special help in doing research with survivors of domestic violence.

The bare minimum for research with survivors is the no-damage approach, meaning maintaining the dignity of and respect for the researched population (Ellsberg & Heise, 2005; Guillemin & Gillam, 2004; Williamson, et al., 2020). In working with and for survivors of gender-based domestic violence, this means integrating research into trauma and its effects into each step of the research design (van der Kolk, 2015). The more intense involvement with the researched we assume, especially in participatory types of research striving for social change (Levenson, 2017; Lewin, 1946), the more salient it is to also pose the following questions.

Firstly, whose need is it to conduct the research, what purposes will it to serve, who will benefit from the research, and at what cost? The issue of cost primarily concerns minimizing further harm to those already harmed but also gauging the effects research may have on the research team, be it as a process, set of relations and/or result of representation choices. Whose actual need is it to conduct the research and why might that be? is the first question we address in each case below.

The second question directly concerns representation: Whose voice/s and in what form will be represented in the research and how do we secure this? Survivors can be integrated into research planning, conducting and analyzing in many ways. They can be only the subjects of such research, but they can also become an integral part of the research team in all research stages (Lykes & Hershberg, 2012). Whose voices are represented? is the second question for each case discussed below.

Comparing the Theatre of the Oppressed, Exploratory Research and Survivor-Centered Research

In the past several years, the two first authors have taken part in three research projects using the methods of the theatre of the oppressed (second author), exploratory research combining various data (both), and survivor-centered research (first author). The main ethical deliberations we focus on in the respective cases concern the situation of the survivors and the ways in which they were involved in the research and in the representation of the results. The concerns we identify here are thus related to the on-going assessment of benefits and risks associated with the research designs as they were carried out. The overall objective of all the research activities was to improve the situation of survivors of gender-based domestic violence, taking a progressive approach with an emphasis on different aspects of that situation: personal, institutional and the interaction of the two.

The first research design based on the theatre of the oppressed started in 2018 in response to the frequently hopeless situation of survivors in Czechia. The theatre of the oppressed and its research aim to empower those who are (made) voiceless and improve their immediate situation by giving them a voice. This meant the researcher (second author) was concerned with the specific situations of the individuals involved.

Second, based on this research by the second author, as well as the first author’s specialization in gender studies and previous participation in mixed-method research into sexualized violence and their shared knowledge of the situation and contacts in the field, in the first year of the COVID-19 pandemic (2020) the first and second author conducted extensive exploratory research, which garnered public attention. We focused here on the institutional framework impacting the survivors.

Third, the response to the exploratory research from institutional representatives and in the mediaFootnote 3 led to two survivor-centered research designs with survivors of intimate partner violence (2021) and with young adults who suffered domestic violence in recent years when they were childrenFootnote 4 (2022-3). We preserve this timeline in the comparison, in which we focus on how the research designs were organized and how particular decisions were interconnected with the societal situation and what impact they had.

Theatre of the oppressed. Theatre of the Oppressed (TO) belongs to a group of action research and participatory research strategies (Clark, 2009; Marín, 2007; Schroeter, 2013). It is predominantly a method that allows participants to identify ways to resolve situations they define as oppressive, but it can also be used as a research method providing a deeper insight into the situation of marginalized people and allowing us to explore a range of strategies for empowerment. In the article, we draw on the second author’s experience with TO techniques and both types of their application.Footnote 5

TO is a method of intervention directed at micro-level social change. It was developed as a method of giving a voice to underprivileged people (Boal, 1979, 1992) and emphasizes learning through dialogue and reflecting on lived experience (Freire, 1992). A group of people experiencing oppression stage a short theatrical performance based on their real stories in cooperation with a “joker” – a person trained in TO. Their performance creates a model situation of oppression intended to provide an opportunity to discuss it in public with the aim of reducing its destructive potential. The play is performed to an audience and the spectators are invited to intervene in the original performance in order to change the story and help the main protagonist. This strategy turns a passive audience into active co-actors, whom Boal calls spect-actors (Boal, 1992). It is very often used as a method for provoking dialogue and social change.

Looking at the field from a research perspective, we can see that participatory and action research most often serves as the basic framework. Within this framework, field observation, ethnography and interviews are primarily used as data collection methods complementing TO workshops and performances (Clark, 2009; Marín, 2007; Schroeter, 2013). However, some voices also claim that research strategies could be more interdisciplinary, which would make it possible to integrate more voices and could lead to a deepening of the theoretical debate (Moree, Vávrová, & Felcmanová, 2017).

We discuss a specific TO performance designed to give a voice to survivors/victims of domestic violence. To be able to compare TO as a research strategy, we look at several aspects of research planning and implementation:

Decision making about topic and method. The impulse to create a TO performance on domestic violence was not motivated by research requirements but was contextual. Gender-based domestic violence is a highly sensitive topic perceived as such by most in the population (Problematika Domácího Násilí Optikou České Populace, 2020). Despite research into issues on the structural level of addressing gender-based domestic violence (Hokr Miholová, Ondrušková, & Dohnal, 2016; Persefona, 2016; Pikálková, et al., 2004), the last substantial legislative change was passed in 2008 with several minor changes triggered by the COVID-19 pandemic. A shortage of shelters, gender-sensitive assistance, and the fact that the institutionalized system of assistanceFootnote 6 to survivors is frequently not able to stop the violence all prevail (Čechová & Jandová, n.d.; Kutálková and Kobová, 2014). This lack of a positive change led a group of women connected to an NGO that provides assistance to survivors, including a shelter, to the decision to create a performance on the topic. The dissatisfaction was both on the part of those assisting survivors and survivors themselves. TO was chosen as a tool for discussing the issue in public and attracting attention to it and because it was seen as a resolution to the ethical dilemma of presenting personal stories while preserving the survivors’ control over the whole creative process and performance while retaining their anonymity – unlike when they feel the need to publicly declare their story.

Group development and preparation of theatre performance. The main motivation for participants was to show the hidden aspects of the situation of survivors in Czechia. They were all unpaid volunteers and only received a small amount of cash after each performance. Originally, there were seven survivors and two NGO employees, all of them women aged between twenty-six and sixty-five.

Research decisions and outcomes. The motivation for the theatrical performance was to publicly point out the weaknesses of the system and start social change. However, one organizer’s (the second author’s) research background meant research strategies were integrated from the very beginning so that if approved by the performers (survivors), the research could contribute to increasing TO’s impact on social change through getting access to a wider audience. The dilemma of integrating the two different aims was resolved through securing the survivors’ on-going consent and openly pointing out the power imbalance in place. Special attention was paid to how best to represent the results, which was resolved through discussions. To help the deliberations, the whole process was recorded in detail: field notes were collected from the first meeting of the group; ethnography and participatory research were applied in the stage of preparing the performance; and, on the first night, the performance was recorded including the reaction of the audience. All of these sources were then presented and discussed with the survivors.

The main topics from the ethnographic part in this group were surprisingly consistent from the very beginning. The main issues were:

  1. 1.

    Being a good girl: all women in the group agreed that they tried their best to keep their family together; the main motivation was fear for their children and fear of the violent partner.

  2. 2.

    Repeated attempts to call for help: all of the women tried asking for help from different institutions such as the police, social care system, schools and others. They repeatedly faced trivialization of their situation which led to staying in the violent relationship (the longest story was 12 years of attempts to leave).

  3. 3.

    Consequent response of the system: all women reported that when they called for help, they repeatedly heard one sentence, which became a symbolic interpretation of their experience: “It’s a he said, she said situation.” To translate this expression into more sophisticated language, the institutions repeatedly expressed their fear that the reality women point out is biased and that they cannot help without “actual” evidence.

All the topics showed the multiple power imbalances experienced by the survivors, which became the focus both of the performance and of the research. In the final performance, the heroine finds the courage to take the first steps to leave her partner. She goes through the whole system, and in the end comes to a shelter where she learns how many obstacles there are (especially financial). In the end, she decides to return to her partner. This decision in the performance mirrors real-life experience of women in the theatre of the oppressed group when they repeatedly tried to leave. None of them was able to leave on the first attempt, which they perceived as a consequence of lack of professional and practical help in such a situation.

Effects of the research. The envisaged effect of TO is the empowerment of survivors. The research project also involved shooting a documentary about the performance, in which the feeling of empowerment became central.Footnote 7 The performance attracted some media attention (broadcast on Czech Radio including survivors’ interviewsFootnote 8). However, the impact in terms of social change was not seen as satisfactory by the researcher and survivors as the situation did not seem to change despite the approachable, interactive and intimate format. This was assessed as an ethical obligation to continue with the research and look for alternative strategies that might trigger change beyond individual survivors’ lives.

Exploratory Research Design

Soon after that the COVID-19 pandemic hit. The first and second author met to research the situation in three stages to cover as much of the system of assistance as possible.

Decision making about topic and method. Choosing an exploratory research design was deeply rooted in the ethical dilemma stemming from the TO (not triggering enough change) and the situation in society at the time. The Czech government failed to integrate domestic violence into its anti-COVID-19 strategy from the beginning with only minor changes since. State representatives did not pay attention to the issue; the only information available was that there were fewer domestic violence cases in the first weeks of the pandemic, based on a statistic recording fewer violent persons evicted.Footnote 9 The High Public Prosecutor’s Office issued a statement pointing out violent acts can be prosecuted more severely if committed under lockdown, yet it ended on this sentence: “The High Public Prosecutor’s Office appeals in these difficult times to everyone’s decency and mutual understanding in general and especially among those close to each other” (Zeman, 2020).

As a consequence of our previous research and contacts with NGOs working with victims and survivors of domestic violence, we learned that their experience was very different. The numbers of cases increased – contacts with victims increased by 30%, but the police and other state organizations did not seem to respond as much as before the pandemic making the statistics look different than reality.

The stark difference between the official information and information from the field combined with expert knowledge and experience with different research strategies led us to develop an exploratory research design. The motivation was firstly to “do one’s job”: a duty we felt as critical social scientists, i.e. to find out as much about the actual situation as possible in the highly constrained conditions of a nation-wide lockdown. Secondly, given our previous experience with trying to start social change and the ethical obligations arising from those attempts, we felt we needed to use the findings to disrupt the silence about domestic violence during the pandemic.

Development of research design. We selected exploratory research because we needed to collect sensitive data in a complex environment with new constraints in place. The motivation was to make the newly silenced (within the pandemic context) victims/survivors heard and raise awareness of the novel aspects of the issue. The dilemma was how to secure the voice: we could not approach the victims directly because of the restrictions and the security risk online contact could pose. It would have also been unethical to use their often only contact with NGOs providing help. Together with three NGOs providing gender-sensitive assistance to survivors in Prague, we therefore took the following steps to mitigate the “costs”:

  1. 1.

    Social workers and psychologists from NGOs working directly with victims started to write their field notes based on ethnographic methodology.

  2. 2.

    NGOs and researchers created a strategy for interviews and asked long-time and former clients not currently in a crisis to be interviewed. To prevent a breach of trust, the interviews were conducted by colleagues from NGOs and not by the researchers.

  3. 3.

    To monitor the situation in other areas, the researchers interviewed people from nine other organizations also working in the field of domestic violence.

We analyzed this data (separately by two different methods to control our outcomes in the team) and presented the results at a press conference in June 2020.

Research decisions and outcomes. The original plan was to stop after analyzing the data about victims/survivors at the beginning of the pandemic. However, the analytical findings were so alarming (increased incidence and brutality of violence) that we decided to continue for ethical reasons, namely a responsibility towards the NGO clients interviewed, the seriousness of the findings from interviews with service providers, and also the visible possibility to challenge the status quo because the COVID-19 pandemic put domestic violence in the spotlight. A second part of the research followed in the summer, involving interviews with the police, social services, intervention centers and NGOs; and a third part in the fall included lawyers and judges. We interviewed: 9 police officers; 9 social workers from social services; 8 social workers from other NGOs; 5 lawyers; and 5 judges. We analyzed approximately 900 pages of transcribed interviews and published the final report in February 2021 (Nyklová & Moree, 2021).

Effects of the research. The research results were presented at two press conferences (June 2020, February 2021) with enormous media coverage – in March 2021 alone, more than 25 articles, public interviews, TV and radio broadcast related its results.

The final report was also repeatedly presented at hearings of the governmental Committee for Prevention of Domestic Violence and Violence against Women with the aim to recommend concrete steps to reduce domestic violence on the level of the state, police and social work. However, despite all the facts presented and institutions approached, concrete changes in the system did not appear before the second wave of the pandemic. In the debates at the Committee and especially in statements from key stakeholders – most ministries and representatives of some intervention centers – two arguments were repeatedly put forward: (1) Children-oriented argument, which makes a seemingly irrefutable ethical claim: Children have a right to have both parents and both parents have a right to spend time with their children. This argument is coupled with severe gender-blindness as the shift in focus from intimate partner violence to children makes degendering of the whole phenomenon easier. (2) Untrustworthy victim: The definition of domestic violence is not unified in its application in Czechia, which means different institutions may not agree on whether concrete behavior is violent or not. The ethical issue of not causing any potential harm is raised, framed as harm to those wrongly accused. As a result, victims are perceived as untrustworthy, which is further exacerbated by lack of policies that would implement trauma- and violence-informed practices in all institutions that deal with the phenomenon, let alone practices for processing gender-based violence.

A third claim put forth by some representatives of the system of assistance but not shared across the board, was that domestic violence definitely did not increase as a result of the pandemic, and that it even subsided in many cases. Trivialization of the phenomenon has been identified as present not just in the pandemic context. However, in this case, our research results were questioned, too. The representatives of two of the organizations claiming the pandemic had no effect wrote letters of complaint to the authors’ respective institutions (addressed to them as well as the heads of the institutions), to the Police Presidium as well as to the Government Commissioner for Human Rights complaining about alleged methodological issues. The authors’ institutional heads on the initiative of one of them wrote a joint response pointing out the claims in the complaint were groundless.

Survivor-Centered Research

Decision-making on topic and methods selection. It was only after using the strategies described aboveFootnote 10 that we decided to include survivors of gender-based domestic violence as research participants in a traditional way of conducting qualitative research. The extent to which this violence, with its gendered and otherwise structurally conditioned nature, were systematically and repeatedly trivialized even under the COVID-19 pandemic called for involving survivors in a traditional type of qualitative research, which should at least in theory be more readily accepted by the institutions in question than less orthodox strategies and strategies not directly involving survivors. We started to conduct two parallel research projects: (1) Research into the barriers faced by survivors of intimate partner violence and providers of social services focused on underprivileged groups, such as homeless, handicapped and/or socially stigmatized people, and (2) Research with young adults who experienced the violent divorce of their parents as minors. We focus here on the research into barriers as it has been completed already.

Research design preparation and method development. Research into the barriers faced by survivors of gender-based intimate partner violence confronted multiple dilemmas and concerned the well-being of research participants/survivors in the research process. The research was done as part of a project awarded to one of the NGOs we had previously worked with. The research team included NGO workers and an academic – the first author. To minimize the risk of re-traumatization of survivors, we first decided to approach clients of the NGO and other NGOs their expert workers believed would benefit from taking part or at least were not at great risk of re-traumatization. The plan was to first interview survivors and select other organizations to talk to based on the analysis of their accounts. However, when the project was launched, a new lockdown was introduced and it was not possible to conduct face-to-face interviews with survivors, the method we deemed safe enough as it comes with some control over the physical conditions of the interview and makes non-verbal communication available. Instead, we first remotely interviewed experts from various organizations providing social, legal, therapeutic, and other assistance to underprivileged and disadvantaged people. Interviews with these organizations partly affected what topics we decided to include in the research with survivors and also confirmed the need to focus on relatively recent cases as the argument that a particular case was “old” and would not be treated the same at present was familiar to our interview partners. This strategic change also made it possible to ask these experts for the possibility to identify survivors who might want to talk to us in the second stage of the research.

To mitigate harm and re-traumatization, we started the research with a brief introduction into crisis intervention so that we could identify when it was needed. We had an experienced therapist on the research team and another one who supervised us in terms of research team needs in order to prevent secondary trauma and re-traumatization in team members with some previous experience with violence as this too was an ethical concern. The interviews proved challenging for some of the research participants, while others commented on them as a positive experience and a chance to turn their experience into something meaningful. In the design, we supported this by openly asking for their recommendations and statements on the situation in the area from their expert position. In line with previous findings (Klein, 2012), survivors perceived the interview setting as a safe space when they used it to talk about their experience as a whole (many commented that it was for the first time ever) and in several instances, to share experience of sexualized violence they had previously not shared with anyone including therapists and psychiatrists.

Research decisions and outcomes. In the first stage, we talked to eleven research participants from five NGOs focused on various types of disadvantages (mental health, physical health, social), workers from two intervention centers, emergency services, a medical doctor, and a journalist specializing in covering domestic violence cases. In the second, we talked to twenty-one women and one man from across Czechia from various backgrounds and different stages of life, although most were middle-aged. We took special precautions to conduct the interviews in a setting either of the participant’s choice or one in which we could control safety (such as premises of an NGO with guaranteed absence of workers). Still, in some cases, the conditions were not ideal: as all of the research participants had children and often faced a challenging work-life situation, the place and time of the interview had to change to a less suitable location, even if one in all cases suggested by the respective research participant. Moreover, different organizations were approached to help recruit a diverse group of survivors to draw attention to challenges stemming from the intersection of multiple axes of inequality (health issues, socioeconomic status – precarity and poverty, ethnicity and related racism). This meant using premises known to the survivors but not to the researchers who were not necessarily able to adjust the setting (room layout, blankets, beverages).

In all cases, the participants were first introduced to the consent form and given space and time to read through it; a contract was signed so that they could receive reimbursement for taking part in the research just like in the case of other experts from the first stage of research; the participants were repeatedly reminded that they could choose not to answer a specific question or comment on a given topic and that they could end the interview at any time. After the analysis, survivors were sent their profiles to be included in the research report for approval (the profiles replace a classic sample description so as not to decontextualize the individual quotes used in the report) and also to be reminded of the fact that the analysis would be published (and they could remove their account from the sample, although nobody opted to do so). After their approval and completion of the final report, both experts from organizations and survivors were sent the report for comments. One participant from one of the organizations was not happy with their verbatim quote used in the report because of its critical context in the report, so the participant suggested a complete rewriting of their original statements and the first quote was removed.

Effects of the research. A research report was prepared with the intention to have it co-published by the NGO and the first author’s academic institution. The reason for this is the bias that exists regarding research done by non-governmental organizations, especially when these identify the violence as gender-based. Co-publishing would thus potentially improve the credibility of the report built on individual accounts of survivors. However, the management of the institution refused co-authorship, claiming unspecified issues with the analysis, which they failed to elaborate upon when asked for details. The complaints against the previous research were partly successful as the institutional management started to be wary of the follow-up research activity, explicitly stating they believe this is a problem of a polarized field rather than anything else. The report was therefore published by the NGO with full attribution to the authors along with their institutional affiliations (Nyklová, Michálková, Gubová, et al., 2022). A press conference was held and several articles about the research were published in the media,Footnote 11 but since the conference took place two months after the beginning of the Russian invasion of Ukraine, many of the questions pertained to that conflict and the role gender-based violence played in it. As for the survivors, the feedback was complex: some were not eager to be further involved in the process, while others were interested in the results and took a lot of time editing their anonymized profiles. For some of the survivors the very possibility to tell their story proved highly positive, for others, the experience of talking about their experience proved more challenging than they had expected. The researchers in some cases witnessed severe maltreatment and neglect by the system of assistance and the research showed that the ethical precautions were appropriate. One team member did experience re-trauma when giving priority to the need expressed by a survivor to relate her account despite not being seen as fit by the NGO that offered her participation. Given the severity of the findings, a further strategy for using the report needs to be devised in order to meet the expectations of all the parties but especially the survivors involved.

Discussion

In the description of the three cases, we focus on how the context of each research affected the next methodological choice and what ethical concerns arose. The general rules of no/minimum harm when doing research with a group as vulnerable as survivors of gender-based domestic violence may often seem both self-evident and obscure at once. We therefore opt for detailing the context and outcome of the steps taken in our individual and joint research.

When we approach the same issue from the perspective of survivors and their participation in research, we can conclude the following: traditional qualitative research and comparatively more participatory research designs such as the theatre of the oppressed differ substantially when it comes to how they work with potential re-traumatization. In traditional qualitative interview research, participants share narratives and information about their cases, but the research situation itself is only descriptive. It does not allow the possibility to change anything; the aim is merely to transmit a story in which powerlessness played a strong role. This is a borderline situation from the perspective of research ethics, partly also because survivors often have high expectations from the research. Our research showed the risk of re-traumatization was high even for researchers. Therefore, meticulous precautions are crucial here, such as conducting research in a team with regular supervision, debriefing after each interview, revision of the interview guide during research, and planning the research context and representation. Our experience with TO suggests that the more participation in research survivors get, the deeper their experience of potential and concrete strategies, which can help to transform powerlessness into empowerment.

However, as we demonstrated in the above examples, applying strongly survivor-oriented research designs is not always possible. In some instances, traditional qualitative research designs are perceived as more relevant by people in positions of power. At that point, the two perspectives – starting with survivors and starting with the system – can reconcile through integrating the issue of meaning into the research decision-making process. When research is properly communicated with the survivors of domestic violence, they can be invited to take part by explaining that although the team has tried all other strategies, it seems that traditional interviewing is the only next step that might help change the whole setting. Survivors are invited to participate in their power – the power of people who have experienced things that are often unimaginable for the public. Their voices are no longer the voice of powerless victims but of experts who are the only owners of very specific experience. If the research can have such meaning to its participants, then it is research representing the voices of survivors. When researchers do not attempt other participative strategies and cannot show genuine interest and willingness to look for change, then our take-away would be: let us think about whose voice and whose need to conduct the research it really is.

Conclusions

We wish to conclude by stressing some take-aways that may not be readily clear from the comparison alone. The global presence of anti-gender initiatives and their local expressions, including reluctance to ratify the Council of Europe Convention on preventing and combating violence against women and domestic violence (Graff & Korolczuk, 2021; Verloo & Paternotte, 2018), might suggest that gender and any intersecting axes of possible marginalization need to take center stage both in research into and action against gender-based domestic violence. What researchers should do when this is clearly not the case may seem trivial – just present results on how the violence is socially conditioned and how it is made possible. However, our examples show it is not clear which method will be the most useful one (when we want to pay attention to detail as well as explore new questions and therefore need to use qualitative or mixed methods). When doing research with a vulnerable group, this means we need to consider all actors concerned.

Gender and the social conditioning of domestic violence in general may get marginalized in research at every stage with different impact. Research participants come from a social and cultural context where gender is seen as a natural part of the social fabric and denied its analytical value at the same time. It may be traumatic to relate this in the research (interview) because it may increase the feelings of disempowerment and injustice in survivors; it may also mean refusal to participate on the part of the system of assistance and law enforcement when the representatives of these institutions themselves hold anti-gender views. The researcher then has to decide how important the accounts are and whether it is justifiable not to disclose one’s expert opinion.

Another potentially difficult set of resolutions pertains to dealing with individuals and their accounts as juxtaposed with what their own institutions say in response to research outcomes. This may involve dealing with participants with decision-making power who suggest completely changing their accounts (all interviews are anonymous), trying to assert the phenomenon is gender neutral despite their own account. This leaves the researcher with an uneasy choice given the relatively small number of decision-makers in the field and the certainty that their decision on complying or not will be shared without context with other gatekeepers in the field. Another dimension is represented by the academic setting of some of the presented research and by considerations as to whether to involve academia as a setting or not in the case of the theatre of the oppressed. This setting holds the promise of giving a voice that will be heard, yet it comes at a price. Writing up research almost always means making ethical choices in terms of deciding what is publishable in a system where quantitative studies are still seen as more objective – or true and relevant – than qualitative ones. This is all the more problematic in a field stemming from feminist activism since the challenge to science as an objective endeavor has most vocally come from feminist thinkers. Moreover, in a setting that is gender-conservative, this also pertains to academic institutions, which may feel threatened by research calling for social change and detailing issues that need fixing. Finally, the question of funding needs to be considered – we do not address it here because most of the research was done pro bono and in collaboration with those supporting gender equality. However, this fact alone also means making sure whose interests are represented and not forgetting that we as researchers are in a position of privilege if we can afford to conduct such research in the first place, and that it comes at a cost.

The take-away here is that doing such sensitive research not yielding to the paranoia is one of the ethical aspects to remember and reflect through the whole process, starting with planning the research and ending with its media presentation. This being said, the possibility of retraumatization of any of the parties involved is by no means to be taken lightly.