Methodological and ethical challenges in violence research

Abstract

Violence is a relevant public health issue. It is recognized as a sensitive topic to research and introduces challenges not usually found when dealing with other research topics. Researchers face a major challenge that is how to identify and measure violence as it occurs in the general population. This paper intends to discuss and raise awareness to some of the main methodological and ethical challenges related to violence research.

Introduction

Violence is an extremely diffuse phenomenon and therefore its definition is also complex to operationalize. Notions of what is acceptable and unacceptable in terms of behaviours and what constitutes harm are culturally influenced and constantly under review as values and social norms evolve.¹ Therefore, the definition of violence is not the result of an exact scientific approach but probably a matter of judgement. The wide diversity of moral codes throughout the world, makes the topic of violence one of the most challenging and politically sensitive to address. However, an effort must be made to reach consensus and set universal standards of behaviour based on human rights in order to protect human life and dignity in our fast-changing world.

As far as public health is concerned, the challenge is to operationalize violence in such a way that it captures the range of perpetrated acts and the subjective experiences of victims, without becoming so broad that it loses meaning. In order to find a global consensus that would allow comparisons between countries, the World Report on Violence and Health defined violence as the intentional use of physical force or power, threatened or actual, against oneself, another person, or against a group or community, that either results in or has high likelihood of resulting in injury, death, psychological harm, maldevelopment or deprivation.¹

This definition covers a diversity of acts, going beyond physical acts to include threats and intimidation, whether they are public or private and whether they are reactive or proactive. However, the complexity and the variety of violence behaviours require an analytical framework to emphasize the common features and linkages between different types of violence, leading to a holistic approach of violence. For that reason, the typology of violence proposed by the World Health Organization includes three broad categories according to the characteristics of those involved in the violent act: self-directed violence; interpersonal violence and collective violence with intention being the common ground.¹ In brief, this categorization differentiates between violence that a person inflicts upon themselves, violence inflicted by another individual or a small group of individuals, and violence inflicted by larger groups such as organized political groups, militia groups and terrorist organizations.¹

Thus, the typology proposed by the World Health Organization¹ provides a useful framework to understand the complex definition of violence and provides some clues for its study. It highlights, for instance, that interpersonal violence may take various forms and occurs most often among known and close people. However, the assessment of the magnitude of interpersonal violence in population-based studies comprises some challenges to researchers. This paper intends to discuss and raise awareness to some of the main methodological and ethical challenges related to violence research.

Assessment of interpersonal violence

Violence is recognized as a sensitive topic to research and a multifaceted problem that introduces challenges not usually found when dealing with other topics of social or health studies. Although there are systems for monitoring non-fatal violent injuries, these are typically restricted to violent injuries presenting to hospital emergency departments,² or to individuals’ reports to authorities or support services. These systems do not detect unreported violence. For that reason, researchers face a major challenge that is how to identify and measure violence as it occurs in the population.

Most of this violence is not possible to measure objectively without asking those involved, directly or indirectly. The evolution of research methods, specific measures and thorough ethical reflections have contributed to the establishment of violence as a global issue, although many challenges remain in the measurement of its scope and nature.³ Therefore, the process of research on violence still raises specific methodological and ethical challenges. Given the stigmatizing nature of violence, over reporting is not common, and thus we expect that prevalence estimates tend to underestimate the true magnitude of interpersonal violence. Nevertheless, the World Health Organization specifically suggested undertaking epidemiological research on these issues; with the major focus being the collection of rigorously sound and internationally comparable quantitative data on interpersonal violence which is the first step taken when adopting a public health approach.

In the research process, data on violence experiences depends on the availability and willingness of the individual to recognize their involvement in a violent situation, which may depend, among other factors, on culture, on how the questions are asked, interviewer training, sensitivity and the setting in which data are collected. All of these factors need to be accounted for and respected by researchers aiming to assess violence.

In countries with strong cultural pressure to keep violence behind closed doors or simply to accept it as natural, non-fatal violence is likely to be underreported. Individuals may be reluctant to discuss violent experiences because of shame, taboos or even fear. Admitting some abusive experiences such as rape may in some countries result in death.¹ In some cultures, the preservation of family honour is a traditional motive for killing women who have been raped.

In epidemiologic research, numerous factors have been shown to influence participation rates and response quality.⁴ The methods for contacting eligible participants, the modes of questionnaire administration and the interviewer characteristics are likely to influence research results.^5–7 Thus, when planning the collection of information on abuse, researchers face critical design decisions that include options on the sampling frame, the structure and the mode of questionnaire administration and also the setting where information collection takes place. There are several recommendations developed in order to assess interpersonal violence, in particular, using structured questionnaires that enable greater consistency in the way questions are asked, the training of interviewers, the development of a research protocol to guide interviewers in problem solving and supervision during data collection.⁸

A systematic review of the instruments used in the assessment of domestic violence, including violence in intimate relationships, shows that researchers tend to choose the instrument according to the method and setting of administration. The variability of instruments used in the evaluation of violence implies additional difficulties in collecting data on the extent of this phenomenon and the respective comparison between results from different studies.⁹ However, most researchers are in agreement that direct questioning about experiences of specific acts of violence over a particular period of time should be used rather than using more open-ended and generic questions.

Regarding the context that frames the conduct of the interview, researchers often opt for the interviews to take place at the participant's home, in order to improve participation rate, as they do not have to move out of an environment that is familiar to them.⁸ However, home interviewing involves greater costs, the possibility of interruptions by telephones or family members, and may put the interviewer, and eventually the respondent, in a situation of greater vulnerability, especially when the abuser lives with the victim.¹⁰ These concerns are especially relevant when addressing violence but few methodological information is available reporting the effect of place of interview on participation and response rates.^11,12 A previous study showed that the interview setting has no influence both in participation rate and in the prevalence estimates of different types of violence in the elderly, at least when describing the reality of a social context similar to the ones found in southern European countries, like Portugal.¹²

The mode of questionnaire administration has been described to influence participation and also disclosure of abuse.⁵ It is possible to choose between face-to-face interview, post mail/self-administered questionnaires or telephone questionnaires, or using some combination of these.¹³ The self-administered questionnaire generally has more advantages than face to face administration or by telephone.⁸ However, the absence of responses tends to be higher in self-administered questionnaires and the participant does not have the opportunity to ask for clarification. Moreover, if the self-administered questionnaire is mailed, it is more likely to get a low response rate.

The least burdensome method is probably the personal face-to-face interview as this only requires the respondent to speak the same language in which the questions are asked, and to have basic verbal and listening skills.⁵ Compared to other modes of data collection, questionnaires administered by interviewers have the advantage of reducing missing items and there is the possibility of helping participants to better understand the items.¹⁴ However, this is a more expensive option and interviewer-interviewee interaction potentiates the effect of the interviewer on the results^8,15–17 which worsens in the evaluation of sensitive topics¹⁸ such as violence. In general, the performance of each method of collecting information depends on the context in which it is administered.

Also, the attributes of the interviewers may affect the participants’ disclosure. There are some recommendations for the recruitment process of interviewers in violence research.^10,19–23 Gender may be one of the most identifiable interviewer characteristics and it is likely that respondents invoke gender-based stereotypes when editing their responses.¹⁸ All of the guidelines for violence research recommend the recruitment of female interviewers²¹ claiming that female interviewers are more likely than males to rate respondents as frank and honest.¹⁸ In fact, some studies show differences in results when interviews were performed by female or male interviewers.²⁴ However, it is still unclear whether there are gender differences in the validity of data collected. Further, the influence in participants’ disclosure of interviewers’ attributes such as gender, personality traits or attitudes is even less known.

Regardless of their attributes, to improve data quality on violence, special attention should be given to interviewer training. All research requires substantial investment in interviewer training in order to provide a common questioning frame and similar strategies to handle unusual or unexpected circumstances during the interviewing process.⁸ Violence research requires particular attention to this process. A two-stage approach is advisable for preparing interviewers to work in violence research. First, interviewers must learn about interpersonal violence and training must include consciousness raising about the topic, causes of violence, myths and facts, diversity and cultural sensitivity, crisis intervention, skills, safety planning, and community resources and supports.^21,22,25 If interviewers are familiar with these topics, they will be better prepared to handle unanticipated situations. Similarly, discussing violence issues during training would increase sensitivity to violence.²⁶ The second stage involves teaching how to conduct the interview. At this stage, interviewers may work through a successive series of practice exercises, including watching model interviews and conducting mock interviews with other team members.²² After this two-stage approach, some of the selected interviewers would not be prepared for fieldwork and therefore the researcher would have to decide who they would be entrusting with this responsibility.²⁵

In the multinational World Health Organization (WHO) study on domestic violence, team-members invested considerable effort training local, lay community women to be interviewers.²¹ However, because of time constraints, they also brought in a group of professional interviewers to help complete data collection. When they compared information collected from these two groups of interviewers, they found that the carefully trained community interviewers obtained higher response rates and disclosure rates for physical and sexual abuse than the professional interviewers did.²¹ Thus, it is recommended that extensive training occurs when implementing successful field projects regardless of the interviewer's previous experience, and also, comparisons of interviewer's personal characteristics may help uncovering the known variability of violence rates across cultures.

In a qualitative study, violence survivors were asked what interviewers should know about rape and how they should interact with participants.²⁵ Results showed that interviewers need to show warmth and compassion, allowing participants to exercise choice and control during the interview process. In fact, it was observed that attitudes and interpersonal skills of interviewers have influence in participants’ willingness to disclose violence.^10,23 A set of criteria for selection of 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.

It is common to incorporate quality control measures into the design of epidemiologic studies to minimize interviewer effects but few researchers report which measures they use, examine the data for interviewer variation or explore the impact of such variation on study findings.²⁷ Even if standard strategies to minimize interviewer effects are incorporated into the study protocol, studies addressing sensitive topics such as those that concern intimate personal behaviours may remain especially prone to interviewer effects.¹⁸ These effects may be consequently attributable to characteristics of the interviewer or the respondent, as well as to interactions between them.

However, some studies have shown that measures taken before and during the investigation to minimize the error introduced by the interviewer does not seem to be sufficient, especially when the topic requires disclosure of more sensitive information.^27–29 In fact, there are aspects that the researcher cannot control at all, such as the personality and attitudes of the interviewer that influence the relationship and empathy between interviewer and interviewee, and as such, the disclosure of information by the respondent.

Arising from this interaction, a specific issue within violence research must be acknowledged: violence is a public crime in some countries, and it is therefore mandatory to report. This may cause some resistance for participants accepting to collaborate in the study and/or depict their experience, and the interviewer has to be prepared to deal with self-reported situations of violence. In the research protocol, particularly when violence is being addressed, the procedure must include guidelines with frequent indication on how the interviewer should advise/inform the participant in order to report the situation and seek help. Generally, the interviewer should know how to deal with such situations calmly, listen to the participant, inform and refer them appropriately.

Ethical issues

Regulatory frameworks, guidelines and guidance for ethical procedures on scientific research have been constructed worldwide over the last decades,^30,31 pointing to four main issues. First, before embarking on any inquiry, researchers must ensure that the information gathering activity is necessary and justified. Second, the benefits of a particular study must be weighed against its risks, both to respondents and to communities, and therefore researchers and ethics boards have the obligation to take every precaution to minimize harm and maximize benefits. Accordingly, researchers should state their engagement in informative and mutually respectful interactions and explain the benefits to those individuals participating in the study when submitting research proposals. Such conduct grounds on the principle of distributive justice, according to which individuals bearing the burden of research should receive an appropriate benefit, and those who stand to benefit most should bear a fair proportion of the risks and the burdens of the study.³² Third, informed consent should be obtained and confidentiality must be protected. Fourth, the safety of respondents and interviewers should be paramount, and infused in all project decisions.

Besides these general ethical principles, violence research poses specific challenges that require particular considerations, namely different legal frameworks that shape research procedures and affect disclosure of experiences of violence; the special training of interviewers allowing them to be able to give assistance to participants after disclosure and the need of on-going support to interviewers. Such issues have been addressed by the World Health Organization, resulting in the development of a set of recommendations for addressing the complex safety and ethical issues associated with researching, monitoring and documenting violence in different contexts.^10,19,20 Additionally, the implications of these guidelines for research on family violence have been discussed in literature.^23,32–35

When planning and designing the study, researchers need to take into account the national legislation on violence. One major issue relates to the mandatory character of reporting to authorities situations of violence. While reporting abuse is optional in the majority of European countries,³⁶ in some countries researchers must report cases of sexual or physical abuse to legal or social-service agencies.³²

The law in a country may shape research procedures and affect disclosure of experiences of violence. Researchers may experience conflicts between the following ethical principles: respect for confidentiality, the need to protect vulnerable populations, and respect for autonomy. Even though most western societies consider violence a crime and ethical guidelines for professionals demand mandatory reporting of such cases, most researchers agree that autonomy and confidentiality should prevail in scientific assessments. This apparent passive position is the subject of an on-going debate, with arguments supporting the need to avoid usurping a participant's right to make autonomous decisions, and the ethical and legal responsibility to act when violence is recognized.³² However, it is consensual that researchers should train fieldworkers to always inform participants of their rights and available help mechanisms.

As shown above, disclosure is related to the skills and sensitivity of the interviewer and consequently the entire research-team should be carefully selected and receive specialized training. When interviewers ask participants to reveal stories of trauma, it can be an opportunity to hear these stories in a sympathetic and non-judgmental way. Interviewers should be prepared to anticipate and to respond appropriately to respondents who may need additional assistance during the interview. Also, they should be clear about the unacceptability of any kind of abuse as a human rights violation and should inform the respondent of their rights under the law. Fieldworkers should be trained to refer respondents requesting assistance to available local services and sources of support. Although interviewers should not take on a role of counsellors, they should be open to assist if asked, but they should not tell participants what to do or to take on.

While respondents may face the greatest risk of harm, interviewers are also at risk when conducting research on violence.³⁷ The emotional toll of listening to repeated stories of participant's despair, physical pain, and degradation constitute the most common risk for fieldworkers.²³ In fact, being involved in violence research is an embodied experience in which it is hard to estimate the emotional effect that research might have on interviewers. Sometimes, it can be an intensely personal and emotional experience that many researchers find difficult and exhausting. This experience can be overwhelming if they have had personal experiences of abuse.³²

How to deal with the emotional costs and dilemmas associated with violence research experiences also needs to be discussed during interviewer training and throughout fieldwork, especially when formal methodological procedures and ethical guidelines seem of limited help.

Conclusions

An important assumption of public health is that effective policies for preventing violence must be firmly grounded in scientific research. Thus, making accurate estimates of violence and consequently valid identification of its determinants are essential to develop programmes, to communicate preventive messages and for policy-making. However, researchers should be aware of the difficulties in assessing the magnitude of interpersonal violence when planning and designing a study on violence. Also, the national legislation on violence and the management of episodes of violence detected by interviewers or researchers should be considered in the study procedures.

Conflicts of interest

The author declares no conflicts of interest.

Acknowledgements

Sílvia Fraga is the recipient of a Post Doc grant from Fundação para a Ciência e Tecnologia (SFRH/BPD/97015/2013). The EPIUnit, Institute of Public Health, University of Porto, Portugal (UID/DTP/04750/2013) is funded by the Portuguese Foundation for Science and Technology.