Some interventions have been implemented to address the risk of family violence during times when women are most at risk. For example, perinatal home visiting programs that include specific intimate partner violence measures. Some have been successful while others have shown sub-optimal effects. Conducting family violence studies entails complex ethical dilemmas, with potential consequences for the data collection, analysis, and findings. This paper summarises empirical evidence on such interventions and discusses key ethical dilemmas in engaging victim survivors in evaluation of family violence programs. While there is no recipe for addressing the inherent ethical dilemmas, a checklist is proposed here as a roadmap for researchers.
Family violence is a public health issue, affecting 1 in 4 families, with children involved in over 30% of cases reported to police [1]. The time of elevated risk is during pregnancy and post-partum [2,3], often from an intimate partner [4]. In a personal safety survey, 25% of those who experienced family violence during pregnancy by a previous partner reported that the violence first occurred during pregnancy. Depression, anxiety, self-harm, cognitive and behavioural difficulties are some of the effects on children and young people who experience family violence [3]. The potential for underreporting of family violence incidents shows that the problem is more prevalent than captured in official statistics [5].
Some interventions have been implemented to address the risk of family violence during times when women are most at risk. For example, perinatal home visiting programs that include specific intimate partner violence (IPV) measures. Some have been successful while others have shown sub-optimal effects [6], warranting a need for continued rigorous investigations. This paper summarises empirical evidence on such interventions and discusses key ethical dilemmas in engaging victim survivors in evaluation of family violence programs and possible solutions.
A randomised controlled trial [7] of a nurse-family partnership home visitation program, in Holland, found that at 32 weeks of pregnancy, women in the intervention group (n=237) self-reported significantly less IPV than women in the control group (n=223) in level 2 psychological aggression (control: 56% versus intervention: 39%), physical assault level 1 (control: 58% versus intervention: 40%) and level 2 (control: 31% versus intervention: 20%), and level 1 sexual coercion (control: 16% versus intervention: 8%). Two years post-birth, IPV was significantly lower in the intervention group for level 1 physical assault (control: 44% versus intervention: 26%), and IPV perpetration was significantly lower for level 1 sexual assault (control: 18% versus intervention: 3%). Multilevel analyses revealed a significant improvement in IPV and perpetration among women in the intervention group at 24 months after birth. Women in the control group received usual care while those in the intervention group had usual care plus nurse home visits periodically during pregnancy and until the child's second birthday.
In a randomised controlled trial [8] involving 643 families with an infant at high risk for child maltreatment in Hawaii, mothers in the intervention group reported lower IPV rates (incidence rate ratio [IRR], 0.86; 95% CI 0.73-1.01), significantly lower rates of perpetration (IRR, 0.82; 95% CI, 0.70-0.96), and significantly lower rates of physical assault.
A randomised trial [9] investigating the effectiveness of an IPV intervention incorporated into a nurse family partnership program found that the intervention on participants differed depending on their baseline experience with IPV. For instance, for physical violence, the intervention reduced IPV at 1 year but only among women who had not experienced past-year physical victimisation at baseline. For sexual violence, women in the intervention group were more likely to report sexual violence at 2-year follow-up, but only among participants who had reported sexual victimisation at baseline.
Using cluster-based, single-blind, randomised clinical trial at 15 sites in 8 United States’ jurisdictions (enrolling 492 socially disadvantaged pregnant women preparing to parent for the first time), Jack, et al. [10] found that augmentation of a nurse home visitation program with a comprehensive IPV intervention, compared with the home visitation program alone, did not significantly improve quality of life at 24 months after delivery. From baseline to 24 months, quality of life improved in both groups (change in World Health Organisation Quality of Life Scale scores from 299.5 [SD, 54.4] to 308.2 [SD, 52.6] in the augmented program group versus from 293.6 [SD, 56.4] to 316.4 [SD, 57.5] in the standard program group. Multilevel growth curve analysis showed there was no statistically significant difference between groups (modelled score difference, -4.9 [95% CI, -16.5 to 6.7]).
In a comparison [11] of the efficacy of Child-Parent Psychotherapy (CPP) for preschool-age children exposed to family violence, involving 75 multi-ethnic preschool mother dyads from diverse socioeconomic backgrounds, children randomly assigned to CPP improved significantly more than children receiving case management plus treatment as usual in the community, both in decreased total behaviour problems and reduced post-traumatic stress disorder (PTSD) symptoms. The treatment group was also significantly less likely to be diagnosed with PTSD. Mothers receiving CPP showed significantly fewer PTSD avoidance symptoms at the end of treatment than comparison group mothers. The results were attributed to CPP’s focus on fostering child mental health by promoting a relational process in which increased maternal responsiveness to the child’s developmental needs strengthens the child’s trust in the mother’s capacity to provide protective care.
CPP is a model of care developed in the US for mother-child dyads aimed at enhancing relationships and minimising trauma. Australia’s National Research Organisation for Women’s Safety Limited is conducting a study on whether the CPP model of care is acceptable to clients and feasible to implement in Australia. With funding from the Safer Families Centre for Research Excellence, the ‘RECOVER study’ will recruit mother-child dyads from regional and rural Victoria and South Australia.
In a case study approach across six hospital antenatal clinics in two states in Australia, New South Wales and Victoria, Hegarty, et al. [12] have explored the complex subject of addressing family violence in antenatal care from multiple perspectives, using an existing Health Systems Implementation Model from the Women’s Input into a Trauma-informed systems model of care in Health settings study. Five of the study sites are in the bottom 40% of disadvantaged communities. At two Victorian sites, a survey completed by 1,219 women showed that family violence is prevalent among pregnant women in antenatal services, and the experience is generally unknown by health providers. Practitioners agreed that the role of screening for domestic violence best fitted with midwives who have an initial role in risk assessment and management, and that social workers are best placed to provide a comprehensive response.
Berry [13] has contended that researchers studying family violence face some of the most complex ethical dilemmas. This is partly because the drive to better understand the causes, processes and consequences of family violence needs to be balanced against the sensitive and potentially distressing nature of the subject for those involved. Often, researchers who share characteristics with the participants may assume that their experiences are alike and that they understand each other and that the participants also perceive them as similar, and therefore trust them.
As Daly [14] notes, "When researchers and the participants operate from shared realities, there may be a tendency to take too much for granted". This has potential consequences for the data collection. First, researchers may overlook certain aspects of participants' realities, risking the loss of certain details that might be important. Second, respondents may withhold information they perceive as too obvious considering the shared reality with the researcher. This dynamic is most obvious in the intimate setting of an in-depth interview, warranting researchers to ask themselves: "What kind of relationship do I have with the participants? What kind of relationship do I want to construct with them through the research?"
Fontes [15] suggests a diverse research team that can harness the richness of differences and similarities between researchers and participants. According to Fontes [15], most research on family violence is ethnocentric: framed according to the dominant culture's views of families, normalcy, violence, trauma, disclosure, and privacy. This limits a researcher's ability to understand the world from the perspectives of participants from disadvantaged populations, making it important for researchers to question their own assumptions and state them openly, particularly as participant observation is almost impossible in family violence.
Based on studies conducted on the Maori in New Zealand, Smith [16] has proposed some questions to consider in family violence research where victims are from diverse sociodemographic groups: Who has helped define the research problem? For whom is this study worthy and relevant? Who says so? Which cultural group will gain new knowledge from this study? To whom is the researcher accountable? Who will gain the most from this study? Answering these questions can enable family violence researchers to design studies based on collaboration. This way the resulting outputs are likely to be more relevant and useful.
Given the violence under study includes a violation of a trusting relationship, the process of obtaining informed consent becomes particularly critical in studies of family violence. The goal of the informed consent is to provide sufficient information to a potential participant in a language that is easily understood, so that they can make the voluntary decision regarding “to” or “not to” participate in the research [17]. While the actual forms used to obtain consent must be clear and jargon-free, it can be helpful to read consent forms in their entirety to participants, especially in settings of low literacy levels or poor fluency in the written than the spoken language [15]. Researchers may be satisfied to obtain nearly a 100% response rate in studies, only to discover that participants have been less than candid in their responses. That is, they felt obligated to participate, but quietly guarded their right to withhold information [15].
Another issue worth considering is what Hardy [18] calls "contemporary theoretical myth of sameness", that is, the risk of exaggerating the differences among groups while minimising their similarities when researchers pose questions in terms of group differences. Sometimes, the differences within each group may be more significant conceptually than the variations between the groups.
Researchers’ familiarity with the culture of study participants can help minimise errors of misunderstanding in all phases of the research, from using alienating instruments to misinterpreting the results to disseminating results in a way that harms the participants. Further, it can be useful to incorporate into the research team people from the cultures. The more people from the culture or cultures being studied are included in the planning, implementation, interpretation, and dissemination of the research the greater the probability that the research will be ethical [15].
Frequently, research participants are asked to use instruments or engage in procedures that are unfamiliar to them. Examples include a true-false format or a Likert scale. Conventional empirical approaches employing standardized quantitative methods may not always be sufficient for culturally sensitive research on family violence. Research methods that account for cultural contexts of the participants is critical. The better the fit, the more accurate the findings are likely to be.
Results of a qualitative study [19] that asked family violence survivors what interviewers should know about rape and how they should interact with participants revealed that interviewers need to show warmth and compassion, allowing participants to exercise choice and control during the interview process. The study found that interviewers’ attitudes and interpersonal skills have influence in participants’ willingness to disclose violence. Based on the results, a set of criteria for selecting interviewers was developed: being able to engage with people of different backgrounds in an empathetic and non-judgmental manner, emotional maturity, skills at building rapport, and ability to deal with sensitive issues.
The potential for research to be emotionally upsetting can arise from other sources other than interviews; errors of omission can feel as upsetting as errors of commission. Fontes [15] has noted instances where victims of rape and sexual abuse described being upset by survey instruments on sexuality that do not give them an opportunity to tell that an experience was an assault; they felt that such surveys forced them to misrepresent themselves as promiscuous or sexually precocious.
To avoid harming participants or their communities, Berry [13] and Morrow and Richards [20] suggest that research teams brainstorm the potential risks of the planned study in a range of categories (emotional, physical, social, and political), and plan ways to reduce the potential for risk in each sphere. Also helpful is enlisting the expertise of the community that will be studied by conducting focus groups in which the research is described, and participants are asked for input on ways to make the study more valid, more beneficial, and less harmful to the participants.
As Fontes [15] notes, it is unethical to conduct research on members of the dominant group only, and then apply the findings to members of all groups. A better approach is to conduct a variety of cross-culture and within-culture studies to learn about violence and identify protective factors.
Researchers such as Berry [13], Ramsey [21], Morrow and Richards [20] have observed that it is never ethical to use children in research where there is no direct benefit for them because they could never properly consent. This then raises the issue of whether parents, as the guardian giving consent for a child to participate in research, can transfer their altruism to their children? The dilemma is that several incidents, including family violence, manifest themselves differently in children compared to adults and, further, that interventions affect children differently [13].
King and Churchill [22] have identified certain ethical principles for research with children or adolescents: scientific soundness; sufficient importance; respect for autonomy; beneficence and non-maleficence; utility; and justice. The principle of sufficient importance stipulates that the research should ask and provide answers to ‘questions important to the welfare of children – or hold substantial promise of benefit to children’. Where research poses more than minimal risk, it is important for the researchers to demonstrate either that the child as a research subject will benefit directly or that the results will substantially further the understanding, treatment, or prevention of a problem. Beneficence and non-maleficence refer to the obligations of acting in a manner that benefits a child and refraining from harm.
Cater and Overlien [23] have identified three ethical dilemmas in research involving children exposed to family violence. The first dilemma concerns whether an individual research project is ethically justified at all. The second concerns the issue of consent. The third dilemma relates to how to handle confidentiality in relation to the risk of unsought disclosure of child maltreatment. Researchers investigating children’s exposure to family violence must always consider the benefits and disbenefits of requiring consent of the child, the victimized parent, and the perpetrating parent. Adults and children may understand and define situations differently and have different reasons for consenting or not.
Cater and Overlien [23] contend that relying on participatory models from disciplines that do not focus on children in vulnerable life situations limits the development of participatory methods in social work research. The authors suggest that research involving children in vulnerable situations requires that special attention be given to research ethics: “researchers must beware of adopting a focus on ‘obtaining ethical clearance’ or ‘child access’, and instead take full responsibility for children’s participation and protection”.
In collaboration with experts in gender-based violence, Innovators for Poverty Action [24], a United States non-for-profit research and policy organisation, has developed a checklist of items for researchers to consider before conducting family violence-associated data collection. Covering the topics of study design and preparation, piloting and training, administering surveys, data storage and security, the checklist includes: communicate study to community framed as a general topic to the non-participants in the community (for example, a survey on intimate partner violence is framed as a study on women’s health); train interviewers to properly ask sensitive questions about violence and how to respond in the case of distress; interviewers are trained to recognise and deal with a respondent’s distress during the interview; informed consent is obtained in all cases; interviews are conducted in a private setting; men and women within the same household are not asked about experience of violence for intimate partner violence research; interviewers and participants are gender-matched; interviewers change questions to non-sensitive subjects if the survey is interrupted; and end the interview on a positive note that emphasises a woman’s strengths [25].
As recognised at the first international statement on research ethics, World Medical Association Declaration of Helsinki 1964 (Article 5), the subject’s welfare must always take precedence over the interests of science and society. Hence, the primary ethical issue is to weigh the value of the possible knowledge gain of a study against its possible harm to the participants. While there is no recipe for addressing the inherent ethical dilemmas, a checklist is proposed here, based on the foregoing discussion, as a roadmap for researchers.
View Supplementary Data
- Royal Commission into Family Violence (2016) Report and Recommendations, Volume 1, March, Melbourne, Australia.
- World Health Organisation (2005) WHO Multi-country Study on Women’s Health and Domestic Violence Against Women. Geneva: WHO.
- Gartland D, Woodhouse H, Mensah FK, Hegarty K, Hiscock H, et al. (2014) The case for early intervention to reduce the impact of intimate partner abuse on child outcomes: Results of an Australian cohort of first-time mothers. Birth 41: 374-383. [Crossref]
- https://www.abs.gov.au/ausstats/abs@.nsf/Lookup/by%20Subject/4533.0~2018~Main%20Features~Personal%20Safety%20Survey~24
- Phillips J, Vandenbroek P (2014) Domestic and family violence in Australia: An overview of the issues. Parliamentary Library Research Paper. Canberra: Parliament of Australia.
- Sharps PW, Campbell J, Baty ML, Walker KS, Blair-Merritt MH (2008) Current evidence on perinatal home visiting and intimate partner violence. J Obstet Gynecol Neonatal Nurs 37: 480-490. [Crossref]
- Mejdoubi J, Silvia CC, van Leerdam FJ, Heymans MH, Hiraising RA, et al. (2013) Effect of nurse home visits vs. usual care on reducing intimate partner violence in young high-risk pregnant women: a randomised controlled trial. PLoS One 8: e78185. [Crossref]
- Bair-Merritt MH, Jennings JM, Chen R, Burrell L, McFarlane E, et al. (2010) Reducing maternal intimate partner violence after birth of a child: a randomised controlled trial of Hawaii Health Start Home Visitation Program. Arch Pediatr Adolesc Med 164: 16-23. [Crossref]
- Feder L, Niolon PH, Campbell J, Whitaker DJ, Brown J, et al. (2018) An intimate partner violence prevention intervention in a nurse home visitation program: A randomised clinical trial. J Womens Health (Larchmt) 27: 1482-1490. [Crossref]
- Jack SM, McKee C, Ford-Gibboe M, Wathen CN, Scribano P, et al. (2019) Effect of addition of an intimate partner violence intervention to a nurse home visitation program on maternal quality of life: A randomised clinical trial. JAMA 321: 1576-1585. [Crossref]
- Lieberman A, Van Horn P, Ippen CG (2005) Toward evidence-based treatment: Child-parent psychotherapy with pre-schoolers exposed to marital violence. J Am Acad Child Adolesc Psychiatry 44: 1241-1248. [Crossref]
- Hegarty K (2020) Sustainability and identification of response to domestic violence in antenatal care. https://www.anrows.org.au/publication/sustainability-of-identification-and-response-to-domestic-violence-in-antenatal-care/
- Berry V (2009) Ethical considerations in conducting family violence research: Research Ethics Review 5: 91-100.
- Daly K (1992) Parenthood as problematic: Insider interviews with couples seeking to adopt. In Gilgun JF, Daly K, Handel G (Eds), Qualitative methods in family research, Newbury Park: Sage, pp: 103-125.
- Fontes LA (1998) Ethics in family violence research: Cross-cultural studies. Family Relations No 1: 53-61.
- Smith LT (1992) Te raapunga I te ao maarama (The search for the world of light). In Hohepa MK, Smith GH (Eds) The issue of research and Maori Monograph # 9. Auckland: Research Unit for Maori Education, University of Auckland pp: 1-12.
- Nijhawan LP, Janodia MD, Muddukrishna BS, Bhat KM, Bairy KL, et al. (2013) Informed consent: Issues and challenges. J Adv Pharm Technol Res 4: 134-140, [Crossref]
- Hardy KV (1993) Implications for practice with ethnic minority families. In Boss PG, Doherty WJ, LaRossa R, Shumm WR, Steinmetz SK (Eds), Sourcebook of family theories and methods, New York: Plenuum, pp. 646-649.
- Fraga S (2016) Methodological and ethical challenges in violence research. Porto Biomedical Journal 1: 77-80. [Crossref]
- Morrow V, Richards M (1996) The ethics of social research with children: an overview. Children & Society 10: 90-105.
- Ramsey P (1970) Consent as a canon of loyalty with special reference to children in medical investigation. In Ramsey P (Ed) The patient as a person. New Haven, CT: Yale University Press pp: 11-12.
- King NMP, Churchill LR (2000) Ethical principles guiding research on child and adolescent subjects. Journal of Interpersonal Violence 15: 710-724.
- Cater A, Overlien C (2014) Children exposed to domestic violence: A discussion about research ethics and researchers’ responsibilities. Nordic Social Work Research 4: 67-79.
- https://www.poverty-action.org/sites/default/files/publications/IPV_Ethical%20Guidance_Public.pdf
- Ellsberg M, Heise L (2005) Researching Violence Against Women: A Practical Guide for Researchers and Activists. Washington DC, United States: World Health Organization, PATH.
