Content uploaded by Mary Ellsberg
Author content
All content in this area was uploaded by Mary Ellsberg
Content may be subject to copyright.
WHO Multi-country
Study on
Women’s Health
and Domestic Violence
against Women
Initial results on
prevalence, health outcomes
and women’s responses
Claudia García-Moreno
Henrica A.F.M. Jansen
Mary Ellsberg
Lori Heise
Charlotte Watts
Contents
Preface vi
Foreword vii
Acknowledgements ix
Executive summary xii
Introduction
Introduction 3
Background to the Study 3
International research on prevalence of violence against women 4
Study objectives 6
Organization of the Study 7
Participating countries 7
References 9
Methods
Definitions and questionnaire development 13
Definitions 13
Formative research 16
Development of the questionnaire 17
Questionnaire structure 17
Maximizing disclosure 17
Country adaptation and translation of the questionnaire 18
References 18
Sample design, ethical and safety considerations, and response rates 19
Sample design 19
Ethical and safety considerations 21
Response rates 22
References 24
Results
Prevalence of violence by intimate partners 27
Physical and sexual violence 28
Acts of physical violence 30
Acts of sexual violence 31
Overlap between physical and sexual violence 32
Demographic factors associated with violence 32
Acts of emotional abuse 35
Controlling behaviour 36
Women’s violence against men 36
Women’s attitudes towards violence 39
Discussion 41
References 42
1
2
3
4
CHAPTERCHAPTERCHAPTERCHAPTER
WHO Library Cataloguing-in-Publication Data
WHO multi-country study on women’s health and domestic violence
against women : initial results on prevalence, health outcomes
and women’s responses / authors: Claudia García-Moreno ... [et al.]
1. Domestic violence 2. Sex offenses 3. Women’s health
4. Cross-cultural comparison 5. Multicenter studies
6. Epidemiologic studies I. García-Moreno,
ISBN 92 4 159358 X (NLM classification: WA 309)
© World Health Organization 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
2476; 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 WHO Press, at the above address (fax: +41 22 791
4806; email: permissions@who.int).
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
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 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.
All reasonable precautions have been taken by the World Health
Organization to verify the information contained in this publication.
However, the published material is being distributed without
warranty of any kind, either express or implied. The responsibility
for the interpretation and use of the material lies with the reader.
In no event shall the World Health Organization be liable for
damages arising from its use.
Designed by: Grundy & Northedge Designers
Printed in Switzerland
Prevalence of violence by perpetrators other than intimate partners since
the age of 15 years 43
Physical violence by non-partners since the age of 15 years 43
Sexual violence by non-partners since the age of 15 years 45
Overall prevalence of non-partner violence since the age of 15 years 45
Non-partner violence compared with partner violence 46
Discussion 46
References 48
Prevalence of sexual abuse in childhood and forced first sexual experience 49
Sexual abuse before 15 years 49
Forced first sex 51
Discussion 52
References 54
Association between violence by intimate partners and women’s physical
and mental health 55
Women’s self-reported health and physical symptoms 55
Injuries caused by physical violence by an intimate partner 57
Mental health 59
Discussion 61
References 62
Associations between violence by intimate partners and women’s sexual and
reproductive health 63
Induced abortion and miscarriage 63
Use of antenatal and postnatal health services 64
Violence during pregnancy 65
Parity 66
Risk of sexually transmitted infections, including HIV 66
Discussion 69
References 71
Women’s coping strategies and responses to physical violence by
intimate partners 73
Who women tell about violence and who helps 73
Agencies or authorities to which women turn 74
Fighting back 76
Women who leave 77
Discussion 79
References 80
5
7
8
9
Conclusions and recommendations
Summary of findings, conclusions, and areas for further research 83
Prevalence and patterns of violence 83
Association of violence with specific health outcomes 85
Women’s responses and use of services 86
Strengths and limitations of the Study 87
Areas for further analysis 88
A basis for action 89
References 89
Recommendations 90
Strengthening national commitment and action 90
Promoting primary prevention 92
Involving the education sector 94
Strengthening the health sector response 95
Supporting women living with violence 96
Sensitizing criminal justice systems 96
Supporting research and collaboration 97
References 98
Annex 1 Methodology 101
Annex 2 Core Research Team and Steering Committee Members 118
Annex 3 Country research team members 119
Annex 4 Questionnaire 127
Statistical appendix 166
Index 198
6
10
11
CHAPTER
CHAPTERCHAPTERCHAPTERCHAPTER
CHAPTERCHAPTER
Contents (continued)
vi
vii
Executive Summary
WHO Multi-country Study on Women’s Health and Domestic Violence
vi
Violence against women is a universal phenomenon that persists in all countries of
the world, and the perpetrators of that violence are often well known to their victims.
Domestic violence, in particular, continues to be frighteningly common and to be accepted
as “normal” within too many societies. Since the World Conference on Human Rights, held
in Vienna in 1993, and the Declaration on the Elimination of Violence against Women in the
same year, civil society and governments have acknowledged that violence against women
is a public policy and human rights concern. While work in this area has resulted in the
establishment of international standards, the task of documenting the magnitude of violence
against women and producing reliable, comparative data to guide policy and monitor
implementation has been exceedingly difficult. The WHO Multi-country Study on Women’s
Health and Domestic Violence against Women is a response to this difficulty.
The Study challenges the perception that home is a safe haven for women by showing
that women are more at risk of experiencing violence in intimate relationships than
anywhere else. According to the Study, it is particularly difficult to respond effectively to this
violence because many women accept such violence as “normal”. Nonetheless, international
human rights law is clear: states have a duty to exercise due diligence to prevent, prosecute
and punish violence against women.
Looking at violence against women from a public health perspective offers a way of
capturing the many dimensions of the phenomenon in order to develop multisectoral
responses. Often the health system is the first point of contact with women who are victims
of violence. Data provided by this Study will contribute to raising awareness among health
policy-makers and care providers of the seriousness of the problem and how it affects the
health of women. Ideally, the findings will inform a more effective response from government,
including the health, justice and social service sectors, as a step towards fulfilling the state’s
obligation to eliminate violence against women under international human rights laws.
Violence against women has a far deeper impact than the immediate harm caused. It has
devastating consequences for the women who experience it, and a traumatic effect on those
who witness it, particularly children. It shames states that fail to prevent it and societies that
tolerate it. Violence against women is a violation of basic human rights that must be eliminated
through political will, and by legal and civil action in all sectors of society.
This report of the WHO Multi-country Study on Women’s Health and Domestic Violence
against Women, along with the recommendations it contains, is an invaluable contribution to
the struggle to eliminate violence against women.
Foreword
vii
Preface
Yakın Ertürk
Special Rapporteur on violence against women, its causes and consequences
Foreword
Violence against women by an intimate partner is a major contributor to the ill-health of
women. This study analyses data from 10 countries and sheds new light on the prevalence of
violence against women in countries where few data were previously available. It also uncovers
the forms and patterns of this violence across different countries and cultures, documenting the
consequences of violence for women’s health. This information has important implications for
prevention, care and mitigation.
The health sector can play a vital role in preventing violence against women, helping to
identify abuse early, providing victims with the necessary treatment, and referring women to
appropriate and informed care. Health services must be places where women feel safe, are
treated with respect, are not stigmatized, and where they can receive quality, informed support.
A comprehensive health sector response to the problem is needed, in particular addressing the
reluctance of abused women to seek help.
The high rates documented by the Study of sexual abuse experienced by girls and women
are of great concern, especially in light of the HIV epidemic. Greater public awareness of
this problem is needed and a strong public health response that focuses on preventing such
violence from occurring in the first place.
The research specialists and the representatives of women’s organizations who carried
out the interviews and dealt so sensitively with the respondents deserve our warmest thanks.
Most of all, I thank the 24 000 women who shared this important information about their lives,
despite the many difficulties involved in talking about it. The fact that so many of them spoke
about their own experience of violence for the first time during this study is both an indictment
of the state of gender relations in our societies, and a spur for action. They, and the countries
that carried out this groundbreaking research have made a vital contribution.
This study will help national authorities to design policies and programmes that begin to
deal with the problem. It will contribute to our understanding of violence against women and
the need to prevent it. Challenging the social norms that condone and therefore perpetuate
violence against women is a responsibility for us all. Supported by WHO, the health sector
must now take a proactive role in responding to the needs of the many women living in violent
relationships. Much greater investment is urgently needed in programmes to reduce violence
against women and to support action on the study’s findings and recommendations.
We must bring the issue of domestic violence out into the open, examine it as we would
the causes of any other preventable health problem, and apply the best remedies available.
LEE Jong-Wook
Director-General, World Health Organization
viii
Executive Summary
ix
Foreword
WHO Multi-country Study on Women’s Health and Domestic Violence
viii
First and foremost, we would like to
acknowledge and thank the more than
24 000 women who participated in the
Study, and who gave their time to answer
our questions and share their life
experiences with us.
We gratefully acknowledge the investigators
and collaborating institutions in the countries,
and the interviewers and other office and
field staff in the countries, who all worked
with immense dedication and commitment
to ensure the successful implementation of
the Study. Particular mention is made of
the investigators:
in Bangladesh, Ruchira Tabassum Naved and
Abbas Bhuiya (ICDDR,B: Centre for Health
and Population Research, Dhaka), Safia Azim
(Naripokkho, Dhaka) and Lars Ake Persson
(Uppsala University, Sweden);
Acknowledgements
in Brazil, Lilia Blima Schraiber, Ana Flavia Lucas
D’Oliveira and Ivan França-Junior (University
of São Paulo, São Paulo), Carmen Simone
Grilo Diniz (Feminist Collective for Health
and Sexuality, São Paulo), Ana Paula Portella
(SOS Corpo Genero e Cidadania, Recife),
Ana Bernarda Ludermir (Federal University of
Pernambuco, Recife);
in Ethiopia, Yemane Berhane, Negussie
Deyessa, Yegomawork Goyasse, Atalay Alem,
Derege Kebede and Alemayehu Negash (Addis
Ababa University, Addis Ababa), Ulf Hogberg,
Gunnar Kullgren and Maria Emmelin (Umeå
University, Sweden), Mary Ellsberg (PATH,
Washington, DC,USA);
in Japan, Mieko Yoshihama (University of
Michigan, Ann Arbor, USA), Saori Kamano
(National Institute of Population and Social
The Study, and this comparative report summarizing the major findings of surveys
conducted in 10 countries, was only possible because of the dedication, commitment and
hard work of all of those involved, both internationally and in the countries concerned.
In addition, the implementation of the Study was supported by many people in all of the
participating institutions. The World Health Organization and the authors would like to
thank all of those who contributed in different ways to making this Study happen, and
apologize to anyone who may inadvertently remain unnamed.
The recommendation for undertaking this research emerged from the WHO
Consultation on Violence against Women, held in 1996. The participants of that meeting,
in particular the late Raquel Tiglao, an advocate for women’s health and for services
for abused women from the Philippines, Mmatshilo Motsei, and Jacquelyn Campbell, all
pioneers in this work, inspired us to action.
The Study was undertaken as a key activity of the Department of Gender, Women
and Health (GWH) of the World Health Organization, and developed and supported by
the Core Research Team which is made up of: Charlotte Watts from the London School
of Hygiene and Tropical Medicine, Mary Ellsberg and Lori Heise of the Program for
Appropriate Technology in Health (PATH) in Washington, DC, and Henrica AFM Jansen
and Claudia García-Moreno (Study Coordinator) from WHO.
Acknowledgements
Each culture has its sayings and songs about the importance of home, and the comfort and
security to be found there. Yet for many women, home is a place of pain and humiliation.
As this report clearly shows, violence against women by their male partners is common,
wide-spread and far-reaching in its impact. For too long hidden behind closed doors and avoided in
public discourse, such violence can no longer be denied as part of everyday life for millions of women.
The research findings presented in this report reinforce the key messages of WHO’s World
Report on Violence and Health in 2002, challenging notions that acts of violence are simply
matters of family privacy, individual choice, or inevitable facts of life. The data collected by WHO
and researchers in 10 countries confirm our understanding that violence against women is an
important social problem. Violence against women is also an important risk factor for women’s
ill-health, and should receive greater attention.
Experience, primarily in industrialized countries, has shown that public health approaches to
violence can make a difference. The health sector has unique potential to deal with violence against
women, particularly through reproductive health services, which most women will access at some
point in their lives. The Study indicates, however, that this potential is far from being realized. This
is partly because stigma and fear make many women reluctant to disclose their suffering. But it is
also because few doctors, nurses or other health personnel have the awareness and the training
to identify violence as the underlying cause of women’s health problems, or can provide help,
particularly in settings where other services for follow-up care or protection are not available. The
health sector can certainly not do this alone, but it should increasingly fulfil its potential to take a
proactive role in violence prevention.
Violence against women is both a consequence and a cause of gender inequality. Primary
prevention programmes that address gender inequality and tackle the many root causes of
violence, changes in legislation, and the provision of services for women living with violence are
all essential. The Millennium Development Goal regarding girls’ education, gender equality and
the empowerment of women reflects the international community’s recognition that health,
development, and gender equality issues are closely interconnected.
WHO regards the prevention of violence in general – and violence against women in particular
– a high priority. It offers technical expertise to countries wishing to work against violence, and urges
international donors to support such work. It continues to emphasize the importance of
action-oriented, ethically based research, such as this Study, to increase our understanding of the
problem and what to do about it. It also strongly urges the health sector to take a more proactive
role in responding to the needs of the many women living in violent relationships.
Joy Phumaphi
Assistant Director-General, Family and Community Health, WHO
x
xi
Statistical appendix
Security Research, Tokyo), Hiroko Akiyama
(University of Tokyo, Tokyo), Fumi Hayashi (Toya
Eiwa University, Tokyo), Tamie Kaino (Ochanomizu
University, Tokyo), Tomoko Yunomae (Japan
Accountability Caucus, Beijing, Tokyo);
in Namibia, Eveline January, Hetty Rose-Junius
and Johan Van Wyk (Ministry of Health and
Social Services, Windhoek), Alvis Weerasinghe
(National Planning Commission, Windhoek);
in Peru, Ana Güezmes García (Centro de
la Mujer Flora Tristan, Lima), Nancy Palomino
Ramirez and Miguel Ramos Padilla (Universidad
Peruana Cayetano Heredia, Lima);
in Samoa, Tina Tauasosi-Posiulai, Tima Levai-
Peteru, Dorothy Counts and Chris McMurray
(Secretariat of the Pacific Community);
in Serbia and Montenegro, Stanislava
Otaševi� and Silvia Koso (Autonomous Women’s
Center Against Sexual Violence, Belgrade),
Viktorija Cucic (University of Belgrade, Belgrade);
in Thailand, Churnrurtai Kanchanachitra,
Kritaya Archavanitkul and Wassana Im-em
(Mahidol University, Bangkok), Usa Lerdsrisanthat
(Foundation for Women, Bangkok);
in the United Republic of Tanzania,
Jessie Mbwambo and Gideon Kwesigabo
(Muhimbili College of Medical Sciences), Joe
Lugalla (University of New Hampshire, Durham,
USA), Sherbanu Kassim (Women’s Research and
Documentation Project, Dar es Salaam).
WHO would also like to thank the members of
the Steering Committee of the Study: Jacquelyn
Campbell, Co-Chair (USA), Lucienne Gillioz
(Switzerland), Rachel Jewkes (South Africa), Ivy
Josiah (Malaysia), Olav Meirik, Co-Chair (Chile/
Norway), Laura Rodrigues (United Kingdom/
Brazil), Irma Saucedo (Mexico), Berit Schei
Finally, the Study would not have been
possible without the generous financial support
given to WHO by the Governments of the
Netherlands, Norway, Sweden, Switzerland and
the United Kingdom. The Rockefeller Foundation
supported the implementation of the Study in
Namibia and the United Republic of Tanzania, and
provided the use of its meeting place in Bellagio
in May 2004, where the initial preparations for
this report with representatives of all the country
research teams took place. This funding enabled
WHO to develop the Study methods and
materials, establish and work with the country
research teams to implement the Study. WHO
provided the funding for the implementation
of the Study in 6 of the 8 initial countries:
Brazil, Japan, Namibia, Peru, Thailand and the
United Republic of Tanzania, with some of these
countries receiving additional funds as detailed
in Annex 3. In Bangladesh, the Study was funded
by the Urban Primary Health Care project of
the Government of Bangladesh; in Ethiopia by
the Swedish Agency for Research Cooperation
with Developing Countries (SAREC/Sida);
in Samoa, by the United Nations Fund for
Population Activities (UNFPA); and in Serbia and
Montenegro, by Trocaire. We also acknowledge
the contribution from the Global Coalition on
Women and AIDS.
About the authors
The authors make up the WHO Core Research
Team for the Study, involved in the development
of the study methodology, questionnaire and
manuals, proving technical and scientific support
to the countries in the study and responsible for
cross-country analysis and reports on the results
of the study.
Claudia García-Moreno is Coordinator in
the WHO Department of Gender, Women and
Health and is the Study Coordinator. She joined
(Norway) and Stig Wall (Sweden). In addition
to their continued support to the Study, they
reviewed and gave valuable input to several drafts
of the report.
The Study would not have been possible
without the support of numerous individuals
within WHO: Tomris Türmen, David Evans,
Nafsiah Mboi, Daniel Makuto, Eva Wallstam
and Joy Phumaphi who, over the period of the
Study, have overseen WHO’s work on gender
and women’s health, under which this Study
was developed and implemented. Particular
thanks are due to colleagues in the Department
of Reproductive Health and Research, in
particular Paul Van Look, Timothy Farley and
Jane Cottingham, for their continuous support
since the Study’s early days. Linda Morison of the
London School of Hygiene and Tropical Medicine,
Timothy Farley of WHO and Stig Wall of Umeå
University provided advice on sampling and other
statistical matters during the planning stages.
Chandrika John, Eva Lustigova, Jenny Perrin, Lesley
Robinson, Lindsay Simmons, Margaret Squadrani
and Ludy Suryantoro provided administrative
support to the Study.
This report also benefited from the
contributions of a number of other people. In
particular, Alexander Butchart, Etienne Krug and
Alison Phinney, in the Department of Injuries and
Violence Prevention, provided valuable comments
on an earlier version of this report. Jose
Bertolote in the Department of Mental Health,
Jack Jones in School Health and Youth Health
Promotion, and Paul Van Look also reviewed
specific sections and provided useful comments.
WHO acknowledges the following contributions
to the production of the report: Andrew Wilson,
for preparing the summary of the report; Angela
Haden and Pat Butler, for editing the manuscript;
Susan Kaplan and Ann Morgan, for proofreading;
Barbara Campanini, for editing the references; Liza
Furnival, for preparing the index; Tilly Northedge,
for the layout and cover design; and Andrew
Dicker, for formatting the report.
Acknowledgements
WHO Multi-country Study on Women’s Health and Domestic Violence
WHO in 1994 and initiated and developed
its work on violence against women. She was
responsible for overseeing the implementation of
the Study, and, with Lori Heise, for developing the
initial proposal for it.
Henrica AFM (Henriette) Jansen is
Epidemiologist to the WHO Multi-country Study
on Women’s Health and Domestic Violence
against Women in the WHO Department of
Gender, Women and Health. She was the lead
person for the final versions of the questionnaire
and data entry and processing programs, and
managed data collection and analysis.
Charlotte Watts is a Senior Lecturer in
Epidemiology and Health Policy in the Health
Policy Unit, Department of Public Health and
Policy, London School of Hygiene and Tropical
Medicine and a Technical Adviser to the WHO
Multi-country Study on Women’s Health
and Domestic Violence against Women. She
developed the initial protocol and questionnaire
for the Study.
Mary Carroll Ellsberg is Senior Adviser for
Gender, Violence and Human Rights at PATH in
Washington, DC, USA. She is an epidemiologist
and has also participated in research on violence
against women in Nicaragua, Indonesia and
Ethiopia. She is the lead author of “Researching
violence against women: a practical guide for
researchers and activists”, which synthesizes the
experience from the WHO Study and other
research on violence against women.
Lori Heise is Director of the Global Campaign
for Microbicides at PATH and a research fellow
in health policy at the London School of Hygiene
and Tropical Medicine. She has worked for over
two decades on intersecting issues of gender,
power, sexuality and violence. She is a co-author
of “Researching violence against women: a practical
guide for researchers and activists”.
xi
xiii
Statistical appendix
WHO Multi-country Study on Women’s Health and Domestic Violence
Organization of the Study
The Study consisted of standardized
population-based household surveys. In five
countries (Bangladesh, Brazil, Peru, Thailand, and
the United Republic of Tanzania), surveys were
conducted in (a) the capital or a large city and
(b) one province or region, usually with urban
and rural populations. One rural setting was used
in Ethiopia, and a single large city was used in
Japan, Namibia, and Serbia and Montenegro. In
Samoa, the whole country was sampled. In this
report, sites are referred to by country name
followed by either “city” or “province”; where
only the country name is used, it should be taken
to refer to both sites.
Work was coordinated by WHO with a
core research team of experts from the London
School of Hygiene and Tropical Medicine
(LSHTM), the Program for Appropriate
Technology in Health (PATH), and WHO
itself. A research team was established in each
country, including representatives from research
organizations and women’s organizations
providing services to abused women. The survey
used female interviewers and supervisors trained
using a standardized 3-week curriculum. Strict
ethical and safety guidelines were adhered to in
each country.
Violence against women by
intimate partners
The results indicate that violence by a male
intimate partner (also called “domestic violence”)
is widespread in all of the countries included
in the Study. However, there was a great deal
of variation from country to country, and from
setting to setting. This indicates that this violence
is not inevitable.
Physical violence by intimate partners
The proportion of ever-partnered women
who had ever suffered physical violence by a
male intimate partner ranged from 13% in
Japan city to 61% in Peru province, with
most sites falling between 23% and 49%.
The prevalence of severe physical violence
(a woman being hit with a fist, kicked, dragged,
Executive summary
choked, burnt on purpose, threatened with
a weapon, or having a weapon used against
her) ranged from 4% in Japan city to 49% in
Peru province. The vast majority of women
physically abused by partners experienced acts
of violence more than once.
Sexual violence by intimate partners
The range of lifetime prevalence of sexual
violence by an intimate partner was between 6%
(Japan city and Serbia and Montenegro city) and
59% (Ethiopia province), with most sites falling
between 10% and 50%. While in most settings
sexual violence was considerably less frequent
than physical violence, sexual violence was
more frequent in Bangladesh province, Ethiopia,
province and Thailand city.
Physical and sexual violence by intimate
partners
For ever-partnered women, the range of
lifetime prevalence of physical or sexual
violence, or both, by an intimate partner
was 15% to 71%, with estimates in most sites
ranging from 30% to 60%. Women in Japan city
were the least likely to have ever experienced
physical or sexual violence, or both, by an
intimate partner, while the greatest amount
of violence was reported by women living in
provincial (for the most part rural) settings
in Bangladesh, Ethiopia, Peru, and the United
Republic of Tanzania. Likewise, regarding
current violence – as defined by one or
more acts of physical or sexual violence
in the year prior to being interviewed – the
range was between 3% (Serbia and
Montenegro city) and 54% (Ethiopia province),
with most sites falling between 20% and 33%.
These findings illustrate the extent to which
violence is a reality in partnered women’s
lives, with a large propor tion of women
having some experience of violence during
their partnership, and many having recent
experiences of abuse.
Emotionally abusive acts and controlling
behaviours
Emotionally abusive acts by a partner included:
being insulted or made to feel bad about
oneself; being humiliated in front of others;
being intimidated or scared on purpose; or
being threatened directly, or through a threat
to someone the respondent cares about.
Across all countries, between 20% and 75%
of women had experienced one or more of
these acts, most within the past 12 months.
Data were also collected about partners’
controlling behaviours, such as: routinely
attempting to restrict a woman’s contact with
her family or friends, insisting on knowing
where she is at all times, and controlling her
access to health care. Significantly, the number
of controlling behaviours by the partner was
associated with the risk of physical or sexual
violence, or both.
Women’s attitudes towards violence
In addition to women’s experience, the Study
investigated women’s attitudes to partner
violence including: (a) the circumstances in
which they believed it was acceptable for a
man to hit or physically mistreat his wife, and
(b) their beliefs about whether and when
a woman may refuse to have sex with her
husband. There was wide variation in women’s
acceptance of different reasons, and indeed of
the idea that violence was ever justified. While
over three quarters of women in the city
sites of Brazil, Japan, Namibia, and Serbia and
Montenegro said no reason justified violence,
less than one quarter thought so in the
provincial settings of Bangladesh, Ethiopia, and
Peru. Acceptance of wife-beating was higher
among women who had experienced abuse
than among those who had not.
Respondents were also asked whether they
believed a woman has a right to refuse to have
xii
Executive summary
xiii
This report of the WHO Multi-country Study on Women’s Health and Domestic Violence
against Women analyses data collected from over 24 000 women in 10 countries
representing diverse cultural, geographical and urban/rural settings: Bangladesh, Brazil,
Ethiopia, Japan, Peru, Namibia, Samoa, Serbia and Montenegro, Thailand, and the United
Republic of Tanzania. The Study was designed to:
estimate the prevalence of physical, sexual and emotional violence against women, with
particular emphasis on violence by intimate partners;
assess the association of partner violence with a range of health outcomes;
identify factors that may either protect or put women at risk of partner violence;
document the strategies and services that women use to cope with violence by an
intimate partner.
This report presents findings on objectives 1, 2, and 4. The third, analysis of risk and
protective factors, will be addressed in a future report.
1
2
3
4
xiv
xv
Statistical appendix
sex with her partner in a number of situations,
including: if she is sick, if she does not want to
have sex, if he is drunk, or if he mistreats her.
In the provinces of Bangladesh, Ethiopia, Peru,
and the United Republic of Tanzania, and in
Samoa, between 10% and 20% of women felt
that women did not have the right to refuse sex
under any of these circumstances.
Non-partner physical and sexual violence
In addition to partner violence, the WHO Study
also collected data on physical and sexual abuse
by perpetrators – male and female – other than
a current or former male partner.
Non-partner physical violence since
age 15 years
Women’s reports of experience of physical
violence by a non-partner since the age of
15 varied widely. By far the highest level of
non-partner physical violence was reported
in Samoa (62%), whereas less than 10% of
women in Ethiopia province, Japan city, Serbia
and Montenegro city, and Thailand repor ted
non-partner physical violence. Commonly
mentioned perpetrators included fathers and
other male or female family members. In some
settings (Bangladesh, Namibia, Samoa, and the
United Republic of Tanzania), teachers were also
frequently mentioned.
Non-partner sexual violence since
age 15 years
The highest levels of sexual violence by
non-partners since age 15 years – between 10%
and 12% – were reported in Peru, Samoa, and
the United Republic of Tanzania city, while levels
below 1% were reported in Bangladesh province
and Ethiopia province. The perpetrators included
strangers, boyfriends and, to a lesser extent, male
family members (excluding fathers) or male
friends of the family.
Comparing partner and non-partner violence
since age 15 years
A common perception is that women are more
at risk of violence from strangers than from
partners or other men they know. The data show
that this is far from the case. In the majority
of settings, over 75% of women physically or
sexually abused by any perpetrator since the age
of 15 years reported abuse by a partner. In only
two settings, Brazil city and Samoa, were at least
40% of women abused only by someone other
than a partner.
Sexual abuse before age 15 years
Early sexual abuse is a highly sensitive issue
that is difficult to explore in a survey. The
Study therefore used a two-stage process
allowing women to report both directly and
anonymously (without having to reveal their
response to the interviewer) whether anyone
had ever touched them sexually, or made
them do something sexual that they did not
want to before the age of 15 years. In all but
one setting, anonymous reporting resulted in
substantially more reports of sexual abuse, and
large differences were recorded in Ethiopia
province (0.2% using direct reporting versus
7% anonymously), Japan city (10% versus 14%),
Namibia city (5% versus 21%), and the United
Republic of Tanzania city (4% versus 11%). “Best
estimates” based on the method that yielded the
higher rate, indicate that prevalence of sexual
abuse before 15 years of age varied from 1%
(Bangladesh province) to 21% (Namibia city).
The most frequently mentioned perpetrators
were male family members other than a father
or stepfather.
Forced first sex
In 10 of the 15 settings, over 5% of women
reported their first sexual experience as forced,
with more than 14% reporting forced first sex
in Bangladesh, Ethiopia province, Peru province,
and the United Republic of Tanzania. In all sites
except Ethiopia province, the younger a woman
at first experience of sex, the greater the
likelihood that this was forced. In more than half
the settings, over 30% of women who reported
first sex before the age of 15 years described
that sexual experience as forced. In some
countries (notably Bangladesh and Ethiopia
province), high levels of forced first sex are likely
to be related to early sexual initiation in the
context of early marriage, rather than being by
perpetrators other than partners.
Violence by intimate partners and
women’s health
Although a cross-sectional survey cannot
establish whether violence causes particular
health problems (with the obvious exception
of injuries), the Study results strongly support
other research which has found clear associations
between partner violence and symptoms of
physical and mental ill-health.
Injury resulting from physical violence
The prevalence of injury among women who
had ever been physically abused by their partner
ranged from 19% in Ethiopia province to 55%
in Peru province and was associated with the
severity of the violence. In Brazil, Peru province,
Samoa, Serbia and Montenegro city, and Thailand,
over 20% of ever-injured women reported
that they had been injured many times. At least
20% of ever-injured women in Namibia, Peru
province, Samoa, Thailand city, and the United
Republic of Tanzania reported injuries to the
eyes and ears.
Physical health
In the majority of settings, women who had ever
experienced partner violence were significantly
more likely to report poor or very poor health
than women who had never experienced
partner violence. Ever-abused women were also
more likely to have had problems walking and
carrying out daily activities, pain, memory loss,
dizziness, and vaginal discharge in the 4 weeks
prior to the interview. An association between
recent ill-health and lifetime experience of violence
suggests that the physical effects of violence
may last a long time after the actual violence has
ended, or that violence over time may have a
cumulative effect.
Mental health and suicide
In all settings, women who had ever experienced
physical or sexual violence, or both, by an
intimate partner reported significantly higher
levels of emotional distress and were more
likely to have thought of suicide, and to have
attempted suicide, than women who had never
experienced partner violence.
Reproductive health and violence during
pregnancy
In the majority of settings, ever-pregnant
women who had experienced partner physical
or sexual violence, or both were significantly
more likely to report having had at least one
induced abortion than women who had never
experienced partner violence. Similar patterns
were found for miscarriage, but the strength of
the association was less.
The proportion of ever-pregnant women
physically abused during at least one pregnancy
exceeded 5% in 11 of the 15 settings. Between
one quarter and one half of women physically
abused in pregnancy were kicked or punched in
the abdomen. In all sites, over 90% were abused
by the biological father of the child the woman
was carrying. The majority of those beaten
during pregnancy had experienced physical
violence before, with between 8% and 34%
reporting that the violence got worse during
the pregnancy. However, from 13% (Ethiopia
province) to about 50% (Brazil city and Serbia
and Montenegro city) were beaten for the first
time during pregnancy.
Executive summary
xv
WHO Multi-country Study on Women’s Health and Domestic Violence
xvi
xvii
Statistical appendix
Risk of HIV and other sexually
transmitted infections
The WHO Study explored the extent to which
women knew whether or not their partner
had had other sexual partners during their
relationship. Across all sites except Ethiopia, a
woman who reported that her intimate par tner
had been physically or sexually violent towards
her was significantly more likely to report that
she knew that her partner was or had been
sexually involved with other women while
being with her.
Women were also asked whether they
had ever used a condom with their partner,
whether they had requested use of condom,
and whether the request had been refused.
The proportion of women who had ever
used a condom with a current or most
recent partner varied greatly across sites.
No significant difference was found in use of
condoms between abused and non-abused
women, with the exception of Thailand
and the United Republic of Tanzania, where
women in a violent relationship were more
likely to have used condoms. However, in a
number of sites (cities in Peru, Namibia, and
the United Republic of Tanzania) women in
violent partnerships were more likely than
non-abused women to have asked their
partner to use condoms. Women in violent
partnerships in these sites, as well as in Brazil
city, Peru province, and Serbia and Montenegro,
were significantly more likely than non-abused
women to report that their par tner had
refused to use a condom. These findings, as well
as the high levels of child sexual abuse, are of
concern in the transmission of HIV and other
STIs, and underline the urgent need to address
this hidden but widespread abuse against
women.
Women’s responses to physical violence
by an intimate partner
Who women talk to
In all countries, the interviewer was the first
person to whom many abused women had ever
talked about their partner’s physical violence.
Two thirds of women who had been physically
abused by their partner in Bangladesh, and about
one half in Samoa and Thailand province, said
they had not told anybody about the violence
prior to the interview. In contrast, about 80% of
physically abused women in Brazil and Namibia
city had told someone, usually family or friends.
But this means that even in these settings, two
out of ten women had kept silent. Relatively few
women in any setting had told staff of formal
services or individuals in a position of authority
about the violence.
Which agencies or authorities women turn to
Over half of physically abused women
(between 55% and 95%) reported that they
had never sought help from formal services
(health services, legal advice, shelter) or
from people in positions of authority (police,
women’s nongovernmental organizations
(NGOs), local leaders, and religious leaders).
Only in Namibia city and Peru had more than
20% of women contacted the police, and only
in Namibia city and the United Republic of
Tanzania city had more than 20% sought help
from health care services.
Low use of formal services reflects in
part their limited availability. However, even
in countries relatively well supplied with
resources for abused women, barriers such
as fear, stigma and the threat of losing their
children stopped many women from seeking
help. In all settings, the most frequently given
reasons for seeking help were related to the
severity of the violence, its impact on the
children, or encouragement from friends
and family to seek help.
WHO Multi-country Study on Women’s Health and Domestic Violence
Leaving or staying with a violent partner
Between 19% and 51% of women who had
been physically abused by their partner had ever
left home for at least one night. Between 8%
and 21% reported leaving 2–5 times. In most
settings, women mainly reported going to their
relatives, and to a lesser extent to friends or
neighbours. Shelters were mentioned only in
Brazil city and Namibia city (by less than 1% of
women who left). Again, these patterns are likely
to reflect both the availability of places of safety
for women and their children, as well as culturally
specific factors relating to the acceptability of
women leaving or staying somewhere without
their partner.
Areas for further analysis
This first report provides descriptive
information on some of the main elements
addressed by the WHO Study. However, it
represents only the first stage of analysis of
an extensive database which has the potential
to address a range of important questions
regarding violence against women. Questions
that will be explored during the next stage
of analysis include risk profiles for violence
in terms of the timing and duration of the
relationship with the violent partner; risk and
protective factors for partner violence and
whether they are context-specific or spanning
all or most contexts; issues around definitions
and prevalence of emotional abuse; more
in-depth analysis of the relationship between
violence and health and of patterns of women’s
responses to violence; and the impact of
violence on other aspects of women’s lives,
including the effect on their children. These
questions are of great relevance to public
health, and exploring them will substantially
improve our understanding of the nature,
causes and consequences of violence, and the
best ways to intervene against it.
Recommendations
In keeping with their responsibility for the
well-being and safety of their citizens, national
governments, in collaboration with NGOs, donors
and international organizations, need to implement
the following recommendations. These are based
on the Study findings, and are grouped by theme.
Strengthening national commitment and action
1. Promote gender equality and women’s
human rights, in line with relevant
international treaties and human rights
mechanisms, including addressing women’s
access to property and assets, and
expanding educational opportunities for
girls and young women.
2. Establish, implement and monitor action
plans to address violence against women,
including violence by intimate partners.
3. Enlist social, political, religious, and other
leaders in speaking out against violence
against women.
4. Enhance capacity and establish systems for
data collection to monitor violence against
women, and the attitudes and beliefs that
perpetuate the practice.
Promoting primary prevention
5. Develop, implement and monitor
programmes aimed at primary prevention
of intimate partner violence and sexual
violence against women. These should
include sustained public awareness activities
aimed at changing the attitudes, beliefs and
values that condone partner violence as
normal and prevent it being challenged or
talked about.
6. Give higher priority to combating sexual
abuse of girls (and boys) in public health
programmes, as well as in responses by other
sectors such as the judiciary, education, and
social services.
xvii
Executive summary
xviii
Introduction7. Integrate responses to violence against women
into existing programmes for the prevention
of HIV and AIDS, and for the promotion
of adolescent health, including to promote
the prevention of sexual violence as well as
intimate-partner violence against women as
an integral part of these programmes.
8. Make physical environments safer for
women, through measures such as identifying
places where violence often occurs,
improving lighting, and increasing police and
other vigilance.
Involving the education sector
9. Make schools safe for girls, by involving
education systems in anti-violence efforts,
including eradicating teacher violence, as well
as engaging in broader anti-violence efforts.
Strengthening the health sector response
10. Develop a comprehensive health sector
response to the various impacts of violence
against women, and in particular address
the barriers and stigma that prevent abused
women from seeking help. This includes
supporting mental health services to
address violence against women as an
important underlying factor in women’s
mental health problems.
11. Use reproductive health services as entry
points for identifying and supporting women
in abusive relationships, and for delivering
referral or support services.
Supporting women living with violence
12. Strengthen formal and informal support
systems for women living with violence.
Sensitizing criminal justice systems
13. Sensitize legal and justice systems to
the particular needs of women victims
of violence.
Supporting further research and collaboration
and increasing donor support
14. Promote and support further research on the
causes and consequences of violence against
women and on effective prevention measures.
15. Increase support to programmes to reduce
and respond to violence against women.
WHO Multi-country Study on Women’s Health and Domestic Violence
Background to the Study
Until recently, most governments and
policy-makers viewed violence against women
as a relatively minor social problem affecting a
limited number of women. The general view was
that cases of violence could be appropriately
addressed through the social welfare and justice
systems. During the past decade, however, the
combined efforts of grass-roots and international
women’s organizations, international experts,
and committed governments have resulted in
a profound transformation in public awareness
regarding this issue (1). Violence against women,
also known as gender-based violence, is now
widely recognized as a serious human rights
abuse, and increasingly also as an important
public health problem that concerns all sectors
of society (2, 3).
Recognition of violence as a health and
rights issue was underscored and strengthened
by agreements and declarations at key
international conferences during the 1990s,
including the World Conference on Human
Rights (Vienna, 1993) (4), the International
Conference on Population and Development
(Cairo, 1994) (5) and the Fourth World
Conference on Women (Beijing, 1995) (6).
Through these international agreements,
governments have increasingly recognized
the need to develop broad multisectoral
approaches for the prevention of and
response to violence against women, and have
committed themselves to implement
the institutional and legislative reforms
necessary to achieve this goal. Despite this
progress, many governments still do not
acknowledge the problem of violence
against women or take measures to prevent
and address it. While the many health
consequences of violence are also increasingly
recognized, the involvement of the health
sector in responding to the problem is still
inadequate in many countries.
Why did WHO embark on a study of violence
against women?
In 1995, the Beijing Platform for Action
identified the lack of adequate information
on the prevalence, nature, causes, and
consequences of violence globally as a serious
obstacle to the development of effective
strategies to address violence. Governments
were urged to invest in research to improve
the relevant knowledge base on the prevalence,
causes, nature, and consequences of violence
against women (6, p.129a).
WHO’s work on gender-based violence
began in 1996 with the convening of an expert
consultation on violence against women. The
consultation brought together researchers, health
care providers and women’s health advocates
from several countries (7). The participants
agreed that there was a dearth of comparable
data, par ticularly from developing countries, that
many governments were reluctant to recognize
violence against women as a problem, and
that health was an important perspective from
which to address this issue. The consultation
recommended that WHO promote and support
international research to explore the dimensions,
health consequences and risk factors of violence
against women. In the same year, the World
Health Assembly declared the prevention of
violence, including violence against women and
children, to be a public health priority needing
urgent action. In response, in 1997, WHO
initiated the development of the Multi-country
Study on Women’s Health and Domestic
Violence against Women (hereafter referred to
as the WHO Study or the Study) (8).
More recently, WHO published the World
report on violence and health (9), which included
a global overview of available information −
including prevalence data − on intimate partner
and sexual violence and their impact on the
health and well-being of women (Chapters
4 and 6). That report recognized the need
for sound and reliable information on the
1Introduction
CHAPTER
This survey should have been conducted
10 years ago. Now I have two daughters. I hope they
will benefit from it.
Woman interviewed in Bangladesh
Thank you so much, I needed to talk to
someone. I have never told anyone what I told you,
but I would like that it happens more often that
someone comes to talk. There should be more
people who come to talk.
Woman interviewed in Peru
4
WHO Multi-country Study on Women’s Health and Domestic Violence
5
magnitude, the nature and the consequences
of violence, as an essential foundation for the
public health approach to violence, including
violence against women. This Study both
informed the WHO report and is an important
contribution to meeting the need for information
on violence, both nationally (in the countries
that participated) and globally. The results of
the Study will also feed into and inform WHO’s
Global Campaign on Violence Prevention, which
was launched in 2002 (for more information, see
http://www.who.int/violence_injury_prevention/
violence/global_campaign/en/).
International research on prevalence of
violence against women
The Declaration on the Elimination of Violence
against Women adopted by the United Nations
General Assembly in 1993 defined violence
against women as “any act of gender-based
violence that results in, or is likely to result in,
physical, sexual or psychological harm or suffering
to women, including threats of such acts, coercion
or arbitrary deprivation of liberty, whether
occurring in public or in private life” (10). It goes
on to define the various forms that this violence
can take. Although intimate-partner violence
and sexual coercion are the most common and
“universal” types of violence affecting women and
girls, in many parts of the world violence takes
on special characteristics according to cultural
and historical conditions, and includes murders
in the name of honour (so-called “honour
killings”), trafficking of women and girls, female
genital mutilation, and violence against women in
situations of armed conflict.
International research conducted over the
past decade has provided increasing evidence
of the extent of violence against women,
particularly that perpetrated by intimate male
partners. The findings show that violence against
women is a much more serious and common
problem than previously suspected. A review
of over 50 population-based studies performed
in 35 countries prior to 1999 indicated that
between 10% and 52% of women around the
world report that they have been physically
abused by an intimate partner at some point in
their lives, and between 10% and 30% that they
have experienced sexual violence by an intimate
partner. Between 10% and 27% of women and
girls reported having been sexually abused, either
as children or as adults (9, 11).
While these studies helped focus attention
on the issue, they also raised many questions
regarding the methods used to obtain estimates
of violence in different countries. There were
many differences in the way violence was defined,
measured and presented. For example, some
studies of partner violence include only physical
violence, while others may also include sexual or
emotional violence. Some studies measure lifetime
experiences of violence, whereas others include
only experiences in the current relationship, or
in a defined period. Studies also differ in other
important respects, such as the definition of the
study population (for example, in terms of the
age range and partnership status of the women),
the forms of violence considered, the range of
questions asked, and whether measures were
taken to ensure privacy and confidentiality of
interviews. Such factors have since been shown to
greatly affect prevalence estimates by influencing
a woman’s willingness to disclose abuse (12, 13).
These methodological differences between
studies have made it difficult to draw meaningful
comparisons or to understand the similarities and
differences in the extent, patterns, and factors
associated with violence in different settings (4).
In response to the methodological and
ethical challenges associated with research on
prevalence of gender-based violence in developing
countries, a group of researchers and advocates
from around the world came together in the
early 1990s to form the International Research
Network on Violence against Women (IRNVAW).
The purpose of the network was to create a
forum for sharing insights and for addressing key
challenges faced by investigators interested in
gender-based violence, such as: how to ensure the
safety of respondents and researchers throughout
the research process, and how to define and
measure violence in a way that allowed results to
be compared across diverse cultural settings (14).
The design and implementation of the WHO
Study incorporated the recommendations of
IRNVAW. It also built on methodological work
and research on violence by partners, carried out
primarily in the United States using the Conflict
Tactics Scale (15, 16), as well as critiques of this
methodology by other researchers (17). Since
the initiation of the WHO Study, a number of
other international research initiatives have also
used population-based surveys to estimate the
prevalence of different forms of violence against
women across countries and cultures. These
include: the World Sur veys of Abuse in Family
Environments (WorldSafe) supported by the
International Clinical Epidemiology Network
(INCLEN) (18), and the International Violence
Against Women Survey (IVAWS) conducted
by the European Institute for Crime Prevention
4
and Control, affiliated with the United Nations
(HEUNI), the United Nations Interregional
Crime and Justice Research Institute (UNICRI)
and Statistics Canada. These studies provide
useful comparisons with aspects of the WHO
Study and, taken together, are beginning to give a
more comprehensive picture of violence against
women around the world.
In addition, the Demographic and Health
Surveys (DHS), supported by MACRO
International and the United States Agency
for International Development (USAID), and
the International Reproductive Health Surveys
(IRHS), suppor ted by the United States Centers
for Disease Control and Prevention (CDC),
now contain a number of questions or a module
on violence against women as part of broader
household surveys on a range of health issues
(19). These surveys offer the advantages of
large sample size, efficiency of data collection,
standardization of measurement instruments
and the possibility of being generalized to the
national population. It has been shown, however,
that focused studies on violence against women
tend to give higher prevalence estimates than
larger health or other surveys which include
only one or a small number of questions on
violence (13). As a result, the DHS have moved
away from single or limited questions to use of
a full violence module in countries that wish to
explore this issue. The module was developed
on the basis of an early draft of the WHO
Study protocol and so provides opportunities
for expanding the database of comparable data.
Furthermore, DHS now recommend the use
of the WHO ethical and safety guidelines when
applying the violence module. This is important,
as the safety of respondents and interviewers
is an important concern when questions about
violence are included in the context of larger
surveys on other issues.
The 1990s also saw rapid growth in the
number of studies exploring the potential
health consequences of violence, particularly
in the United States and other industrialized
countries. For years, clinicians and policy-makers
had focused on injury as the primary health
outcome of violence – if they considered health
outcomes at all. Then, research began to draw
attention to a range of other health-related
conditions associated with intimate-partner
violence and sexual abuse of women, such
as chronic pain syndromes, drug and alcohol
abuse, complications of pregnancy, increased
risk of unwanted pregnancy and sexually
transmitted infections, mental health problems,
gynaecological problems, and decreased
physical functioning (20–23).These studies
suggested that, in addition to causing injury and
other immediate sequelae, violence increased
women’s risk of future ill-health. Awareness of
this is causing a significant shift in the way health
professionals conceptualize violence. Rather
than being seen as just a health problem in and
of itself, violence can also be understood as a
risk factor that – like smoking or unsafe sex –
increases women’s risk of a variety of diseases
and conditions (24, 25).
During the 1990s, researchers and
practitioners also began exploring patterns of
violence in different settings. Data increasingly
suggested that the level of partner violence
against women varied substantially, both between
and within countries (26). This raised the
question of what combination of factors could
best explain the variation. What insights could
be gained from this analysis that would advance
violence theory and intervention?
Increasingly, researchers and practitioners
– as well as WHO – are using an “ecological
framework” to understand the interplay of
personal, situational, and sociocultural factors
that combine to cause interpersonal violence
(9, 27). Introduced in the late 1970s, the
ecological model was first applied to child abuse
(28, 29), and subsequently to youth violence
(30, 31). More recently, it has been used to
understand intimate partner violence (32) and
abuse of the elderly (33, 34). In the ecological
model, interpersonal violence results from the
interaction of factors at different levels of the
social environment.
The model can best be conceptualized as
four nested circles (Figure 1.1). The innermost
circle represents the biological and personal
history that each individual brings to his or her
behaviour in relationships. The second circle
represents the immediate context in which
violence takes place – frequently the family or
other intimate or acquaintance relationship. The
third circle represents the institutions and social
structures, both formal and informal, in which
relationships are embedded – neighbourhood,
workplace, social networks, and peer groups. The
fourth, outermost circle is the economic and
social environment, including cultural norms.
The WHO Study incorporates an ecological
model for understanding partner violence by
including, at each level of the social ecology,
variables hypothesized to increase or decrease a
woman’s risk of partner violence.
Analyses at national and international
level comparing settings with high and low
prevalence of partner violence provide an
Chapter 1 Introduction
5
opportunity to identify potential individual,
community and societal factors associated with
its occurrence. Comparative analysis could
be used to test whether there are identifiable
risk factors within the immediate and larger
community that could possibly be reduced
through community activities.
To date, the lack of comparability among
studies has made this type of analysis difficult,
if not impossible. To explore potential risk and
protective factors with any rigour requires
a study that minimizes all methodologically
induced variation among sites. Although there
will always be sources of variation that cannot
be fully controlled (such as cultural variation in
women’s willingness to disclose violence), the
WHO Study included a variety of measures
designed to maximize the comparability of data
across sites (see Annex 1).
In future analyses, the data from this study
will be used to explore individual, household,
and community risk and protective factors
in greater depth. Greater insights into the
situations and contexts in which violence
does and does not occur will be sought
through multivariate and multilevel analysis
of possible combinations of factors acting at
different levels (35, 36).
Clearly, if the potentially modifiable risk
factors – and potentially protective factors
– could be identified, this would have important
implications for the development of preventive
interventions both locally and internationally.
Study objectives
The WHO Multi-country Study on Women’s
Health and Domestic Violence against Women
was designed to address some of the major
gaps in the international literature on violence
against women, especially related to intimate-
partner violence in developing country settings
and its impact on women’s health. It attempted
to overcome the obstacles to comparability
encountered in previous studies by carrying out
population-based surveys using a standardized
questionnaire, with standardized training and
procedures across sites.
The WHO Study’s objectives were as follows:
• to obtain valid estimates of the prevalence
and frequency of different forms of physical,
sexual and emotional violence against
women, with particular emphasis on violence
perpetrated by intimate male partners;
• to assess the extent to which violence by
intimate partners is associated with a range
of health outcomes;
• to identify factors that may protect or put
women at risk for intimate-partner violence;
• to document and compare the strategies and
services that women use to deal with the
violence they experience.
The study aimed to provide a strong
evidence base for informing policy and
action at the national and international level.
Additional goals included: developing and
testing new instruments for measuring violence
cross-culturally; increasing national capacity
and collaboration among researchers and
women’s organizations working on violence;
and increasing sensitivity to violence among
researchers, policy-makers and health care
providers. To achieve these goals, WHO adopted
an action-oriented model of research that
encouraged the active engagement of women’s
organizations with expertise on violence against
women. The model also gave priority to ensuring
women’s safety and well-being.
6
WHO Multi-country Study on Women’s Health and Domestic Violence
This first report describes the findings
related to three of the four study objectives: to
assess prevalence, determine health outcomes,
and document women’s coping strategies.
Analysis of risk and protective factors for
violence will be addressed in a future report.
More in-depth multivariate and multilevel analysis
of study outcomes will be explored in individual
papers to be submitted for publication in the
peer-reviewed scientific literature.
The original plan for the WHO Study
included a survey of men. However this was not
implemented (see Box 1.1).
Organization of the Study
The study was implemented by WHO
through a core research team made up of
international experts from WHO (including
the study coordinator), the London School
of Hygiene and Tropical Medicine, and the
Program for Appropriate Technology in Health
in Washington, DC (see Annex 2 for a list of
participants in the core research team). This
core research team had overall responsibility
for designing the study, and supporting its
implementation and analysis. WHO also
established an expert steering committee that
included internationally known epidemiologists,
advocates and researchers on violence against
women, from different regions of the world.
This steering committee provided technical
and scientific oversight to the study, and met
periodically to review the progress and outputs
of the study (see Annex 2 for a list of members
of the steering committee).
Within each participating country, a
collaborative research team was established to
implement the study. This generally consisted
of representatives of research organizations
experienced in conducting survey research,
a women’s organization with experience of
providing services to women experiencing
violence and, in some places, government and
national statistics offices (see Annex 3 for a list of
country par ticipants).
Each country research team also
established an advisory group to support the
implementation of the study and ensure the
dissemination of the results. The membership
of the groups differed between countries,
but generally included key decision-makers,
representatives of women’s organizations and
researchers. The study also aimed to ensure that
representatives from relevant divisions within the
ministry of health and other concerned ministries
or bodies were included. Where possible,
existing multisectoral committees on violence
against women formed the core membership
of the advisory group. Members of the country
research teams met regularly with the advisory
group to review progress and to discuss
emerging issues.
Participating countries
Participating countries were identified, following
discussions with the WHO regional offices, on
the basis of the following criteria:
• presence of local women’s groups working
on violence against women that could
use the data generated for advocacy and
policy reform;
• absence of existing population-based data on
violence against women;
• presence of strong potential partner
organizations known to WHO;
• a political environment receptive to taking up
the issue;
• absence of recent war-related conflict;
• representation of the different WHO regions.
Chapter 1 Introduction
7
Figure 1.1 Ecological model for understanding violence
Society Community Relationship Individual
Source: Reproduced from reference 9.
The original plan for the WHO Study included
interviews with a subpopulation of men about
their experiences and perpetration of violence,
including partner violence. This would have
allowed researchers to compare men’s and
women’s accounts of violence in intimate
relationships and would have yielded data
to investigate the extent to which men are
physically or sexually abused by their female
partners. On the advice of the Study Steering
Committee, it was decided to include men only
in the qualitative, formative component of the
study and not in the quantitative survey.
This decision was taken for two reasons.
First, it was considered unsafe to interview men
and women in the same household, because
this could have potentially put a woman at risk
of future violence by alerting her par tner to the
nature of the questions. Second, to carry out
an equivalent number of interviews in separate
households was deemed too expensive.
Nevertheless, it is recognized that men’s
experiences of partner violence, as well as
the reasons why men perpetrate violence
against women, need to be explored in future
research. Extreme caution should be used
in any study of partner violence that seeks
to compile prevalence data on men as well
as women at the same time because of the
potential safety implications.
Box 1.1 Studying men
8
9
Even before the data were available, the WHO
Study brought about several positive changes at
different levels.
• The WHO Study contributed to increased
awareness among researchers, interviewers and
others involved in doing the research, as well as
among the women interviewed. Most importantly, a
pool of over 500 trained interviewers, researchers
and other staff have been sensitized to the problem
of violence against women and have acquired
understanding and skills to investigate it. A large
number of the female staff have reported making
major changes in their personal or professional lives
as a result of their involvement in the Study. Many of
those involved in the Study, both men and women,
continue to be actively engaged in working to
address violence against women in their countries.
• The WHO Study contributed to the inclusion
of violence by intimate partners in several
policies and educational programmes of the
partner universities and ministries of health. In
Peru, for example, violence against women has
Box 1.2 Preliminary impact of the WHO Multi-country Study on Women’s Health and
Domestic Violence against Women
been incorporated into the Masters course
on reproductive health and sexuality in the
Faculty of Public Health of the Cayetano
Heredia University and has been discussed with
local community leaders in the provincial site.
In Brazil, medical and social science students
were involved in the study, and violence against
women has been included in postgraduate
training at the University of São Paulo.
• The WHO Study prompted further research.
For example: one of the researchers in Peru is
now doing a study on men and violence against
women; researchers in Brazil have done a study
on women attending health centres in São Paulo,
using the same instrument as in the WHO Study;
researchers in Thailand and the United Republic
of Tanzania report using the ethical and safety
guidelines for research on other issues.
• At the grass-roots level, networks of service
providers have been established or identified,
and information on local organizations has been
compiled and distributed widely.
The first countries selected were:
Bangladesh, Brazil, Japan, Namibia, Peru, Samoa,
Thailand, and the United Republic of Tanzania. A
second group of countries later replicated the
study: Ethiopia, New Zealand, and Serbia and
Montenegro. Other countries, including Chile,
China, Indonesia, and Viet Nam, have adapted or
used parts of the study questionnaire.
This first report presents the findings from
the countries that participated in the first round
of the study, conducted between 2000 and
2003 – Bangladesh, Brazil, Japan, Namibia, Peru,
Samoa, Thailand, and the United Republic of
Tanzania – as well as from two countries that
participated in the second round – Ethiopia
and Serbia and Montenegro.1 In combination,
the results provide evidence of the extent of
physical and sexual violence from 15 sites in
10 geographically, culturally and economically
diverse countries (Figure 1.2).
Figure 1.2 Countries participating in the WHO Multi-country Study on Women’s Health
and Domestic Violence against Women
Countries in first round
Countries in second round
Serbia and Montenegro
New Zealand
Namibia
Peru
Ethiopia
United Republic
of Tanzania
Samoa Bangladesh
Thailand
Japan
Brazil
In each country, the findings from the national
analysis have already been written up as a country
report, and disseminated at the local and national
level in a variety of ways. The dissemination
activities were coordinated by the country
research teams, and drew on the experience
and resources made available by each country’s
advisory group and WHO. Where possible,
the findings are being fed into advocacy and
intervention activities concerned with violence
against women – such as the 16 days of action
against violence against women in Namibia, the
development of the national plan of action for
the elimination of violence against women and
children in Thailand, and the development of the
national policy and plan of action for violence
prevention in Brazil. In addition, the study has
already resulted in various important changes
(Box 1.2). WHO country offices and relevant
ministries, together with the researchers, are
helping to disseminate the findings to different
sectors, and to the donor community.
References
Chapter 1 Introduction
WHO Multi-country Study on Women’s Health and Domestic Violence
1 The data set from New
Zealand was not available when
this report was being prepared.
However, the first results from
New Zealand have recently
been published (37).
1. Heise L. Violence against women: global organizing
for change. In: Edleson JL, Eisikovits ZC, eds. Future
interventions with battered women and their families.
Thousand Oaks, CA, Sage Publications, 1996:7–33.
2. Joachim J. Shaping the human rights agenda: the case
of violence against women. In: Meyer MK, Prugl E, eds.
Gender politics in global governance. Lanham, MD,
Rowman and Littlefield Publishers Inc., 2000:142–160.
3. Mayhew S, Watts C. Global rhetoric and individual
realities: linking violence against women and
reproductive health. In: Lee K, Buse K, Fustukian S,
eds. Health policy in a globalising world. Cambridge,
Cambridge University Press, 2002:159–180.
4. Vienna Declaration and Programme of Action.
Adopted by the World Conference on Human Rights,
Vienna, 14–25 June 1993. New York, NY, United
Nations, 1993 (document A/CONF.157/23).
5. International Conference on Population and
Development (ICPD), Cairo, Egypt, 5–13 September
1994. New York, NY, United Nations, 1994
(document A/CONF.171/13).
6. The Fourth World Conference on Women, Beijing,
China, 4–15 September 1995. New York, NY, United
Nations, 1995 (document A/CONF.177/20).
7. Violence against women: WHO Consultation,
Geneva, 5–7 Februar y 1996. Geneva, World Health
Organization, 1996 (document FRH/WHD/96.27,
available at: http://whqlibdoc.who.int/hq/1996/FRH_
WHD_96.27.pdf, accessed 18 March 2005).
8. WHO Multi-countr y Study on Women’s Health and
Domestic Violence against Women: study protocol.
Geneva, World Health Organization, 2004.
9. Krug EG et al. eds. World report on violence and
health. Geneva, World Health Organization, 2002.
10. Declaration on the elimination of violence against
women. New York, NY, United Nations, 1993 (United
Nations General Assembly resolution, document
A/RES/48/104).
11. Heise L, Ellsberg M, Gottemoeller M. Ending
violence against women. Baltimore, MD, Johns
Hopkins University Press, 1999.
12. Koss MP. Detecting the scope of rape: a review of
prevalence research methods. Journal of Interpersonal
Violence, 1993, 8:198–222.
13. Ellsberg M et al. Researching domestic violence against
women: methodological and ethical considerations.
Studies in Family Planning, 2001, 32:1–16.
14. Measuring violence against women cross-culturally:
notes from a meeting. Takoma Park, MD, Health and
Development Policy Project, 1995.
15. Straus MA. Measuring intrafamily conflict and
violence: the Conflict Tactics Scale (CTS). Journal of
Marriage and the Family, 1979, 41:75–88.
16. Straus MA et al. The revised Conflict Tactics Scale
(CTS2). Journal of Family Issues, 1996, 17:283–316.
17. Dobash RE, Dobash RD. The myth of sexual
symmetry in marital violence. Social Problems, 1992,
39:71–91.
18. Hassan F et al. Physical intimate partner violence in
Chile, Egypt, India and the Philippines. Injury Control
and Safety Promotion, 2004, 11:111–116.
19. Kishor S, Johnson K. Domestic violence in nine
developing countries: a comparative study. Calverton,
MD, MACRO International, 2004.
20. Campbell J et al. Intimate partner violence and
physical health consequences. Archives of Internal
Medicine, 2002, 162:1157–1163.
21. Gazmararian JA et al. The relationship between
pregnancy intendedness and physical violence in
mothers of newborns. The PRAMS Working Group.
Obstetrics and Gynecology, 1995, 85:1031–1038.
22. Golding J. Sexual assault history and women’s
reproductive and sexual health. Psychology of Women
Quarterly, 1996, 20:101–121.
23. Murphy CC et al. Abuse: a risk factor for low birth
weight? A systematic review and meta-analysis. Canadian
Medical Association Journal, 2001, 164:1567–1572.
24. Campbell JC. Health consequences of intimate
partner violence. Lancet, 2002, 359:1331–1336.
25. Counts D, Brown JK, Campbell JC, eds. To have
and to hit, 2nd ed. Chicago, IL, University of Chicago
Press, 1999.
26. Levinson D. Violence in cross cultural perspective.
Newbury Park, CA, Sage Publications, 1989.
27. Bronfenbrenner V. The ecology of human
development: experiments by nature and design.
Cambridge, MA, Harvard University Press, 1979.
28. Garbarino J, Crouter A. Defining the community
10
WHO Multi-country Study on Women’s Health and Domestic Violence
This is an “A” heading here
Methods
context for parent–child relations: the correlates of child
maltreatment. Child Development, 1978, 49:604–616.
29. Belsky J. Child maltreatment: an ecological
integration. American Psychologist 1980;35:320–335.
30. Tolan PH, Guerra NG. What works in reducing
adolescent violence: an empirical review of the field.
Boulder, CO, University of Colorado, Center for the
Study and Prevention of Violence, 1994.
31. Chaulk R, King PA. Violence in families: assessing
prevention and treatment programs. Washington, DC,
National Academy Press, 1998.
32. Heise L. Violence against women: an integrated,
ecological framework. Violence Against Women, 1998,
4:262–290.
33. Schiamberg LB, Gans D. An ecological framework for
contextual risk factors in elder abuse by adult children.
Journal of Elder Abuse and Neglect, 1999, 11:79–103.
34. Carp RM. Elder abuse in the family: an interdisciplinar y
model for research. New York, NY, Springer, 2000.
35. O’Campo P et al. Violence by male partners against
women during the childbearing year: a contextual
analysis. American Journal of Public Health, 1995,
85:1092–1097.
36. Koenig MA et al. Women’s status and domestic
violence in rural Bangladesh: individual- and
community-level effects. Demography, 2003, 40:269–
288.
37. Fanslow J, Robinson E. Violence against women in
New Zealand: prevalence and health consequences.
New Zealand Medical Journal, 2004, 117:1173–1184.
CHAPTER
15
2
Definitions
One of the main challenges facing international
researchers on violence against women is to
develop clear operational definitions of different
types of violence and tools for measuring
violence that permit meaningful comparisons
among diverse settings.
Researchers have used many criteria to
define violence. A common method is to
classify violence according to the type of act:
for example, physical violence (e.g. slapping,
hitting, kicking, and beating), sexual violence (e.g.
forced intercourse and other forms of coerced
sex), and emotional or psychological violence
(e.g. intimidation and humiliation). Violence can
also be defined by the relationship between the
victim and perpetrator; for example, intimate
partner violence, incest, sexual assault by a
stranger, date rape or acquaintance rape.
In the World report on violence and health
(1), WHO adopted a typology that categorizes
violence in three broad categories, according to
those committing the violent act:
• self-directed violence,
• interpersonal violence,
• collective violence.
These categories are each divided further to
reflect specific types of violence (Figure 2.1).
Measuring violence
The WHO Study focused primarily on “domestic
violence”,1 or violence by an intimate partner,
experienced by women. Included in this were acts of
physical, sexual and emotional abuse by a current or
former intimate male partner, whether cohabiting or
not.2 In addition, it looked at controlling behaviours,
including acts to constrain a woman’s mobility or
her access to friends and relatives, extreme jealousy,
etc. The Study also included physical and sexual
violence against women, before and after
15 years of age, by perpetrators other than intimate
partners. Definitions of each of these aspects of
violence were operationalized in the study using
a range of behaviour-specific questions related
Definitions and questionnaire development
Figure 2.1 A typology of violence
Source: Reproduced from reference 1.
Interpersonal CollectiveSelf-directed
Acquaintance Stranger
Child
Nature of
violence
Partner Elder
Family/
partner Community Social Political Economic
Suicidal
behaviour Self-abuse
Deprivation
or neglect
Psychological
Sexual
Physical
Violence
1 The term “intimate-partner
violence” is now used in
preference to the term
“domestic violence”, which
is not specific and could
include child abuse, intimate
partner violence and abuse of
the elderly. This report uses
intimate-par tner or partner
violence, except in the name
of the Study, which was
agreed before the appearance
of the World report on violence
and health (1).
2 The Study focused on
violence by male par tners
only, mainly because most
intimate partners of women
throughout the world
are male. Indeed, in some
countries it would not be
culturally acceptable to
ask about female–female
relationships. In addition,
the Study was intended
as a contribution to the
understanding of gender-based
violence as an expression of
gender inequality in relations
between women and men.
15
to each type of violence (Annex 4). The study did
not attempt to document an exhaustive list of acts
of violence, but instead asked a limited number
of questions about specific acts that commonly
occur in violent partnerships. This approach has
been used widely in studies of partner violence
in the United States and elsewhere, and has been
shown to encourage greater disclosure of violence
than approaches that require respondents to
identify themselves as abused or battered (2, 3).
Given that the conceptualization of violence differs
between individuals and communities, a fairly
conservative definition of violence was used. Thus
the prevalence estimated in this manner is more
likely to underestimate rather than overestimate
the true prevalence of violence. The acts used
to define each type of violence measured in the
Study are summarized in Box 2.1.
Violence by intimate partners
While there is widespread agreement, and
some standardization, regarding what acts are
included as physical violence, this is less true for
sexual violence. There is even less agreement
on how to define and measure psychological or
emotional abuse, especially in a cross-cultural
perspective, because the acts that are perceived
as abusive are likely to vary between countries
and between socioeconomic and ethnic groups,
and according to the overall level of violence
in the group. Because of the complexity of
defining and measuring emotional abuse in a way
that is relevant and meaningful across cultures,
the questions regarding emotional violence
and controlling behaviour in the WHO Study
questionnaire should be considered as a
starting-point, rather than a comprehensive
measure of all forms of emotional abuse.
The questions on physical partner violence
were divided into those related to “moderate”
violence, and those considered “severe” violence
(Box 2.2). The distinction between moderate
and severe violence is based on the likelihood of
physical injury. For each act of physical, sexual, or
emotional abuse that the respondent reported
as having happened to her, she was asked
whether it had happened ever or in the past
12 months, and with what frequency (once or
twice, a few times, or many times) (Questions
704, 705, 706). The answers to these questions
made it possible to assess the level of sexual or
physical violence by current or former partners.
Physical violence by an intimate partner
• Was slapped or had something thrown at her
that could hurt her
• Was pushed or shoved
• Was hit with fist or something else that could hurt
• Was kicked, dragged or beaten up
• Was choked or burnt on purpose
• Perpetrator threatened to use or actually used a
gun, knife or other weapon against her
Sexual violence by an intimate partner
• Was physically forced to have sexual intercourse
when she did not want to
• Had sexual intercourse when she did not
want to because she was afraid of what partner
might do
• Was forced to do something sexual that she
found degrading or humiliating
Emotional abuse by an intimate partner
• Was insulted or made to feel bad about herself
• Was belittled or humiliated in front of
other people
• Perpetrator had done things to scare or
intimidate her on purpose, e.g. by the way he
looked at her, by yelling or smashing things
• Perpetrator had threatened to hurt someone
she cared about
Box 2.1 Operational definitions of violence used in the WHO Multi-country Study on
Women’s Health and Domestic Violence against Women
Controlling behaviours by an intimate
partner
• He tried to keep her from seeing friends
• He tried to restrict contact with her family of birth
• He insisted on knowing where she was at all times
• He ignored her and treated her indifferently
• He got angry if she spoke with another man
• He was often suspicious that she was unfaithful
• He expected her to ask permission before
seeking health care for herself
Physical violence in pregnancy
• Was slapped, hit or beaten while pregnant
• Was punched or kicked in the abdomen
while pregnant
Physical violence since age 15 years by
others (non-partners)
• Since age 15 years someone other than partner
beat or physically mistreated her
Sexual violence since age 15 years by
others (non-partners)
• Since age 15 years someone other than partner
forced her to have sex or to perform a sexual act
when she did not want to
Childhood sexual abuse (before age 15 years)
• Before age 15 years someone had touched her
sexually or made her do something sexual that
she did not want to
14
WHO Multi-country Study on Women’s Health and Domestic Violence
Psychometric analysis was performed on the
violence questions used in the Study to ascertain
the appropriateness of the behavioural items
included in the different measures of physical,
emotional and sexual violence. In general, there
was good internal consistency among the items
for each measure, indicating that the instrument
provided a reliable and valid measure for each of
the types of violence.
An exposure chart (Question 716) was used
to collect information about the timing of the
onset of physical or sexual violence by an intimate
partner and when such violence last occurred. This
was an important aspect of the data collection,
which partly addressed the inherent limitations
of the cross-sectional study design, as information
about the timing of different forms of violence can
be compared with details about the timing of the
start and end of the relationship or marriage. This
information allows for analysis of the extent to
which different forms of violence occur prior to or
during marriage or cohabitation, or after separation.
The data can also be used to understand how
women’s risk of intimate-partner violence changes
over the duration of the relationship.
Ever-partnered women
The definition of “ever-partnered women”
is central to the study, because it defines the
population that could potentially be at risk
of partner violence (and hence becomes the
denominator for prevalence figures). Although the
study tried to maintain the highest possible level
of standardization across countries, it was agreed
that the same definition could not be used in all
the countries, because the concept of “partner”
is culturally or legally defined. In developing the
country-specific definitions of “ever-partnered
women”, the study researchers were aware of
the need to use a broad definition of partnership,
since any woman who had been in a relationship
with an intimate partner, whether or not they had
been married, could have been exposed to the
risk of violence. It was also recognized that the
definition of ever-partnered women would need
to be narrower in some contexts than others.
For example, in Bangladesh it was considered
inappropriate to ask unmarried women about
non-marital partners; in any case, an unmarried
woman in Bangladesh cohabiting with a partner
would most likely have identified herself as being
married and so be included in the study population.
In general, the definition of “ever-partnered
women” included women who were or had ever
been married or in a common-law relationship. In
countries where premarital sexual relationships
are common, the definition covered dating
relationships – defined as regular sexual partners,
not living together. Former dating partners were
not included, except in Japan, Namibia and Peru,
where many women never live with regular sexual
partners, even if they have children by them.
Box 2.3 gives the definitions of “ever partnered”
used in the countries taking part in the WHO Study.
Violence by non-partners
The survey also explored the extent to which
women report experiencing violence by
perpetrators other than a current or former
male partner. It included questions on physically
abusive behaviour by such perpetrators since
the age of 15 years, in different contexts (at
school or work, by a friend or neighbour or
anyone else). Follow-up questions explored the
frequency of violence for each perpetrator.
Severity scale used for level of
violence (see Question 705 of
the WHO Study questionnaire)
“Moderate” violence: respondent
answers “yes” to one or more of the
following questions regarding her
intimate partner (and does not answer
“yes” to questions c–f below):
(a) [Has he] slapped you or thrown something
at you that could hurt you?
(b) [Has he] pushed you or shoved you?
“Severe” violence: respondent answers
“yes” to one or more of the following
questions regarding her intimate partner:
(c) [Has he] hit you with his fist or with
something else that could hurt you?
(d) [Has he] kicked you, dragged you or beaten
you up?
(e) [Has he] choked or burnt you on purpose?
(f) [Has he] threatened to use or actually used
a gun, knife or other weapon against you?
Box 2.2
Country-specific definitions
of “ever-partnered women”
Bangladesh
Brazil, Ethiopia,
Serbia and Montenegro,
Thailand, United
Republic of Tanzania
Japan, Namibia, Peru
Samoa
Box 2.3
Ever married
Ever married, ever
lived with a man,
currently with a
regular sexual partner
Ever married, ever
lived with a man, ever
with a regular sexual
partner
Ever married, ever
lived with a man
Chapter 2 Definitions and questionnaire development
16
WHO Multi-country Study on Women’s Health and Domestic Violence
17
Likewise, the survey explored the extent
to which the women had been sexually abused
by others, including before age 15 years (child
sexual abuse). As this is a highly sensitive issue,
four methods were used to enhance disclosure
of different forms of abuse. Respondents were
asked whether, since the age of 15 years, any
person other than their partner or husband had
forced them to have sex or to perform a sexual
act when they did not want to (Question 1002).
Again, probing questions were used to explore
the different contexts in which this might have
occurred. For respondents who reported having
experienced this type of abuse, information was
collected about the perpetrator and the frequency.
Second, respondents were asked whether, before
the age of 15 years, anyone had ever touched
them sexually or made them do something sexual
that they did not want to do (Question 1003).
Follow-on questions asked about the perpetrator,
the ages of the respondent and the perpetrator
at the time, and the frequency. Third, respondents
were asked how old they were at their first sexual
experience (Question 1004), and whether it had
been something they had wanted to happen,
something they had not wanted but that had
happened anyway, or something that they had
been forced into (Question 1005). Finally, at the
end of each interview the respondent was offered
an opportunity to indicate in a hidden manner
whether anyone had ever touched her sexually
or made her do something sexual against her will
before the age of 15 years, without having
to disclose her reply to the interviewer
(Question 1201). For this question, respondents
were handed a card that had a pictorial
representation for yes and no and asked to record
their response in private (Figure 2.2). In most sites,
the respondent then folded the card and placed
it in an envelope or a bag containing other cards
before handing it back to the interviewer, thus
keeping her answer secret from the interviewer. In
Serbia and Montenegro and the United Republic
of Tanzania, the sealed envelope with the card
was attached to the questionnaire to allow the
information to be linked to the individual woman
at the time of data entry. The use of a card was
intended to increase the likelihood of obtaining a
more complete estimate of the prevalence of
childhood sexual abuse.
Formative research
The WHO Study incorporated formative research,
including research on definitional issues, in each
of the country sites. The aim of this work was to
gain insights that could be used in designing and
translating the questionnaire, and in interpreting the
survey findings. The research included: interviews
with key informants; in-depth interviews with
survivors of violence; and focus group discussions
with women and men of different age groups.
Key informants
Informants included representatives from
nongovernmental organizations focusing on
areas such as violence against women,
HIV/AIDS, women’s health, women’s rights and
their awareness of those rights, or women’s
education and development.
In-depth interviews with survivors
In each country, in-depth semi-structured
interviews were held with at least five women
who were known to have been abused by their
partners or former partners. Participants were
recruited through different support services, by
means of “snowball” techniques. These interviews
were used to gain a better understanding of how
women describe their experiences of domestic
violence, the ways in which they have responded,
and how such violence has influenced their lives.
The structure of the interviews reflected the
forms of information to be collected during
the survey. The women’s narratives helped
inform the development and translation of the
relevant modules within the core and country
questionnaires. The information is also being
used to help in interpreting the quantitative
research findings, and to supplement the
quantitative data obtained.
During the interviews, careful attention was
given to the ethical and safety issues associated
with the study (see Chapter 3). This included
recognizing that the interviews might be
distressing, and ensuring that adequate follow-up
support was provided. Care was also taken to
Figure 2.2 Sample response card
Pictorial representation of response to Question 1201
concerning sexual abuse before 15 years of age:
tearful face indicates “yes”; smiling face indicates “no”
ensure that strict confidentiality was maintained,
and that the respondent could not be identified in
follow-up dissemination activities. Each interview
aimed to end on a positive note, identifying the
respondent’s strengths and abilities. All tapes were
erased once transcripts had been made.
Focus group discussions
Focus group discussions were held with women
and men, young and old, in both urban and rural
settings. The aim was again to explore local
views and language about violence and obtain
descriptions of different forms of violence. Focus
group discussions were conducted using a script
and short scenarios; participants were left to
complete the story-line.
Development of the questionnaire
The study questionnaire was the outcome of a long
process of discussion and consultation. Following
an extensive review of a range of pre-existing
study instruments, and consultation with technical
experts in specific areas (including violence against
women, reproductive health, mental health, and
tobacco and alcohol use), the core research team
developed a first draft of the questionnaire. This was
then reviewed by the expert steering committee
and experts in relevant fields, and suggestions
for revision were incorporated. The revised
questionnaire was then reviewed by the country
teams during an international meeting. Discussion
focused on incorporating country priorities,
and achieving a balance between exhaustively
exploring specific issues and compiling less detailed
information on a range of issues.
The questionnaire was then translated
(see Box 2.4) and pretested in six countries
(Bangladesh, Brazil, Namibia, Samoa, Thailand, and
the United Republic of Tanzania). The experiences
from these pretests were reviewed at the third
meeting of the research teams, and used to make
further revisions to the questionnaire.
Following a final pretest, the questionnaire
for the Study was completed as version 9.9
(Annex 4), and was used in Bangladesh, Brazil,
Ethiopia, Japan, Namibia, Peru, Samoa, Thailand,
and the United Republic of Tanzania. An updated
version of the questionnaire (version 10), which
incorporates the experience in the first eight
countries, was used in Serbia and Montenegro.
Questionnaire structure
The questionnaire consisted of an administration
form, a household selection form a household
questionnaire, a women’s questionnaire, and
a reference sheet. The women’s questionnaire
included an individual consent form and
12 sections designed to obtain details about the
respondent and her community, her general and
reproductive health, her financial autonomy, her
children, her par tner, her experiences of partner
and non-partner violence, and the impact of
violence on her life (see Box 2.5 for an outline of
the questionnaire).
Maximizing disclosure
From the outset of the study it was recognized
that violence is a highly sensitive issue, and that
there was a danger that women would not
The working language for the development
of the questionnaire was English. Before
pre-testing in each country, the questionnaire
was professionally translated into the relevant
local languages. The formative research was used
to guide the forms of language and expressions
used, with the focus being on using words and
expressions that were widely understood in
the study sites. In settings where a number of
languages were in use, questionnaires were
developed in each language.
Previous research experience in South Africa
and Zimbabwe found that professional
back-translations were not a reliable way to
check the accuracy of questions on violence and
its consequences. For this reason, the translated
questionnaire was first checked by local
researchers involved in the study who compared
the English and translated versions. Lengthy
oral back-translation sessions with step-by-step
discussion of each question were conducted
with people not familiar with the questionnaire
but fluent in the language and with people who
understood the questionnaire and violence issues.
The main purpose of this exercise was to identify
differences in translations that could alter the
meaning of questions and to establish cognitive
understanding of the items in the questionnaire.
Adjustments were made where needed. Once
the translation had been finalized, the questions
were again discussed during interviewer-training
sessions on the basis of a question-by-question
description of the questionnaire. Having
interviewers from various cultural backgrounds
aided in ascertaining whether wording used was
culturally acceptable. During the training itself,
further revisions to the translated questionnaires
were made. Final minor modifications to fine-tune
the translated questionnaire were usually made
during the pilot survey in the field, in the third
week of interviewer training.
Box 2.4 Translation of the
questionnaire
Chapter 2 Definitions and questionnaire development
18
WHO Multi-country Study on Women’s Health and Domestic Violence
CHAPTER
disclose their experiences of violence. For this
reason, in designing the questionnaire, an attempt
was made to ensure that women would feel
able to disclose any experiences of violence.
The questionnaire was structured so that early
sections collected information on less sensitive
issues, and that more sensitive issues, including
the nature and extent of partner and non-
partner violence, were explored later, once
a rapport had been established between the
interviewer and the respondent.
Partner violence often carries a stigma, and
women may be blamed, or blame themselves,
for the violence they experience. For this
reason, all questions about violence and its
consequences were phrased in a supportive
and non-judgemental manner. The word
“violence” itself was avoided throughout the
questionnaire. In addition, careful attention was
paid to the wording used to introduce the
different questions on violence. These sections
forewarned the respondent about the sensitivity
of the forthcoming questions, assured her that
the questions referred to events that many
women experience, highlighted the confidentiality
of her responses, and reminded her that she
could choose not to answer any question or to
stop the interview at any point. For example,
the wording used to introduce the section on
intimate-partner violence was:
“When two people marry or live
together, they usually share both good and
bad moments. I would now like to ask you
some questions about your current and past
relationships and how your husband/partner
treats (treated) you. If anyone interrupts us I
will change the topic of conversation. I would
again like to assure you that your answers will
be kept secret, and that you do not have to
answer any questions that you do not want to.
May I continue?”
This form of introduction also ensured that
women were given a second opportunity (in
addition to the informed consent) to decline to
answer questions about violence.
Country adaptation and translation of
the questionnaire
Once the questionnaire had been finalized,
country teams were able to make minor
adaptations. Country modifications generally
involved either adding a limited number of
questions to explore country-specific issues
or modifying the response categories used to
make them appropriate to the particular setting.
To ensure that cross-country comparability
was not jeopardized, all proposed changes
were reviewed by the core research team.
Relatively significant changes were made to the
questionnaire only in Ethiopia, Japan, and Serbia
and Montenegro (see Annex 1).
1. Krug EG et al. World report on violence and health.
Geneva, World Health Organization, 2002.
2. Straus MA, Gelles RJ. Societal change and change
in family violence from 1975 to 1985 as revealed
by two national surveys. Journal of Marriage and
the Family, 1986, 48:465–480.
3. Straus MA et al. The revised Conflict Tactics Scales
(CTS2). Journal of Family Issues, 1996, 17:283–316.
References
3
This chapter contains basic information
on sample design, the ethical and safety
considerations in the study methodology,
and the response rates in the study sites.
Details on the following subjects are given in
Annex 1 Methodology:
1. Ensuring comparability across sites and
sampling strategies
2. Enhancing data quality
3. Interviewer selection and training
4. Respondents’ satisfaction with the interview
5. Data processing and analysis
6. Characteristics of respondents
7. Representativeness of the sample.
Sample design
In each country, the quantitative component
of the study consisted of a cross-sectional
population-based household survey conducted in
one or two sites (Box 3.1).
In Bangladesh, Brazil, Peru, Thailand, and
the United Republic of Tanzania, surveys were
conducted in two sites: one in the capital or
a large city; and one in a province or region,
usually with urban and rural populations. One
rural setting was used in Ethiopia, and a single
large city in Japan, Namibia, and Serbia and
Montenegro. In Samoa the whole country was
sampled. In this repor t, sites are referred to
by country name followed by either “city” or
“province”; where only the country name is used,
it should be taken to refer to both sites.
The following criteria were used to help
select an appropriate province:
• availability of, or the possibility of establishing,
support services for women who, through
the course of the survey, were identified as
having experienced some form of violence
and needing support;
• location broadly representative of the
country as a whole, in terms of the range of
communities, ethnic groups and religions;
• population not marginalized, and not
perceived as being likely to have higher
levels of partner violence than in the rest of
the country.
In general, a woman was considered eligible
for the study if she was aged between 15 and
49 years, and if she fulfilled one of the following
three conditions:
• she normally lived in the household;
• she was a domestic servant who slept for
five nights a week or more in the household;
• she was a visitor who had slept in the
household for at least the past 4 weeks.
In Japan, where for legal reasons it was not
feasible to interview women under 18 years of
age, women aged 18–49 years were sampled.
The initial sample size calculations suggested
that an obtained sample size of 1500 women in
each site would give sufficient power to meet
the study objectives (see Chapter 1). In order to
make up for losses to the sample as a result of
households without eligible women, refusals to
participate, or incomplete interviews, the initial
number of households to be visited was set
approximately 20–30% higher than the target
sample size in most sites. Appendix Table 1 shows
details of the sample sizes obtained.
For most sites, a two-stage cluster sampling
scheme was used to select households. In
settings where the site (city or province) was
very large, a multistage procedure was used
in which districts (or analogous administrative
units) were first selected, and then clusters
were selected from within the chosen districts.
Either explicit or implicit stratification by an
appropriate socioeconomic indicator was used
to ensure that the sample was representative
of all socioeconomic groups. Depending on the
sampling frame, between 22 and 200 clusters
were selected from each of the sites participating
in the study.
Sample design, ethical and safety
considerations, and response rates
WHO Multi-country Study
on Women’s Health and
Domestic Violence against
Women: topics covered by the
women’s questionnaire
Section 1: Characteristics of the respondent
and her community
Section 2: General health
Section 3: Reproductive health
Section 4: Information regarding children
Section 5: Characteristics of current or most
recent partner
Section 6: Attitudes towards gender roles
Section 7: Experiences of partner violence
Section 8: Injuries resulting from partner
violence
Section 9: Impact of partner violence and
coping mechanisms used by women
who experience partner violence
Section 10: Non-partner violence
Section 11: Financial autonomy
Section 12: Anonymous reporting of childhood
sexual abuse; respondent feedback
Box 2.5