photo by hafm jansen - path.azureedge.net · based violence research.4 the guidelines were based on...

2PHO

TO B

Y H

AFM

JAN

SEN

[The experience] that most affected me was with a girl my age, maybe 22 years old...She toldme all about how her husband beat her while she was washing clothes in the back patio.Her mother-in-law would spy on her and tell her son things so that he would punish her. Shewas very afraid, and her voice trembled as she spoke, but she really wanted to tell me abouther tragedy. She kept looking over to where her mother-in-law was watching us. She askedme for help and I told her about the Women’s Police Station. When her mother-in-law got upto go to the latrine, I quickly gave her a copy of the pamphlet and she hid it. She thanked mewhen I left and I ended up crying in the street because I couldn’t stand to see such a younggirl being so mistreated… Nicaraguan interviewer. (Ellsberg et al, 2001.19)

A Practical Guide for Researchers and Activists 35

that transcend those in other areas becauseof the potentially threatening and traumaticnature of the subject matter. In the case ofviolence, the safety and even the lives ofwomen respondents and interviewers maybe at risk.1

In 1991, the Council for InternationalOrganization of Medical Sciences (CIOMS)

In many ways, researching violenceagainst women is similar to researchingother sensitive topics. There are issues ofconfidentiality, problems of disclosure, andthe need to ensure adequate and informedconsent. As the previous quote from aninterviewer illustrates, however, there areaspects of gender-based violence research

C H A P T E R T W O

Ethical Considerations for Researching ViolenceAgainst Women*

Topics covered in this chapter:

Respect for persons at all stages of the research processMinimizing harm to respondents and research staffMaximizing benefits to participants and communities (beneficence)Justice: Balancing risks and benefits of research on violence against women

* This chapter was adapted from Ellsberg and Heise, 2002.1

36 Researching Violence Against Women

C H A P T E R T W O

presented a set of International Guidelinesfor Ethical Review of EpidemiologicalStudies.3 These guidelines apply the basicethical principles of biomedical researchinvolving human subjects to the field ofepidemiology: respect for persons, non-maleficence (minimizing harm), benefi-cence (maximizing benefits), and justice.In 1999, the World Health Organization(WHO) published guidelines for address-ing ethical and safety issues in gender-based violence research.4 The guidelineswere based on the experiences of theInternational Research Network on ViolenceAgainst Women (IRNVAW) and weredesigned to inform the WHO Multi-countryStudy on Women’s Health and DomesticViolence Against Women. (See Box 2.1 for adescription of the main points.) The authorsargue that these ethical guidelines are criti-cal, not only to protecting the safety ofrespondents and researchers, but also toensuring data quality.

This chapter examines each of thebasic principles mentioned in the CIOMS

guidelines in turn and explores the chal-lenges of applying them to the specialcase of conducting research on domesticand sexual violence.

RESPECT FOR PERSONS AT ALL STAGES OF THERESEARCH PROCESS

Informed consent for respondents The principle of respect for persons incor-porates two fundamental ethical principles:respect for autonomy and protection ofvulnerable persons. These are commonlyaddressed by individual informed consentprocedures that ensure that respondentsunderstand the purpose of the researchand that their participation is voluntary.

There is still no consensus on whetherthe informed consent process for VAW stud-ies should explicitly acknowledge that thestudy will include questions on violence orwhether it is sufficient to warn participantsthat sensitive topics will be raised. TheWHO VAW study used an oral consentprocess that referred to the survey as astudy on women’s health and life experi-ences.5 Women were advised that, “Some ofthe topics discussed may be personal anddifficult to talk about, but many womenhave found it useful to have the opportunityto talk.” Women were told that they couldend the interview at any time or skip anyquestion they did not want to answer. (SeeBox 2.3 for an example of the informedconsent form used in the WHO VAW study.)A more detailed explanation of the natureof the questions on violence was provideddirectly before the violence questions, andrespondents were asked whether theywanted to continue and were againreminded of their option not to answer. It isa good idea to prepare a list of responsesfor questions that a woman might ask aboutthe study, such as how she was selected forthe study, what will the study be used for,and how her responses will be kept secret.

■ The safety of respondents and the research team is paramount and should infuseall project decisions.

■ Prevalence studies need to be methodologically sound and to build upon currentresearch experience about how to minimize the underreporting of abuse.

■ Protecting confidentiality is essential to ensure both women’s safety and dataquality.

■ All research team members should be carefully selected and receive specializedtraining and ongoing support.

■ The study design must include a number of actions aimed at reducing anypossible distress caused to the participants by the research.

■ Fieldworkers should be trained to refer women requesting assistance to availablesources of support. Where few resources exist, it may be necessary for the studyto create short-term support mechanisms.

■ Researchers and donors have an ethical obligation to help ensure that theirfindings are properly interpreted and used to advance policy and interventiondevelopment.

■ Violence questions should be incorporated into surveys designed for other purposes only when ethical and methodological requirements can be met.

(From WHO, 1999.4)

BOX 2.1 ETHICAL AND SAFETY RECOMMENDATIONS FOR DOMEST IC V IOLENCE RESEARCH


because children are generally consideredmore vulnerable and less able to act ontheir own behalf. The dilemma is particu-larly acute in settings where there are noeffective services to assist troubled families,or where reporting is likely to trigger a cas-cade of events that might put the child ateven greater risk (such as being removedfrom his/her home and placed in an institu-tion). The WHO VAW study specificallyexcluded questions about child abuse, butrequired teams to develop local protocols

E T H I C A L C O N S I D E R AT I O N S F O R R E S E A R C H I N G V I O L E N C E A G A I N S T W O M E N

Mandatory reporting of abuse Some countries have laws that require cer-tain kinds of professionals to report casesof suspected abuse to authorities or socialservice agencies. Such laws raise difficultissues for researchers because they throwinto conflict several key ethical principles:respect for confidentiality, the need to pro-tect vulnerable populations, and respect forautonomy. In the case of adult women,there is consensus among most researchersthat the principles of autonomy and confi-dentiality should prevail and thatresearchers should do everything withintheir power to avoid usurping a woman’sright to make autonomous decisions abouther life. (Of course if a woman seeks sup-port in reporting her abuse, researchersshould oblige.)

The dilemma of whether to comply withlegal reporting requirements is particularlyproblematic when dealing with child abuse.There is no consensus internationally abouthow to handle cases of child abuse

Researchers involved in the WHO Multi-country Studyon Women’s Health and Domestic Violence AgainstWomen debated at length the value of mentioningviolence directly in the initial consent process versusadding a second-order consent process immediatelybefore the questions on abuse. Some researchersargued that it was important to alert women up frontas to the true nature of the questions whereas othersfelt it was preferable to postpone introducing thenotion of violence until immediately prior to the actualabuse-related questions. This would allow some rap-port to develop, but still give a woman an opportu-nity to opt out of the violence-related questions.

The consent process was well received by respon-dents in all countries except Japan. During pilot test-ing, several Japanese respondents expressed asense of betrayal because they had not beeninformed that the interview contained questionsabout violence.6 As a result, the Japan team modi-fied its consent language to explicitly acknowledgeviolence up front. This is an excellent example ofhow ethical principles and actual experience cancombine to guide practice.

BOX 2.2 ADAPTING ETHICAL GUIDEL INES TO LOCAL SETT INGS

Used in the WHO Multi-country Study on Women’s Health and Domestic ViolenceAgainst Women

Hello, my name is [*]. I work for [*]. We are conducting a survey in [study loca-tion] to learn about women’s health and life experiences. You have been chosen bychance (as in a lottery/raffle) to participate in the study.

I want to assure you that all of your answers will be kept strictly secret. I will notkeep a record of your name or address. You have the right to stop the interview atany time, or to skip any questions that you don’t want to answer. There are no rightor wrong answers. Some of the topics may be difficult to discuss, but many womenhave found it useful to have the opportunity to talk.

Your participation is completely voluntary but your experiences could be very helpfulto other women in [country].

Do you have any questions?

(The interview takes approximately [*] minutes to complete). Do you agree to beinterviewed?

NOTE WHETHER RESPONDENT AGREES TO INTERVIEW.

[ ] DOES NOT AGREE TO BE INTERVIEWED

THANK PARTICIPANT FOR HER TIME AND END INTERACTION.

[ ] AGREES TO BE INTERVIEWED.Is now a good time to talk?It’s very important that we talk in private. Is this a good place to hold the inter-view, or is there somewhere else that you would like to go?

TO BE COMPLETED BY INTERVIEWER

I CERTIFY THAT I HAVE READ THE ABOVE CONSENT PROCEDURE TO THEPARTICIPANT.

SIGNED: ____________________________________________________________

(From WHO, 2004.5)

BOX 2.3 INDIVIDUAL CONSENT FORM


for handling cases of child abuse that inter-viewers might nonetheless come to knowabout. The guiding principle of these proto-cols was to act in “the best interests of thechild,” a standard that each team opera-tionalized locally, based on advice from keyagencies about prevailing conditions.

Community agreement In many countries, it is also important toobtain community support for research, aswell as individual consent. (Communityconsent, however, should never replaceindividual consent.) This is often sought bymeeting with community leaders to explainthe overall objectives of the research. Forsafety reasons, when obtaining communitysupport for VAW research, it is important toframe the study in general terms—such asa study on women’s health or life experi-ences rather than mention violence orabuse directly. If it becomes well known inthe community that women are beingquestioned about violence, men may pro-hibit their partners from participating ormay retaliate against them for their partici-pation. In addition to potentially jeopardiz-ing the safety of respondents, this couldalso undermine the study objectives anddata accuracy.

MINIMIZING HARM TORESPONDENTS ANDRESEARCH STAFF

Ensuring participant safety The primary ethical concern related toresearching VAW is the potential for inflict-ing harm to respondents through their par-ticipation in the study. A respondent maysuffer physical harm if a partner finds outthat she has been talking to others abouther relationship with him. Because manyviolent partners control the actions of theirspouses closely, even the act of speakingto another person without his permissionmay trigger a beating.

No systematic studies have been per-formed to determine how often womensuffer negative consequences from partici-pating in research on violence. However,several VAW researchers have recordedchilling examples of experiences wherewomen have been placed at risk as a resultof inadequate attention to safety issues.8

For example, researchers from Chiapas,Mexico, describe how, when they firstbegan researching domestic violence, theywere not fully aware of the risks involved.They included a small set of questions ondomestic violence within a larger study onreproductive health without taking anyspecial precautions regarding safety ofrespondents. They were shocked to learnlater that three respondents were beaten bytheir partners because they had partici-pated in the survey.9

The WHO guidelines provide a numberof suggestions about how to minimize risksto respondents, including:

■ Interviewing only one woman per house-hold (to avoid alerting other women whomay communicate the nature of the studyback to potential abusers).

■ Not informing the wider community thatthe survey includes questions on violence.

■ Not conducting any research on violencewith men in the same clusters wherewomen have been interviewed.4

Protecting privacy and confidentialityHis mother and sisters kept passing by, andwould peek in the doorway to see what wewere talking about, so we would have tospeak really softly…and the girl said to me,“Ay, don’t ask me anything in front ofthem.” (Nicaraguan interviewer) 2

Protecting privacy is important in itsown right and is also an essential elementin ensuring women’s safety. In addition to

C H A P T E R T W O


supervisors and even drivers can also playa role in distracting household memberswho are intent on listening to the inter-view. In one instance in Zimbabwe, field-workers entered into lengthy negotiationsto purchase a chicken from the husband ofa respondent so that she could be inter-viewed in private.10 Other researchers havecarried candy and coloring books to keepchildren busy during interviews.

Indeed, the Japanese team for theWHO VAW study found it so difficultto achieve privacy in Japan’scrowded apartments that they had todepart from the protocol and useself-response booklets for especiallysensitive questions. In this highly lit-erate population, women were ableto read and record their answerswithout the questions having to beread aloud.6

Ensuring privacy may be evenmore problematic in telephone sur-veys. Interviewers for the VAW sur-vey in Canada were trained to detectwhether anyone else was in theroom or listening on another line,and to ask whether they should callback at another time. They providedrespondents with a toll free numberto call back if they wanted to verify


interviewing only one woman per house-hold, the WHO recommendations adviseresearchers to conduct violence-relatedinterviews in complete privacy, with theexception of children under the age oftwo. In cases where privacy cannot beensured, interviewers should be encour-aged to reschedule the interview for a dif-ferent time or place. Achieving this level ofprivacy is difficult and may require moreresources than might be needed forresearch on less sensitive topics.

Researchers have developed a variety ofcreative methods for ensuring privacy.Interviewers in Zimbabwe and Nicaraguaoften held interviews outside or accompa-nied women to the river as they washedclothes. Many studies have successfullyused “dummy” questionnaires, containingunthreatening questions on issues such asbreastfeeding or reproductive health.Respondents are forewarned that if some-one enters the room, the interviewer willchange the topic of conversation byswitching to a dummy questionnaire. Othermembers of the research team such as

■ Interview only one woman per household.

■ Don’t inform the wider community that the survey includes questions on violence.

■ Don’t interview men about violence in the same households or clusters wherewomen have been asked about violence.

■ Interviews should be conducted in complete privacy.

■ Dummy questionnaires may be used if others enter the room during the interview.

■ Candy and games may be used to distract children during interviews.

■ Use of self-response questionnaires for some portions of the interview may beuseful for literate populations.

■ Train interviewers to recognize and deal with a respondent’s distress during theinterview.

■ End the interview on a positive note that emphasizes a woman’s strengths.

BOX 2.4 SUGGEST IONS FOR MINIMIZ ING HARM TO WOMEN PART IC IPAT ING IN RESEARCH

Interview in Thailand

PHO

TO B

Y H

AFM

JAN

SEN


that the interview was legitimate, or in casethey needed to hang up quickly. About1,000 out of a sample of 12,000 womencalled back, and 15 percent of the callswere to finish interrupted interviews.11

Minimizing participant distressInterviews on sensitive topics can provokepowerful emotional responses in some par-ticipants. The interview may cause awoman to relive painful and frighteningevents, and this in itself can be distressingif she does not have a supportive socialenvironment.12 Interviewers therefore needto be trained to be aware of the effects thatthe questions may have on informants andhow best to respond, based on a woman’slevel of distress.

Most women who become emotionalduring an interview actively choose to pro-ceed, after being given a moment to collectthemselves. Interviewer training shouldinclude practice sessions on how to iden-tify and respond appropriately to symp-toms of distress as well as how toterminate an interview if the impact of thequestions becomes too negative.

Interviewer training should also includeexplicit exercises to help field staff exam-

ine their own attitudes and beliefs aroundrape and other forms of violence.Interviewers frequently share many of thesame stereotypes and biases about victimsthat are dominant in the society at large.Left unchallenged, these beliefs can lead tovictim-blaming and other destructive atti-tudes that can undermine both the respon-dent’s self esteem and the interviewer’sability to obtain quality data.

Referrals for care and supportAt a minimum, the WHO guidelines sug-gest that researchers have an ethical obliga-tion to provide a respondent withinformation or services that can help hersituation. In areas where specific violence-related services are available, researchteams have developed detailed directoriesthat interviewers can use to make referrals.In Canada’s VAW survey, for example, thecomputer program used by telephoneinterviewers had a pop-up screen thatlisted resources near the respondent, basedon her mail code. In Zimbabwe, Brazil,Peru, and South Africa, researchers devel-oped small pamphlets for respondents thatlisted resources for victims along with ahost of other health and social serviceagencies.10 All women were offered thepamphlet after being asked if it would besafe for them to receive it (cases have beenreported where women have been beatenwhen a partner found informational mate-rial addressing violence). In Zimbabwe,interviewers carried a referral directory andwrote out addresses on physician referralpads so that the referral would not attractsuspicion if discovered. Ideally, contactshould be made in advance with the serv-ices so that they are prepared to receivereferrals from the study.

In settings where resources are scarce ornonexistent, researchers have developedinterim support measures. For example, astudy on violence against women per-formed in rural Indonesia brought in a

C H A P T E R T W O

Respondent inTanzania tells childrento go play beforestarting her interview

PHO

TO B

Y H

AFM

JAN

SEN


counselor to the field once a week to meetwith respondents.13 In Ethiopia, the studyhired mental health nurses to work in theclosest health center for the duration of thefieldwork.14 The number of women whoactually make use of such services is oftenquite low, but subsequent interviews withwomen indicate that they appreciate know-ing that services are available if needed.11

In Peru and in Bangladesh, the WHO VAWteam has used the study as an opportunityto train local health promoters in basiccounseling and support skills. In this way,the team will leave behind a permanentresource for the community.

Bearing witness to violence The image of these stories affects you, to seehow these women suffer, and especially thefeeling that no one supports them. These areexperiences that you never forget…(Nicaraguan interviewer)2

Although preventing harm to respon-dents is of primary importance, researchersalso have an ethical obligation to minimizepossible risks to field staff and researchers.Sources of risk include threats to physicalsafety either as a result of having to travelin dangerous neighborhoods or fromunplanned encounters with abusive indi-viduals who object to the study. Some

strategies to reduce the first source of riskinclude removing extremely dangerousneighborhoods from the sampling framebefore drawing the sample (for examplethose controlled by narco-traffickers); out-fitting teams with cell phones; and havingmale drivers accompany female interview-ers into dangerous areas.

Abusive partners have also been knownto threaten interviewers with physical harm.In a South African study, for example, aman came home from a bar in the middleof his partner’s interview and pulled a gunon the fieldworker, demanding to see thequestionnaire. Because of prior training, theinterviewer had the presence of mind togive the man an English version of thequestionnaire, which he was unable toread.10 “Dummy” questionnaires would alsohave been helpful in this situation.

The most common risk for fieldwork-ers, however, is the emotional toll of listen-ing to women’s repeated stories of despair,physical pain, and degradation. It is hardto overestimate the emotional impact thatresearch on violence may have on field-


Researchers in a study performed in Cambodia found a young woman who washeld prisoner in her own home by her husband. When the research team arrivedto interview her, they found the woman locked in her house, with only a peepholewhere a chain was threaded through a crudely cut hole in the door. The womanconducted the interview through the peephole. During the interview, the husbandappeared and was suspicious about their activity. The team gave him a falseexplanation for their visit and then left the home.

The next day, the team sought help from the Ministry of Women’s Affairs, which co-sponsored the study. Secretariat staff informed the researchers that the woman’s hus-band had stormed into their office the preceding afternoon, dragging his wife bythe arm. He demanded to know who had been at his door. He told the Secretariatpersonnel that if they couldn’t confirm her explanation, then his wife would suffer.They readily confirmed her story. She was safe for the moment, but the researchersrealized that it would be too dangerous to ever approach this woman again.

The team made several overtures with different government officials and the policeto help get the woman freed, but everyone was afraid to intervene because thewoman’s husband had an important position. Researchers described the frustrationthat the team felt at not being able to free the woman and the guilt they felt at having put the woman in greater danger.

(From Zimmerman, 1995.8)

BOX 2.5 PROTECT ING RESPONDENT SAFETY IN CAMBODIA

Interview in Bangladesh

PHO

TO B

Y H

AFM

JAN

SEN


workers and researchers. As the narrativefrom a Nicaraguan fieldworker presented atthe beginning of this chapter illustrates, astudy on violence often becomes anintensely personal and emotional journeyfor which many researchers are not pre-pared. Particularly when field staff havehad personal experiences of abuse, theexperience can be overwhelming. JudithHerman, in her work on psychologicaltrauma in survivors of political and domes-tic violence, describes this as a commonexperience for those who study violence:

To study psychological trauma is to comeface to face both with human vulnerabilityin the natural world and with the capacityfor evil in human nature. To study psycho-logical trauma means bearing witness tohorrible events.15

Including discussions of violence ininterviewer training is crucial for reducingdistress during fieldwork. During fieldwork,another important measure is to provideinterviewers and research staff with regularopportunities for emotional debriefing, orwhen necessary, individual counseling.Researchers have used a variety of creativestrategies for protecting the emotionalhealth of their staff. In Peru, for example,the WHO multi-country team employed aprofessional counselor to lead weekly sup-port sessions that incorporated guidedimagery and relaxation techniques.Experience has repeatedly demonstratedthat emotional support for fieldworkers isessential. Not only does it help interview-ers withstand the demands of the field-work, but it also improves their ability togather quality data.

Transcripts of debriefing sessions withinterviewers who participated in studieswithout adequate support illustrate thispoint:

…When I heard stories about womenbeing beaten and tied up, I would leavethere feeling desperate… I would be awreck, and my supervisor would tell me

“get a hold of yourself, you cry for every lit-tle thing.” But how could I control myself? Icouldn’t stand it… I would try, but some-times it was impossible, and I would burstinto tears during the next interview…(Nicaraguan interviewer) 2

Other interviewers commented that theyfelt extremely drained and distracted by theinterviews where women reported vio-lence. One woman reported that she hadstopped working for the study because shecould not bear to listen to women’s storiesof abuse. 2

Experience has shown that trauma-related stress is not confined to field staffwho are directly involved with respon-dents. Field supervisors, transcribers, driv-ers, and even data entry personnel may beaffected. In one study in Belize, a tran-scriber broke down after hours of listeningto in-depth qualitative interviews with sur-vivors of abuse.16

It is particularly important to provideopportunities during training for inter-viewers to address their own experiencesof abuse. Given the high prevalence ofgender-based violence globally, it is likelythat a substantial proportion of interview-ers will have experienced gender-basedviolence themselves at some point. Theseexperiences need to be taken into consid-eration. Most people learn to cope withpainful past experiences, and usually donot dwell on them in their everyday lives.However, when trainees are confrontedwith the subject matter the informationmay awaken disturbing images and oremotions. For many trainees, simplyacknowledging the fact that these reac-tions are normal and providing timelyopportunities to discuss them will be suffi-cient to help them complete the trainingand participate successfully in fieldwork.In those rare cases where feelings becometoo overwhelming, trainees should besupported in their decision to withdrawfrom the study.

C H A P T E R T W O


MAXIMIZING BENEFITS TO PARTIC IPANTS AND COMMUNIT IES(BENEF ICENCE)

The principle of beneficence refers to theethical obligation to maximize possiblebenefits to study participants and thegroup of individuals to which they belong.This principle gives rise to norms requiringthat the risks of research be reasonable inlight of the expected benefits, that theresearch design be sound, and that theinvestigators be competent both to conductthe research and to ensure the well-beingof participants.

The interview as an interventionAsking women to reveal stories of traumacan be a transforming experience for bothresearchers and respondents. Indeed, thereis ample evidence that most women wel-come the opportunity to tell their stories if they are asked in a sympathetic, non-judgmental way. In our experience,women rarely refuse to answer questionson violence.

Many women who disclose violence insurveys have never told anyone abouttheir situations.17 Many studies find thatparticipants find the experience to be sohelpful that they ask fieldworkers to “inter-view” a friend or relative who has a storyto tell. As Herman notes, “rememberingand telling the truth about terrible eventsare prerequisites both for the restoration ofthe social order and for the healing ofindividual victims.”15

Even the act of telling her story canoffer a woman some small way of trans-forming her personal ordeal into a way tohelp others. Indeed, researchers sensitiveto this issue encourage interviewers andfield staff to take hope and satisfactionfrom their participation in the process ofgiving a voice to women’s suffering.

A qualitative study of survivors of

abuse who had visited a women’s crisiscenter in Nicaragua found that a centralpart of women’s process of recovery andpersonal as well as collective empower-ment came not only from increasedknowledge of their rights, but also fromthe opportunity to share their experiencesand to help other women in similar situa-tions.18 In this sense, asking women aboutexperiences of violence may be seen asan intervention in itself. At the very least,asking conveys the message that violenceis a topic worthy of study, and not ashameful or unimportant issue.

In this same vein, many fieldworkers inthe León, Nicaragua, research describedthe experience of listening to women’s sto-ries, as well as the opportunity to tell theirown stories in the debriefing sessions, as aprofoundly healing experience. One inter-viewer who had never before discussedher experiences said,

[when I joined this study] I felt that I hadfinally found someone I couldtell everything to, someone withwhom I could share my bur-den, because it’s horrible to feelso alone. Now I feel that aweight has been taken offme…I feel relieved…19

The interview is also anopportunity to provide womenwith information on gender-based violence.Many studies have issued small cards thatcan be easily hidden in a shoe or inside ablouse with information about localresources for abused women and messagessuch as, “If you are being abused, thereare ways out” or “Violence is never justi-fied.” Such messages may enable womento see experiences in a new light or toidentify violence in others close to them.

Researchers also stress the importance ofending the interview on a note thatemphasizes women’s strengths and tries tominimize distress, particularly as a respon-dent may have revealed information that


Remembering and telling the truth about terrible eventsare prerequisites both for the restoration of the socialorder and for the healing of individual victims. (Herman,

1992.15)


made her feel vulnerable.20 A number ofstudies have carefully scripted such end-ings to ensure that the interview finisheswith clear statements that explicitlyacknowledge the abuse, highlight theunacceptability of the violence, andemphasise the respondent’s strengths inenduring and/or ending the violence. TheWHO study ends each interview with thewords, “From what you have told me, I cantell that you have had some very difficulttimes in your life. No one has the right totreat someone else in that way. However,from what you have told me I can also seethat you are strong and have survivedthrough some difficult circumstances.”7

One indication of how women haveviewed the interview process can beobtained by assessing respondents’ satisfac-tion with the interview. At the end of theWHO interview, respondents were askedthe following question: “I have askedabout many difficult things. How has talk-ing about these things made you feel?” Theanswers were written down verbatim andcoded by the interviewer into the followingthree categories: good/better, bad/worse,and same/no difference. The majority(between 60 and 95 percent in seven sites)of women who had experienced physicalor sexual partner violence reported that

they felt good/better at the end of theinterview. In most countries, the range wassimilar between women who had or hadnot experienced partner violence. Very fewwomen reported feeling worse after beinginterviewed. Between 0.5 and 8.4 percentof women reporting partner violence ever(highest in Peru) and between zero and 3.2percent of women with no history of part-ner violence felt worse.17

Assuring scientific soundnessThe CIOMS guidelines note: “A study thatis scientifically unsound is unethical in thatit exposes subjects to risk or inconveniencewhile achieving no benefit in knowledge.”3

This principle is particularly important inthe area of gender-based violence wherewomen are asked to disclose difficult andpainful experiences and where the act ofresearch itself may put women at furtherrisk of abuse. Thus the WHO guidelinesnote that violence researchers have an ethi-cal responsibility to ensure the soundnessof their work by selecting a large enoughsample size to permit conclusions to bedrawn, and by building upon currentknowledge about how to minimize under-reporting of violence. (See Chapter 7 formore discussion of sampling techniques.)Underreporting of violence will dilute asso-ciations between potential risk factors andhealth outcomes, leading to falsely nega-tive results. Underestimating the dimen-sions of violence may also preventviolence intervention programs fromreceiving the priority they deserve in theallocation of resources.

Research demonstrates that disclosurerates of violence are highly influenced bythe design and wording of questions, thetraining of interviewers, and the imple-mentation of the study.2 In Chapter 6, wediscuss this issue in much greater depthand outline the variety of measures thathave been developed to enhance disclo-sure of violence.

C H A P T E R T W O

Looking for householdsin Samoa

PHO

TO B

Y H

AFM

JAN

SEN


Using study results for social changeIt is important to feed research findingsinto ongoing advocacy, policy making,and intervention activities. Too often criti-cal research findings never reach the atten-tion of the policy makers and advocatesbest positioned to use them. The enor-mous personal, social, and health-relatedcosts of violence against women place amoral obligation on researchers anddonors to try to ensure that study findingsare applied in the real world. It is alsoimportant that the study communityreceives early feedback on the results ofthe research in which it has participated.Chapter 14 addresses this issue in moredetail and describes several successfulexamples of how research findings havebeen used to contribute to changing lawsand policies on domestic violence.

One way to improve the relevance ofresearch projects is, from the outset, toinvolve organizations that carry out advo-cacy and direct support for survivors ofviolence, either as full partners in theresearch or as members of an advisorycommittee. Such committees can play animportant role in helping guide the studydesign, advise on the wording of ques-tions, assist with interviewer training, andgive guidance on possible forms of analy-sis and the interpretation of results. Thesegroups also have a central role to play inpublicizing and applying the project’sfindings.

JUST ICE : BALANCINGRISKS AND BENEFITS OFRESEARCH ON VIOLENCEAGAINST WOMEN

Research, like any endeavor that touchespeople’s lives, involves inherent risks. Theprinciple of distributive justice demandsthat the class of individuals bearing theburden of research should receive anappropriate benefit, and those who stand


to benefit most should bear a fair propor-tion of the risks and burdens of the study.

In the case of gender-based violenceresearch, the risks are potentially large, butso too are the risks of ignorance, silence,and inaction. Researchers and ethicalreview boards must constantly balance thisreality. Lisa Fontes cites the case of a col-league from India who wanted to studywives who were hospitalized after havingbeen burned by their husbands in disputesover dowry. She ultimatelydecided not to conduct theresearch for fear that theresearch would put women atfurther risk. As Fontesobserves, “Her decision elimi-nated the research-related riskto the participants, but alsoeliminated the potential benefitof reducing the terrible isola-tion and vulnerability of thesevictims.”21

It is possible to conductresearch on violence with full respect forethical and safety considerations if propercare and resources are devoted to this end.We must remember that women living withviolence are already at risk. Researcherscannot eliminate this reality, just as theycannot fully eliminate the possibility thatfurther harm will be caused by their study.The obligation of researchers is to carefullyweigh the risks and benefits of any studyand to take every measure possible to limitpossible harm and to maximize possiblebenefit. At the very least, we must ensurethat when women take risks to share theirstories, we honor that risk by using thefindings for social change.

Women would ask me whatthis survey was for, and howit would help them. I wouldtell them that we won’t seethe solution tomorrow or thenext year. Our daughtersand granddaughters will seethe fruits of this work, maybethings will be better by then.Nicaraguan fieldworker. (From

Ellsberg, et al, 2000.19)


C H A P T E R T W O

1. Ellsberg M, Heise L. Bearing witness: Ethics indomestic violence research. Lancet.2002;359(9317):1599-1604.

2. Ellsberg M, Heise L, Peña R, Agurto S, Winkvist A.Researching domestic violence against women:Methodological and ethical considerations. Studiesin Family Planning. 2001;32(1):1-16.

3. Council for International Organizations of MedicalSciences. International Guidelines for EthicalReview of Epidemiological Studies. Geneva:CIOMS; 1991.

4. World Health Organization. Putting Women’sSafety First: Ethical and Safety Recommendationsfor Research on Domestic Violence AgainstWomen. Geneva: Global Programme on Evidencefor Health Policy, World Health Organization;1999. Report No.: WHO/EIP/GPE/99.2.

5. World Health Organization. WHO Multi-countryStudy on Women’s Health and Domestic ViolenceAgainst Women: Study Protocol. Geneva: WorldHealth Organization; 2004.

6. Yoshihama M. Personal Communication. AnnArbor, Michigan. Washington, DC. 2004.

7. World Health Organization. WHO Multi-countryStudy on Women’s Health and Domestic ViolenceAgainst Women: Study Questionnaire V10.Geneva: World Health Organization; 2004.

8. Zimmerman K. Plates in a Basket Will Rattle:Domestic Violence in Cambodia, a Summary.Phnom Penh, Cambodia: Project AgainstDomestic Violence; 1995.

9. Health and Development Policy Project.Measuring Violence Against Women Cross-cultur-ally: Notes from a Meeting. Takoma Park,Maryland: Health and Development PolicyProject; 1995.

10.Jewkes R, Watts C, Abrahams N, Penn-Kekana L,Garcia-Moreno C. Ethical and methodologicalissues in conducting research on gender-basedviolence in Southern Africa. Reproductive HealthMatters. 2000;8(15):93-103.

11.Johnson H. Dangerous Domains: ViolenceAgainst Women in Canada. Ontario, Canada:International Thomson Publishing; 1996.

12.Finkelhor D, Hotaling GT, Yllo K. Special EthicalConcerns in Family Violence Research. In:Finkelhor D, Hotaling GT, Yllo K, editors.Stopping Family Violence: Research Priorities forthe Coming Decade. London: Sage; 1988.

13.Hakimi M, Nur Hayati E, Ellsberg M, Winkvist A.Silence for the Sake of Harmony: DomesticViolence and Health in Central Java, Indonesia.Yogyakarta, Indonesia: Gadjah MadaUniversity;PATH, Rifka Annisa, Umeå University;2002.

14.Gossaye Y, Deyessa N, Berhane Y, et al. Women’shealth and life events study in rural Ethiopia.Ethiopian Journal of Health Development.2003;17(Second Special Issue):1-49.

15.Herman J. Trauma and Recovery: The Aftermathof Violence: From Domestic Abuse to PoliticalTerror. New York: Basic Books; 1992.

16.Shrader E. Personal Communication. Washington,DC; 2000.

17.Jansen HAFM, Watts C, Ellsberg M, Heise L,Garcia-Moreno C. Interviewer training in theWHO Multi-country Study on Women’s Healthand Domestic Violence Against Women. Violenceagainst Women. 2004;10(7):831-849.

18.Wessel L, Campbell J. Providing sanctuary for bat-tered women: Nicaragua’s Casas de la Mujer.Issues in Mental Health Nursing. 1997;18:455-476.

19.Ellsberg M. Candies in Hell: Research and Actionon Domestic Violence in Nicaragua [DoctoralDissertation]. Umeå, Sweden: Umeå University;2000.

20.Parker B, Ulrich Y. A protocol of safety: Researchon abuse of women. Nursing Research.1990;38:248-250.

21.Fontes LA. Ethics in family violence research:Cross-cultural issues. Family Relations.1998;47:53-61.

photo by hafm jansen - path.azureedge.net · based violence research.4 the guidelines were based on...

Documents