Ethics in Violence and Abuse Research - a Positive Empowerment Approach

by Julia Downes, Liz Kelly and Nicole Westmarland
School of Applied Social Sciences; London Metropolitan University; Durham University

Sociological Research Online, 19 (1) 2
<http://www.socresonline.org.uk/19/1/2.html>
10.5153/sro.3140

Received: 5 Feb 2013     Accepted: 4 Jul 2013    Published: 28 Feb 2014

Abstract

Research governance, including research ethics committees and data protection legislation, is invested in protecting the individual rights of participants in social care and health research. Increasingly funders expect evidence of outcomes that engage with 'service users', making research critical in supporting social interventions to compete for scant resources in an economic climate marked by 'austerity' (Sullivan 2011). This article focuses on the tensions that can arise from the research governance of violence and abuse research. We argue that increased scrutiny of violence and abuse as a 'sensitive' topic that involves 'vulnerable' groups has made ethical clearance more challenging, which in turn can lead to a dangerous lack of evidence. This can have a harmful impact upon women and children and leave specialised violence and abuse services facing a precarious future. Drawing on recent debates we describe the 'positive empowerment' approach used to engage victim-survivors and perpetrators of domestic violence in Project Mirabal. We conclude with recommendations for ethical decision-making in violence and abuse research: (i) to reconsider participants as active agents and stakeholders; (ii) to prioritise the development of skilled researchers; (iii) to develop situated processes of informed consent and confidentiality; and (iv) to continue to discuss and share practical experiences of feminist research practice that seeks to deliver justice and social change.

Keywords: Research Governance, Research Ethics, Research Design, Confidentiality, Informed Consent, Violence, Abuse

Introduction

1.1 Following several decades of feminist practice and activism, the elimination of gendered violence and abuse has become a global ambition at the United Nations, European Union and Council of Europe levels (European Commission 2010). The last 15 years have seen increasing demands from policy makers, funders, and practitioners for a rigorous evidence base to assess the effectiveness of social interventions which aim to reduce and prevent such violence (Sullivan 2011; Gondolf 2012). In England and Wales particularly, the necessity of evidencing outcomes and the 'value' of such interventions is likely to increase further as the combination of the newly introduced Police and Crime Commissioners, local health commissioning, and the implementation of 'austerity' measures require third sector providers to compete against one another for diminishing resources. The promotion of evidence-based policy and funding has led to increased scrutiny of social interventions and an influx of academics, researchers and students within the everyday working practices, sites and client groups of domestic violence services. This raises the question of the appropriate balance between ethical and professional responsibilities in collaborative research pursuits (see Gondolf et al. 1997). On one hand, research involvement can enable practitioners and client groups to inform social policy and guide the future direction of service provision. On the other hand, there is a clear need for ethical guidance to ensure that the rights of research participants are promoted and protected.

1.2 This article starts by surveying the competing paradigms at work in research governance as they apply to violence and abuse research, followed by a description of and rationale behind recent calls to expand the ethical lens. We then describe the 'positive empowerment' approach our current project (Project Mirabal, which investigates what domestic violence perpetrator programmes add to a coordinated community response) has taken with victim-survivors and perpetrators of domestic violence. The article concludes with recommendations for ethical decision-making in the recruitment and involvement of domestic violence victim-survivors and perpetrators in social research.

Research governance and violence and abuse research: ethical tensions

2.1 Research governance, including research ethics committees and the Data Protection Acts 1998 and 2003, is invested in protecting the rights of individual participants in research. Research ethics committees (RECs or Institutional Review Boards – IRBs – in the United States) that sit in higher education institutions have a crucial gatekeeping role in health and social care research. In essence RECs decide which academic research is conducted, with whom and how it is carried out. The international growth of RECs has been shaped by a biomedical paradigm based on lessons learned from medical trials and principles defined by the World Medical Association and the World Health Organisation (Department of Health 2005; Smith 2008; Edwards 2010; Fisher 2012). These principles of autonomy, beneficence, non-maleficence and distributive justice reflect ethical obligations to respect the decisions that people make, to help others, to avoid harm, and to contribute to social justice and equity.

2.2 In 2002 Julie Kent and others highlighted, in this journal, that ethical review for social science research was rare. Ten years on, universities require all research projects involving human subjects to go through a REC. Where research involves health and social care services further additional processes are required and the issue of ethics has taken greater prominence in application forms for research funding. We explore whether this enhanced scrutiny is being implemented in ways that may have differential impacts on research with marginalised groups at a time when the involvement of 'service users' and 'clients' in developing a reliable evidence-base is increasingly emphasised in social policy (Smith 2008). This connects to a recent observation that access to 'vulnerable' groups has become increasingly difficult for researchers often leading to small and/or unrepresentative samples (Department of Education 2012). The impact of which reinforces a view among policy makers that we lack a rigorous knowledge base on the experiences of marginalised groups. The biomedical principles and requirements of some RECs have been criticised as inapplicable to the ethical challenges that face social researchers (Antle & Regehr 2003; Schrag 2011). In undertaking social research with victim-survivors and perpetrators of domestic violence researchers can encounter tensions between the ethical requirements of RECs, which protect individual rights and freedoms and institutional reputations, and their own professional judgement and commitment to promote social justice for women and children.

2.3 Violence and abuse researchers often face complex ethical dilemmas and difficulties in relation to women and children's safety which are rarely addressed in ethical guidelines. Ethical practice is, thus, guided by the judgements of experienced research teams and publication of ethical protocols and procedures (Berry 2009). In this section we focus on the practical and ethical challenges that confront domestic violence researchers faced with conventional understandings of informed consent and confidentiality rooted in biomedical research that assumed direct face to face contact with health practitioners and the researchers in a clinical setting. Typical guidelines (Antle & Regehr 2003; Edwards 2010; WMA 2008) indicate that the core elements of informed consent should involve:

2.4 We now explore how tensions between biomedical and social justice paradigms can play out in relation to these requirements. In particular we examine: the use of information sheets and consent forms that presume a signature; assumptions of 'capacity'; blanket assurances of confidentiality; time-consuming resubmission processes; and, broader ethical responsibilities, including the impact of research findings. Each of these has the potential to undermine women and children's safety and/or feminist social justice principles to empower marginalised women.

Consent as a formal contract

2.5 In our experience RECs expect to see forms, which will be signed by participants, as evidence of how informed consent will be sought and obtained. This 'contractual ethics' model of 'informed consent' is often a challenge for researchers in violence and abuse. Informed consent forms can be interpreted as official documents to protect institutions and researchers rather than participants (Riessman 2005; McNutt et al. 2008), and the processes are suited to contexts in which individuals have regular face to face contact with researchers in a clinical setting. These conditions are rare in social research, and the issues in violence and abuse research are especially complex. Women deploy daily strategies in order to make themselves and their children safer within violent relationships and abusive family regimes (Cavanagh 2003). This complex management of living with violence and abuse can be difficult for outsiders to understand, requiring researchers to both develop awareness of the unique circumstances of participants and be able to adapt procedures and respond flexibly to fluctuations in risk and safety. The unanticipated nature of such situations means that all potential risks and dangers of research participation cannot be assessed during ethical clearance and the onus rests on the expertise of the violence and abuse research team to maximise safety for women and children. For instance, where women are still living with, or having regular contact with, their abuser it may not be safe for them to sign forms or have participant information sheets sent to or left in their homes. A victim-survivor's choice to take part may be contested by their partner and may result in harmful repercussions for them: this makes confidentiality an absolute requirement (Holtzworth-Munroe et al. 2003). Both reassurance and processes that demonstrate the researchers' on-going awareness of these risks are often key factors in victim-survivors' decisions of whether to take part; here a requirement to sign their name on a form appears for some to contradict this. Mindful of this very immediate potential for further harm, some researchers argue for verbal consent only or use of unique identification codes instead of real names to avoid the detection of research participation (WHO 2001).

Defining the 'victim': 'Sensitive' research topics, 'vulnerability', and the 'capacity' for research participation

2.6 Most RECs view research on violence and abuse as 'sensitive', and may also define the participant group as 'vulnerable', both assumptions make the study more likely to be carefully scrutinised. We explore both concepts in turn and how they can play out in research governance. What is and is not a 'sensitive topic' in social research has been debated. Whatever the definition, the fields of violence, abuse and power invariably come within the remit given their clear potential for harm as research inevitably requires exploring the personal with an accompanying potential for distress (Lee & Renzetti 1990). Where a research study is identified as being on a 'sensitive topic' this seems to carry an extra level of responsibility for a research team. Such research projects are presumed to make extra demands on participants beyond the normal costs of time and inconvenience. Concerns here centre on the potential of 'psychic' costs – guilt, shame or embarrassment – or the more serious concerns about 're-traumatisation' for those who have been victimised and the potential consequences for those who have perpetrated violence of revealing on-going abuse that places a woman or child at risk. In the latter case this raises potential conflicts with confidentiality for researchers which do need to be addressed in research protocols. However some researchers have questioned why there should be extra ethical considerations for 'sensitive' research topics. Lisa DeMarni Cromer and Elana Newman (2011) argue that there are no clear grounds for holding violence and abuse researchers to a different or higher standard, and they dispute any need for victim-survivors to have unique protections. All research with human subjects shares the responsibility to respect the rights, welfare and dignity of its participants. Arguably these could be placed in more jeopardy when the topic appears, on the surface, not to be about violence or abuse. For example, a study on consumer choices may lead to disclosures about violence and abuse that were not anticipated by the research team and may not have been considered relevant in research design and protocols. Therefore an awareness, not only of potential disclosures of violence and abuse, but that any aspect of research, including seemingly innocuous questions, has the potential to resonate in specific ways for victim-survivors with a potential for harm is an issue for all social research.

2.7 Domestic violence is acknowledged as being a global pandemic that affects a large proportion of the population. Following this logic, and dominant biomedical foundations of research governance, it may be worthwhile to apply the 'universal precautions' principle developed during the 1980s HIV/AIDS crisis whereby every patient was treated as though they were HIV positive. This minimised risk to all healthcare workers and patients and tackled the discrimination of HIV positive individuals in accessing healthcare (CDC 1987; Gostin & Webber 1998). Given that at least one in four women will have experienced physical and/or sexual violence in their lifetime (Hall & Smith 2011; ONS 2013) the reality is that all research with human subjects has the potential to involve people who have experienced violence and abuse, with a similar, although not totally parallel (due to some being serial offenders), issue with respect to perpetrators. Therefore all social research carries a potential risk of identification, repercussions and harm to victim-survivors and perpetrators. However, for violence and abuse researchers, it is important to challenge the biomedical model, classification of 'sensitive' topics and concept of vulnerability that set into play negative stereotypes of victim-survivors of violence and abuse.

2.8 Locating anyone who has been victimised in a class of 'vulnerable' people (alongside prisoners, pregnant women and children) who may be more likely to face coercion and exploitation in their everyday lives (Clark & Walker 2011) highlights how 'vulnerability' has an increasingly slippery and ambiguous meaning. Positioning victim-survivors who have been raped or abused as 'vulnerable' carries an assumption of capacity. This means that victim-survivors are assumed to be unable to comprehend the potential risks and consequences that may unfold as a result of participating in research and the process of informed consent may not be possible for them. This is connected to wider debates about the concept of 'vulnerability'. For instance, in some contexts, including health research, vulnerability has become a euphemism for inequalities, raising the critical question of whether vulnerability applies to persons or situations (Coleman 2009). This variability of definition creates ambiguities and uncertainties, with the potential for inconsistency in the decision-making in RECs. If women who have experienced violence and abuse are included in the minorities category [sic] of vulnerable social groups, along with children, long-term ill, recently bereaved or institutionalised, this effectively places the majority of persons in this category. The key contestation here is whether social groups should be automatically allocated into a category that is defined against an implicit norm of persons in relation to which they are deemed too deficient, and thus less capable of protecting their own interests. In an interesting reflexive commentary Lancaster University caution that a 'co-production' may be occurring here, in so far as the categories are defined and filled with meaning by research governance processes, with potentials for discriminatory practice and outcomes[1].

2.9 The medical model, based as it is in testing new forms of treatment, has to presume the potential for harm: but the point of studies is to look for benefits in the trial groups and for a wider group of patients. There has been relatively little exploration of the potential benefits of participation in social research, especially in research contexts that might maximise this. Whilst participation in a research project on experiences of violence and abuse can present a risk of violent repercussions (Holtzworth-Munroe et al. 2003), other studies have documented positive outcomes of research project participation for victim-survivors (Kelly 1988; Newman et al. 1999; Cromer et al. 2006; DePrince & Chu 2008). It is axiomatic that individuals should be treated as full human beings rather than 'objects' for study, and that researchers should be interested in facilitating and listening to what they have to say. There is evidence that shows the benefits of taking part in well designed, safety conscious, violence and abuse research. These include the role of 'bearing witness' for victim-survivors – to be living proof that domestic violence and abuse exists, has had real impacts on that person's life and that they have sought to address these in a variety of ways. For example a participant in one study on sexual violence explained: 'Actually, I'm quite surprised, I've found it really helpful. I can't think about it so talking is the only way of admitting it ever happened… I have never talked in that concentrated way before… I think I like myself a lot more, I feel quite brave really' (Kelly 1988: 13). Participants have also reported feeling valued and listened to, being treated with respect and dignity, gaining personal insight, and finding the experience meaningful (Newman et al 1999; Cromer et al. 2006; DePrince & Chu 2008). For some victim-survivors taking part in a research project can be an act of resistance and/or an opportunity to use one's own experience in order to help others. Here the connection to feminist and anti-rape movements with the focus on consciousness-raising and empowerment begins to emerge.

2.10 What is insufficiently explored is what the potential for greater harm is within the research design. Situating vulnerability as an inherent and shared property or static category of personhood is problematic for social research, not least since it includes a potential for 'othering' participants. Little thought would be given, for example, to the fact that those who have already experienced harm and/or exploitation might be more alert to its potential, thereby more able to make decisions about participation. Similar to findings in research projects with other so-called 'vulnerable' social groups, sophisticated forms of resistance and agency, reluctance and refusal to take part in research can be regarded as a form of agency (Fisher 2012). The fact that it can be difficult to recruit into domestic violence studies suggests that many women exercise their agency with relative ease. We concur with domestic violence researchers who argue that the blanket designation of 'vulnerable' is a paternalistic practice that undermines the agency, capacity and diversity of victim-survivors and their decisions (Alcoff & Gray 1993; Lamb 1999; Mulla & Hlavka 2011).

2.11 In Project Mirabal researchers encounter victim-survivors and perpetrators of violence and abuse at different time points; some are living with on-going violence and abuse (and even here there are variations in frequency and forms of violence) whilst others have been separated for considerable lengths of time. We are thus meeting with women and men at different stages in the process of dealing with violence. This has led us to consider when the term 'vulnerable victim' is mobilised, by whom, and to what effect in research governance. This touches on well-worn feminist debates about the meanings of 'victim' and 'survivor', whether and when either should be used (Kelly et al. 1993). Some argue that these categories can only be self-defined (Alcoff & Gray 1993), others that the continued deployment of the term 'victim' by researchers and experts is paternalistic and diminishes the agency of the participant (Lamb 1999). We use 'victim-survivor' to indicate that both victimisation and survival are present in the experience of violence and its aftermath, since women have to find ways to cope with the myriad consequences of abuse in their lives. It also reflects the changes and legacies individuals talk about when reflecting on the place violence has in their lives, their sense of self and relations with others.

2.12 If victim-survivors are, by definition, too fragile to give informed consent or participate in research there will be increased reluctance to grant permission for research involving them. This will curtail studies on traumatic life events, including understandings of what enables the re-building of selves and lives. This knowledge gap could be harmful to victim-survivors of violence and abuse and could even signal an avoidance of responsibility for social research to better understand how women and children can achieve safety and freedom (Stark 2007). As Joan Sieber and Barbara Stanley argue:

Sensitive research addresses some of society's most pressing social issues and policy questions. Although ignoring the ethical issues in sensitive research is not a responsible approach to science, shying away from controversial topics, simply because they are controversial, is also an avoidance of responsibility (1988: 55).

2.13 Whilst recognising that appropriate research governance protects the human rights of participants we are troubled by the designation of 'sensitive' and 'vulnerable' to all violence and abuse research, and the risk averse decision making that this can lead to. We are not talking here about disallowing projects as such, but rather anxieties about asking certain questions or in-depth exploration of violence and abuse, its impacts and meanings. Discussions within our research networks evidence this 'caution through categorisation' – defining all victim-survivors as 'vulnerable' – runs the risk of inappropriately restricting new knowledge and perspectives on the widespread social problem of domestic and sexual violence. A shrinking evidence base also runs the risk of further marginalising victim-survivors and undercutting the social significance of specialist support services.

Blanket assurances of confidentiality

2.14 Some RECs require a blanket assurance of confidentiality for participants, particularly in studies of 'sensitive' topics. This means that participants are able to open up and discuss a topic anonymously without fear of repercussions associated with research participation. Not to underplay the risks that victim-survivors and perpetrators face; a number of studies have found that at the point of actual or possible separation women are at an increased risk of harming themselves, killing or being killed by their violent partner (Jones 2000; Stark 2007). Furthermore, children living with domestic violence are also at an increased risk of a range of harms (Peled 2000; Radford & Hester 2006). Therefore any disclosures of serious harm towards women and children within a research project that operates a blanket assurance of confidentiality cannot be dealt with by the research team. This leaves a research team in a dubious ethical situation; with the knowledge of a harmful situation and an inability to respond to make the situation safer. For instance, Lisa Aronson Fontes (2004) described how some researchers have developed techniques to ensure the participant in domestic violence research remains completely anonymous thereby bypassing the ethical dilemmas of having to consider reporting concerns to an agency. However we concur that this amounts to a dangerous avoidance of responsibility and lack of care and concern for research participants. We seek to balance the need to promote openness during interviews and genuine care and concern about the wellbeing and safety of all research participants. Therefore RECs need to reconsider blanket assurances of confidentiality in violence and abuse research. Ethical and safety protocols that emphasise a limit to confidentiality in violence and abuse research and engage with (rather than ignore) unanticipated ethical dilemmas will maximise the safety and wellbeing of research participants throughout the research process.

Resubmission of ethical protocols and fieldwork materials

2.15 Some RECs require comprehensive ethical and safety protocols to be submitted in advance of contact with the participant group to adequately demonstrate how the research team will protect the participants from risk of harm. However, as with the open-ended nature of social research, within our area of research it is impossible to produce a protocol which anticipates and addresses every possible risk. Instead abuse and violence researchers rely on training, experience and practitioner expertise to respond appropriately to unanticipated dilemmas which arise within the research process (Sullivan & Cain 2004). The deviations from traditional protocol required to conduct safe research on topics of violence and abuse (e.g. not signing names on consent forms etc.) can therefore conflict with the standard requirements of a research ethics committee. Projects often face an extended process of ethical approval involving the resubmission of protocols and data collection tools and accompanying justifications. Although some research teams have found this process useful (Berry 2009), it can result in severe delays to research timelines with the potential to undermine relationships between academics and their external funders.

2.16 Recent debates in violence and abuse research have discussed the need to 'widen the ethical lens', to shift from rule-based participant-focused ethical protocols to a focus on researchers as 'ethical thinkers' capable of responding appropriately to the unexpected throughout the research process (Clark & Walker 2011). Experienced violence and abuse researchers and practitioners are acutely aware of the complexities of power, control and resistance managed in regimes of abuse and violence (Cavanagh 2003; Stark 2007). Violence and abuse is embedded in everyday lives and the actions of 'perpetrators' and 'victims' are naturalised in ways that are significant to a particular relationship and/or family (Stark 2007; Morris 2009; Williamson 2010). This lends weight to the importance of staff training and supervision to ensure that the researchers 'have the knowledge and skills to handle unanticipated situations' (Sullivan & Cain 2004: 607; see also WHO 2001).

From participant-centred ethics to broader ethical responsibilities

2.17 In assessing the ethical integrity of projects some RECs focus on the robustness of procedures that manage the consequences of participation in social research, with relatively limited attention paid to the potential of research findings to cause harm. A growing number of violence and abuse researchers (e.g. special issue of Violence Against Women, December 2011) have flagged up the need to consider ethical consequences for the research team and wider society including service provision, social policy, criminal justice system and legal frameworks. Some have argued for a consideration of consequences for the research team, broadly defined to include researchers, transcribers, supervisors, readers/viewers of publications and presentations, practitioners and professionals (McCosker et al. 2001; Johnson & Clarke 2003; Grinyer 2005).

2.18 In addition the fit of research findings with feminist principles to improve the safety and freedom of women and children and possible impact on social groups and services not directly involved in the research project need to be considered (WHO 2001). For instance research findings, even motivated by good intentions, have the potential to further stigmatise an already 'vulnerable' population. An instructive example here is the creation of the 'battered women syndrome' (Walker 2000) as a psychological and medical model in order to enable women access counselling through their health insurance in the US. It was subsequently used by legal defence teams in cases where women had killed their abuser. Rooted in the notion of 'learned helplessness' feminist critics (see, for example, Ferraro 2003; Stubbs & Tolmie 1999) illuminate how it has created an exclusionary stereotype based on conventional white middle-class femininity, reinforcing gendered discourses of women as passive, emotionally fragile and dependent. In the legal context the result has been that women who act assertively and are not cowed by abuse fail to fit the 'victim' profile. Furthermore, viewing domestic violence as an individual pathology obscures the complicity of state agencies – the police, health and social services – with violent men, in so far as they have historically ignored it and/or failed to hold perpetrators accountable (Romito 2008). This raises the question of the wider ethical responsibilities of researchers and RECs: it is not just the research process that has the potential to cause more harm.

The positive empowerment approach: core ethical values of Project Mirabal

3.1 Project Mirabal is a longitudinal mixed-method study that investigates what domestic violence perpetrator programmes (DVPPs) add to a coordinated community response to domestic violence. DVPPs are assessed using nuanced indicators of success, developed in a pilot study, that focus on change beyond a reduction in physical violence. These indicators consider change in: communication, women's space for action, violence and abuse, parenting, men's awareness of self and others and children's well-being (see Westmarland et al. 2010; Westmarland & Kelly 2012). Quantitative data about men's behaviour is collected from reports of partners and ex-partners in five questionnaires completed by phone over the course of twelve months. This involves following a large sample of women at different stages of experiencing violence and abuse who are in contact with a support service. Some are living with on-going violence and abuse, whilst others have been separated for a considerable length of time (in these cases a man's attendance at a DVPP is linked to applications for child contact). Women are recruited from the integrated partner support service at fourteen community-based DVPPs and twelve Freedom Programmes – a widespread programme for women who have experienced domestic violence and abuse – situated across England, Wales and Scotland. Comparison of the lives of women whose partner or ex-partner is on a DVPP with the lives of Freedom Programme women who live in areas where a DVPP is unavailable means that Project Mirabal is designed to identify what an intervention for perpetrators adds to women's services. In addition qualitative data about programme impact and abusive behaviour is generated from a sub-sample of 60 women and 60 men who take part in two semi-structured interviews at the beginning and end of DVPP involvement.

3.2 The development of an ethical protocol, in which researchers are invested in protecting victim-survivors of violence and abuse whilst also maximising the capacity for self-determination and autonomy, within Project Mirabal has enabled us to articulate core ethical values that underpin a positive empowerment approach in violence and abuse research. These core values include, (i) conceptualising victim-survivors and perpetrators as active agents, (ii) empowering participants to make choices about taking action to improve their lives and, (iii) maximising opportunities for positive experiences and impacts of research.

Victim-survivors and perpetrators as active agents

3.3 The fundamental core value of our approach is that both victim-survivors and perpetrators are active agents in making decisions about whether and how to take part in Project Mirabal. In contrast to the stereotype of a victim of domestic violence as passive, women who live in (or have lived through) violent and abusive situations are often highly skilled in managing risk and harm to themselves and their children (Cavanagh 2003). We do not want to undermine the risks involved in research participation in violence and abuse. Re-visiting traumatic events can have unanticipated consequences for individuals. However a perception of research participation as inevitably negative for victim-survivors should not be presumed and where it does occur it should be addressed. For example, in Project Mirabal one woman, whose current partner was attending a DVPP, did not report any distress immediately after the interview but did after a few days. The woman was engaging with a women's support service associated with the programme and the researcher liaised with her support worker to enable her to withdraw from the research project and seek support. This is a clear example of how 'vulnerability' is not a constant, and that ethics need to extend into how consent is re-negotiated during research.

3.4 Researchers note that both violent men and non-specialist organisations have been resistant to take the step to participate in research on domestic violence (Hearn 1993, 1998; Gondolf & Deemer 2004); a situation that is likely to be exacerbated as 'austerity' measures reduce the extent and range of such services (Walby & Towers 2012). Men's encounters with DVPPs are saturated with expectations and pressures, especially in cases where child contact and custody disputes are on-going and the outcomes are linked to men's completion of a programme and/or when either/both parties hope to maintain or resume their relationship. Such reluctance may, in part, be linked to mistrust of researchers whilst individuals are under the surveillance of agencies which are often the route to recruitment into social research (Fisher 2012). However in Project Mirabal men have been keen to be involved in the in-depth interviews and we quickly met our quota of 60 men. One participant explained that he felt able to take part because the positive image of a father, used on the leaflets given to men, demonstrated to him that Project Mirabal thought of perpetrators differently – as fathers and men capable of positive change. He perceived the interviews as an opportunity to speak openly about his experiences in a non-judgemental forum and help other men.

3.5 A different set of issues arise for researchers interviewing violent men, where the potential exists for them to view the researcher as supporting their rationalisations or justifications of violence against a partner or ex-partner (Hearn et al. 2007). The potential for harm here extends beyond the individual to the other members of their family or household; possibly exacerbating the abuse and danger that women experience in their daily lives. This is particularly the case in longitudinal research that requires multiple contacts with participants over time (Gondolf 2000). This is one of the reasons why men are only involved in the in-depth interviews in Project Mirabal, since these require more reflection and examination of thoughts, feelings and actions.

Empower participants to make choices about taking action to improve their lives

3.6 Stemming from our core value above we believe that research on violence and abuse has a responsibility to support victim-survivors and perpetrators to make changes and choices in their lives. Therefore it is important to develop multiple opportunities for victim-survivors and perpetrators to reflect on their lives and actions throughout the research process. For example, Project Mirabal was embedded in a wider network of support services. Researchers had a list of national and local support services that they could signpost participants to. Due to the closure of many local services as an impact of the 'austerity' measures, and the longitudinal design of the research, some participants were no longer getting support from a specialist domestic violence service. Therefore involvement in Project Mirabal provided opportunities to access alternative local and national support services.

Maximise opportunities for positive experiences and impacts of research for participants, domestic violence services and social justice

3.7 Where women have been asked to reflect on their participation in research – as part of ethical practice – what they tend to regard positively is a research process in which they are able to 'tell their story' and/or explore experience through the lens of the interview focus and questions (see, for example, Campbell et al. 2010). For example, in Project Mirabal one woman reported that she decided to take part in the research to help researchers understand more about the different types of domestic abuse, to demonstrate how domestic violence continues after separation and make people more aware of domestic violence services. She also found taking part in Project Mirabal was a useful way for her to track her own change over time. This suggests that it is not just 'empathy', but the content and design of research instruments which offers potentials to participants, and we spent many months re-drafting and piloting the survey and interview guides so that they not only addressed our research questions, but provided opportunities for reflection, insight and feedback for participants.

3.8 In addition Project Mirabal is unique in that it is a collaborative research project that is grounded in the needs and interests of domestic violence support services. The research was initiated not by academics but by the third sector organisation Respect, the national umbrella organisation for domestic violence perpetrator programmes in the UK. Although the research questions and design have been extensively refined by the research team, the key questions Project Mirabal seeks to answer have very much come from perpetrator programme workers themselves. A sub-group of the advisory group, including chair, Mary MacLeod, CEO of the National Family and Parenting Institute, recruited the researchers. A high level advisory group was then recruited and first fully convened in April 2009. The advisory group is made up of key academics, policy makers and practitioners including Prof Edward Gondolf (University of Indiana), Nicola Harwin (Women's Aid), Dr Lorraine Radford (NSPCC) and Jo Todd (Respect). This group meets regularly throughout the research providing guidance and scrutiny to the research process but with no control over the research design or analysis. As reported by Gondolf et al. (1997) collaborative research design, that involves third sector, policy makers, academics and researchers, maximises opportunities for research to have a positive impact on participants, domestic violence services and wider society.

3.9 As already mentioned above, in the 'positive empowerment' model researchers are invested in both protecting and safeguarding victim-survivors of domestic violence and maximising their capacity for self-determination and autonomy within the research process. We approach women and men as individuals who are able to understand the potential harm, risk and benefits of their participation at numerous points throughout the research process. For us vulnerability and risk is a continuum that can change over the course of a research project. This requires full disclosure of uncertainties, risks and potential benefits – to enable those who choose to take part to manage their participation throughout the research process as active agents. Researchers need to practice 'responsible advocacy' in engaging and recruiting victim-survivors and perpetrators into the research process (Ulrich et al. 2002; Smith 2008). We concur with other feminist researchers in advocating a move towards viewing informed consent as a dynamic, adaptive and situated process (Fontes 2004). We have taken a flexible and situated approach to recruitment that recognises and respects each potential participant's capacity and agency in deciding whether or not to become part of the study. In the rest of this paper we describe the informed consent process used in Project Mirabal, to illustrate ethical practice which requires departures from conventions to ensure the safety and freedom of women and children living with domestic violence.

Informed consent within a 'positive empowerment' approach

Participants and support service involvement

4.1 All participants were receiving services from a specialist domestic violence project. Men had been assessed and accepted onto a community-based DVPP. Consistent with the practices of a DVPP to provide partner support, the partners and ex-partners of men on the DVPP were contacted by a women's support worker and offered support. This support could range from text updates (to confirm the man's attendance at the programme), phone support, a place in a women's group, or regular one-to-one sessions with a women's support worker. Women in the comparison group were involved in a domestic violence support group in areas where there was no community-based DVPP.

Initial contact and the role of support workers

4.2 Men and women were initially approached by a project worker to ask if they were happy for their contact information to be shared with researchers, one of whom would contact them with more detailed information about the project. A 'Consent to Pass Contact Details' form was created for staff members to adapt and use if appropriate in relation to their in-house data protection policies and procedures. This form also asked participants to highlight the best times and days for the research team to call, specify if it was safe to leave voicemail messages and/or send text messages, and indicate whether they currently lived with their violent partner or ex-partner; and asked the women's support workers for further information on risk. In addition the pro forma emphasised that providing contact details did not mean that anyone was then obligated to take part and also highlighted that the decision of whether they then took part or not would not affect the service they received. The timing of when men and women were approached was left to the discretion and professional judgement of support workers. In addition the means by which workers communicated the contact details (i.e. by phone, post or email) was in line with in-house data protection policies and the personal situations of individuals. In this way the risks were minimised, since they were negotiated by workers who knew them in time and place. Research protocols adapted to each agency were developed collaboratively with each research site, enhancing the awareness, credibility and reputation of the study in those agencies (Logan et al. 2008). Rather than a 'one size fits all' we worked with already established working practices, conventions and expertise of each DVPP and women's group, recognising the skills, knowledge and needs of workers. For example, some women support workers were happy to introduce the research to women during an individual or group support session, gather names and phone numbers, and telephone or email the research team to pass on women's contact details immediately after, whilst others preferred to post out the 'Consent to Pass on Contact Details' form, gain a woman's signature and send this form to the research team by post.

The provision of information

4.3 Leaflets and participant information sheets were made available to women and men (https://www.dur.ac.uk/criva/projectmirabal/). These leaflets and information sheets directly invited them to take part in the research project and provided more information in accessible language. They also contained the contact details for the researchers in case anyone wanted to contact researchers directly to ask questions, or bypass their support worker and find out more information themselves. The information sheets clearly described the reasons why the research project was happening, why they had been asked to take part and what they would be asked to do if they agreed to participate. Each step of the research process was detailed including what will happen to the audio-recordings of interviews, assurances of anonymity in transcripts, future reports and publications and an internet link for participants to access the research findings. Attempts were made to highlight the voluntary nature of the research project and independence from the DVPP or support service. The participant information sheet emphasised that men and women did not have to take part and that they could withdraw at a later point. The possible risks and benefits of taking part were outlined (see section on confidentiality below) and information was given about who was conducting and funding the research, in order to make it clear that this was an independent academic research study and therefore not directly associated with the host third sector organisation and/or local authorities. The leaflets and information sheets were posted out or given to men and women during individual and/or group sessions. Women were able to take the leaflets and information sheets home with them if it was safe to do so. In cases where it was not safe for a woman to have this information at home the information was worked through during or after a face-to-face session, or over the phone, and kept at the group premises. In this sense, we balanced REC 'ideals' with the safety issues faced by participants.

First contact phone call

4.4 On receipt of a potential participant's contact details a researcher called the woman or man at a time that had been identified as suitable. In this first contact conversation the information sheets were worked through verbally to ensure the potential participant had understood what the research involved. Researchers also used their professional judgement in taking more time to explain what the research involved in more detail, explore any barriers or anxieties regarding participation and answer questions. In particular women with partners and ex-partners on DVPPs were offered a choice of what they could be involved in – five phone surveys over the course of one year and/or two in-depth qualitative interviews. Researchers were also able to post out information sheets (if safe to do so) and give the potential participant more time to think about their involvement.

The consent form

4.5 An informed consent form (https://www.dur.ac.uk/criva/projectmirabal/) was developed as another check to ensure that the potential participant understood what the research project involved for them. The consent form was completed when the woman or man felt comfortable: either at the first conversation, a later conversation when arrangements were made to meet, or before the first phone survey or in-depth interview. There was no expectation that making an arrangement to be interviewed meant that informed consent had been achieved. For example, one researcher met with a woman to carry out an in-depth interview and during the consent form process the woman stated that she did not want to be recorded and did not consent to take part. The researcher thanked her for her interest, talked about the phone surveys and gave her more time to think about being involved in the research project in a different way. This woman was phoned back and declined to take part in phone surveys, she was reassured that this decision would not affect the service she received at women's support service. Crucially, given the multi-site nature of this study, the consent form did not have to be signed and could be completed verbally over the telephone. Each participant was offered the option of keeping a paper copy of their consent form, although many had safety concerns about this. Emphasis has traditionally been placed on a participant's ownership of the consent form as a useful document to make sure that they know what to expect and what their rights are throughout the research process. This project suggests there at least two ethical problems with this: that for some individuals keeping a copy may place them at risk; and as consent is constantly open to re-negotiation the form ceases to represent the reality of individual situations (Fontes 2004). Consent was re-checked at the beginning of each phone survey and interview interaction.

Consent is given freely and without coercion

4.6 Underpinning Project Mirabal is a commitment to recognise that the safety and freedom of women and children is compromised by abusive household regimes of coercive control. It was, therefore, crucial for researchers on Project Mirabal to avoid using personal influence or professional power to make it difficult for a potential participant to decline involvement. Equally we took great care not to permit men to make the assumption that their involvement might help them to make a positive impression in court proceedings. That some women found the opportunity to help other women and children an important factor in deciding to participate meant managing possibly unrealistic expectations of what the long-term results might achieve. At the same this placed a responsibility on the researchers to think carefully about the ethics involved in how we write up and present the research findings.

4.7 In line with some violence and abuse researchers Project Mirabal decided to compensate women for the time they give and its possible emotional toll (Sullivan & Cain 2004). The quantitative and qualitative components both involved discussion of personal and traumatic life events. The value of the incentive/compensation therefore needed to be high enough to respect women's participation but not too high that it became the main reason women took part. In Project Mirabal women were given £60 in high street shopping vouchers at the end of the phone surveys and £15 of high street shopping vouchers at each qualitative interview session. Women were made aware that vouchers were available for participation during initial conversations and this was noted in the information sheets. To minimise the possibility of this influencing their decision the actual amounts were not mentioned until after individuals had agreed to take part. We took the decision not to offer similar compensations to the men, since it created equity with women that they had not respected within the relationship. For the research team, this abuse of power disqualified them from any benefit, other than the opportunity to think and reflect with us on abuse and violence. Additional internal ethical clearance needed to be gained from two third sector organisations (Barnardos and Cafcass). One of them questioned our practice of not giving vouchers to men for their participation and raised concerns that the value of the vouchers for women could act as a form of coercion into participation. In this case we explained the reason why vouchers were only for women and emphasised that the women would only know the value of the vouchers until after they had consented to take part in the research project. This meant that the information sheets and leaflets for women did not contain the financial value of the vouchers but mentioned that they would be given vouchers to thank them for their time. After these discussions and amendments to fieldwork materials ethical clearance was granted.

Considerations of confidentiality and safeguarding

4.8 One important departure from a complete assurance of confidentiality was the fact that the researchers might discover current risks to women and children. The risk and safety concerns of women and the domestic violence services about the confidentiality of their contact details warranted careful consideration. The contact details and names of all women and men involved in the project were kept confidential and participants were assigned unique codes to be used on all questionnaires and audio-recordings. Hard copies of contact details and signed consent forms were kept in a separate location to the questionnaires and contact details were stored in a secure database on password protected computers. The ethical standpoint of Project Mirabal required researchers to develop a rapport and enable participants to give open accounts of their experiences. However researchers also had to be upfront about their duty to report any serious safety concerns. The following text was included in the information sheets and formed part of the informed consent process.

Whatever you say in this interview is confidential unless you tell the researcher that you or someone else is in immediate danger of serious harm, or the researcher sees or is told about something that is likely to cause serious harm. If that happens, the researcher will raise this with you during the interview and tell you about what could happen if you continue to talk about it and explore how you would prefer to deal with the situation. The researcher will encourage you to seek support from your programme or other local services to help you make the situation safer. If the researcher feels unsure that you will go and get support, they will talk to you about what they need to do and what might happen next. In an extreme case where a child is at serious risk, and you choose not to seek help/advice the researcher has a duty to disclose this to the relevant agencies.

4.9 This text was also part a much more comprehensive safety protocol[2], which went through many iterations in the early stages of the study, on how to manage any information that revealed an intention or threat to harm a partner or ex-partner, child or any other person. Where a researcher had concerns and was anxious that the person might not seek support this was raised with their line manager in conversation, and a decision across a range of options was reached on what to do next. Where women were in regular contact with a support service that used pro-active contact, it was relatively easy to share concerns and ask that they be prioritised for urgent follow up, however there were only two instances when we needed to do this. In the longitudinal phone survey component it was common for researchers to be in contact with women who were no longer receiving support, here it was vital that the project had up to date information on other services in the area, to signpost participants towards these services or to national helplines. Concern for the safety and welfare of participants is consistent with the practice of 'responsible advocacy' that was embedded within the research process (Ulrich et al. 2002; Smith 2008). Here the 'positive empowerment' orientation involved conversations in which concerns were shared with the woman and her right and need for support explored. We sought to be an enabler, to enable women to choose a course of action that might improve her situation.

Conclusion: Recommendations for ethical decision-making in domestic violence research

5.1 This article has discussed the ethical challenges and dilemmas that face domestic violence researchers in a national collaborative project that recruits and engages victim-survivors and perpetrators. These challenges are linked to wider debates and issues that have impacted upon violence and abuse research teams. For instance, the pressure to provide a rigorous evidence base to justify access to diminishing public funds in an era of 'austerity' combined with increased scrutiny of social science research by RECs can restrict both the production of knowledge about experiences of victim-survivors and perpetrators and practical understandings of how services can best work towards the safety and freedom of women and children. Building on critiques by others (Clark & Walker 2011; Mulla & Hlavka 2011) we have discussed how the biomedical paradigm embedded in social science research governance can exist in tension with the values and practices of violence and abuse researchers. In particular we have highlighted how the contractual model of informed consent, categories of 'vulnerable' and 'sensitive' research topics, blanket assurances of confidentiality and ethical protocol resubmission processes can constrain research and crucial considerations of women and children's safety.

5.2 In response to this we have outlined our 'positive empowerment' approach to ethics in violence and abuse research. Our core ethical value focuses on victim-survivors and perpetrators as active agents and stakeholders in research. We advocate recognition of their agency, diversity and capacity and illustrate how violence and abuse research can, for some, represent a positive and empowering opportunity for victim-survivors to enact active citizenship (Smith 2008; Fisher 2012). At the level of research design a 'positive empowerment' approach should include opportunities to empower women and men to make their own choices about taking action to improve their lives, and, maximise the likelihood that the research process will have positive impacts on participants, domestic violence services and wider society. Our ethical procedures were tailored to the realities of enhancing agency in complicated lives, including: verbal consent; negotiating protocols which synchronised with multiple agency policies and procedures; working with consent as an on-going negotiation; and including a clause about the limits of confidentiality in our safety protocol to protect women and children from further serious harm. Our experience in Project Mirabal means we join with those who have called for a move away from rule-based and participant-focused ethical protocols and procedures towards a focus on researchers as 'ethical thinkers' (Clark & Walker 2011) who act with 'responsible advocacy' (Ulrich et al. 2002;Smith 2008). It is crucial that this thoughtful research design needs to be implemented by skilled and experienced researchers (Sullivan & Cain 2004).

5.3 Given the limited resources to assist violence and abuse research teams or RECs, as advice tends to come from informal sharing of protocols and experiences (see, for example, Berry 2009), we have shared an example of how we have put our 'positive empowerment' approach into practice in our informed consent process. In this example we shift away from understanding informed consent as a one-off contractual obligation to view it as a dynamic, adaptive and situated process between the participant and researchers. We offer our experience as a contribution to ongoing discussions amongst researchers and those involved in research governance about the ethical challenges and dilemmas in social research.

Recommendations for RECs: The Project Mirabal Approach
  • Be flexible with contractual ethical requirements: Weigh up the suitability of conventional informed consent procedures (in particular the use of forms and signatures) and blanket assurances of confidentiality with the potential risks this may represent to women and children's safety and freedom. Encourage research projects that proactively engage with ethical dilemmas and take action to safeguard women and children's safety.
  • Acknowledge the active agency of all participants: Reconsider categories of 'vulnerable' social groups and 'sensitive' research topics and the blanket use of these terms. Encourage research projects that acknowledge the agency and diversity of victim-survivors and incorporate opportunities for research participation to be an empowering and positive experience for marginalised groups.
  • Consider the wider ethical responsibilities of social sciences: Weigh up the potential impacts of disallowing research projects on the lives of victim-survivors, perpetrators, children, violence and abuse services and society.
  • Value the experience of Principal Investigators and their research teams: Take previous research projects, professional practice, awards, expert and advisory roles into consideration as evidence of valuable expertise and experience of being an 'ethical thinker' in violence and abuse research.

Funding

This work was supported by the Economic and Social Research Council [grant number ES/H038086] and the Northern Rock Foundation [grant number 20080739].

Acknowledgments

We are grateful for the helpful comments from two anonymous peer reviewers. Research material can be accessed from <http://www.dur.ac.uk/criva/projectmirabal/researchtools>.

Notes

1http://www.lancs.ac.uk/researchethics/4-1-intro.html (Accessed 23 April 2013)

2This document also included a detailed section on the safety and welfare of researchers.

References

ALCOFF, L. & Gray, L. (1993) 'Survivor discourse: Transgression or recuperation?' Signs, 18 p. 260–290.

ANTLE, B. J. & Regehr, C. (2003) 'Beyond Individual Rights and Freedoms: Metaethics in Social Work Research', Social Work, 48(1) p. 135–144.

BERRY, V. (2009) 'Ethical considerations in conducting family violence research', Research Ethics Review, 5(3) p. 91–100.

CAMPBELL, R., Adams, A.E., Wasco, S.M., Ahrens, C.E. & Sefl, T. (2010) '"What has it been like for you to talk with me today?": The impact of participating in interview research on rape survivors', Violence Against Women, 16(1) p. 60–83.

CAVANAGH, K. (2003) 'Understanding women's responses to domestic violence', Qualitative Social Work, 2(3) p. 229–249.

CENTERS FOR DISEASE CONTROL (CDC) (1987) 'Recommendations for prevention of HIV transmission in health-care settings', Morbidity and Mortality Weekly Report, 36(suppl no. 2S)

CLARK, J.J. & Walker, R. (2011) 'Research ethics in victimization studies: Widening the lens', Violence Against Women, 17(12) p. 1489–1508.

COLEMAN, C. (2009) 'Vulnerability as a regulatory category in human subject research', Journal of Law, Medicine and Ethics, 37(1) p.12–18.

CROMER, L.D. & Newman, E. (2011) 'Research ethics in victimization studies: Widening the lens.' Violence Against Women, 17(12) p. 1536–1548.

CROMER, L.D., Freyd, J.J., Binder, A.K., DePrince, A.P. & Becker-Blease, K. (2006) 'What's the risk in asking? Participant reaction to trauma history questions compared with reaction to other personal questions', Ethics & Behavior, 16 p. 347–362.

DEPARTMENT OF EDUCATION (2012) Decision-making Within a Child's Timeframe: An Overview of Current Research Evidence for Family Justice Professionals Concerning Child Development and the Impact of Maltreatment. Child Wellbeing Research Centre. London: Department of Education.

DEPARTMENT OF HEALTH (2005) Research Framework Governance for Health and Social Care (2nd edition). London: Department of Health.

DEPRINCE, A.P. & Chu, A. (2008) 'Perceived benefits in trauma research: Examining methodological and individual difference factors in responses to research participation', Journal of Empirical Research on Human Research Ethics, 3 p. 35–47.

EDWARDS, S.J.L. (2010) 'The role, remit and function of the Research Ethics Committee - 5. Limits to consent?', Research Ethics Review, 6(4) p. 159–163.

EUROPEAN COMMISSION (2010) Feasibility Study to Assess the Possibilities, Opportunities and Needs to Standardise National Legislation on Violence Against Women, Violence Against Children and Sexual-orientation-based Violence. Brussels: European Commission. Daphne Feasibility Study

FERRARO, K.J. (2003) 'The words change, but the melody lingers: The persistence of the battered woman syndrome in criminal cases involving battered women', Violence Against Women, 9(1) p. 110–129.

FISHER, P. (2012) 'Ethics in qualitative research: "Vulnerability", citizenship and human rights', Ethics and Social Welfare, 6(1) p. 2–17.

FONTES, L.A. (2004) 'Ethics in violence against women research: The sensitive, the dangerous, and the overlooked', Ethics & Behavior, 14(2) p. 141–174.

GONDOLF, E.W. (2000) 'Human subject issues in batterer program evaluation', Journal of Aggression, Maltreatment and Trauma, 4(1) p. 273–297.

GONDOLF, E.W. (2012) The Future of Batterer Programs: Reassessing Evidence-based Practice. Boston: Northeastern University Press.

GONDOLF, E.W. & Deemer, C. (2004) 'Phoning logistics in a longitudinal follow-up of batterers and their partners', Journal of Interpersonal Violence, 18(10) p. 1–19.

GONDOLF, E.W., Yllo, K. & Campbell, J. (1997) Collaboration between researchers and advocates. In Kaufman, G.K. & Jasinski, J. (Eds.) Out of Darkness: Contemporary Research Perspectives on Family Violence (p. 255–261). Thousand Oaks, CA: Sage.

GOSTIN, L.O. & Webber, D.W. (1998) 'HIV Infection and AIDS in public health and health care systems: the role of law and litigation', The Journal of the American Medical Association, 279(14) p. 1108–1113

GRINYER, A. (2005) Personal agendas in emotionally demanding research. Social Research Update, 46 <http://sru.soc.surrey.ac.uk/SRU46.html>

HALL, P. & Smith, K. (Eds.)(2011) British Crime Survey: Intimate personal violence splits ample experiment. London: Home Office.

HEARN, J. (Ed.) (1993) Researching Men and Researching Men's Violences. Research Paper No. 4. Research Unit on Violence, Abuse and Gender Relations, University of Bradford.

HEARN, J. (1998) The Violences of Men: How Men Talk About and How Agencies Respond to Men's Violence to Women. London: Sage.

HEARN, J., Andersson, K. & Cowburn, M. (2007) Background Paper on Guidelines for Researchers on Doing Research with Perpetrators of Sexual Violence. Sexual Violence Research Initiative, Pretoria, South Africa.

HOLTZWORTH-MUNROE, A., Meehan, J.C., Herron, K., Rehman, U. and Stuart, G. L. (2003) 'Do subtypes of maritally violent men continue to differ over time?' Journal of Consulting and Clinical Psychology, 71(4) p. 728–740.

JOHNSON, B. & Clarke, J.M. (2003) 'Collecting sensitive data: The impact on researchers', Qualitative Health Research, 13(3) p. 421–434.

JONES, A. (2000) Next Time, She'll Be Dead. Boston, MA: Beacon Press.

KELLY, L. (1988) Surviving Sexual Violence. London: Polity Press.

KELLY, L., Burton, S. & Regan, L. (1993) 'Beyond victim to survivor: the implications of knowledge about children's resistance and avoidance strategies', in Ferguson, H., Gilligan, R. & Torode, R. (Eds.), Surviving Childhood Adversity: Issues for Policy and Practice. Dublin: Social Studies Press.

KENT, J., Williamson, E., Goodenough, T. & Ashcroft, R. (2002) Social science gets the ethics treatment: research governance and ethical review, Sociological Research Online, 7(4) < http://www.socresonline.org.uk/7/4/williamson.html>

LAMB, S. (1999) (Ed.) New Versions of Victims: Feminists Struggle with the Concept. New York: New York University Press.

LEE, R.M. & Renzetti, C.M. (1990) 'The problems of researching sensitive topics: An overview and introduction', American Behavioral Scientist, 33(5) p. 510–528.

LOGAN, T.K., Walker, R., Shannon, L. & Cole, J. (2008) 'Combining ethical considerations with recruitment and follow-up strategies for partner violence victimization research', Violence Against Women, 14(11) p. 1226–1251.

MCCOSKER, H., Barnard, A. & Gerber, R. (2001) 'Undertaking sensitive research: Issues and strategies for meeting the safety needs of all participants', Forum Qualitative Social Research, 2(1) Art. 22 February.

MCNUTT, L., Waltermaurer, E., Bednarczyk, R.A., Carlson, B.E., Kotval, J., McCauley, J., Campbell, J.C. & Ford, D.E. (2008) 'Are we misjudging how well informed consent forms are read?' Journal of Empirical Research on Human Research Ethics, 3(1) p. 89–97.

MORRIS, A. (2009) 'Gendered dynamics of abuse and violence in families: Considering the Abusive household gender regime', Child Abuse Review, 18(6) p. 414–427.

MULLA, S.A. & Hlavka, H.R. (2011) 'Gendered violence and the ethics of social science research', Violence Against Women, 17(12) p. 1509–1520.

NEWMAN, E., Walker, E.A., & Gefland, A. (1999) 'Assessing the ethical costs and benefits of trauma-focused research', General Hospital Psychiatry, 21 p. 187–196.

ONS (2013) Violent crime and sexual offences 2011/2012. < http://www.ons.gov.uk/ons/dcp171778_298904.pdf> [accessed April 23 2013]

PELED, E. (2000) 'Parenting by men who abuse women: Issues and dilemmas', British Journal of Social Work, 30(1) p. 25–36.

RADFORD, L. & Hester, M. (2006) Mothering Through Domestic Violence. London: Kingsley.

RIESSMAN, C.K. (2005) 'Exporting ethics: A narrative about narrative research in South India', Health, 9(4) p. 473–490.

ROMITO, P. (2008) A Deafening Silence: Hidden Violence Against Women and Children. Bristol: Policy Press.

SCHRAG, Z. M. (2011) 'The case against ethics review in the Social Sciences', Research Ethics, 7(4) p. 120–131.

SIEBER, J.E. & Stanley, B (1988) 'Ethical and professional dimensions of socially sensitive research', American Psychologist, 42 p. 735–741.

SMITH, L.J. (2008) 'How ethical is ethical research? Recruiting marginalized, vulnerable groups into health services research', Journal of Advanced Nursing, 62(2) p. 248–257.

STARK, E. (2007) Coercive Control: How Men Entrap Women in Personal Life. Oxford: Oxford University Press.

STUBBS, J. & Tolmi, J (1999) 'Falling short of the challenge? A comparative assessment of the Australian use of expert evidence on the battered woman syndrome', Melbourne University Law Review, 23(3) p. 709–748

SULLIVAN, C. M. and Cain, D. (2004) 'Ethical and safety considerations when obtaining information from or about battered women for research purposes', Journal of Interpersonal Violence, 19(5) p. 603–618.

SULLIVAN, C.M. (2011) 'Evaluating domestic violence support service programs: Waste of time, necessary evil, or opportunity for growth?' Aggression and Violent Behavior, 16 p. 354–360.

ULRICH, C., Wallen, G. & Grady C. (2002) 'Research vulnerability and patient advocacy: Balance-seeking perspectives for the clinical nurse scientist? Editorial', Nursing Research, 51(2) p. 71–72.

WALBY, S. & Towers, J. (2012) Measuring the Impact of Cuts in Public Expenditure on the Provision of Services to Prevent Violence Against Women and Girls. Safe: The Domestic Abuse Quarterly, 41: 14-17.

WALKER, L. (2000) The Battered Woman Syndrome. New York: Springer.

WESTMARLAND, N., Kelly, L. & Chalder-Mills, J. (2010) What Counts as Success? London: Respect.

WESTMARLAND, N. & Kelly, L. (2012) 'Why extending measurements of "success" in domestic violence perpetrator programmes matters for social work', British Journal of Social Work Advance Access < http://bjsw.oxfordjournals.org/content/early/2013/04/18/bjsw.bcs049>

WILLIAMSON, E. (2010) 'Living in the world of the domestic violence perpetrator: Negotiating the unreality of coercive control', Violence Against Women, 16(12) p. 1412–1423.

WHO (2001) Putting Women First: Ethical and Safety Recommendations for Research on Domestic Violence Against Women. Geneva, Switzerland: World Health Organisation.

WMA (2008) Declaration of Helsinki – Ethical principles for medical research involving human subjects. World Medical Association < http://www.wma.net/en/30publications/10policies/b3/index.html>