Measuring Intimate Partner Violence TEchnIcal BrIEf
Measuring intimate partner violenceTechnical Brief JULY 2017Authors: Lori Heise and Mazeda HossainThis brief provides guidance for the non-expert onhow to collect valid quantitative data on partnerviolence in an ethically and methodologically soundmanner. Specifically, it addresses:Data available to fully articulate the relationship betweenviolence and the uptake and use of services and preventionand treatment options are still relatively sparse. However,investigators could learn much by regularly including questionson violence in their on-going research. Definitions Ethical and safety obligations (informed consent,privacy) Methods for increasing disclosure amongresearch participants Minimum items necessary for measuring intimatepartner violence Defining IPV as an outcome or exposure variableIndeed, an ever-widening array of researchers have expressedinterest in collecting data on women’s experience of violence.Generally, such studies are designed to serve other ends:evaluating the effectiveness of an intervention, exploringfactors that predict women’s access to and control overincome or monitoring clinical trials for social harms.Regardless of the study’s main purpose, the fields of violenceprevention, international development and global health wouldgreatly benefit if researchers from other disciplines could,reliably and ethically, insert questions on violence – as eithera potential explanatory variable or an outcome of interest– into their quantitative surveys. This brief is intended as acontribution towards this end.Gender-based violence – including physical, sexual, emotionaland economic violence and abuse – is widespread globally.The most pervasive form of gender-based violence is intimatepartner violence (IPV), also known as domestic violence orpartner/spouse abuse. On average, 30% of women worldwidewill experience at least one episode of sexual and/or physicalIPV within her lifetime; the incidence and prevalence ofviolence in relationships, however, varies greatly both betweencountries and regions, and between neighbourhoods andvillages.1, 2DefinitionsIPV can be defined as a pattern of behaviour within anintimate relationship that includes physical or sexuallyviolent acts, often accompanied by emotional aggressionand controlling behaviours, enacted by a current or formerintimate partner (i.e. spouse, boyfriend/girlfriend, datingpartner, or ongoing sexual partner).28Decades of research have demonstrated that the healthconsequences of violence are cumulative and long term andextend far beyond injury. They include immediate and longerterm physical and mental health outcomes such as chronicpelvic pain, HIV and other sexually transmitted infections(STIs), unwanted pregnancy, adverse pregnancy outcomes,suicidal ideation, depression and increased risk of homicide.3-11Partner violence includes: physical aggression – such as slapping hitting, kickingbeating forced intercourse or other forms of sexual coercion psychological abuse – such as intimidation, constantbelittling and humiliation controlling behaviours – such as isolating a personfrom their family and friends, monitoring theirmovements and restricting their access to assistanceor information.Partner violence also affects a range of other developmentoutcomes, including infant and under-5 child mortality,women’s political participation and women’s ability tobetter their economic prospects through micro-finance andsavings schemes.12-15 As a result, there has been increasinginterest in exploring whether, and under what conditions,violence in relationships may be associated with other health,development and socio-economic outcomes.IPV is part of a larger category of abuses called violence againstwomen (VAW) or gender-based violence (GBV). The acronymsVAW and GBV are often used interchangeably because mostgender-based violence is perpetrated by men against womenand girls. Violence against women includes a range of abusesthat extends throughout the life cycle, from sex selectiveabortion and child sexual abuse, to so-called ‘honour’ killingsand female genital cutting (see Figures 1 and 2).In the context of HIV, partner violence has emerged as abarrier to use of services and to the uptake and consistentuse of various treatment and prevention strategies. Partnerviolence is associated, for example, with reluctance to seekHIV testing, fear of disclosing one’s HIV status, difficultyin attending clinic visits and reduced adherence to lifesaving HIV prevention methods and treatment regimens.16-181
More recently, people have begun to expand the termGBV to include violence directed at anyone based on theirbiological sex, their actual or perceived gender identity or theirperceived adherence to socially defined norms of masculinityor femininity (a concept known as gender expression)19. Thisexpands the category of GBV to include violence directedat lesbian, gay, bisexual or transgender (LGBT) individuals ifsomeone targets them explicitly for failing to conform to existingnorms around gender expression.Surveys (DHS) also now include a standard module on partnerviolence that has generated population-based data on IPV fromover 50 countries in the Global South.22Ethical and safety issuesIn 1999, investigators involved in the WHO Multi-country Studyon Women’s Health and Domestic Violence developed a set ofethical guidelines for the conduct of population-based surveysof physical and sexual abuse of women.23 These guidelinesargued that attention to safety and privacy is essential both forensuring data quality and for the ethical conduct of research.Previous experience had shown that women were willingto disclose abuse if given a safe environment to do so. Thisincludes, at a minimum, conducting interviews in completeprivacy; ensuring confidentiality; providing specialised trainingfor interviewers and other field staff on violence against women,trauma, and safety concerns; and making provision for supportservices in case women are in immediate danger or need followup emotional support. Research has demonstrated that withoutthese assurances, violence is under-reported and women’ssafety can be compromised.23This brief focuses on IPV both because of its ubiquity andbecause researchers have a better grasp on how to get reliableself-report data on partner violence than on more stigmatisedforms of violence such as child sexual abuse.Measuring IPVOfficial statistics, such as police reports, crime statistics, andhospital records, capture only the most severe cases of IPV. Asa result, service provider data substantially under-estimate theprevalence of violence within relationships. A recent reviewof women’s reporting behaviour in 24 low and middle-incomecountries found that while 40% of women had disclosed theviolence to someone in the past, only 7% of women had reportedit to a formal authority such as a health worker, religious leader,social worker or the police. This study finds that prevalencefigures based on health systems data or on police reports mayunderestimate the total prevalence of GBV from 11- to 128-fold,depending on the region and type of reporting.20Additional ethics guidanceGuidance on researching VAW has since been issued andreaders are strongly encouraged to consult these: Researching Violence against Women: A Practical Guidefor Researchers and Activists 24 Ethical and Safety Recommendations for InterventionResearch on Violence Against Women 25 Ethical and Safety Guidelines for Researching thePerpetration of Sexual Violence2 6 Ethical principles, dilemmas and risks in collecting dataon violence against children: A review of availableliterature 27The first ever global study on IPV using standardised definitions,methods and study design was conducted by the World HealthOrganization (WHO) between 1997 and 2004 in 15 different studysites.21 This study set forth an approach that has been widelyadapted, allowing for the collection of similar data across avariety of settings by using standardised definitions, samplingframes and training guidelines. The Demographic and HealthFigure 1: Gender-based violence throughout the life cycleElderlyElder/widow abuseEconomic abusePre-birthReproductive agePre-natal sexselectionFemicideDowry-related violenceIPVNon-partner sexual assaultSo-called ‘honour’ crimesSexual harassmentPolitical violenceEconomic abuseTraffickingSexualviolence andpsychologicalabuseInfancyFemale infanticideNeglect (health,care, nutrition)AdolescenceFemicideFGM/CForced marriageForced sex (includinginititation)IPV/dating violenceTraffickingChildhoodChild abuseChild marriageMalnutritionFGM/CTraffickingSource: Adapted from Ellsberg and Heise (2005) Researching Violence against Women: A Practical Guide for Researchers and Activists. WHO and PATH, Geneva: 10; adapted from Watts andZimmerman. 2002. ‘Violence against Women: Global Scope and Magnitude’, The Lancet 359 (9313): 1233, and Shane and Ellsberg. 2002. Violence against Women: Effects on Reproductive Health,Report No 20 (1), PATH, UNFPA, Washington: 2.2
Figure 2: Proportional venn diagram depicting overlap of types of violence among 24,000 womeninterviewed as part of WHO multi-country study on domestic violence and women’s healthno history of violencePhysical intimatepartner violenceonlyPhysical andsexual partnerviolenceSexual intimatepartner violence onlySexual assault since age15 by someone otherthan an intimate partnerChild sexual abuse(before age 15)Source: Adapted from Lori Heise (2005) WHO multi-country study on women’s health and domestic violence against women.often limited by violence exposure definitions that constitutethe legal definition of a crime.30These resources provide practical tips and recommendationsfor achieving safety and providing follow-up support to fieldstaff and respondents in the context of violence research.Such strategies include planning diversionary activitiesfor children and other family members to facilitate privacy;preparing a set of benign, default questions to switch to ifsomeone enters the room; conducting routine de-briefing andstress reduction exercises for field staff; and having a trainedcounsellor accompany the field team in settings where localreferral systems are inadequate.28Likewise, surveys that give women multiple opportunitiesto disclose yield consistently higher rates of violence.Accordingly, investigators strongly discourage broad ‘gateway’questions that ask a general question on violence and thenskip further abuse-related questions if the answer is no (e.g.Has your partner ever abused you?).24, 31, 32,Arguably the most important aspect for maximising disclosureis the selection and preparation of the field staff, especiallywhen using face-to-face interviews. In the Serbian site ofthe WHO Multi-country Study, for example, levels of violencedisclosed to field staff who received only 3 days of trainingwere significantly lower than those who received the study’snormal 2–3 week training course (21% vs. 26%; p .05). Theassassination of the Serbian president forced investigators tohire new interviewers, mid-stream, to speed up fieldwork priorto the special election called to replace the president. Thiscreated a ‘natural experiment’ that demonstrated the impact ofreduced interviewer preparation on levels of disclosure.33Encouraging disclosureResearch has demonstrated that multiple factors can influencelevels of disclosure in domestic violence surveys, including: the wording and framing of the questionsthe number of opportunities to disclosethe preparation and skills of the interviewersmode of survey delivery.Experience has shown that questions that inquire aboutbehaviourally specific acts, rather than using emotionallyladen terms such as rape or abuse, consistently yield higherlevels of disclosure.29 Asking whether your husband has everpushed or shoved you, for example, is emotionally easier forwomen to answer than questions that force them to labeltheir experience as ‘rape’ or ‘abuse’. These types of questionstap actual behaviours rather than ascribed meanings orinterpretations; behaviourally specific questions also helpmake responses comparable between women and acrosssettings.A final strategy to maximize disclosure is to shift to anonymousmethods of data capture, such as self-completed itemson paper, tablets or ACASI (Audio-Computer Assisted SelfInterview) devices. In low and middle-income settings,confidential methods have been shown to substantiallyincrease reporting of highly stigmatised forms of violencesuch as child sexual abuse.21, 34 However, other studies havesuggested that face-to-face interviews may encourage moredisclosure and yield more reliable data on less stigmatisedforms of violence.35 For example, studies done in India haveshown that women reported forcibly being touched36 anddomestic violence less often in ACASI compared to traditionalface-to-face-interviewing methods37. More studies are neededbefore conclusions can be drawn regarding the impact ofACASI on data quality, disclosure and bias for different types ofviolence.How surveys frame questions on violence can also influenceresults. Questions asked in the context of crime victimisationsurveys, for example, tend to yield lower estimates of partnerviolence than surveys about family life or relationshipsbecause they cue women to think only of events they feelcomfortable characterising as ‘crimes’. Questionnaires are3
Question wording and presentationexperienced in the last 12 months. In some cases, it may makesense to focus on a different timeframe, especially whenevaluating interventions. In a humanitarian crisis, for example,the occurrence of violence in the short term (i.e. the last threemonths) may be important for capturing programme effects ortime periods before and after a particular crisis.Physical partner violenceBox 1 provides an example of behaviourally specific questionson acts of physical violence by an intimate partner. Thesequestions from the WHO Multi-country study and the DHS havebeen shown to be valid and reliable in many settings.38, 39The format of the questions can be adapted to suit differentliteracy levels or delivery styles. The layout in Box 2, forexample, makes it easier for respondents to complete thequestions anonymously on a tablet device.Significantly, the layout also captures the timing and frequencyof acts. Standard practice in violence prevalence surveys isto report both lifetime experiences of violence and violenceBox 1: Timing and frequency of actsRespondent and her husband/partner“When two people marry or live together, they usually share both good and bad moments. I would now like to askyou some questions about your current and past relationships and how your husband/partner treats (treated) you. Ifanyone interrupts us, I will change the topic of conversation. I would again like to assure you that your answers willbe kept confidential, and that you do not have to answer any questions that you do not want to. May I continue?”705Has he or any other partnerever A)(If YEScontinuewith B. IfNO skip tonext item)B)Has thishappened inthe past 12months?(If YES ask Cand D. If NOask D only)C)In the past 12 monthswould you say thatthis has happenedonce, a few times ormany times?D)Did this happen before thepast 12 months?If YES: would you say thatthis has happened once, afew times or many times?YesNoYesNoOneFewManyNoOneFewManya) Slapped you or thrownsomething at you that could hurtyou?12121230123b) Pushed you or shoved you orpulled your hair?12121230123c) Hit you with his fist or withsomething else that could hurtyou?12121230123d) Kicked you, dragged you orbeaten you up?12121230123e) Choked or burnt you onpurpose?12121230123f) Threatened you with oractually used a gun, knife orother weapon against you?12121230123705g Who did the things you just mentioned?(Mention acts reported in 705) Was it yourcurrent or most recent husband/partner, aprevious husband or partner or both?Current/most recent husband/partner.1Previous husband/partner . 2Both . 3Don’t know/don’t remember . 8Refused/no answer . 94
Box 2: Optional format for presenting questions for self-completion on a tablet deviceWhen answering the following questions, I want you to think specificallyabout the last 12 months onlyNeverOnceFewManyIn the past 12 months, how many times has a current or previous husband orpartner slapped you or thrown something at you which could hurt you?1234In the past 12 months, how many times has a current or previous husband orpartner pushed or shoved you?1234In the past 12 months, how many times has a current or previous husband orpartner hit you with a fist or with something else that could hurt you?1234In the past 12 months, how many times has a current or previous husband orpartner kicked, dragged, or beaten you?1234In the past 12 months, how many times has a current or previous husband orpartner choked or burnt you on purpose?1234In the past 12 months, how many times has a current or previous husbandor partner threatened to use or actually used a gun, knife or other weaponagainst you?1234Sexual coercion and forced sexSurveys generally include three questions that are aggregatedto get a measure of sexual violence:Questions on forced or coerced sex in relationships areless standardised than physical violence items. Box 3 offersexamples that could be adapted. These may be laid out in theWHO format or more simply as above. one probing unwanted sexual intercourse exacted throughphysical force one capturing intercourse through non-physical forms ofcoercion one that asks about unwanted sexual acts other thanintercourse.Box 3: Examples of questions for measuring sexual violence or coercion by an intimate partnerChoose one from each columnPhysical forceNon-physical coercionOther unwanted sexual actsIn the past 12 months, how many timeshas a current or previous husband orboyfriend physically forced you to havesex when you did not want to?In the past 12 months, how many timeshave you had sexual intercourse whenyou did not want to because you wereafraid your husband or partner mighthurt or abandon you?In the past 12 months, how many timeshas a current or previous husband/partner made you do sexual things thatyou found degrading or humiliating?Did your current husband/partneror any other husband/partner everphysically force you to have sexualintercourse when you did not want to,for example, by twisting your arm orholding you down?Did you ever have sexual intercoursewhen you did not want to because youwere afraid of what your partner mightdo if you refused?Did your husband ever force you toperform sexual acts (other than vaginalintercourse) when you did not want to?Did you husband/partner ever forceyou to have vaginal intercourse withhim even when you did not want to?In the past 12 months, how manytimes has your partner used threatsor intimidation (but not physical force)to get you to have sexual intercoursewhen you did not want to? Did he use physical force (likeholding you down or hitting you)?Never, once, few times or manytimes? Did you give in because you felt youhad no other choice?5
Constructing outcome and explanatorymeasuresyou with his fist or with something that could hurt you; Kickedyou, dragged you or beaten you up; Choked or burnt you onpurpose; Threatened you with or actually used a gun, knife orother weapon against you). 41Depending on the purpose of your research, you will needto construct an overall measure of violence to use in youranalysis. The best way to do this will depend largely on yourresearch question and on whether you are using partnerviolence as an outcome or explanatory variable.Research from multiple settings confirm that, on average,women who have experienced only acts of moderat
Gender-based violence – including physical, sexual, emotional and economic violence and abuse – is widespread globally. The most pervasive form of gender-based violence is intimate partner violence (IPV), also known as domestic violence or partner
Kew Words: Domestic violence, intimate partner violence, education, children, outcomes, Colombia. I. NTRODUCTION. Due to its endemic nature, intimate partner violence (IPV) - violence perpetrated by a male against his female partner - is increasingly being recognized as a human development problem worldwide (Kishor and Johnson 2004).
intimate partner violence may be an issue.2 Allegations of intimate partner violence are common among custody‐litigating families, and custody evaluators face special challenges when conducting evaluations in this context. Model Standard 5.11 states that evaluations
Violence-Related Impact . Women As reported in the NISVS 2015 Data Brief (Smith et al., 2018), during their lifetime, about 1 in 4 women (25.1% or 30.0 million) in the U.S. experienced contact sexual violence, physical violence, and/or stalking by an intimate partner and reported at least
Vic Health (2004), The health costs of violence: Measuring the burden of disease caused by intimate partner violence. 13. Women are most at risk of violence generally, in: a) Dark streets b) Hotels and pubs c) Their home d) Cities e) Public transport Reference: Flood 2008 Violence against women: facts and figures White Ribbon Campaign
Prevalence of violence by an intimate partner in childhood and adolescence . 169. 7 . Physical and Mental Health Conditions by Victimization History . . . . . . . . . . . 177. Physical and Mental Health Conditions among U.S. Women and Men with and without a History of Contact Sexual Violence or Stalking by any Perpetrator, or Physical .
Abbreviations GBV: Gender-Based Violence IPV: Intimate Partner Violence DV: Domestic Violence VAW: Violence Against Women LGBTQIA: Lesbian, Gay, Bisexual, Transgender, Queer, Intersex, Asexual SAPS: South African Police Service Creative handling of a sensitive story can not only retain the dignity and respect of those involved, but can be used as a
Preventing Intimate Partner Violence Across the Lifespan: A Technical Package of Programs, Policies, and Practices. 5. Acknowledgements. We would like to thank the following individuals who contribu
Banking Standard Implementation Requirements, PSD2, the RTS and the CMA Order. Where relevant, it provides a regulatory reference (as per the CMA Order, PSD2/PSRs and the RTS on SCA and CSC). These are marked as either mandatory, optional or conditional in line with the definitions used across the Open Banking Standards. For TPPs, certifying against the CEG Checklist is considered as a signal .
