Researching Vaw - Practical Guide

Please see the Table of Contents for access to the entire publication.
Researching Violence
Against Women
A PRACTICAL GUIDE
FOR RESEARCHERS AND ACTIVISTS
Researching Violence
Against Women
A PRACTICAL GUIDE
FOR RESEARCHERS AND ACTIVISTS
Mary Ellsberg ■ Lori Heise

WHO Library Cataloguing-in-Publication Data
Ellsberg, Mary Carroll, Heise, Lori.

Researching Violence Against Women: A Practical Guide for Researchers and Activists/
Mary Ellsberg, Lori Heise.
Suggested citation: Ellsberg M, and Heise L. Researching Violence Against Women: A

Practical Guide for Researchers and Activists. Washington DC, United States: World Health
Organization, PATH; 2005.
1. Domestic violence 2. Spouse abuse 3. Women 4. Research design 5. Manuals

I. Title
ISBN 92 4 154647 6 (LC/NLM classification: HV 6556)
© World Health Organization and Program for Appropriate Technology in Health (PATH)
2005. All rights reserved. Publications of the World Health Organization can be obtained
from WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27,
Switzerland (tel: +41 22 791 3264; fax: +41 22 791 4857; email: bookorders@who.int).
Requests for permission to reproduce or translate WHO publications – whether for sale or
for noncommercial distribution – should be addressed to Publications, at the above address
(fax: +41 22 791 4806; email: permissions@who.int). Publications of PATH can be obtained
from publications@path.org.
The designations employed and the presentation of the material in this publication do
not imply the expression of any opinion whatsoever on the part of the World Health
Organization or PATH concerning the legal status of any country, territory, city, or area or of
its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on
maps represent approximate border lines for which there may not yet be full agreement.
The mention of specific companies or of certain manufacturers’ products does not imply
that they are endorsed or recommended by the World Health Organization or PATH in
preference to others of a similar nature that are not mentioned. Errors and omissions
excepted, the names of proprietary products are distinguished by initial capital letters.
Neither the World Health Organization nor PATH warrants that the information contained in
this publication is complete and correct and neither shall be liable for any damages
incurred as a result of its use.
The named authors alone are responsible for the views expressed in this publication.
Printed in the United States.
PATH creates sustainable, culturally relevant solutions that enable communities worldwide to
break longstanding cycles of poor health. By collaborating with diverse public- and private-
sector partners, we help provide appropriate health technologies and vital strategies that
change the way people think and act. Our work improves global health and well-being.
For more information, please visit www.who.int/gender/en or www.path.org.

Please click inside the blue boxes below for access to the chapters within.
Table of Contents
ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
CHAPTER ONE: VIOLENCE AGAINST WOMEN

A S A H E A LT H A N D D E V E L O P M E N T I S S U E . . . . . . . . . . . . . . . 8
Definitions of violence against women
Prevalence of intimate partner violence
The patterning of intimate partner violence
Prevalence and characteristics of sexual coercion and abuse
The effects of violence on women’s health
Explaining gender-based violence
How do women respond to abuse?
Challenges for international research on gender-based violence
C H A P T E R T W O : E T H I C A L C O N S I D E R AT I O N S F O R
RESEARCHING VIOLENCE AGAINST WOMEN . . . . . . . . . . . . 34
Respect for persons at all stages of the research process
Minimizing harm to respondents and research staff
Maximizing benefits to participants and communities (beneficence)
Justice: Balancing risks and benefits of research on violence against women
C H A P T E R T H R E E : D E V E L O P I N G A R E S E A R C H S T R AT E G Y . . . . . 4 8
Different types of research
The research process
Choosing a research topic and objectives
Formulating your research questions
Choosing a research design
Quantitative or qualitative methods?
Population- or service-based research
Collaboration between researchers and activists
Drafting the protocol
C H A P T E R F O U R : Q U A N T I TAT I V E A P P R O A C H E S
TO RESEARCH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Cross-sectional surveys
Cohort studies
Case control studies
C H A P T E R F I V E : Q U A L I TAT I V E A P P R O A C H E S T O R E S E A R C H . . . 7 2
Rapid Assessments
In-depth qualitative studies
CHAPTER SIX: THE CHALLENGE OF DEFINING AND

M E A S U R I N G V I O L E N C E I N Q U A N T I TAT I V E R E S E A R C H . . . . . . 84
Estimating the prevalence of violence
The study population: Choosing interview subjects
Definitions of violence
Enhancing disclosure of violence
Common tools for measuring violence
C H A P T E R S E V E N : D E V E L O P I N G A S A M P L I N G S T R AT E G Y . . . 104
Sampling considerations in qualitative studies
Sampling issues in quantitative research surveys
C H A P T E R E I G H T: T O O L S F O R C O L L E C T I N G
Q U A N T I TAT I V E D ATA . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 6
Developing the conceptual framework
Operationalizing the main variables
Formulating your questions
Formatting your questionnaire
Translating the instrument
Pre-testing the instrument
CHAPTER NINE: TOOLS FOR COLLECTING

Q U A L I TAT I V E D ATA . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2 8
Personal interviews
Focus group discussions
Observation
Free listing
Ranking
Pair-wise ranking
Timelines and seasonal calendars
Causal flow analysis
Open-ended stories
Genograms
Circular or Venn diagrams
Community mapping
Role playing
Body mapping
Photo voice
s CWhat
HAPTER TEN: BUILDING YOUR RESEARCH
is special about research on gender-based violence?
TEAM . . . . . . . . 154
Building the research field team

Training fieldworkers
Remuneration of interviewers
CHAPTER ELEVEN: IN THE FIELD . . . . . . . . . . . . . . . . . . . 168

m Organizing the fieldwork
Negotiating community access
Protecting the safety and well-being of respondents and fieldworkers
Finalizing the procedures and instruments
Managing nonresponse
Data quality control
Data entry
C H A P T E R T W E LV E : A N A LY Z I N G Q U A N T I TAT I V E D ATA . . . . . 186

Y Basic analysis of survey data on violence against women
Looking at associations between violence and other variables
Assessing the validity of survey results
Interpreting the results
z CApproaches
H A P T E R T H I R T E E N : A N A LY Z I N G
to qualitative analysis
Q U A L I TAT I V E D ATA . . . . . 202
Data coding
Using a computer for coding and categorizing
Data reduction and data displays
Ensuring rigor in qualitative research
CHAPTER FOURTEEN: FROM RESEARCH TO ACTION . . . . . . 216

o Outreach to key constituencies
Matching your message to your audience
Sharing findings with the community
Reaching beyond your borders
1 AWPOPME NE ND I XI N1S :T RTUHME EWNHT O. V. I.O. L.E N. C. E. A. G. A. I.N. S.T . . . . . . . . . . . . 230
APPENDIX 2: TRAINING EXERCISES FOR INTERVIEWERS . . . 240
APPENDIX 3: SUGGESTED RESOURCES . . . . . . . . . . . . . . . 244

Acknowledgements
T his manual was truly a collective

effort, as can be seen by the extensive list
USA; Srilatha Batliwala, India; Devi
Bhuyan, India; Rebecca Calder, Lesotho;
of contributors. The authors would like to Jacquelyn C. Campbell, USA; Cheng Yimin,
thank the many individuals who gener- China; Desiree Daniels, South Africa; Nisha
ously contributed their time and energy to Dhawan, India; Joy Dladla, South Africa;
enrich this manual over the course of Nata Duvvury, India; Gillian Fawcett,
nearly a decade. Mexico; Rezina Ferdous, Bangladesh;
We are especially indebted to Elizabeth Fariyal Fikree, Pakistan; Liz Frank, Namibia;
Shrader, who drafted an early version of Soledad Gonzalez Montes, Mexico; Mary
this manual as a consultant to the Center Goodwin, USA; Lorelei Goodyear, USA;
for Health and Gender Equity (CHANGE). Nicole Haberland, USA; Muhammad Haj-
Her pioneering research on violence in Yahia, Israel; Penn Handwerker, USA;
Mexico, and later in the Ruta Crítica Study, Pamela Hartigan, USA; Zanele Hlatswayo,
inspired a generation of researchers, and South Africa; Dianne Hubbard, Namibia;
helped inform the early work of the Wanda M. Hunter, USA; Susan Igras, USA;
International Research Network on Surinder K.P. Jaswal, India; Carol Jenkins,
Violence Against Women (IRNVAW). Bangladesh/USA; Peggy Jennings, USA;
This project was initiated by CHANGE. Rachel Jewkes, South Africa; Holly
We are therefore indebted to CHANGE Johnson, Canada; Pumzile Kedama, South
and the current Executive Director, Jodi Africa; Barbara Kenyon, South Africa;
Jacobson, for support of the manual dur- Mpefe Ketihapile, South Africa; Julia Kim,
ing its initial stages. South Africa; Sunita Kishor, USA; Mary
We also wish to acknowledge the indi- Koss, USA; Shubhada Maitra, India;
vidual participants of IRNVAW whose Suzanne Maman, USA; Lorna Martin, South
efforts and critical reflections form the basis Africa; Sandra Martin, USA; Bongiwe
of many of the recommendations con- Masilela, Swaziland; Raymond Matsi, South
tained herein. Africa; Donna McCarraher, USA; Lori
Participants in IRNVAW meetings whose Michau, Tanzania/USA; Claudia Garcia
work and reflections contributed to this Moreno, Mexico/Switzerland; Andrew
manual include: Morrison, USA; Caroline Moser, USA;
Naeema Abrahams, South Africa; Ann Oswaldo Montoya, Nicaragua; Khosi
Adair, USA; Zeinab Abdi Ahmed, Kenya; Mthethwa, Swaziland; Bernadette Muthien;
Cherub Antwi-Nsiah, Ghana; Lilian Artz, Shenaaz Nair, South Africa; Dorothy
South Africa; Jill Astbury, Australia; Safia Nairne, South Africa; Mavis Ndlovu,
Azim, Bangladesh; Suzanna Stout Banwell, Zimbabwe; Mzikazi Nduna, South Africa;
A Practical Guide for Researchers and Activists 1

Sydia Nduna, Tanzania; Erin Nelson, USA; ticularly indebted to Lars Åke Persson and
Nguyen Thi Hoai Duc, Vietnam; Tara Stig Wall for graciously allowing us to draw
Nutley, USA; Kwadzanai Nyanyungo, heavily from their book on epidemiology
Zimbabwe; Maria Beatriz Orlando, USA; and field methods, as well as to Lars
Loveday Penn-Kekana, South Africa; Jasjit Dalgren, Anna Winkvist, and Maria
Purewal, India; Hnin Hnin Pyne, USA; Emmelin for the inspiration we received
Regan Ralph, USA; Juan Carlos Ramirez from their book on qualitative research
Rodriguez, Mexico; Laurie Ramiro, methods.
Philippines; Koketso Rantona, Botswana; Likewise, many of the examples used in
Matsie Ratsaka, South Africa; Heather this manual have come from the WHO
Robinson, USA; Louise Robinson, Lesotho; Multi-country Study on Women’s Health
Mariana Romero, Argentina; Stephanie and Domestic Violence. In particular,
Rosseti, Botswana; Laura Sadowski, USA; Claudia García Moreno, coordinator of the
Irma Saucedo Gonzalez, Mexico; Lynn study, has supported this project from its
Short, USA; Patrick Sikana, Zambia; Sawera inception and has provided invaluable
Singh, South Africa; Pinky Singh Rana, financial, technical, and moral support
Nepal; Paige Hall Smith, USA; Shobha throughout. We are also grateful to Henrica
Srinivasan, India/USA; Shana Swiss, USA; Jansen for permission to use her beautiful
Imani Tafari-Ama, Jamaica; Kathryn Tolbert, photographs from the WHO study. We
Mexico; Tran Anh Vinh, Vietnam; Ian would also like to acknowledge the able
Tweedie, USA; Sandi Tyler, USA; Lisa guidance of the steering committee of the
Vetten, South Africa; Penny Ward, South WHO multi-country study, and the valu-
Africa; Carole Warshaw, USA; Charlotte able contributions of all the researchers
Watts, Zimbabwe/UK; Linda Williams, USA; involved in this study:
Kate Wood, UK; Pamela Wyville-Staples,
South Africa; Mieko Yoshihama, Core Research Team
Japan/USA; Cathy Zimmerman, Claudia Garcia-Moreno, World Health
Cambodia/USA. Organization, Geneva, Switzerland
We have also drawn heavily on experi- (Study Coordinator),
ences and insights emerging from our par- Charlotte Watts, London School of Hygiene
ticipation in research collaborations with and Tropical Medicine, London, UK,
colleagues from the Department of Lori Heise, PATH, Washington, DC, USA,
Epidemiology and Public Health, Umeå Mary Ellsberg, PATH, Washington DC, USA,
University, Sweden, particularly Lars Åke Henrica A.F.M. Jansen, World Health
Persson, Stig Wall, Lars Dalgren, Anna Organization, Geneva, Switzerland
Winkvist, Maria Emmelin, Jerker Liljestrand,
Gunnar Kullgren and Ulf Högberg, as well Country Researchers
as colleagues from Addis Ababa University, Bangladesh: Ruchira Tabassum Naved,
Ethiopia: Yemane Berhane, Negussie ICCDR,B, Dhaka; Safia Azim, Naripokkho,
Deyessa, Yegomawork Gossaye, and Atalay Dhaka; Abbas Bhuiya, ICCDR,B, Dhaka; Lars-
Alem; from Gadjah Mada University and Åke Persson, Uppsala University, Sweden.
Rifka Annisa Women’s Crisis Center in
Yogyakarta, Indonesia: Mohammad Brazil: Lilia Blima Schraiber, University of
Hakimi and Elli Nur Hayati; and from the Sao Paulo – Faculty of Medicine, Sao
Autonomous Nicaraguan University at León Paulo; Ana Flavia Lucas D’Oliveira,
(UNAN-León): Rodolfo Peña, Andrés University of Sao Paulo – Faculty of
Herrera, and Eliette Valladares. We are par- Medicine, Sao Paulo; Ivan Franca-Junior,
2 Researching Violence Against Women

University of Sao Paulo – School of Public Samoa: Tina Tauasosi-Posiulai, Tima Levai-
Health, Sao Paulo; Carmen Simone Grilo Peteru, Dorothy Counts and Chris
Diniz, Feminist Collective for Health and McMurray, Secretariat of the Pacific
Sexuality, Sao Paulo; Ana Paula Portella, Community.
SOS Corpo, Genero e Cidadania, Recife,
Pernambuco; Ana Bernarda Ludermir, Serbia and Montenegro: Stanislava
Federal University of Pernambuco, Recife. Otasevic, Autonomous Women’s Center,
Belgrade; Silvia Koso, Autonomous
Ethiopia: Yemane Berhane, Addis Ababa Women’s Center, Belgrade; Katarina
University, Addis Ababa; Ulf Högberg, Bogavac, Autonomous Women’s Center,
Umeå University, Umeå, Sweden; Gunnar Belgrade; Dragisa Bjeloglav, Strategic
Kullgren, Umeå University, Umeå, Sweden; Marketing, Belgrade; Viktorija Cucic,
Negussie Deyessa, Addis Ababa University, University of Belgrade, Belgrade.
Addis Ababa; Maria Emmelin, Umeå
University, Umeå, Sweden; Yegomawork Thailand: Churnrurtai Kanchanachitra,
Gossaye, Addis Ababa University, Addis Mahidol University, Bangkok; Kritaya
Ababa; Mary Ellsberg, PATH, Washington, Archavanitkul, Mahidol University,
DC, USA; Atalay Alem, Addis Ababa Bangkok; Wassana Im-em, Mahidol
University, Addis Ababa; Derege Kebede, University, Bangkok; Usa Lerdsrisanthat,
Addis Ababa University, Addis Ababa; Foundation for Women, Bangkok.
Alemayehu Negash, Addis Ababa
University, Addis Ababa. United Republic of Tanzania: Jessie
Mbwambo, Muhimbili University College of
Japan: Mieko Yoshihama, University of Health Sciences, Dar es Salaam; Gideon
Michigan, Ann Arbor, USA; Saori Kamano, Kwesigabo, Muhimbili University College
National Institute of Population and Social of Health Sciences, Dar es Salaam; Joe
Security Research, Tokyo; Tamie Kaino, Lugalla, University of New Hampshire,
Ochanomizu University, Tokyo; Fumi Durham, USA; Sherbanu Kassim, University
Hayashi, Toyo Eiwa Women’s University, of Dar es Salaam, Dar es Salaam.
Tokyo; Hiroko Akiyama, University of
Tokyo, Tokyo; Tomoko Yunomae, Japan WHO Steering Committee Members:
Accountability Caucus. Jacquelyn Campbell, Johns Hopkins
University (Co-Chair), Baltimore, USA;
Namibia: Eveline January, Ministry of Lucienne Gillioz, Bureau d’Egalite, Geneva,
Health and Social Services, Windhoek; Switzerland; Irma Saucedo Gonzalez, El
Hetty Rose-Junius, Ministry of Health and Colegio de Mexico, Mexico City, Mexico;
Social Services, Windhoek; Johan Van Wyk, Rachel Jewkes, Medical Research Council,
Ministry of Health and Social Services, Pretoria, South Africa; Ivy Josiah, Women’s
Windhoek; Alves Weerasinghe, National AID Organisation, Selangor, Malaysia; Olav
Planning Commission, Windhoek. Meirik, Instituto Chileno de Medicina
Reproductiva (ICMER) (Co-Chair), Santiago,
Peru: Ana Güezmes Garcia, Centro de la Chile; Laura Rodrigues, London School of
Mujer Peruana Flora Tristan, Lima; Nancy Hygiene and Tropical Medicine, London,
Palomino Ramirez, Universidad Peruana UK; Berit Schei, Norwegian University of
Cayetano Heredia, Lima; Miguel Ramos Science and Technology, Trondheim,
Padilla, Universidad Peruana Cayetano Norway; Stig Wall, Umeå University, Umeå,
Heredia, Lima. Sweden.

We are deeply grateful to all those indi-
viduals who reviewed the manuscript at
different stages of preparation and in
many cases provided extensive comments,
research examples, and field experiences
that greatly enriched the manual. We
are particularly grateful to the following
individuals:
Elizabeth Shrader, Sara Bott, Lori
Michau, Claudia Garcia Moreno, Christine
Bradley, Jacquelyn Campbell, Victoria Frye,
Henrica Jansen, Alessandra Guedes,
Shireen Jeejeebhoy, Erica Keogh, Mary
Koss, Linda Morison, Elaine Murphy,
Monica O’Connor, Vijayendra Rao, Stig
Wall, Rachel Jewkes, Naeema Abrahams,
Charlotte Watts, Michelle Folsom, Mieko
Yoshihama, Anna Winkvist.
We also gratefully acknowledge the serv-
ices of the following individuals and their
support in the production of the book:
Editing by Cheryl Silver.

Graphic Design by Gretchen Maxwell of
GLM Design.
Support in production and editing pro-
vided by Rani Boehlke, Rebeca Quiroga,
and Tricia Klosky at PATH.
Proofreading by Janet Saulsbury at PATH
and Lenore Jackson.
Cover graphic by Liliana Gutierrez Lopez
of Lapiz y Papel, Quito, Ecuador
This manual was produced with the

generous financial support of WHO, the
Ford Foundation, The Moriah Fund,
CHANGE, and The Swedish Agency for
International Development Cooperation
(Sida).

Introduction
T wenty years ago, violence against

women was not considered an issue wor-
international data on physical and sexual
abuse. It outlines some of the method-
thy of international attention or concern. ological and ethical challenges of conduct-
Victims of violence suffered in silence, ing research on violence against women
with little public recognition of their and describes a range of innovative tech-
plight. This began to change in the 1980s niques that have been used to address
as women’s groups organized locally and these challenges. We hope that the man-
internationally to demand attention to the ual will be useful for those interested in
physical, psychological, and economic pursuing research on violence, especially
abuse of women. Gradually, violence in developing countries and other
against women has come to be recognized resource-poor settings.
as a legitimate human rights issue and as a The manual draws on the collective
significant threat to women’s health and experiences and insights of many individu-
well-being. als, most notably the members of the
Now that international attention is International Research Network on
focused on gender-based violence, Violence Against Women (IRNVAW), an ad
methodologically rigorous research is hoc group of researchers and activists that
needed to guide the formulation and meets periodically to share experiences
implementation of effective interventions, regarding research on violence. The
policies, and prevention strategies. Until Network arose out of a two-day meeting
fairly recently, the majority of research on on methodology and research ethics organ-
violence consisted of anecdotal accounts or ized in June 1995 by the Center for Health
exploratory studies performed on nonrep- and Gender Equity. To date the IRNVAW
resentative samples of women, such as has sponsored four international meetings
those attending services for battered and several members have collaborated
women. While this research has played a with the World Health Organization
critical role in bringing to light the issues (WHO) in the design and implementation
of wife abuse, rape, trafficking, incest, and of a multi-country study on women’s
other manifestations of gender-based vio- health and domestic violence. Many of the
lence, it is less useful for understanding the examples and insights included in this
dimensions or characteristics of abuse manual come from the pioneering work of
among the broader population. IRNVAW members, as well as the WHO
This manual has been developed in Multi-country Study on Women’s Health
response to the growing need to improve and Domestic Violence, a household sur-
the quality, quantity, and comparability of vey of women that has been conducted in

at least ten countries to date. We have also The manual is directed particularly to
drawn extensively from our own research those researchers interested in the intersec-
experiences, primarily in Nicaragua, tion of violence and health in develop-
Indonesia, and Ethiopia. ing countries, given the clear impact that
gender violence has on women’s health
Readership status. However, much of the information
This manual is written for anyone inter- presented in the manual is applicable to
ested in the application of social science violence research as it relates to other
and public health research methods to the issues, such as human rights, the well-
study of gender-based violence. The man- being of families and children, and eco-
ual assumes a certain level of familiarity nomic development. Similarly, the lessons
with the logic of research and is not a sub- for developing countries may be relevant
stitute for training in research or research to some violence research undertaken in
methodologies. industrialized countries, particularly among
It is designed for researchers who economically marginalized and/or politi-
want to know more about adapting tradi- cally disenfranchised populations.
tional research techniques to the special Finally, the manual advances an ethic of
case of investigating physical and sexual research that is action-oriented, account-
abuse. And it is designed for activists, able to the antiviolence movement, and
community workers, and service responsive to the needs of women living
providers who want to become conver- with violence. It strongly encourages col-
sant in methodological issues. One of the laboration between researchers and those
goals of this manual is to facilitate collabo- working directly on violence as activists
rations between researchers and commu- and/or practitioners. Recent experiences in
nity-based workers and activists by countries as diverse as Canada, Zimbabwe,
providing practitioners with an introduction Indonesia, South Africa, Nicaragua, and
to the tools and language of research, and Cambodia have shown that powerful syn-
by giving researchers greater insight into ergies can be achieved from partnerships
the specific issues that accompany research between researchers and advocates.
on violence. Whereas researchers help to ensure that
the endeavor is grounded in the principles
Focus of the manual of scientific inquiry, the involvement of
For the sake of brevity, this manual advocates and service providers helps
focuses primarily on the issue of violence ensure that the right questions are asked in
against women by their intimate part- the right way, and that the knowledge gen-
ners. Gender-based violence assumes erated is used for social change.
many forms, including rape, sexual assault
and coercion, stalking, incest, sexual
harassment, female genital mutilation, and
trafficking in women. Although many of
the insights presented herein will apply to
these other types of violence, no single
manual could exhaustively address all
forms of abuse. Additionally, we concen-
trate on applied research, as opposed to
research designed to advance theory or to
address questions of primarily academic
relevance.

This document was based on the contributions of thousands of women from around the
world who shared their stories and personal experiences in the hopes that their voices
would contribute to diminishing the suffering of future generations of women from
violence. The publication is dedicated to them.

1
PHOTO BY HAFM JANSEN
CHAPTER ONE
Violence Against Women as a

Health and Development Issue*
Topics covered in this chapter:
Definitions of violence against women

Prevalence of intimate partner violence
The patterning of intimate partner violence
Prevalence and characteristics of sexual coercion and abuse
The effects of violence on women’s health
Explaining gender-based violence
How do women respond to abuse?
Challenges for international research on gender-based violence
V iolence against women is the most per-

vasive yet underrecognized human rights vio-
society in the world legitimize, obscure, and
deny abuse. The same acts that would be
lation in the world. It is also a profound punished if directed at an employer, a neigh-
health problem that saps women’s energy, bor, or an acquaintance often go unchal-
compromises their physical and mental lenged when men direct them at women,
health, and erodes their self-esteem. In addi- especially within the family.
tion to causing injury, violence increases For over three decades, women’s advo-
women’s long-term risk of a number of other cacy groups around the world have been
health problems, including chronic pain, working to draw more attention to the
physical disability, drug and alcohol abuse, physical, psychological, and sexual abuse
and depression.1, 2 Women with a history of of women and to stimulate action. They
physical or sexual abuse are also at increased have provided abused women with shelter,
risk for unintended pregnancy, sexually lobbied for legal reforms, and challenged
transmitted infections, and miscarriages.3-5 the widespread attitudes and beliefs that
Despite the high costs of violence against support violence against women.2
women, social institutions in almost every Increasingly, these efforts are having
* Parts of this chapter are reprinted from Heise, Ellsberg and Gottemoeller, 19992 (available online at
http://www.infoforhealth.org/pr/l11edsum.shtml).

CHAPTER ONE
FIGURE 1.1 THE LIFE CYCLE OF VIOLENCE AGAINST WOMEN
Pre-birth Elderly
Elder/widow abuse
Sex-selective abortion
Reproductive Age
Honor killing
Infancy Dowry killing
Female infanticide Intimate partner violence
Neglect (health care, Sexual assault by non-partner
nutrition) Homicide/Femicide
Sex trafficking
Sexual harassment
Childhood Adolescence
Child abuse Forced prostitution
Malnutrition Trafficking
FGM Forced early marriage
Psychological abuse
Rape
(Adapted from Watts and Zimmerman, 20026 and Shane and Ellsberg, 2002.7)
results. Today, international institutions are the characteristics of violence most com-
speaking out against gender-based vio- monly committed against women differ in
lence. Surveys and studies are collecting critical respects from violence commonly
more information about the prevalence and committed against men. Men are more
nature of abuse. More organizations, serv- likely to be killed or injured in wars or
ice providers, and policy makers are recog- youth- and gang-related violence than
nizing that violence against women has women, and they are more likely to be
serious adverse consequences for women’s physically assaulted or killed on the street
health and for society. by a stranger. Men are also more likely to
This chapter provides a brief overview be the perpetrators of violence, regardless
of the issue of violence against women, of the sex of the victim.1 In contrast,
including definitions, international preva- women are more likely to be physically
lence, the documented health conse- assaulted or murdered by someone they
quences of abuse, and evidence regarding know, often a family member or intimate
causation and women’s experiences of partner.2 They are also at greater risk of
abuse. We include this information here for being sexually assaulted or exploited,
individuals who may be new to the topic either in childhood, adolescence, or as
and/or for those who are writing research adults. Women are vulnerable to different
proposals and may not have easy access to types of violence at different moments in
the international literature. their lives (see Figure 1.1).
There is still no universally agreed-upon
DEFINITIONS OF VIOLENCE terminology for referring to violence against
AGAINST WOMEN women. Many of the most commonly used
terms have different meanings in different
Although both men and women can be regions, and are derived from diverse theo-
victims as well as perpetrators of violence, retical perspectives and disciplines.

VIOLENCE AGAINST WOMEN AS A HEALTH AND DEVELOPMENT ISSUE
One frequently used model for under- BOX 1.1 UNITED NATIONS DEFINITION OF VIOLENCE AGAINST WOMEN
standing intimate partner abuse and sexual
abuse of girls is the “family violence” The term “violence against women” means any act of gender-based violence that
results in, or is likely to result in, physical, sexual or psychological harm or suffer-
framework, which has been developed pri- ing to women, including threats of such acts, coercion or arbitrary deprivation of
marily from the fields of sociology and liberty, whether occurring in public or private life. Accordingly, violence against
psychology.8, 9 “Family violence” refers to women encompasses but is not limited to the following:
all forms of abuse within the family regard- a) Physical, sexual and psychological violence occurring in the family, including
less of the age or sex of the victim or the battering, sexual abuse of female children in the household, dowry-related vio-
lence, marital rape, female genital mutilation and other traditional practices
perpetrator. Although women are fre- harmful to women, non-spousal violence and violence related to exploitation;
quently victimized by a spouse, parent, or b) Physical, sexual and psychological violence occurring within the general com-
other family member, the concept of “fam- munity, including rape, sexual abuse, sexual harassment and intimidation at
ily violence” does not encompass the many work, in educational institutions and elsewhere, trafficking in women and forced
types of violence to which women are prostitution;
exposed outside the home, such as sexual c) Physical, sexual and psychological violence perpetrated or condoned by the
State, wherever it occurs.
assault and harassment in the workplace.
Acts of violence against women also include forced sterilization and forced
Moreover, feminist researchers find the
abortion, coercive/forced use of contraceptives, female infanticide and prenatal
assumption of gender neutrality in the term sex selection.
“family violence” problematic because it
(From United Nations, 1993.10)
fails to highlight that violence in the family
is mostly perpetrated by men against
women and children. by current or former male intimate part-
There is increasing international consen- ners.11, 12 However, in some regions, includ-
sus that the abuse of women and girls, ing Latin America, “domestic violence”
regardless of where it occurs, should be refers to any violence that takes place in
considered as “gender-based violence,” as it the home, including violence against chil-
largely stems from women’s subordinate sta- dren and the elderly.13, 14 The term “bat-
tus in society with regard to men (Figure tered women” emerged in the 1970s and is
1.2). The official United Nations definition widely used in the United States and
of gender-based violence was first presented Europe to describe women who experi-
in 1993 when the General Assembly passed ence a pattern of systematic domination
the Declaration on the Elimination of and physical assault by their male part-
Violence against Women.10 According to this ners.15 The terms “spouse abuse,” “sexual-
definition, gender-based violence includes a ized violence,” “intimate partner violence,”
host of harmful behaviors directed at and “wife abuse” or “wife assault” are gen-
women and girls because of their sex, erally used interchangeably, although each
including wife abuse, sexual assault, dowry- term has weaknesses. “Spouse abuse” and
related murder, marital rape, selective mal- “intimate partner violence” do not make
nourishment of female children, forced explicit that the victims are generally
prostitution, female genital mutilation, and women, whereas “wife abuse” and “wife
sexual abuse of female children (see Box assault” can be read to exclude common-
1.1 for the complete definition).10 law unions and dating violence.
Even when the abuse of women by For the purposes of this manual, we use
male partners is conceptualized as gender- the terms “violence against women” (VAW)
based violence, the terms used to describe and “gender-based violence” (GBV) inter-
this type of violence are not consistent. In changeably to refer to the full range of
many parts of the world, the term “domes- abuses recognized by the UN Declaration
tic violence” refers to the abuse of women and other international agreements. We use

CHAPTER ONE
FIGURE 1.2 THE OVERLAP BETWEEN GENDER-BASED VIOLENCE AND Researchers find considerable variation
FAMILY/DOMESTIC VIOLENCE in the prevalence of partner violence from
country to country, and among studies
within a country. Unfortunately, lack of
Gender-based consistency in study methods, study
violence Family violence design, and presentation of results makes it
Intimate For example: difficult to explore the causes and conse-
For example:
partner ■ Child abuse quences of violence. As a result, it is often
■ Rape by strangers
violence ■ Elder abuse
■ Female genital difficult to compare results even between
mutilation Sexual abuse studies performed in the same country.
■ Sexual harassament of women
in the workplace
Partly to address this shortcoming, the
and girls
■ Selective malnutrition in the family
World Health Organization worked with
of girls collaborating institutions in 15 sites in ten
countries between 1998 and 2004 to imple-
ment a multi-country study of domestic
violence and women’s health. The WHO
Multi-country Study on Women’s Health
and Domestic Violence Against Women—
the terms “intimate partner violence,” “wife also referred to here as the WHO VAW
abuse” and “domestic violence” inter- Study—was the first ever to produce truly
changeably to refer to the range of sexu- comparable data on physical and sexual
ally, psychologically, and physically abuse across settings.16 This research proj-
coercive acts used against adult and ado- ect sought to minimize differences related
lescent women by current or former male to methods by employing standardized
intimate partners. questionnaires and procedures, as well as a
common approach to interviewer training.
P R E VA L E N C E O F I N T I M AT E We will return to the WHO VAW Study
PA R T N E R V I O L E N C E many times throughout the manual to
highlight some of the challenges posed by
International research consistently demon- this project and how they were resolved.
strates that a woman is more likely to be
assaulted, injured, raped, or killed by a cur- T H E PAT T E R N I N G O F
rent or former partner than by any other I N T I M AT E PA R T N E R
person. Table 1.1 presents findings from VIOLENCE
nearly 80 population-based studies carried
out in more than 50 countries. These studies The WHO VAW Study also provided a rare
indicate that between 10 percent and 60 per- opportunity to examine the “patterning” of
cent of women who have ever been married violence across settings. Does physical vio-
or partnered have experienced at least one lence occur together with other types of
incident of physical violence from a current violence? Do violent acts tend to escalate
or former intimate partner. Most studies esti- over time? Are women most at risk from
mate a lifetime prevalence of partner vio- partners or from others in their lives?
lence between 20 percent and 50 percent. The WHO VAW Study findings confirm
Although women can also be violent, and that most women who suffer physical or
abuse exists in some same-sex relationships, sexual abuse by a partner generally experi-
the vast majority of partner abuse is perpe- ence multiple acts over time. Likewise, phys-
trated by men against their female partners. ical and sexual abuse tend to co-occur in

TABLE 1.1 PHYSICAL ASSAULTS ON WOMEN BY AN INTIMATE MALE PARTNER, SELECTED POPULATION-BASED STUDIES, 1982–2004
Proportion of women
physically assaulted
Year of Sample Study* by a partner (%)
Country Ref study Coverage size population* Age (years) last 12 mo Ever
Africa
Ethiopia ◆ 17
2002 Meskanena Woreda 2261 III 15–49 29 49
Kenya 18
1984–87 Kisii District 612 V >15 42d
■
19
2003 National 3856 III 15–49 24 40
Namibia ◆ 20
2002 Winhoek 1367 III 15–49 16 31
South Africa 21
1998 Eastern Cape 396 III 18–49 11 27
1998 Mpumalanga 419 III 18–49 12 28
1998 Northern Province 464 III 18–49 5 19
■
22
1998 National 10,190 II 15–49 6 13
Tanzania ◆ 20
2002 Dar es Salaam 1442 III 15–49 15 33
◆ 20
2002 Mbeya 1256 III 15–49 19 47
Uganda 23
1995–1996 Lira & Masaka 1660 II 20–44 41d
Zambia ■
24
2001–2002 National 3792 III 15–49 27 49
Zimbabwe 25
1996 Midlands Province 966 I >18 17b
Latin America and the Caribbean
Barbados 26
1990 National 264 I 20–45 30a,c
Brazil ◆ 20
2001 Sao Paulo 940 III 15–49 8 27
◆ 20
2001 Pernambuco 1188 III 15–49 13 34
Chile 27
1993 Santiago Province 1000 II 22–55 26d
28
1997 Santiago 310 II 15–49 23
● 29
2004p Santa Rosa 422 IV 15–49 4 25
Colombia ■
30
1995 National 6097 II 15–49 19d
■
31
2000 National 7602 III 15–49 3 44
Dominican Republic ■
24
2002 National 6807 III 15–49 11 22
Ecuador ▲ 32
1995 National 11,657 II 15–49 12
El Salvador ▲ 33
2002 National 10,689 III 15–49 6 20d
Guatemala ▲ 34
2002 National 6595f VI 15–49 9
Honduras ▲ 35
2001 National 6827 VI 15–49 6 10
Haiti ■
24
2000 National 2347 III 15–49 21 29
Mexico 36
1996 Guadalajara 650 III >15 27
37
1996 p
Monterrey 1064 III >15 17
38
2003 National 34,184 II >15 9
Nicaragua 39
1995 Leon 360 III 15–49 27 52
40
1997 Managua 378 III 15–49 33 69
■
41
1998 National 8507 III 15–49 13 30
Paraguay ▲ 42
1995–1996 National 5940 III 15–49 10
▲ 43
2004 National 5070 III 15–44 7 19
Peru ■
24
2000 National 17,369 III 15–49 2 42

CHAPTER ONE
Proportion of women
Latin America and the Caribbean (continued)

Peru (continued) ◆ 20
2001 Lima 1019 III 15–49 17 50
◆ 20
2001 Cusco 1497 III 15–49 25 62
Puerto Rico ▲ 44
1995–1996 National 4755 III 15–49 13e
Uruguay 45
1997 National 545 IIk 22–55 10c
North America
Canada 46
1993 National 12,300 I >18 3b,c 29b,c
47
1999 National 8356 III >15 3 8g
United States 48
1995–1996 National 8000 I >18 1a 22a
Asia and Western Pacific
Australia *
49
1996 National 6300 I 3b 8b,d
50
2002–2003 National 6438 III 18–69 3 31
Bangladesh 51
1992 National (villages) 1225 II <50 19 47
52
1993 Two rural regions 10,368 II 15–49 42d
◆ 20
2003 Dhaka 1373 III 15–49 19 40
◆ 20
2003 Matlab 1329 III 15–49 16 42
Cambodia 53
1996 Six regions 1374 III 15–49 16
■
24
2000 National 2403 III 15–49 15 18
China 54
1999–2000 National 1665 II 20–64 15
India ■
24
1998–1999 National 90,303 III 15–49 10 19
53
1999 Six states 9938 III 15–49 14 40
● 29
2004 p
Lucknow 506 IV 15–49 25 35
● 29
2004 p
Trivandrum 700 IV 15–49 20 43
● 29
2004 p
Vellore 716 IV 15–49 16 31
Indonesia 55
2000 Central Java 765 IV 15–49 2 11
Japan ◆ 20
2001 Yokohama 1276 III 18–49 3 13
New Zealand ◆ 56
2002 Auckland 1309 III 18–64 5 30
◆ 56
2002 North Waikato 1360 III 18–64 34
Papua New Guinea 57
1982 National, rural villages 628 IIIk 67
Philippines ■
58
1993 National 8481 IV 15–49 10
59
1998 Cagayan de Oro 1660 II 15–49 26
City & Bukidnon
● 29
2004p Paco 1000 IV 15–49 6 21
Republic of Korea 60
1989 National 707 II >20 38
Samoa ◆ 20
2000 National 1204 III 15–49 18 41
Thailand ◆ 20
2002 Bangkok 1048 III 15–49 8 23
20
2002 Nakonsawan 1024 III 15–49 13 34
Vietnam 61
2004 Ha Tay province 1090 III 15–60 14 25

Proportion of women
Europe
Albania ▲ 62
2002 National 4049 III 15–44 5 8
Azerbaijan ▲ 63
2001 National 5533 III 15–44 8 20
Finland *
64
1997 National 4955 I 18–74 30
France *
65
2002 National 5908 II >18 3 9i
Georgia ▲ 66
1999 National 5694 III 15–44 2 5
Germany *
67
2003 National 10,264 III 16–85 23b
Lithuania *
68
1999 National 1010 II 18–74 42b,d,h
Netherlands 69
1986 National 989 I 20–60 21a
Norway 70
1989 Trondheim 111 III 20–49 18
*
71
2003 National 2143 III 20–56 6 27
Republic of Moldova ▲ 72
1997 National 4790 III 15–44 8 15
Romania ▲ 73
1999 National 5322 III 15–44 10 29
Russia ▲ 74
2000 Three provinces 5482 III 15–44 7 22
Serbia/Montenegro ◆ 20
2003 Belgrade 1189 III 15–49 3 23
Sweden *
75
2000 National 5868 III 18–64 4e 18e
Switzerland 76
1994–1996 National 1500 II 20–60 6c 21c
*
77
2003 National 1882 III >18 10
Turkey 78
1998 E & SE Anatolia 599 I 14–75 58a
Ukraine ▲ 79
1999 National 5596 III 15–44 7 19
United Kingdom 80
1993 p
North London 430 I >16 12 a
30a
81
2001 National 12,226 I 16–59 3 19j
Eastern Mediterranean
Egypt ■
82
1995–1996 National 7123 III 15–49 13 34
● 29
2004p El–Sheik Zayed 631 IV 15–49 11 11
Israel 83
1997 Arab population 1826 II 19–67 32
West Bank and
Gaza Strip 84
1994 Palestinian population 2410 II 17–65 52
Key ■ DHS survey data24 ● INCLEN data85 ▲ CDC study ◆ WHO study20 * International Violence Against Women (IVAWS) Study
* Study population: I = all women; b

Although sample included all women, rate of g
Within the last five years.
II = currently married/partnered women; abuse is shown for ever–married/partnered h
Includes threats.
III = ever–married/partnered women; women (number not given).
IV = women with a pregnancy outcome;
i
Since the age of 18.
c
Physical or sexual assault.
V = married women – half with pregnancy j
Since the age of 16.
outcome, half without; VI women who had a
d
During current relationship.
k
Nonrandom sampling methods used.
partner within the last 12 months. e
Rate of partner abuse among ever–
married/partnered women recalculated
p
Publication date (field work dates not
a
Sample group included women who had
from authors’ data. reported).
never been in a relationship and therefore
were not in exposed group. f
Weighted for national representativity.
(Updated from Heise et al, 1999.2)

CHAPTER ONE
FIGURE 1.3 PREVALENCE OF PHYSICAL VIOLENCE AND/OR SEXUAL PARTNER VIOLENCE IN TEN COUNTRIES
Percent
80 -
71 69
70 -
62 62
60 - 59
56
53 52
50 49 50
50 - 47 47 47 46
42 41 41 41
40
40 - 37 37 36
34 34 33
31 30 31
29 29
30 - 27
23 23 23 23 24
20
20 - 17
14 13 15
10
10 - 6 6
0-
y
ce
ce
ce
ity
ity
ce
rbia
a
Cit
Cit
Cit
Cit
Cit
inc
inc
mo
uC
dC
vin
vin
vin
vin
Se
rov
rov
Sa
esh
zil
an
ia
ia
Pro
Pro
Pro
Pro
Per
ilan
mib
zan
Bra
uP
dP
Jap
lad
esh
zil
ia
ia
Tha
Per
Na
Tan
ilan
iop
zan
Bra
ng
lad
Ba
Tha
Eth
Tan
ng
Ba
■ % women who have ever experienced physical violence by a partner

■ % women who have ever experienced sexual violence by a partner
■ % women who have ever experienced physical and/or sexual violence by a partner
(From WHO, 2005.20)
many relationships. Figure 1.3 summarizes suggest that, although this pattern is main-
the proportion of women who have experi- tained in many countries, a few sites
enced violence by an intimate partner demonstrate a significant departure. In
among ever-partnered women aged 15 to 49 both the capital and province of Thailand,
in the various sites included in the study. a substantial portion of women who expe-
The first bar portrays the percentage of rience partner violence, experience sexual
women in each setting who have experi- violence only (Figure 1.4). In Bangkok, 44
enced physical violence by a partner; the percent of all cases of lifetime partner vio-
second bar portrays sexual violence by a lence have experienced only sexual vio-
partner; and the third bar represents the per- lence. The corresponding statistic in the
centage of ever-partnered women who have Thai province is 29 percent of cases. A
experienced either physical and/or sexual similarly high percentage of cases of vio-
violence by a partner in their lifetime. lence in Bangladesh province (32 percent)
Until recently, it was believed that few and Ethiopia province (31 percent) involve
women exclusively experienced sexual vio- sexual violence only.
lence by an intimate partner. Available stud- These results speak to the importance of
ies from North and Central America had developing a broader international research
indicated that sexual violence was generally base on violence against women. Insights
accompanied by physical abuse and by emo- derived exclusively from the North
tional violence and controlling behaviors.2 American literature may not reflect the real-
The findings from the WHO VAW Study ity of women’s experiences in other settings.

FIGURE 1.4 INTIMATE PARTNER VIOLENCE ACCORDING TO TYPES OF

VIOLENCE (AMONG EVER-ABUSED WOMEN)
Percent
100 -
29 30 25 32 28 32 35 23 30
80 - 44 48 39 39
51 57
60 - 28
39 45
40 - 30 19 62 54 45 73 58
65 61 17 56
32
20 - 44
26 32 31 29 21
6 13 14 4 11 17 4 12
0- 8
y
ce
ce
ce
ity
ity
ce
rbia
a
Cit
Cit
Cit
Cit
Cit
inc
inc
mo
vin
vin
vin
uC
dC
vin
Se
rov
rov
esh
zil
an
ia
Sa
ia
Pro
Pro
Pro
Pro
Per
ilan
mib
Bra
zan
Jap
uP
dP
lad
esh
zil
ia
ia
Tha
Na
Per
Tan
ilan
iop
Bra
ng
zan
lad
Ba
Eth
Tha
Tan
ng
Ba
■ sexual violence only ■ physical violence only ■ sexual and physical violence
(From WHO, 2005.20)
P R E VA L E N C E A N D Much sexual coercion also takes place

CHARACTERISTICS against children and adolescents in both
OF SEXUAL COERCION industrial and developing countries.
AND ABUSE Between one-third and two-thirds of
known sexual assault victims are age 15 or
For many women and girls, sexual coercion younger, according to justice system statis-
and abuse are defining features of their tics and information from rape crisis cen-
lives. Forced sexual contact can take place ters in Chile, Peru, Malaysia, Mexico,
at any time in a woman’s life and includes Panama, Papua New Guinea, and the
a range of behaviors, from forcible rape to United States.2
nonphysical forms of pressure that compel Sexual exploitation of children is wide-
girls and women to engage in sex against spread in virtually all societies. Child sex-
their will. The touchstone of coercion is ual abuse refers to any sexual act that
that a woman lacks choice and faces severe occurs between an adult or older adoles-
physical, social, or economic consequences cent and a child, and any nonconsensual
if she resists sexual advances. sexual contact between a child and a peer.
Studies indicate that the majority of non- Laws generally consider the issue of con-
consensual sex takes place among individu- sent to be irrelevant in cases of sexual
als who know each other—spouses, family contact by an adult with a child, defined
members, dating partners, or acquain- variously as someone under 13, 14, 15, or
tances.86, 87 In fact, much nonconsensual sex 16 years of age.
takes place within consensual unions and Because of the taboo nature of the
includes a woman being compelled to have topic, it is difficult to collect reliable figures
sex when she does not want it, or to on the prevalence of sexual abuse in child-
engage in types of sexual activity that she hood. Nonetheless, the few representative
finds degrading or humiliating.1, 88, 89 sample surveys provide cause for concern.

CHAPTER ONE
A review of 25 studies worldwide indicates Rwanda, Liberia, Sierra Leone, and

that 0 to 32 percent of women report that Uganda.94-96 These reports have highlighted
they experienced sexual abuse in child- the extent to which rape has been used as
hood (see Table 1.2). Although both girls a deliberate strategy to “destabilize popula-
and boys can be victims of sexual abuse, tion, advance ethnic cleansing, express
most studies report that the prevalence of hatred for the enemy or supply combatants
abuse among girls is at least 1.5 to 3 times with sexual services.”96 In 2002, the
higher than among boys.90 Abuse among International Criminal Tribunal in The
boys may be underreported compared with Hague recognized the seriousness of sex-
abuse among girls, however. ual offences in war as a crime against
Further data reveal that coercion may be humanity. International relief agencies are
an element in many young girls’ initiation also calling attention to the precarious situ-
into sexual life. An increasing number of ation of women in refugee settings where
studies have begun to document that a rape, child sexual abuse, intimate partner
substantial number of young women’s violence, and other forms of sexual
first sexual experiences are forced or exploitation are widespread.
unwanted, especially among younger ado-
lescents. Table 1.3 summarizes data from a THE EFFECTS OF VIOLENCE
number of population-based surveys on O N W O M E N ’ S H E A LT H
the prevalence of forced first sex, including
data emerging from the WHO VAW Study. Gender-based violence is associated with
A plethora of studies now confirm that the serious health problems affecting both
younger a girl is when she first has sex, women and children, including injuries,
the more likely she is to report her sexual gynecological disorders, mental health dis-
debut as forced.91 orders, adverse pregnancy outcomes, and
Trafficking in women and girls for sexually transmitted infections (STIs)
forced labor and sexual exploitation is (Figure 1.5). Violence can have direct con-
another type of gender-based violence that sequences for women’s health, and it can
has grown rapidly during the past decade, increase women’s risk of future ill health.
largely as a result of war, displacement, Therefore, victimization, like tobacco or
and economic and social inequities alcohol use, can best be conceptualized as
between and within countries. Although a risk factor for a variety of diseases and
reliable statistics on the number of women conditions, rather than primarily as a health
and children who are trafficked are lack- problem in and of itself.2, 4
ing, rough estimates suggest that from Both population-based research and
700,000 to 2 million women and girls are studies of emergency room visits in the
trafficked across international borders every United States indicate that physical abuse is
year.6, 92, 93 These women face many risks, an important cause of injury among
including physical violence and rape, both women.97 Documented injuries sustained
in their work and when trying to negotiate from such physical abuse include contu-
safer-sex practices. sions, concussions, lacerations, fractures,
Another aspect of gender-based violence and gunshot wounds. Population-based
that has been largely overlooked until studies indicate that 40 to 75 percent of
recently is violence against women in situ- women who are physically abused by a
ations of armed conflict. Recent reports partner report injuries due to violence at
have documented systematic rape in many some point in their life.2
conflicts, including the former Yugoslavia, Nevertheless, injury is not the most

TABLE 1.2 PREVALENCE OF CHILD SEXUAL ABUSE: SELECTED STUDIES, 1990–2003
Definition of Child
Country & Year (Ref. No.) Study Method & Sample Sexual Abuse Prevalence
Australia 1997 98 ■ Retrospective study of 710 ■ Sexual contact before the age ■ 20% of women report abuse
women of 12 with perpetrator 5+
years older; or unwanted sex-
ual activity at ages 12-16
Bangladesh 2002 20 ■ Population-based survey of ■ Unwanted sexual activity, con- ■ In Dhaka 7% of women; in
women ages 15–49 (Dhaka tact and noncontact before the Matlab 1% of women report
1602, Matlab 1527) age of 15 abuse
Barbados 1993 99 ■ National random sample of ■ Sexual contact that is ■ 30% of women report abuse
264 women unwanted or with a biological
relative; or before the age of
16 with perpetrator 5+ years
older
Brazil 2002 20 ■ Population-based survey of ■ Unwanted sexual activity, con- ■ In Sao Paulo 8% of women; in
women ages 15–49 (Sao tact and noncontact before the Pernambuco 6% of women
Paulo 1172, Pernambuco age of 15 report abuse
1473)
Canada 1990 100 ■ Population survey of 9953 ■ Unwanted sexual activity, con- ■ 13% of women, 4% of men
men and women age 15+ tact and noncontact, while report abuse
growing up
Costa Rica 1992 101 ■ Retrospective survey of ■ Unwanted sexual activity, con- ■ 32% of women, 13% of men
university students tact and noncontact; no ages report abuse
specified
Ethiopia 2002 20 ■ Population-based survey of ■ Unwanted sexual activity, con- ■ 0.2% of women report abuse
3014 women ages 15–49 tact and noncontact before the
age of 15
Germany 1992 102 ■ Multiple-screen questionnaire ■ Distressing sexual activity, con- ■ In Würzburg 16% of girls, 6%
answered by 2,151 students tact and noncontact, before boys; in Leipzig 10% of girls,
in Würzburg and Leipzig the age of 14; or with perpe- 6% of boys report abuse
trator 5+ years older
Japan 2002 20 ■ Population-based survey of ■ Unwanted sexual activity, con- ■ 10% of women report abuse
age of 15
Malaysia 1996 103 ■ Retrospective self-administered ■ Vaginal or anal penetration, or ■ 8% of women, 2% of men
questionnaire answered by unsolicited sexual contact, or report abuse
616 paramedical students witnessing exhibitionism before
the age of 18
Namibia 2002 20 ■ Population-based survey of ■ Unwanted sexual activity, con- ■ 5% of women report abuse
age of 15
New Zealand 1997 104 ■ Birth cohort of 520 girls, ■ Unwanted sexual activity, con- ■ 14% of girls report contact
studied from birth to age 18 tact and noncontact, before abuse; 17% report any abuse
the age of 16
Nicaragua 1997 105 ■ Anonymous self-administered ■ Sexual contact, including ■ 26% of women, 20% of men
questionnaire answered by attempted penetration, before report abuse
134 men and 202 women the age of 13 with perpetrator
ages 25–44 drawn from 5+ years older; or nonconsen-
population-based sample sual activity over the age of 12
Norway (Oslo) 1996 106 ■ Population-based sample of ■ Sexual contact, including ■ 17% of girls, 1% of boys
465 adolescents, ages "intercourse after pressure," report abuse
13–19, followed for 6 years occurring between a child
before the age of 13 and an
adult over the age of 17; or
involving force

CHAPTER ONE
TABLE 1.2 PREVALENCE OF CHILD SEXUAL ABUSE: SELECTED STUDIES, 1990–2003
Definition of Child
Country & Year (Ref. No.) Study Method & Sample Sexual Abuse Prevalence
Peru 2002 20 ■ Population-based survey of ■ Unwanted sexual activity, con- ■ In Lima 20% of women; in
women ages 15–49 (Lima tact and noncontact before the Cusco 8% of women report
1414, Cusco 1837) age of 15. abuse
Samoa 2000 20 ■ Population-based survey of ■ Unwanted sexual activity, con- ■ 2% of women report abuse
age of 15.
Serbia & Montenegro 2003 20 ■ Population-based survey of ■ Unwanted sexual activity, con- ■ 2% of women report abuse
age of 15.
Spain 1995 107 ■ Face-to-face interviews and ■ Unwanted sexual activity, con- ■ 22% of women and 15% of
self-administered question- tact and noncontact before the men report abuse
naires answered by 895 age of 17.
adults ages 18–60
Switzerland (Geneva) 1996 108 ■ Self-administered questionnaire ■ Unwanted sexual activity, con- ■ 20% of girls, 3% of boys
answered by 1193 9th grade tact and noncontact. report contact abuse; 34% of
students girls, 11% of boys report any
abuse
Switzerland (National) 1998 109 ■ National survey of 3993 girls, ■ “Sexual victimization,” defined ■ 19% of girls report abuse
ages 15–20, enrolled in as “when someone in your
schools or professional training family, or someone else,
programs touches you in a place you
didn’t want to be touched, or
does something to you sexu-
ally which they shouldn't have
done.”
Thailand 2002 20 ■ Population-based survey of ■ Unwanted sexual activity, con- ■ In Bangkok 7.6% of women;
women ages 15–49 tact and noncontact before the in Nakhonsawan 4.7% of
(Bangkok 1534, age of 15. women report abuse
Nakhonsawan 1280)
Tanzania 2002 20 ■ Population-based survey of ■ Unwanted sexual activity, con- ■ In Dar es Salaam 4% of
women ages 15–49 (Dar es tact and noncontact before the women; in Mbeya 4% of
Salaam 1816, Mbeya 1443) age of 15. women report abuse
United States 1997 110 ■ National 10-year longitudinal ■ Unwanted sexual activity, con- ■ 21% of women report abuse
study of women's drinking that tact and noncontact, before
included questions about sex- the age of 18; or before the
ual abuse, answered by age of 13 with perpetrator 5+
1099 women years older.
United States (Midwest) 1997 111 ■ Self-administered questionnaire ■ “Sexual abuse,” defined as ■ 12% of girls, 4% of boys
answered by 42,568 students “when someone in your family report abuse
in grades 7–12 or another person does sexual
things to you or makes you do
sexual things to them that you
don’t want to do.”
United States (Washington State) ■ Multiple-choice survey of ■ “Sexual abuse,” defined as ■ 23% of all girls; 18% of 8th
1997 112 3128 girls in grades 8,10 “when someone in your family graders, 24% of 10th
and 12 or someone else touches you graders, 28% of 12th
in a sexual way in a place graders report abuse
you didn’t want to be
touched, or does something to
you sexually which they
shouldn’t have done.”
(Updated and adapted from Heise et al, 19992 and WHO, 2002.1)

TABLE 1.3 PERCENTAGE OF MEN AND WOMEN REPORTING FORCED SEXUAL INITIATION:
SELECTED POPULATION-BASED SURVEYS, 1993—2003
Percentage reporting
Age first sexual intercourse
Country or Study Sample Group as forced
Area Population Year Size (years) Females Males
Bangladesh Dhaka 2002 1369 15–49 24

Bangladesh Matlab 2002 1326 15–49 30
Brazil Sao Paulo 2002 1051 15–49 3
Brazil Pernambuco 2002 1234 15–49 4
Cameroon Bamenda 1995 646 12–25 37 30
Caribbean Nine countries 1997–1998 15,695 10–18 48 32
Ethiopia Gurage 2002 2238 15–49 17
Ghana Three urban towns 1996 750 12–24 21 5
Japan Yokohama 2002 1116 15–49 0
Mozambique Maputo 1999 1659 13–18 19 7
Namibia Windhoek 2002 1357 15–49 2
New Zealand Dunedin 1993–1994 935 Birth cohort 7 0
Peru Lima 1995 611 16–17 40 11
Peru Lima 2002 1103 15–49 7
Peru Cusco 2002 1557 15–49 24
Samoa National 2002 1317 15–49 8
Serbia & Montenegro Belgrade 2002 1310 15–49 1
South Africa Transkei 1994–1995 1975 15–18 28 6
Tanzania Dar es Salaam 2002 1556 15–49 14
Tanzania Mbeya 2002 1287 15–49 17
Tanzania Mwanza 1996 892 12–19 29 7
Thailand Bangkok 2002 1051 15–49 4
Thailand Nakhonsawan 2002 1028 15–49 5
United States National 1995 2042 15–24 9 —
(From World Health Organization, 20021, 2005.20)
common physical health outcome of gen- women who have not been abused.113-116
der-based abuse. More common are “func- For many women, the psychological
tional disorders”—ailments that frequently consequences of abuse are even more seri-
have no identifiable cause, such as irritable ous than its physical effects. The experi-
bowel syndrome; gastrointestinal disorders; ence of abuse often erodes women’s
and various chronic pain syndromes, self-esteem and puts them at greater risk of
including chronic pelvic pain. Studies con- a variety of mental health problems,
sistently link such disorders with a history including depression, anxiety, phobias,
of physical or sexual abuse. Women who post-traumatic stress disorder, and alcohol
have been abused also tend to experience and drug abuse.2
poorer physical functioning, more physical Violence and sexual abuse also lie
symptoms, and more days in bed than do behind some of the most intractable

CHAPTER ONE
reproductive health issues of our times— from China, Egypt, Ethiopia, Mexico, India,
unwanted pregnancies, HIV and other Nicaragua, Pakistan, Saudi Arabia, and
STIs, and complications of pregnancy. South Africa.3
Physical violence and sexual abuse can Violence during pregnancy can have
put women at risk of infection and serious health consequences for women
unwanted pregnancies directly, if women and their children.2 Documented effects
are forced to have sex, for example, or if include delayed prenatal care, inadequate
they fear using contraception or condoms weight gain, increased smoking and sub-
because of their partner’s reaction. A his- stance abuse, STIs, vaginal and cervical
tory of sexual abuse in childhood also can infections, kidney infections, miscarriages
lead to unwanted pregnancies and STIs and abortions, premature labor, fetal dis-
indirectly by increasing sexual risk-taking tress, and bleeding during pregnancy.4
in adolescence and adulthood. There is a Recent research has focused on the rela-
growing body of research indicating that tionship between violence in pregnancy
violence may increase women’s suscepti- and low birth weight, a leading cause of
bility to HIV infection.117-120 Studies carried infant deaths in the developing world.
out in Tanzania and South Africa found Although research is still emerging, find-
that seropositive women were more likely ings of six different studies performed in
than their seronegative peers to report the United States, Mexico, and Nicaragua
physical partner abuse. The results indi- suggest that violence during pregnancy
cate that women with violent or control- contributes to low birth weight, pre-term
ling male partners are at increased risk of delivery, and to fetal growth retardation, at
HIV infection. There is little information least in some settings.121, 123 A recent meta
as yet to indicate how violence increases analysis of existing studies confirms that
women’s risk for HIV. Dunkle and col- intimate partner violence during pregnancy
leagues suggest that abusive men are is indeed associated with a significant,
more likely to have HIV and impose albeit small, reduction in birth weight.124
risky sexual practices on their partners. In its most extreme form, violence kills
There are also indications that disclosure women. Worldwide, an estimated 40 to
of HIV status may put women at risk for more than 70 percent of homicides of
violence.118 women are perpetrated by intimate part-
Violence can also be a risk factor during ners, frequently in the context of an abu-
pregnancy. Studies from around the world sive relationship.125 By contrast, only a small
demonstrate that violence during preg- percentage of men who are murdered are
nancy is not a rare phenomenon. Within killed by their female partners, and in many
the United States, for example, between 1 such cases, the women are defending
percent and 20 percent of currently preg- themselves or retaliating against abusive
nant women report physical violence, with men.126 A study of female homicide in
the majority of findings between 4 percent South Africa found that intimate femicide
and 8 percent.5 The differences are due (female murder by an intimate partner)
partly to differences in the way women accounted for 41 percent of all female
were asked about violence.3, 5, 121, 122 A recent homicides. This study estimated that a
review found that the prevalence of abuse woman is killed by her intimate partner in
during pregnancy is 3 to 11 percent in South Africa every six hours.127 Violence is
industrialized countries outside of North also a significant risk factor for suicide.
America and between 4 and 32 percent in Studies in numerous countries have found
developing countries, including studies that women who have suffered domestic

FIGURE 1.5 HEALTH OUTCOMES OF VIOLENCE AGAINST WOMEN
Gender-Based Victimization
Child sexual abuse
Sexual assault
Physical abuse
Nonfatal Outcomes Fatal Outcomes

Direct & Indirect
Physical Health Mental Health

■ Injury ■ Post traumatic stress ■ Femicide
■ Functional impairment ■ Depression ■ Suicide
■ Physical symptoms ■ Anxiety ■ Maternal mortality
■ Poor subjective health ■ Phobias/panic disorders ■ AIDS-related
■ Permanent disability ■ Eating disorders
■ Sexual dysfunction
Injurious Health Behaviors ■ Low self-esteem
■ Smoking ■ Mental distress
■ Alcohol and drug use ■ Substance abuse disorders
■ Sexual risk-taking
■ Physical inactivity
■ Overeating
Functional Disorders
■ Chronic pain syndromes
■ Irritable bowel syndrome
■ Gastrointestinal disorders
■ Somatic complaints
■ Fibromyalgia
Reproductive Health
■ Unwanted pregnancy
■ STIs/HIV
■ Gynecological disorders
■ Unsafe abortion
■ Pregnancy complications
■ Miscarriage/low birth weight
■ Pelvic inflammatory disease
(From Heise et al, 1999.2)

CHAPTER ONE
violence or sexual assault are much more violence in the home, had an absent or
likely to have had suicidal thoughts, or to rejecting father, or frequently uses alco-
have attempted to kill themselves.19 hol. A recent review of nationally repre-
sentative surveys in nine countries
EXPLAINING GENDER- found that for women, low educational
BASED VIOLENCE attainment, being under 25 years of age,
having witnessed her father’s violence
Violence against women is widespread, but against her mother, living in an urban
it is not universal. Anthropologists have area, and low socio-economic status
documented small-scale societies—such as were consistently associated with an
the Wape of Papua New Guinea—where increased risk of abuse.24
domestic violence is virtually absent.128, 129
This reality stands as testament to the fact ■ At the level of the family and rela-
that social relations can be organized to tionship, the male controls wealth and
minimize abuse. decision making within the family and
Why is violence more widespread in marital conflict is frequent.
some places than in others? Increasingly,
researchers are using an “ecological frame- ■ At the community level, women are
work” to understand the interplay of per- isolated with reduced mobility and lack
sonal, situational, and socio-cultural factors of social support. Male peer groups
that combine to cause abuse.21, 130-133 In this condone and legitimize men’s violence.
framework, violence against women results
from the interaction of factors at different ■ At the societal level, gender roles are
levels of the social environment (Figure 1.6). rigidly defined and enforced and the
The framework can best be visualized as concept of masculinity is linked to
four concentric circles. The innermost cir- toughness, male honor, or dominance.
cle represents the biological and personal The prevailing culture tolerates physical
history that each individual brings to his or punishment of women and children,
her behavior in relationships. The second accepts violence as a means to settle
circle represents the immediate context in interpersonal disputes, and perpetuates
which abuse takes place: frequently the the notion that men “own” women.
family or other intimate or acquaintance
relationship. The third circle represents the The ecological framework combines
institutions and social structures, both for- individual level risk factors with family,
mal and informal, in which relationships community, and society level factors identi-
are embedded, such as neighborhoods, the fied through cross cultural studies, and
workplace, social networks, and peer helps explain why some societies and
groups. The fourth, outermost circle is the some individuals are more violent than
economic and social environment, includ- others, and why women, especially wives,
ing cultural norms. are so much more likely to be the victims
A wide range of studies shows that of violence within the family. Other factors
several factors at each of these levels combine to protect some women. For
increase the likelihood that a man will example, women who have authority and
abuse his partner: power outside the family tend to experi-
ence lower levels of abuse in intimate part-
■ At the individual level, the male was nerships. Likewise, when family members
abused as a child or witnessed marital and friends intervene promptly, they

appear to reduce the likelihood of domes- individual household income levels.135 A

tic violence. In contrast, wives are more study in Bangladesh found that some
frequently abused in cultures where family aspects of women’s status could either
affairs are considered “private” and outside increase or decrease a woman’s risk of
public scrutiny. being beaten, depending on the socio-cul-
Justifications for violence frequently tural conditions of the community she lives
evolve from gender norms, that is, social in. In one site, characterized by more con-
norms about the proper roles and respon- servative norms regarding women’s roles
sibilities of men and women. Many cul- and status, women with greater personal
tures hold that a man has the right to autonomy and those who participated for a
control his wife’s behavior and that women short time in savings and credit groups
who challenge that right—even by asking experienced more violence than women
for household money or by expressing the with less autonomy. Community-level
needs of the children—may be punished. measures of women’s status had no effect
In countries as different as Bangladesh, on the risk of violence.
Cambodia, India, Mexico, Nigeria, Pakistan, The opposite was true in the less conser-
Papua New Guinea, Nicaragua, Tanzania, vative setting where women had better
and Zimbabwe, studies find that violence is overall status. In this site, individual meas-
frequently viewed as physical chastise- ures of autonomy and participation in credit
ment—the husband’s right to “correct” an schemes had no impact on the risk of vio-
erring wife.2 As one husband said in a lence, whereas living in a community where
focus group discussion in Tamil Nadu, more women participated in credit groups
India, “If it is a great mistake, then the hus- and where women had a higher status over-
band is justified in beating his wife. Why all had a protective effect. These findings
not? A cow will not be obedient without suggest that the same condition (mobility or
beatings.”134 participating in a credit group) may have
Worldwide, studies identify a consistent completely different effects on a woman’s
list of events that are said to “trigger” vio- risk of violence, according to whether the
lence.130 These include: not obeying the activity is seen as acceptable by community
husband, talking back, not having food norms. These findings underscore the com-
ready on time, failing to care adequately plexity of these issues and the dangers in
for the children or home, questioning him applying knowledge gained from one site to
about money or girlfriends, going some- another without understanding of the
where without his permission, refusing him broader cultural context.136
sex, or expressing suspicions of infidelity.
All of these represent transgressions of HOW DO WOMEN
dominant gender norms in many societies. RESPOND TO ABUSE?
Although the ecological framework has
gained broad acceptance for conceptualiz- Most abused women are not passive vic-
ing violence, there have been few attempts tims, but use active strategies to maximize
to explore how individual and community their safety and that of their children.
level risk factors relate to each other and Some women resist, others flee, and still
ultimately influence women’s vulnerability others attempt to keep the peace by capit-
to violence. One study performed in the ulating to their husband’s demands. What
United States found that the socio-eco- may seem to an observer to be lack of
nomic status of the neighborhood had a response to living with violence may in
greater impact on the risk of violence than fact be a woman’s strategic assessment of

CHAPTER ONE
FIGURE 1.6. AN ECOLOGICAL FRAMEWORK FOR EXPLAINING GENDER-BASED VIOLENCE
Societal Community Family Individual
■ Norms and laws ■ Isolation of ■ Marital conflict ■ Being male

granting men control women and ■ Male control of ■ Witnessing marital
over female behavior family wealth and conflict as a child
■ Violence accepted for ■ Delinquent peer decision making ■ Absent or rejecting
resolving conflict groups in the family father
■ Masculinity linked to ■ Low socio - ■ Poverty ■ Being abused as
dominance, honor or economic status ■ Unemployment a child
agression ■ Alcohol use
(From Heise, 1998.130)
what it takes to survive and to protect her- Despite the obstacles, many women
self and her children. eventually do leave violent partners—even
A woman’s response to abuse is often if after many years. In a study in León,
limited by the options available to her. Nicaragua, for example, 70 percent of
Women consistently cite similar reasons for abused women eventually left their
remaining in abusive relationships: fear of abusers. The median time that women
retribution, lack of other means of eco- spent in a violent relationship was six
nomic support, concern for the children, years. Younger women were likely to leave
emotional dependence, lack of support sooner than older women.137
from family and friends, and an abiding Studies suggest a consistent set of factors
hope that “he will change.” In some coun- that propel a woman to leave an abusive
tries, women say that the social unaccept- relationship: The violence gets more severe
ability of being single or divorced poses an and triggers a realization that her partner is
additional barrier that keeps them from not going to change, or the violence
leaving destructive marriages.2 begins to take a toll on the children.
At the same time, denial and fear of Women also cite emotional and logistical
social stigma often prevent women from support from family or friends as pivotal in
reaching out for help. In numerous sur- their decision to leave.2
veys, for example, from 22 to almost 70 Leaving an abusive relationship is a multi-
percent of abused women say that until stage process. The process often includes
the interview they never told anyone about periods of denial, self-blame, and endurance
their abuse. Those who reach out do so before women recognize the abuse as a
primarily to family members and friends. pattern and identify with other women in
Few have ever contacted the police.1, 20 the same situation, thereby beginning to

disengage and recover. Most women leave consensus around violence research
and return several times before they finally methods that allow us to make mean-
leave once and for all.138 Leaving does not ingful comparisons between studies.
necessarily guarantee a woman’s safety, Methodological consistency refers not
however, because violence may continue only to defining violence using similar
even after a woman leaves. In fact, a criteria, but also the use of measures to
woman’s risk of being murdered by her minimize underreporting of violence,
abuser is often greatest immediately after such as ensuring privacy during the
separation.139 interview and providing interviewers
with special training on violence.139
CHALLENGES FOR
I N T E R N AT I O N A L ■ Research on violence may put
RESEARCH ON GENDER- women at risk. Many researchers point
BASED VIOLENCE out that research on violence involves a
number of inherent risks to both respon-
Nearly 30 years of groundbreaking dents and interviewers.140 The World
research in the field of gender-based vio- Health Organization has developed a set
lence has greatly expanded international of guidelines to minimize the risk of
awareness of the dimensions and dynamics harm to researchers and participants.141
of violence. However, there are still many However, these guidelines are just now
gaps in our current state of knowledge. being incorporated more widely into
Researchers interested in gender-based vio- international research practice.
lence from a public health perspective face
a number of important challenges. ■ More public health research is
needed to understand how violence
■ The scarcity of population-based affects the health of women and
data limits our understanding of children in different settings. Studies
how violence affects different of battered women consistently demon-
groups of women. Until very recently, strate the negative impact of abuse on
the majority of research was been car- women’s psychological status and repro-
ried out with nonrepresentative samples ductive health, and emerging epidemio-
of women, often those who have logical studies indicate that violence
attended shelters or other services for towards mothers may even affect infant
victims. Although these studies are use- birth weight and survival. However,
ful for understanding the dynamics of more research is needed to determine
abuse, they do not tell us how many what proportion of women’s overall
women overall are affected, nor provide mental and physical health problems is
information about individuals who do associated with violence and to investi-
not seek services. According to most gate the mechanisms through which
estimates, these women greatly outnum- violence affects health.
ber those who seek help.
■ More cross-cultural research is
■ Most international prevalence figures needed to reveal how societal norms
on violence are not comparable. This and institutions promote or discour-
is due mainly to inconsistencies in the age violent behavior. Most researchers
way that violence is conceptualized and agree that cultural norms can greatly
measured. Researchers need to develop affect the extent and characteristics of

CHAPTER ONE
violence, as well as the way that spe- 1. World Health Organization. World Report on
cific acts are interpreted in different Violence and Health. Geneva, Switzerland:
societies. Nonetheless, there have been World Health Organization; 2002.
2. Heise L, Ellsberg M, Gottemoeller M. Ending
few systematic attempts to compare
Violence Against Women. Baltimore: John’s
these issues in different settings. Most Hopkins University School of Public Health;
theories about the dynamics of abuse Population Information Program; 1999. Report
have been based on the experiences of No.: Series L, No. 11.
US and European women, and it is 3. Campbell J, Garcia-Moreno C, Sharps P. Abuse
unclear how relevant these are to during pregnancy in industrialized and devel-
oping countries. Violence against Women.
women from other cultures.
2004;10(7):770-789.
4. Campbell JC. Health consequences of inti-
■ Research evaluating different mate partner violence. Lancet.
approaches to violence prevention is 2002;359(9314):1331-1336.
scarce. Although there has been an 5. Gazmararian JA, Lazorick S, Spitz AM, et al.
enormous increase in both community Prevalence of violence against pregnant
women. Journal of the American Medical
and clinic-based programs to prevent
Association. 1996;275(24):1915-1920.
violence and to support abused women 6. Watts C, Zimmerman C. Violence against
and girls, few programs have been sys- women: Global scope and magnitude. Lancet.
tematically documented or evaluated. 2002;359(9313):1232-1237.
For example, many activists and profes- 7. Shane B, Ellsberg M. Violence Against Women:
sional associations in the United States Effects on Reproductive Health. Seattle,
Washington: PATH, UNFPA; 2002. Report No.:
currently encourage health providers to
20 (1).
ask each woman at every visit whether 8. Denzin NK. Toward a phenomenology of
she has been abused. However, there is domestic family violence. American Journal of
little information about what happens to Sociology. 1984;90:483-513.
women after disclosing violence, or 9. Straus MA, Gelles RJ. Societal change and
whether asking women is an effective change in family violence from 1975 to 1985 as
revealed by two national surveys. Journal of
tool for enhancing women’s safety. In
Marriage and the Family. 1986;48:465-480.
particular, we need to develop criteria 10. United Nations General Assembly. Declaration
for assessing whether practices that are on the Elimination of Violence Against Women.
effective in one setting are likely to be In: 85th Plenary Meeting. December 20, 1993.
relevant or feasible in another, very dif- Geneva, Switzerland; 1993.
ferent setting. 11. Fischbach RL, Herbert B. Domestic violence
and mental health: Correlates and conundrums
within and across cultures. Social Science and
The greatest challenge facing Medicine. 1997;45(8):1161-1176.
researchers in the field of violence is to 12. Johnson J, Sacco V. Researching violence
learn from past mistakes, to identify “best against women: Statistics Canada’s national
practices,” and to find out what makes survey. Canadian Journal of Criminology.
them successful so that we can channel 1995;37:281-304.
13. Kornblit AL. Domestic violence: An emerging
resources and efforts where they are most
health issue. Social Science and Medicine.
likely to make a difference. 1994;39:1181-1188.
14. Claramunt MC. Casitas Quebradas: El Problema
de la Violencia Doméstica en Costa Rica. San
José: Editorial Universidad Estatal a Distancia;
1997.
15. Walker L. The Battered Woman. New York:
Harper and Row; 1979.

16. Garcia Moreno C, Watts C, Jansen H, Ellsberg 29. Hassan F, Sadowski L, Shrikant B, et al.
M, Heise L. Responding to violence against Physical intimate partner violence in Chile,
women: WHO’s Multi-country Study on Egypt, India and the Philippines. Injury Control
Women’s Health and Domestic Violence. and Safety Promotion. 2004;2:111-116.
Health and Human Rights. 2003;6(2):112-127. 30. PROFAMILIA. Encuesta Nacional de
17. Gossaye Y, Deyessa N, Berhane Y, et al. Demografia y Salud 1995. Bogotá, Colombia:
Women’s health and life events study in rural PROFAMILIA and Macro International; 1995.
Ethiopia. Ethiopian Journal of Health 31. PROFAMILIA. Salud Sexual y Reproductiva:
Development. 2003;17(Second Special Issue):1-49. Resultados Encuesta Nacional de Demografía y
18. Raikes A. Pregnancy, Birthing and Family Salud 2000. Bogotá: Asociación Probienestar
Planning in Kenya: Changing Patterns of de la Familia Colombiana; 2000.
Behaviour: A Health Service Utilization Study in 32. CEPAR C. ENDEMAIN-94: Encuesta
Kisii District. Copenhagen, Denmark: Centre for Demográfica y de Salud Materna e Infantil:
Development Research; 1990. Informe General. Quito, Ecuador: Centers for
19. Central Bureau of Statistics (CBS) [Kenya], Disease Control, Centro de Estudios de
Ministry of Health [Kenya], ORC Macro. Kenya Población y Desarrollo Social; 1995.
Demographic and Health Survey 2003. 33. Asociación Demográfica Salvadoreña. Encuesta
Calverton, Maryland: CBS, MOH, and ORC Nacional de Salud Familiar de 2002-2003. San
Macro; 2004. Salvador, El Salvador: ADS, Centers for Disease
20. World Health Organization. WHO Multi-country Control; 2002.
Study on Women’s Health and Domestic 34. Ministerio de Salud Publica y Asistencia Social,
Violence Against Women: Report on the First Centers for Disease Control and Prevention.
Results. Geneva, Switzerland: WHO; 2005. Guatemala, Encuesta Nacional de Salud
21. Jewkes R, Levin J, Penn-Kekana L. Risk factors Materno Infantil 2002. Guatemala City,
for domestic violence: Findings from a South Guatemala: MSPAS, CDC; 2003.
African cross-sectional study. Social Science 35. Secretaría de Salud Honduras. Encuesta
and Medicine. 2002;55(9):1603.-1617. Nacional de Epidemiología y Salud Familiar,
22. Macro International, South Africa Department Encuesta Nacional de Salud Masculina, 2001.
of Health. South Africa Demographic and Atlanta, Georgia: Centers for Disease Control;
Health Survey 1998: Preliminary Report. DHS 2002.
survey. Calverton, Maryland: Macro 36. Ramirez J, et al. Mujeres de Guadalajara y vio-
International; 1998. lencia doméstica: Resultados de un estudio
23. Blanc AK, Wolff B, Gage AJ, et al. Negotiating piloto. Cadernos de Saude Pública.
Reproductive Outcomes in Uganda. DHS 1996;12(3):405-409.
Survey: Institute of Statistics and Applied 37. Granados M. Salud Reproductiva y Violencia
Economics and Macro International Inc; 1996. Contra la Mujer: Un Análisis Desde la
24. Kishor S, Johnson K. Domestic Violence in Nine Perspectiva de Género. Nuevo León, Mexico:
Developing Countries: A Comparative Study. Asociación Mexicana de Población (AMEP),
Calverton, Maryland: Macro International; 2004. Consejo Estatal de Población, Nuevo León
25. Watts C, Ndlovu M, Keogh E, Kwaramb R. (COESPO), El Colegio de México; 1996.
Withholding of sex and forced sex: Dimensions 38. Instituto Nacional de Estadística Geografía e
of violence against Zimbabwean women. Informática. Encuesta Nacional sobre la
Reproductive Health Matters. 1998;6:57-65. Dinámica de las Relaciones en los Hogares
26. Handwerker WP. Power and gender: Violence 2003 (ENDIREH). Distrito Federal, México:
and affection experienced by children in Instituto Nacional de Estadística, Geografía e
Barbados, West Indies. Med Anthropol. Informática, Instituto Nacional de las Mujeres,
1996;17(2):101-128. Fondo de Población de Naciones Unidas; 2004.
27. Larrain SH. Violencia Puertas Adentro: La 39. Ellsberg MC, Peña R, Herrera A, Liljestrand J,
Mujer Golpeada. Santiago, Chile: Editorial Winkvist A. Wife abuse among women of
Universitaria; 1994. childbearing age in Nicaragua. American
28. Morrison AR, Orlando MB. Social and Journal of Public Health. 1999;89(2):241-244.
Economic Costs of Domestic Violence: Chile 40. Ellsberg M, Heise L, Peña R, Agurto S, Winkvist
and Nicaragua. In: Morrison AR, Biehl ML, edi- A. Researching domestic violence against women:
tors. Too Close to Home: Domestic Violence in Methodological and ethical considerations. Studies
the Americas. Washington, DC: Inter-American in Family Planning. 2001;32(1):1-16.
Development Bank; 1999. 51-80.

CHAPTER ONE
41. Rosales J, Loaiza E, Primante D, et al. Encuesta 54. Parish W, Wang T, Laumann E, Pan S, Luo Y.
Nicaraguense de Demografia y Salud, 1998. Intimate partner violence in China: National
Managua, Nicaragua: Instituto Nacional de prevalence, risk factors and associated health
Estadisticas y Censos (INEC); 1999. problems. International Family Planning
42. CEPEP. Encuesta Nacional de Demografia y Perspectives. 2004;30(4):174-181.
Salud Reproductiva, 1995-1996. Asunción, 55. Hakimi M, Nur Hayati E, Ellsberg M, Winkvist
Paraguay: Centro Paraguayo de Estudios de A. Silence for the Sake of Harmony: Domestic
Población, Centers for Disease Control and Violence and Health in Central Java, Indonesia.
Prevention, USAID; 1997. Yogyakarta, Indonesia: Gadjah Mada University,
43. CEPEP. Encuesta Nacional de Demografía y PATH, Rifka Annisa, Umeå Univeristy; 2002.
Salud Sexual y Reproductiva 2004: ENDSSR 56. Fanslow J, Robinson EM. Violence against
2004. Informe Resumido. Asunción, Paraguay: women in New Zealand: Prevalence and health
Centro Paraguayo de Estudios de Población, consequences. The New Zealand Medical
Centers for Disease Control and Prevention, Journal. 2004;117(1206):1173-1184.
USAID; 2004. 57. Toft S, Bonnell S. Marriage and Domestic
44. Dávila AL, Ramos G, Mattei H. Encuesta de Violence in Rural Papua New Guinea,
Salud Reproductiva: Puerto Rico, 1995-96. San Occasional Paper No. 18, 1985. Boroko, Papua
Juan, Puerto Rico: CDC; 1998. New Guinea: Law Reform Commission; 1985.
45. Traverso MT. Violencia en la Pareja: La Cara 58. Macro International, National Statistics Office of
Oculta de la Relación. Washington, DC: IDB; Philippines. National Safe Motherhood Survey,
1999. 1993. Manila, Philippines: National Statistics
46. Johnson H. Dangerous Domains: Violence Office, Macro International; 1994.
Against Women in Canada. Ontario, Canada: 59. Cabaraban M, Morales B. Social and Economic
International Thomson Publishing; 1996. Consequences for Family Planning Use in
47. Statistics Canada. Family Violence in Canada: A Southern Philippines. Cagayan de Oro City,
Statistical Profile 2000. Ottawa, Canada: Philippines: Research Institute for Mindanao
Statistics Canada; 2000. Culture; 1998.
48. Tjaden P, Thoennes N. Extent, Nature and 60. Kim K-I, Cho Y-G. Epidemiological Survey of
Consequences of Intimate Partner Violence: Spousal Abuse in Korea. In: Viano EC, ed.
Findings from the National Violence Against Intimate Violence: Interdisciplinary Perspectives.
Women Survey. Washington, DC: National Washington, DC: Hemisphere Publishing Corp.;
Institute of Justice, Centers for Disease Control 1992. p. 277-282.
and Prevention; 2000. 61. Krantz G. Domestic violence against women: A
49. Australian Statistics Bureau (ASB). Women’s population-based study in Vietnam. Stockholm;
Safety: Australia. Belconnen, Australia: ASB; 2003.
1996. 62. Herold J, Seither R, Ylli A, et al. Albania
50. Mouzos J, Makkai T. Women’s Experience of Reproductive Health Survey 2002: Preliminary
Male Violence: Findings from the Australian Report. Tirana, Albania: Institute of Public
Component of the International Violence Against Health, Albania Ministry of Health, Institute of
Women Survey (IVAWS). Canberra, Australia: Statistics, CDC-Altanta; 2003.
Australian Institute of Criminology; 2004. 63. Serbanescu F, Morris L, Rahimova S, Stupp P.
51. Schuler SR, Hashemi SM, Riley AP, Akhter S. Reproductive Health Survey, Azerbaijan, 2001.
Credit programs, patriarchy and men’s violence Final Report. Atlanta, Georgia: Azerbaijan
against women in rural Bangladesh. Social Ministry of Health and Centers for Disease
Science and Medicine. 1996;43(12):1729-1742. Control and Prevention; 2003.
52. Steele F, Amin S, Naved RT. Savings/credit 64. Heiskanen M, Piisspa M. Faith, Hope,
group formation and change in contraception. Battering: A survey of men’s violence against
Demography. 2001;38(2):267-282. women in Finland. Helsinki, Finland: Statistics
53. Nelson E, Zimmerman C. Household Survey on Finland, Council for Equality; 1998.
Domestic Violence in Cambodia. Phnom Penh, 65. Jaspard M, Brown E, Condon S, et al. Les
Cambodia: Ministry of Women’s Affairs, Project Violences enver les Femmes en France: Une
Against Domestic Violence (PADV); 1996. Enquete Nationale. Paris, France: Idup, Ined,
CNRS, Universite de Paris Dauphine; 2001.

66. Serbanescu F, Morris L, Nutsubidze N, Imnadze 77. Killias M, Simonin M, De Puy J. Violence
P, Shaknazarova M. Reproductive Health Survey Experienced by Women in Switzerland over
Georgia, 1999-2000. Final Report. Atlanta, their Lifespan. Results of the International
Georgia: Georgian National Center for Disease Violence against Women Survey (IVAWS). Berne,
Control, Centers for Disease Control and Switzerland: Staempfli Publishers Ltd; 2005.
Prevention; 2001. 78. Ilkkaracan P. Exploring the context of women’s
67. Federal Ministry for Family Affairs, Senior sexuality in eastern Turkey. Reproductive
Citizens, Women and Youth (BMFSFJ). Health, Health Matters. 1998;6(12).
Well-being and Personal Safety of Women in 79. KIIS, CDC, USAID. 1999 Ukraine Reproductive
Germany: A Representative Study of Women in Health Survey. Kiev, Ukraine: Kiev International
Germany. Bonn, Germany: BMFSFJ; 2004. Institute of Sociology, Centers for Disease
68. Women’s Issues Information Centre (WHC), Control and Prevention, U.S.Agency for
United Nations Development Fund for Women International Development; 2001.
(UNIFEM). Violence against Women in Lithuania. 80. Mooney J. The Hidden Figure: Domestic
Vilnius, Lithuania: WHC, UNIFEM; 1999. Violence in North London. London, UK:
69. Römkens R. Prevalence of wife abuse in the Middlesex University; 1993.
Netherlands: Combining quantitative and quali- 81. Walby S, Allen J. Domestic Violence, Sexual
tative methods in survey research. Journal of Assault and Stalking: Findings from the British
Interpersonal Violence. 1997;12:99-125. Crime Survey. London, U.K.: Home Office
70. Schei B, Bakketeig LS. Gynaecological impact Research, Development and Statistics
of sexual and physical abuse by spouse. A Directorate; 2004.
study of a random sample of Norwegian 82. El-Zanaty F, Hussein EM, Shawky GA, Way AA,
women. British Journal of Obstetrics and Kishor S. Egypt Demographic and Health
Gynaecology. 1989;96(12):1379-1383. Survey 1995. Calverton, Maryland: Macro
71. Schei B. Report from the First National International; 1996.
Norwegian Study on Violence against Women. 83. Haj-Yahia MM. The First National Survey of
Oslo, Norway: Statistics Norway; forthcoming. Abuse and Battering Against Arab Women from
72. Serbanescu F, Morris L, Stratila M, Bivol O. Israel: Preliminary Results: Unpublished; 1997.
Reproductive Health Survey, Moldova, 1997. 84. Haj-Yahia MM. The Incidence of Wife Abuse
Atlanta, Georgia: Institute for Mother and Child and Battering and Some Socio-demographic
Health Care and Centers for Disease Control Correlates as Revealed in Two National Surveys
and Prevention; 1998. in Palestinian Society. Ramallah, The
73. Serbanescu F, Morris L, Marin M. Reproductive Palestinian Authority: Besir Center for Research
Health Survey, Romania, 1999. Final Report. and Development; 1998.
Atlanta, Georgia: Romanian Association of 85. Sadowski L, Hunter W, Bangdiwala S, Munoz S.
Public Health and Management and Centers for The world studies of abuse in the family envi-
Disease Control and Prevention; 2001. ronment (WorldSAFE): A model of a multi-
74. Russian Center for Public Opinion and Market national study of family violence. Injury
Research, Centers for Disease Control and Control and Safety Promotion. 2004;11(2):81-90.
Prevention. 1999 Russian Women’s Reproductive 86. Heise L, Moore K, Toubia N. Sexual Coercion
Health Survey: A Follow-up of Three Sites. Final and Women’s Reproductive Health: A Focus on
Report. Atlanta, Georgia: Centers for Disease Research. New York, New York: Population
Control and Prevention; 2000. Council; 1995.
75. Lundgren E, Heimer G, Westerstand J, 87. World Health Organization. Violence Against
Kalliokoski A-M. Captured Queen: Men’s Women: A Priority Health Issue. Fact sheets.
Violence Against Women in “Equal” Sweden: A Geneva, Switzerland: World Health
Prevalence Study. Umeå, Sweden: Fritzes Organization; 1997.
Offentliga Publikationer; 2001. 88. Jewkes R, Levin J, Mbananga N, Bradshaw D.
76. Gillioz L, De Puy J, Ducret V. Domination et Rape of girls in South Africa. Lancet.
Violence Envers la Femme dans le Couple. 2002;359(9303):319-320.
Geneva, Switzerland: Editions Payot Lausanne; 89. Jewkes R, Abrahams N. The epidemiology of
1997. rape and sexual coercion in South Africa: An
overview. Social Science and Medicine.
2002;55(7):1231-1244.

CHAPTER ONE
90. Finkelhor D. The international epidemiology of 103. Singh HS, Yiing WW, Nurani HN. Prevalence of
child sexual abuse. Child Abuse and Neglect. childhood sexual abuse among Malaysian para-
1994;18(5):409-417. medical students. Child Abuse and Neglect.
91. Jejeebhoy S, Bott S. Non-consensual Sexual 1996;20(6):487-492.
Experiences of Young People: A Review of the 104. Fergusson DM, Horwood LJ, Lynskey MT.
Evidence from Developing Countries. South and Childhood sexual abuse, adolescent sexual
East Asia Regional Working Papers. New Delhi, behaviors and sexual revictimization. Child
India: Population Council; 2003. Report No.: 16. Abuse and Neglect. 1997;21(8):789-803.
92. Zimmerman C, Yun K, Shvab I, et al. The 105. Olsson A, Ellsberg M, Berglund S, et al. Sexual
Health Risks and Consequences of Trafficking abuse during childhood and adolescence
in Women and Adolescents: Findings from a among Nicaraguan men and women: A popula-
European Study. London, United Kingdom: tion-based anonymous survey. Child Abuse
London School of Hygiene and Tropical and Neglect. 2000;24(12):1579-1589.
Medicine (LSHTM); 2003. 106. Pedersen W, Skrondal A. Alcohol and sexual
93. Orhant M, Murphy E. Trafficking in Persons. In: victimization: A longitudinal study of
Murphy E, Ringheim K, editors. Reproductive Norwegian girls. Addiction. 1996;91(4):565-581.
Health and Rights: Reaching the Hardly 107. López F, Carpintero E, Hernandez A, Martin MJ,
Reached. Washington, DC: PATH; 2002. Fuertes A. Prevalence and sequelae of child-
94. Swiss S, Jennings PJ, Aryee GV, et al. Violence hood sexual abuse in Spain. Child Abuse and
against women during the Liberian civil con- Neglect. 1995;19(9):1039-1050.
flict. Journal of the American Medical 108. Halperin DS, Bouvier P, Jaffe PD, et al.
Association. 1998;279:625-629. Prevalence of child sexual abuse among ado-
95. Ward J. If Not Now, When? Addressing Gender- lescents in Geneva: Results of a cross sectional
based Violence in Refugee, Internally Displaced study. British Medical Journal.
and Post-conflict Settings: A Global Overview. 1996;312(7042):1326-1329.
New York, New York: Reproductive Health 109. Tschumper A, Narring F, Meier C, Michaud PA.
Response in Conflict Consortium; 2002. Sexual victimization in adolescent girls (age 15-
96. Ward J, Vann B. Gender-based violence in 20 years) enrolled in post-mandatory schools or
refugee settings. Lancet. 2002;360 Suppl:s13-14. professional training programmes in Switzerland.
97. Kyriacou DN, Anglin D, Taliaferro E, et al. Risk Acta Paediatrica. 1998;87(2):212-217.
factors for injury to women from domestic vio- 110. Wilsnack SC, Vogeltanz ND, Klassen AD, Harris
lence against women. New England Journal of TR. Childhood sexual abuse and women’s sub-
Medicine. 1999;341(25):1892-1898. stance abuse: National survey findings. J Stud
98. Fleming JM. Prevalence of childhood sexual Alcohol. 1997;58(3):264-271.
abuse in a community sample of Australian 111. Luster T, Small SA. Sexual abuse history and
women. Medical Journal of Australia. problems in adolescence: Exploring the effects
1997;166(2):65-68. of moderating variables. Journal of Marriage
99. Handwerker WP. Gender power differences and the Family. 1997;59:131-142.
between parents and high-risk sexual behavior 112. Stock JL, Bell MA, Boyer DK, Connell FA.
by their children: AIDS/STD risk factors extend Adolescent pregnancy and sexual risk-taking
to a prior generation. Journal of Women’s among sexually abused girls. Family Planning
Health. 1993;2(3):301-316. Perspectives. 1997;29(5):200-203, 227.
100. MacMillan HL, Fleming JE, Trocme N, et al. 113. Golding J. Sexual assault history and women’s
Prevalence of child physical and sexual abuse reproductive and sexual health. Psychology of
in the community. Results from the Ontario Women Quarterly. 1996;20:101-121.
Health Supplement. Journal of the American 114. Walker EA, Katon WJ, Roy-Byrne PP, Jemelka
Medical Association. 1997;278(2):131-135. RP, Russo J. Histories of sexual victimization in
101. Krugman S, Mata L, Krugman R. Sexual abuse patients with irritable bowel syndrome or
and corporal punishment during childhood: A inflammatory bowel disease. American Journal
pilot retrospective survey of university students of Psychiatry. 1993;150(10):1502-1506.
in Costa Rica. Pediatrics. 1992;90(1 Pt 2):157-161. 115. Golding JM. Sexual assault history and limita-
102. Schotensack K, Elliger T, Gross A, Nissen G. tions in physical functioning in two general
Prevalence of sexual abuse of children in population samples. Research in Nursing and
Germany. Acta Paedopsychiatrica. Health. 1996;19(1):33-44.
1992;55(4):211-216.

116. Campbell J, Jones AS, Dienemann J, et al. 129. Levinson D. Violence in Cross-cultural
Intimate partner violence and physical health Perspective. Newbury Park, California: Sage
consequences. Archives of Internal Medicine. Publishers; 1989.
2002;162(10):1157-1163. 130. Heise L. Violence against women: An integrated,
117. Garcia-Moreno C, Watts C. Violence against ecological framework. Violence against Women.
women: Its importance for HIV/AIDS. AIDS. 1998;4(3):262-290.
2000;14(Suppl 3):S253-265. 131. Jewkes R. Intimate partner violence: Causes and
118. Maman S, Campbell J, Sweat MD, Gielen AC. prevention. Lancet. 2002;359(9315):1423-1429.
The intersections of HIV and violence: 132. Koenig MA, Lutalo T, Zhao F, et al. Coercive
Directions for future research and interventions. sex in rural Uganda: Prevalence and associated
Social Science and Medicine. 2000;50(4):459-478. risk factors. Social Science and Medicine.
119. Maman S, Mbwambo JK, Hogan NM, et al. HIV- 2004;58:787-798.
positive women report more lifetime partner 133. Koenig M, Lutalo T, Zhao F, et al. Domestic vio-
violence: Findings from a voluntary counseling lence in rural Uganda: Evidence from a com-
and testing clinic in Dar es Salaam, Tanzania. munity-based study. Bulletin of the World
American Journal of Public Health. Health Organization. 2003;81:53-60.
2002;92(8):1331-1337. 134. Jejeebhoy SJ. Wife-beating in rural India: A hus-
120. Dunkle KL, Jewkes RK, Brown HC, et al. band’s right? Economic and Political Weekly
Gender-based violence, relationship power, and (India). 1998;23(15):855-862.
risk of HIV infection in women attending ante- 135. O’Campo P, Gielen AC, Faden RR, et al.
natal clinics in South Africa. Lancet. Violence by male partners against women dur-
2004;363(9419):1415-1421. ing the childbearing year: A contextual analysis.
121. Petersen R, Gazmararian JA, Spitz AM, et al. American Journal of Public Health.
Violence and adverse pregnancy outcomes: A 1995;85(8):1092-1097.
review of the literature and directions for future 136. Koenig MA, Ahmed S, Hossain MB, Khorshed
research. American Journal of Preventive Alam Mozumder AB. Women’s status and
Medicine. 1997;13(5):366-373. domestic violence in rural Bangladesh: individ-
122. Nasir K. Violence against pregnant women in ual- and community-level effects. Demography.
developing countries. European Journal of 2003;40(2):269-288.
Public Health. 2003;13(2):105-107. 137. Ellsberg MC, Winkvist A, Peña R, Stenlund H.
123. Valladares E, Ellsberg M, Peña R, Högberg U, Women’s strategic responses to violence in
Persson L-Å. Physical partner abuse during Nicaragua. Journal of Epidemiology and
pregnancy: A risk factor for low birth weight in Community Health. 2001;55(8):547-555.
Nicaragua. Obstetrics and Gynecology. 138. Landenburger K. A process of entrapment in
2002;100(4):100-105. and recovery from an abusive relationship.
124. Murphy D, Schei B, Myhr T, Du Mont J. Abuse: Issues in Mental Health Nursing. 1989;10(3-
A risk factor for low birth weight? A systematic 4):209-227.
review and meta-analysis. 2001;164:1567-1572. 139. Campbell J. Assessing Dangerousness: Violence
125. Bailey JE, Kellermann AL, Somes GW, et al. Risk by Sexual Offenders, Batterers, and Child
factors for violent death of women in the Abusers. Thousand Oaks, California: Sage
home. Archives of Internal Medicine. Publications; 1995.
1997;157(7):777-782. 140. Ellsberg M, Heise L. Bearing witness: Ethics in
126. Smith PH, Moracco KE, Butts JD. Partner homi- domestic violence research. Lancet.
cide in context: A population-based perspec- 2002;359(9317):1599-1604.
tive. Homicide Studies. 1998;2(4):400-421. 141. World Health Organization. Putting Women’s
127. Mathews S, Abrahams N, Martin L, et al. Every Safety First: Ethical and Safety
Six Hours a Woman is Killed by her Intimate Recommendations for Research on Domestic
Partner: A National Study of Female Homicide Violence Against Women. Geneva, Switzerland:
in South Africa. Pretoria, South Africa: Gender Global Programme on Evidence for Health
and Health Research Group, Medical Research Policy, World Health Organization; 1999. Report
Council, South Africa; 2004. No.: WHO/EIP/GPE/99.2.
128. Counts D, Brown JK, Campbell JC. To Have
and To Hit. 2nd ed. Chicago, Illinois: University
of Chicago Press; 1999.

2
CHAPTER TWO
Ethical Considerations
for Researching Violence
Against Women*
Topics covered in this chapter:
Respect for persons at all stages of the research process

Minimizing harm to respondents and research staff
Maximizing benefits to participants and communities (beneficence)
Justice: Balancing risks and benefits of research on violence against women
[The experience] that most affected me was with a girl my age, maybe 22 years old...She told
me 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. She
was very afraid, and her voice trembled as she spoke, but she really wanted to tell me about
her tragedy. She kept looking over to where her mother-in-law was watching us. She asked
me for help and I told her about the Women’s Police Station. When her mother-in-law got up
to go to the latrine, I quickly gave her a copy of the pamphlet and she hid it. She thanked me
when I left and I ended up crying in the street because I couldn’t stand to see such a young
girl being so mistreated… Nicaraguan interviewer. (Ellsberg et al, 2001.19)
I n many ways, researching violence

against women is similar to researching
that transcend those in other areas because
of the potentially threatening and traumatic
other sensitive topics. There are issues of nature of the subject matter. In the case of
confidentiality, problems of disclosure, and violence, the safety and even the lives of
the need to ensure adequate and informed women respondents and interviewers may
consent. As the previous quote from an be at risk.1
interviewer illustrates, however, there are In 1991, the Council for International
aspects of gender-based violence research Organization of Medical Sciences (CIOMS)
* This chapter was adapted from Ellsberg and Heise, 2002.1

CHAPTER TWO
presented a set of International Guidelines guidelines in turn and explores the chal-
for Ethical Review of Epidemiological lenges of applying them to the special
Studies.3 These guidelines apply the basic case of conducting research on domestic
ethical principles of biomedical research and sexual violence.
involving human subjects to the field of
epidemiology: respect for persons, non- RESPECT FOR PERSONS
maleficence (minimizing harm), benefi- AT A L L S TA G E S O F T H E
cence (maximizing benefits), and justice. RESEARCH PROCESS
In 1999, the World Health Organization
(WHO) published guidelines for address- Informed consent for respondents
ing ethical and safety issues in gender- The principle of respect for persons incor-
based violence research.4 The guidelines porates two fundamental ethical principles:
were based on the experiences of the respect for autonomy and protection of
International Research Network on Violence vulnerable persons. These are commonly
Against Women (IRNVAW) and were addressed by individual informed consent
designed to inform the WHO Multi-country procedures that ensure that respondents
Study on Women’s Health and Domestic understand the purpose of the research
Violence Against Women. (See Box 2.1 for a and that their participation is voluntary.
description of the main points.) The authors There is still no consensus on whether
argue that these ethical guidelines are criti- the informed consent process for VAW stud-
cal, not only to protecting the safety of ies should explicitly acknowledge that the
respondents and researchers, but also to study will include questions on violence or
ensuring data quality. whether it is sufficient to warn participants
This chapter examines each of the that sensitive topics will be raised. The
basic principles mentioned in the CIOMS WHO VAW study used an oral consent
process that referred to the survey as a
B O X 2 . 1 E T H I C A L A N D S A F E T Y R E C O M M E N D AT I O N S
FOR DOMESTIC VIOLENCE RESEARCH study on women’s health and life experi-
ences.5 Women were advised that, “Some of
■ The safety of respondents and the research team is paramount and should infuse the topics discussed may be personal and
all project decisions.
difficult to talk about, but many women
■ Prevalence studies need to be methodologically sound and to build upon current
research experience about how to minimize the underreporting of abuse.
have found it useful to have the opportunity
to talk.” Women were told that they could
■ Protecting confidentiality is essential to ensure both women’s safety and data
quality. end the interview at any time or skip any
■ All research team members should be carefully selected and receive specialized question they did not want to answer. (See
training and ongoing support. Box 2.3 for an example of the informed
■ The study design must include a number of actions aimed at reducing any consent form used in the WHO VAW study.)
possible distress caused to the participants by the research. A more detailed explanation of the nature
■ Fieldworkers should be trained to refer women requesting assistance to available of the questions on violence was provided
sources of support. Where few resources exist, it may be necessary for the study directly before the violence questions, and
to create short-term support mechanisms.
respondents were asked whether they
■ Researchers and donors have an ethical obligation to help ensure that their
wanted to continue and were again
findings are properly interpreted and used to advance policy and intervention
development. reminded of their option not to answer. It is
■ Violence questions should be incorporated into surveys designed for other a good idea to prepare a list of responses
purposes only when ethical and methodological requirements can be met. for questions that a woman might ask about
the study, such as how she was selected for
(From WHO, 1999.4) the study, what will the study be used for,
and how her responses will be kept secret.

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
BOX 2.2 ADAPTING ETHICAL BOX 2.3 INDIVIDUAL CONSENT FORM

GUIDELINES TO LOCAL SETTINGS
Used in the WHO Multi-country Study on Women’s Health and Domestic Violence
Researchers involved in the WHO Multi-country Study Against Women
on Women’s Health and Domestic Violence Against
Women debated at length the value of mentioning Hello, my name is [*]. I work for [*]. We are conducting a survey in [study loca-
violence directly in the initial consent process versus tion] to learn about women’s health and life experiences. You have been chosen by
adding a second-order consent process immediately chance (as in a lottery/raffle) to participate in the study.
before the questions on abuse. Some researchers
argued that it was important to alert women up front I want to assure you that all of your answers will be kept strictly secret. I will not
as to the true nature of the questions whereas others keep a record of your name or address. You have the right to stop the interview at
felt it was preferable to postpone introducing the any time, or to skip any questions that you don’t want to answer. There are no right
notion of violence until immediately prior to the actual or wrong answers. Some of the topics may be difficult to discuss, but many women
abuse-related questions. This would allow some rap- have found it useful to have the opportunity to talk.
port to develop, but still give a woman an opportu- Your participation is completely voluntary but your experiences could be very helpful
nity to opt out of the violence-related questions. to other women in [country].
The consent process was well received by respon- Do you have any questions?
dents in all countries except Japan. During pilot test-
(The interview takes approximately [*] minutes to complete). Do you agree to be
ing, several Japanese respondents expressed a
interviewed?
sense of betrayal because they had not been
informed that the interview contained questions
about violence.6 As a result, the Japan team modi- NOTE WHETHER RESPONDENT AGREES TO INTERVIEW.
fied its consent language to explicitly acknowledge
[ ] DOES NOT AGREE TO BE INTERVIEWED
violence up front. This is an excellent example of
how ethical principles and actual experience can THANK PARTICIPANT FOR HER TIME AND END INTERACTION.
combine to guide practice.
[ ] AGREES TO BE INTERVIEWED.
Is now a good time to talk?
Mandatory reporting of abuse It’s very important that we talk in private. Is this a good place to hold the inter-
Some countries have laws that require cer- view, or is there somewhere else that you would like to go?
tain kinds of professionals to report cases
of suspected abuse to authorities or social TO BE COMPLETED BY INTERVIEWER
service agencies. Such laws raise difficult
issues for researchers because they throw I CERTIFY THAT I HAVE READ THE ABOVE CONSENT PROCEDURE TO THE
PARTICIPANT.
into conflict several key ethical principles:
respect for confidentiality, the need to pro- SIGNED: ____________________________________________________________
tect vulnerable populations, and respect for
autonomy. In the case of adult women,
(From WHO, 2004.5)
there is consensus among most researchers
that the principles of autonomy and confi-
dentiality should prevail and that because children are generally considered
researchers should do everything within more vulnerable and less able to act on
their power to avoid usurping a woman’s their own behalf. The dilemma is particu-
right to make autonomous decisions about larly acute in settings where there are no
her life. (Of course if a woman seeks sup- effective services to assist troubled families,
port in reporting her abuse, researchers or where reporting is likely to trigger a cas-
should oblige.) cade of events that might put the child at
The dilemma of whether to comply with even greater risk (such as being removed
legal reporting requirements is particularly from his/her home and placed in an institu-
problematic when dealing with child abuse. tion). The WHO VAW study specifically
There is no consensus internationally about excluded questions about child abuse, but
how to handle cases of child abuse required teams to develop local protocols

CHAPTER TWO
for handling cases of child abuse that inter- No systematic studies have been per-
viewers might nonetheless come to know formed to determine how often women
about. The guiding principle of these proto- suffer negative consequences from partici-
cols was to act in “the best interests of the pating in research on violence. However,
child,” a standard that each team opera- several VAW researchers have recorded
tionalized locally, based on advice from key chilling examples of experiences where
agencies about prevailing conditions. women have been placed at risk as a result
of inadequate attention to safety issues.8
Community agreement For example, researchers from Chiapas,
In many countries, it is also important to Mexico, describe how, when they first
obtain community support for research, as began researching domestic violence, they
well as individual consent. (Community were not fully aware of the risks involved.
consent, however, should never replace They included a small set of questions on
individual consent.) This is often sought by domestic violence within a larger study on
meeting with community leaders to explain reproductive health without taking any
the overall objectives of the research. For special precautions regarding safety of
safety reasons, when obtaining community respondents. They were shocked to learn
support for VAW research, it is important to later that three respondents were beaten by
frame the study in general terms—such as their partners because they had partici-
a study on women’s health or life experi- pated in the survey.9
ences rather than mention violence or The WHO guidelines provide a number
abuse directly. If it becomes well known in of suggestions about how to minimize risks
the community that women are being to respondents, including:
questioned about violence, men may pro-
hibit their partners from participating or ■ Interviewing only one woman per house-
may retaliate against them for their partici- hold (to avoid alerting other women who
pation. In addition to potentially jeopardiz- may communicate the nature of the study
ing the safety of respondents, this could back to potential abusers).
also undermine the study objectives and
data accuracy. ■ Not informing the wider community that
the survey includes questions on violence.
MINIMIZING HARM TO
RESPONDENTS AND ■ Not conducting any research on violence
R E S E A R C H S TA F F with men in the same clusters where
women have been interviewed.4
Ensuring participant safety
The primary ethical concern related to Protecting privacy and
researching VAW is the potential for inflict- confidentiality
ing harm to respondents through their par- His mother and sisters kept passing by, and
ticipation in the study. A respondent may would peek in the doorway to see what we
suffer physical harm if a partner finds out were talking about, so we would have to
that she has been talking to others about speak really softly…and the girl said to me,
her relationship with him. Because many “Ay, don’t ask me anything in front of
violent partners control the actions of their them.” (Nicaraguan interviewer) 2
spouses closely, even the act of speaking Protecting privacy is important in its
to another person without his permission own right and is also an essential element
may trigger a beating. in ensuring women’s safety. In addition to

interviewing only one woman per house- BOX 2.4 SUGGESTIONS FOR MINIMIZING HARM
hold, the WHO recommendations advise T O W O M E N PA RT I C I PAT I N G I N R E S E A R C H
researchers to conduct violence-related
■ Interview only one woman per household.
interviews in complete privacy, with the
exception of children under the age of ■ Don’t inform the wider community that the survey includes questions on violence.
two. In cases where privacy cannot be ■ Don’t interview men about violence in the same households or clusters where
women have been asked about violence.
ensured, interviewers should be encour-
■ Interviews should be conducted in complete privacy.
aged to reschedule the interview for a dif-
ferent time or place. Achieving this level of ■ Dummy questionnaires may be used if others enter the room during the interview.
privacy is difficult and may require more ■ Candy and games may be used to distract children during interviews.
resources than might be needed for ■ Use of self-response questionnaires for some portions of the interview may be
useful for literate populations.
research on less sensitive topics.
Researchers have developed a variety of ■ Train interviewers to recognize and deal with a respondent’s distress during the
interview.
creative methods for ensuring privacy.
■ End the interview on a positive note that emphasizes a woman’s strengths.
Interviewers in Zimbabwe and Nicaragua
often held interviews outside or accompa-
nied women to the river as they washed supervisors and even drivers can also play
clothes. Many studies have successfully a role in distracting household members
used “dummy” questionnaires, containing who are intent on listening to the inter-
unthreatening questions on issues such as view. In one instance in Zimbabwe, field-
breastfeeding or reproductive health. workers entered into lengthy negotiations
Respondents are forewarned that if some- to purchase a chicken from the husband of
one enters the room, the interviewer will a respondent so that she could be inter-
change the topic of conversation by viewed in private.10 Other researchers have
switching to a dummy questionnaire. Other carried candy and coloring books to keep
members of the research team such as children busy during interviews.
Indeed, the Japanese team for the
WHO VAW study found it so difficult
to achieve privacy in Japan’s
crowded apartments that they had to
depart from the protocol and use
self-response booklets for especially
sensitive questions. In this highly lit-
erate population, women were able
to read and record their answers
without the questions having to be
read aloud.6
Ensuring privacy may be even
more problematic in telephone sur-
veys. Interviewers for the VAW sur-
vey in Canada were trained to detect
whether anyone else was in the

room or listening on another line,
and to ask whether they should call
back at another time. They provided
respondents with a toll free number
Interview in Thailand to call back if they wanted to verify

CHAPTER TWO
ine their own attitudes and beliefs around

rape and other forms of violence.
Interviewers frequently share many of the
same stereotypes and biases about victims
that are dominant in the society at large.
Left unchallenged, these beliefs can lead to
victim-blaming and other destructive atti-
tudes that can undermine both the respon-
dent’s self esteem and the interviewer’s
ability to obtain quality data.
Referrals for care and support

At a minimum, the WHO guidelines sug-
gest that researchers have an ethical obliga-

tion to provide a respondent with
information or services that can help her
situation. In areas where specific violence-
related services are available, research
Respondent in that the interview was legitimate, or in case teams have developed detailed directories
Tanzania tells children they needed to hang up quickly. About that interviewers can use to make referrals.
to go play before 1,000 out of a sample of 12,000 women In Canada’s VAW survey, for example, the
starting her interview
called back, and 15 percent of the calls computer program used by telephone
were to finish interrupted interviews.11 interviewers had a pop-up screen that
listed resources near the respondent, based
Minimizing participant distress on her mail code. In Zimbabwe, Brazil,
Interviews on sensitive topics can provoke Peru, and South Africa, researchers devel-
powerful emotional responses in some par- oped small pamphlets for respondents that
ticipants. The interview may cause a listed resources for victims along with a
woman to relive painful and frightening host of other health and social service
events, and this in itself can be distressing agencies.10 All women were offered the
if she does not have a supportive social pamphlet after being asked if it would be
environment.12 Interviewers therefore need safe for them to receive it (cases have been
to be trained to be aware of the effects that reported where women have been beaten
the questions may have on informants and when a partner found informational mate-
how best to respond, based on a woman’s rial addressing violence). In Zimbabwe,
level of distress. interviewers carried a referral directory and
Most women who become emotional wrote out addresses on physician referral
during an interview actively choose to pro- pads so that the referral would not attract
ceed, after being given a moment to collect suspicion if discovered. Ideally, contact
themselves. Interviewer training should should be made in advance with the serv-
include practice sessions on how to iden- ices so that they are prepared to receive
tify and respond appropriately to symp- referrals from the study.
toms of distress as well as how to In settings where resources are scarce or
terminate an interview if the impact of the nonexistent, researchers have developed
questions becomes too negative. interim support measures. For example, a
Interviewer training should also include study on violence against women per-
explicit exercises to help field staff exam- formed in rural Indonesia brought in a

counselor to the field once a week to meet BOX 2.5 PROTECTING RESPONDENT SAFETY IN CAMBODIA
with respondents.13 In Ethiopia, the study
hired mental health nurses to work in the Researchers in a study performed in Cambodia found a young woman who was
closest health center for the duration of the held prisoner in her own home by her husband. When the research team arrived
to interview her, they found the woman locked in her house, with only a peephole
fieldwork.14 The number of women who where a chain was threaded through a crudely cut hole in the door. The woman
actually make use of such services is often conducted the interview through the peephole. During the interview, the husband
quite low, but subsequent interviews with appeared and was suspicious about their activity. The team gave him a false
explanation for their visit and then left the home.
women indicate that they appreciate know-
ing that services are available if needed.11 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-
In Peru and in Bangladesh, the WHO VAW band had stormed into their office the preceding afternoon, dragging his wife by
team has used the study as an opportunity the arm. He demanded to know who had been at his door. He told the Secretariat
to train local health promoters in basic personnel 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 researchers
counseling and support skills. In this way,
realized that it would be too dangerous to ever approach this woman again.
the team will leave behind a permanent
The team made several overtures with different government officials and the police
resource for the community. to help get the woman freed, but everyone was afraid to intervene because the
woman’s husband had an important position. Researchers described the frustration
Bearing witness to violence that the team felt at not being able to free the woman and the guilt they felt at
having put the woman in greater danger.
The image of these stories affects you, to see
how these women suffer, and especially the
(From Zimmerman, 1995.8)
feeling that no one supports them. These are
experiences that you never forget…
(Nicaraguan interviewer)2 strategies to reduce the first source of risk
Although preventing harm to respon- include removing extremely dangerous
dents is of primary importance, researchers neighborhoods from the sampling frame
also have an ethical obligation to minimize before drawing the sample (for example
possible risks to field staff and researchers. those controlled by narco-traffickers); out-
Sources of risk include threats to physical fitting teams with cell phones; and having
safety either as a result of having to travel male drivers accompany female interview-
in dangerous neighborhoods or from ers into dangerous areas.
unplanned encounters with abusive indi- Abusive partners have also been known
viduals who object to the study. Some to threaten interviewers with physical harm.
In a South African study, for example, a
man came home from a bar in the middle
of his partner’s interview and pulled a gun
on the fieldworker, demanding to see the
questionnaire. Because of prior training, the
interviewer had the presence of mind to
give the man an English version of the
questionnaire, which he was unable to
read.10 “Dummy” questionnaires would also
have 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 hard
to overestimate the emotional impact that
Interview in Bangladesh research on violence may have on field-

CHAPTER TWO
workers and researchers. As the narrative “get a hold of yourself, you cry for every lit-
from a Nicaraguan fieldworker presented at tle thing.” But how could I control myself? I
the beginning of this chapter illustrates, a couldn’t stand it… I would try, but some-
study on violence often becomes an times it was impossible, and I would burst
intensely personal and emotional journey into tears during the next interview…
for which many researchers are not pre- (Nicaraguan interviewer) 2
pared. Particularly when field staff have Other interviewers commented that they
had personal experiences of abuse, the felt extremely drained and distracted by the
experience can be overwhelming. Judith interviews where women reported vio-
Herman, in her work on psychological lence. One woman reported that she had
trauma in survivors of political and domes- stopped working for the study because she
tic violence, describes this as a common could not bear to listen to women’s stories
experience for those who study violence: of abuse. 2
To study psychological trauma is to come Experience has shown that trauma-
face to face both with human vulnerability related stress is not confined to field staff
in the natural world and with the capacity who are directly involved with respon-
for evil in human nature. To study psycho- dents. Field supervisors, transcribers, driv-
logical trauma means bearing witness to ers, and even data entry personnel may be
horrible events.15 affected. In one study in Belize, a tran-
Including discussions of violence in scriber broke down after hours of listening
interviewer training is crucial for reducing to in-depth qualitative interviews with sur-
distress during fieldwork. During fieldwork, vivors of abuse.16
another important measure is to provide It is particularly important to provide
interviewers and research staff with regular opportunities during training for inter-
opportunities for emotional debriefing, or viewers to address their own experiences
when necessary, individual counseling. of abuse. Given the high prevalence of
Researchers have used a variety of creative gender-based violence globally, it is likely
strategies for protecting the emotional that a substantial proportion of interview-
health of their staff. In Peru, for example, ers will have experienced gender-based
the WHO multi-country team employed a violence themselves at some point. These
professional counselor to lead weekly sup- experiences need to be taken into consid-
port sessions that incorporated guided eration. Most people learn to cope with
imagery and relaxation techniques. painful past experiences, and usually do
Experience has repeatedly demonstrated not dwell on them in their everyday lives.
that emotional support for fieldworkers is However, when trainees are confronted
essential. Not only does it help interview- with the subject matter the information
ers withstand the demands of the field- may awaken disturbing images and or
work, but it also improves their ability to emotions. For many trainees, simply
gather quality data. acknowledging the fact that these reac-
Transcripts of debriefing sessions with tions are normal and providing timely
interviewers who participated in studies opportunities to discuss them will be suffi-
without adequate support illustrate this cient to help them complete the training
point: and participate successfully in fieldwork.
…When I heard stories about women In those rare cases where feelings become
being beaten and tied up, I would leave too overwhelming, trainees should be
there feeling desperate… I would be a supported in their decision to withdraw
wreck, and my supervisor would tell me from the study.

MAXIMIZING BENEFITS abuse who had visited a women’s crisis

T O PA R T I C I PA N T S center in Nicaragua found that a central
AND COMMUNITIES part of women’s process of recovery and
(BENEFICENCE) personal as well as collective empower-
ment came not only from increased
The principle of beneficence refers to the knowledge of their rights, but also from
ethical obligation to maximize possible the opportunity to share their experiences
benefits to study participants and the and to help other women in similar situa-
group of individuals to which they belong. tions.18 In this sense, asking women about
This principle gives rise to norms requiring experiences of violence may be seen as
that the risks of research be reasonable in an intervention in itself. At the very least,
light of the expected benefits, that the asking conveys the message that violence
research design be sound, and that the is a topic worthy of study, and not a
investigators be competent both to conduct shameful or unimportant issue.
the research and to ensure the well-being In this same vein, many fieldworkers in
of participants. the León, Nicaragua, research described
the experience of listening to women’s sto-
The interview as an intervention ries, as well as the opportunity to tell their
Asking women to reveal stories of trauma own stories in the debriefing sessions, as a
can be a transforming experience for both profoundly healing experience. One inter-
researchers and respondents. Indeed, there viewer who had never before discussed
is ample evidence that most women wel- her experiences said,
come the opportunity to tell their stories [when I joined this study] I felt that I had
if they are asked in a sympathetic, non- finally found someone I could
judgmental way. In our experience, tell everything to, someone with Remembering and telling
women rarely refuse to answer questions whom I could share my bur- the truth about terrible events
on violence. den, because it’s horrible to feel are prerequisites both for
Many women who disclose violence in so alone. Now I feel that a the restoration of the social
surveys have never told anyone about weight has been taken off order and for the healing
their situations.17 Many studies find that me…I feel relieved… 19 of individual victims. (Herman,
participants find the experience to be so The interview is also an 1992. )15
helpful that they ask fieldworkers to “inter- opportunity to provide women

view” a friend or relative who has a story with information on gender-based violence.
to tell. As Herman notes, “remembering Many studies have issued small cards that
and telling the truth about terrible events can be easily hidden in a shoe or inside a
are prerequisites both for the restoration of blouse with information about local
the social order and for the healing of resources for abused women and messages
individual victims.”15 such as, “If you are being abused, there
Even the act of telling her story can are ways out” or “Violence is never justi-
offer a woman some small way of trans- fied.” Such messages may enable women
forming her personal ordeal into a way to to see experiences in a new light or to
help others. Indeed, researchers sensitive identify violence in others close to them.
to this issue encourage interviewers and Researchers also stress the importance of
field staff to take hope and satisfaction ending the interview on a note that
from their participation in the process of emphasizes women’s strengths and tries to
giving a voice to women’s suffering. minimize distress, particularly as a respon-
A qualitative study of survivors of dent may have revealed information that

CHAPTER TWO
made her feel vulnerable.20 A number of they felt good/better at the end of the
studies have carefully scripted such end- interview. In most countries, the range was
ings to ensure that the interview finishes similar between women who had or had
with clear statements that explicitly not experienced partner violence. Very few
acknowledge the abuse, highlight the women reported feeling worse after being
unacceptability of the violence, and interviewed. Between 0.5 and 8.4 percent
emphasise the respondent’s strengths in of women reporting partner violence ever
enduring and/or ending the violence. The (highest in Peru) and between zero and 3.2
WHO study ends each interview with the percent of women with no history of part-
words, “From what you have told me, I can ner violence felt worse.17
tell that you have had some very difficult
times in your life. No one has the right to Assuring scientific soundness
treat someone else in that way. However, The CIOMS guidelines note: “A study that
from what you have told me I can also see is scientifically unsound is unethical in that
that you are strong and have survived it exposes subjects to risk or inconvenience
through some difficult circumstances.”7 while achieving no benefit in knowledge.”3
One indication of how women have This principle is particularly important in
viewed the interview process can be the area of gender-based violence where
obtained by assessing respondents’ satisfac- women are asked to disclose difficult and
tion with the interview. At the end of the painful experiences and where the act of
WHO interview, respondents were asked research itself may put women at further
the following question: “I have asked risk of abuse. Thus the WHO guidelines
about many difficult things. How has talk- note that violence researchers have an ethi-
ing about these things made you feel?” The cal responsibility to ensure the soundness
answers were written down verbatim and of their work by selecting a large enough
coded by the interviewer into the following sample size to permit conclusions to be
three categories: good/better, bad/worse, drawn, and by building upon current
and same/no difference. The majority knowledge about how to minimize under-
(between 60 and 95 percent in seven sites) reporting of violence. (See Chapter 7 for
Looking for households of women who had experienced physical more discussion of sampling techniques.)
in Samoa or sexual partner violence reported that Underreporting of violence will dilute asso-
ciations between potential risk factors and
health outcomes, leading to falsely nega-
tive results. Underestimating the dimen-
sions of violence may also prevent
violence intervention programs from
receiving the priority they deserve in the
allocation of resources.
Research demonstrates that disclosure
rates of violence are highly influenced by
the design and wording of questions, the
training of interviewers, and the imple-
mentation of the study.2 In Chapter 6, we

discuss this issue in much greater depth
and outline the variety of measures that
have been developed to enhance disclo-
sure of violence.

Using study results for social change to benefit most should bear a fair propor-
It is important to feed research findings tion of the risks and burdens of the study.
into ongoing advocacy, policy making, In the case of gender-based violence
and intervention activities. Too often criti- research, the risks are potentially large, but
cal research findings never reach the atten- so too are the risks of ignorance, silence,
tion of the policy makers and advocates and inaction. Researchers and ethical
best positioned to use them. The enor- review boards must constantly balance this
mous personal, social, and health-related reality. Lisa Fontes cites the case of a col-
costs of violence against women place a league from India who wanted to study
moral obligation on researchers and wives who were hospitalized after having
donors to try to ensure that study findings been burned by their husbands in disputes
are applied in the real world. It is also over dowry. She ultimately
important that the study community decided not to conduct the Women would ask me what
receives early feedback on the results of research for fear that the this survey was for, and how
the research in which it has participated. research would put women at it would help them. I would
Chapter 14 addresses this issue in more further risk. As Fontes tell them that we won’t see
detail and describes several successful observes, “Her decision elimi- the solution tomorrow or the
examples of how research findings have nated the research-related risk next year. Our daughters
been used to contribute to changing laws to the participants, but also and granddaughters will see
and policies on domestic violence. eliminated the potential benefit the fruits of this work, maybe
One way to improve the relevance of of reducing the terrible isola- things will be better by then.
research projects is, from the outset, to tion and vulnerability of these Nicaraguan fieldworker. (From
involve organizations that carry out advo- victims.”21 Ellsberg, et al, 2000. ) 19
cacy and direct support for survivors of It is possible to conduct

violence, either as full partners in the research on violence with full respect for
research or as members of an advisory ethical and safety considerations if proper
committee. Such committees can play an care and resources are devoted to this end.
important role in helping guide the study We must remember that women living with
design, advise on the wording of ques- violence are already at risk. Researchers
tions, assist with interviewer training, and cannot eliminate this reality, just as they
give guidance on possible forms of analy- cannot fully eliminate the possibility that
sis and the interpretation of results. These further harm will be caused by their study.
groups also have a central role to play in The obligation of researchers is to carefully
publicizing and applying the project’s weigh the risks and benefits of any study
findings. and to take every measure possible to limit
possible harm and to maximize possible
JUSTICE: BALANCING benefit. At the very least, we must ensure
RISKS AND BENEFITS OF that when women take risks to share their
RESEARCH ON VIOLENCE stories, we honor that risk by using the
AGAINST WOMEN findings for social change.
Research, like any endeavor that touches

people’s lives, involves inherent risks. The
principle of distributive justice demands
that the class of individuals bearing the
burden of research should receive an
appropriate benefit, and those who stand

CHAPTER TWO
1. Ellsberg M, Heise L. Bearing witness: Ethics in 14.Gossaye Y, Deyessa N, Berhane Y, et al. Women’s
domestic violence research. Lancet. health and life events study in rural Ethiopia.
2002;359(9317):1599-1604. Ethiopian Journal of Health Development.
2. Ellsberg M, Heise L, Peña R, Agurto S, Winkvist A. 2003;17(Second Special Issue):1-49.
Researching domestic violence against women: 15.Herman J. Trauma and Recovery: The Aftermath
Methodological and ethical considerations. Studies of Violence: From Domestic Abuse to Political
in Family Planning. 2001;32(1):1-16. Terror. New York: Basic Books; 1992.
3. Council for International Organizations of Medical 16.Shrader E. Personal Communication. Washington,
Sciences. International Guidelines for Ethical DC; 2000.
Review of Epidemiological Studies. Geneva: 17.Jansen HAFM, Watts C, Ellsberg M, Heise L,
CIOMS; 1991. Garcia-Moreno C. Interviewer training in the
4. World Health Organization. Putting Women’s WHO Multi-country Study on Women’s Health
Safety First: Ethical and Safety Recommendations and Domestic Violence Against Women. Violence
for Research on Domestic Violence Against against Women. 2004;10(7):831-849.
Women. Geneva: Global Programme on Evidence 18.Wessel L, Campbell J. Providing sanctuary for bat-
for Health Policy, World Health Organization; tered women: Nicaragua’s Casas de la Mujer.
1999. Report No.: WHO/EIP/GPE/99.2. Issues in Mental Health Nursing. 1997;18:455-476.
5. World Health Organization. WHO Multi-country 19.Ellsberg M. Candies in Hell: Research and Action
Study on Women’s Health and Domestic Violence on Domestic Violence in Nicaragua [Doctoral
Against Women: Study Protocol. Geneva: World Dissertation]. Umeå, Sweden: Umeå University;
Health Organization; 2004. 2000.
6. Yoshihama M. Personal Communication. Ann 20.Parker B, Ulrich Y. A protocol of safety: Research
Arbor, Michigan. Washington, DC. 2004. on abuse of women. Nursing Research.
7. World Health Organization. WHO Multi-country 1990;38:248-250.
Study on Women’s Health and Domestic Violence 21.Fontes LA. Ethics in family violence research:
Against Women: Study Questionnaire V10. Cross-cultural issues. Family Relations.
Geneva: World Health Organization; 2004. 1998;47:53-61.
8. Zimmerman K. Plates in a Basket Will Rattle:
Domestic Violence in Cambodia, a Summary.
Phnom Penh, Cambodia: Project Against
Domestic 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 Policy
Project; 1995.
10.Jewkes R, Watts C, Abrahams N, Penn-Kekana L,
Garcia-Moreno C. Ethical and methodological
issues in conducting research on gender-based
violence in Southern Africa. Reproductive Health
Matters. 2000;8(15):93-103.
11.Johnson H. Dangerous Domains: Violence
Against Women in Canada. Ontario, Canada:
International Thomson Publishing; 1996.
12.Finkelhor D, Hotaling GT, Yllo K. Special Ethical
Concerns in Family Violence Research. In:
Finkelhor D, Hotaling GT, Yllo K, editors.
Stopping Family Violence: Research Priorities for
the Coming Decade. London: Sage; 1988.
13.Hakimi M, Nur Hayati E, Ellsberg M, Winkvist A.
Silence for the Sake of Harmony: Domestic
Violence and Health in Central Java, Indonesia.
Yogyakarta, Indonesia: Gadjah Mada
University;PATH, Rifka Annisa, Umeå University;
2002.

Researching Vaw - Practical Guide

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Researching Vaw - Practical Guide

Uploaded by

Copyright:

Available Formats

Please see the Table of Contents for access to the entire publication.

Mary Ellsberg ■ Lori Heise

Ellsberg, Mary Carroll, Heise, Lori.

Suggested citation: Ellsberg M, and Heise L. Researching Violence Against Women: A

1. Domestic violence 2. Spouse abuse 3. Women 4. Research design 5. Manuals

ISBN 92 4 154647 6 (LC/NLM classification: HV 6556)

Printed in the United States.

For more information, please visit www.who.int/gender/en or www.path.org.

CHAPTER ONE: VIOLENCE AGAINST WOMEN

CHAPTER SIX: THE CHALLENGE OF DEFINING AND

CHAPTER NINE: TOOLS FOR COLLECTING

Building the research field team

CHAPTER ELEVEN: IN THE FIELD . . . . . . . . . . . . . . . . . . . 168

C H A P T E R T W E LV E : A N A LY Z I N G Q U A N T I TAT I V E D ATA . . . . . 186

CHAPTER FOURTEEN: FROM RESEARCH TO ACTION . . . . . . 216

1 AWPOPME NE ND I XI N1S :T RTUHME EWNHT O. V. I.O. L.E N. C. E. A. G. A. I.N. S.T . . . . . . . . . . . . 230

APPENDIX 2: TRAINING EXERCISES FOR INTERVIEWERS . . . 240

APPENDIX 3: SUGGESTED RESOURCES . . . . . . . . . . . . . . . 244

T his manual was truly a collective

A Practical Guide for Researchers and Activists 1

2 Researching Violence Against Women

A Practical Guide for Researchers and Activists 3

Editing by Cheryl Silver.

This manual was produced with the

4 Researching Violence Against Women

T wenty years ago, violence against

A Practical Guide for Researchers and Activists 5

6 Researching Violence Against Women

A Practical Guide for Researchers and Activists 7

Violence Against Women as a

Topics covered in this chapter:

Definitions of violence against women

V iolence against women is the most per-

A Practical Guide for Researchers and Activists 9

FIGURE 1.1 THE LIFE CYCLE OF VIOLENCE AGAINST WOMEN

10 Researching Violence Against Women

A Practical Guide for Researchers and Activists 11

12 Researching Violence Against Women

A Practical Guide for Researchers and Activists 13

Latin America and the Caribbean (continued)

14 Researching Violence Against Women

* Study population: I = all women; b

(Updated from Heise et al, 1999.2)

A Practical Guide for Researchers and Activists 15

■ % women who have ever experienced physical violence by a partner

(From WHO, 2005.20)

16 Researching Violence Against Women

FIGURE 1.4 INTIMATE PARTNER VIOLENCE ACCORDING TO TYPES OF

(From WHO, 2005.20)

P R E VA L E N C E A N D Much sexual coercion also takes place

A Practical Guide for Researchers and Activists 17

A review of 25 studies worldwide indicates Rwanda, Liberia, Sierra Leone, and

18 Researching Violence Against Women

TABLE 1.2 PREVALENCE OF CHILD SEXUAL ABUSE: SELECTED STUDIES, 1990–2003

A Practical Guide for Researchers and Activists 19

TABLE 1.2 PREVALENCE OF CHILD SEXUAL ABUSE: SELECTED STUDIES, 1990–2003

20 Researching Violence Against Women

Bangladesh Dhaka 2002 1369 15–49 24

(From World Health Organization, 20021, 2005.20)

A Practical Guide for Researchers and Activists 21

22 Researching Violence Against Women

FIGURE 1.5 HEALTH OUTCOMES OF VIOLENCE AGAINST WOMEN

Nonfatal Outcomes Fatal Outcomes