MEASURE DHS+ assists countries worldwide in the collection and use of data to monitor and evaluate popula-

tion, health, and nutrition programs. Funded by the U.S. Agency for International Development (USAID),
MEASURE DHS+ is implemented by ORC Macro in Calverton, Maryland.

The main objectives of the MEASURE DHS+ project are:

1) to provide decisionmakers in survey countries with information useful for informed policy choices,
2) to expand the international population and health database,
3) to advance survey methodology, and
4) to develop in participating countries the skills and resources necessary to conduct high-quality demographic
and health surveys.

Information about the MEASURE DHS+ project or the status of MEASURE DHS+ surveys is available on the
Internet at http://www.measuredhs.com or by contacting:

ORC Macro
11785 Beltsville Drive,
Suite 300
Calverton, MD 20705 USA
Telephone: 301-572-0200
Fax: 301-572-0999
E-mail: reports@orcmacro.com
Profiling Domestic
Violence

A Multi-Country Study

Sunita Kishor

Kiersten Johnson

June 2004

ORC Macro
Editor: Megan Meline
Series design: Katherine Senzee
Report production: Kaye Mitchell

Suggested citation:

Kishor, Sunita and Kiersten Johnson. 2004. Profiling Domestic Violence – A Multi-Country Study.
Calverton, Maryland: ORC Macro.
 Contents

Tables and Figures v

Preface xi

Acknowledgments xiii

Executive Summary xv

Introduction 1

1.1 Background 1
1.2 Domestic Violence Measurement in the DHS Program 4

Prevalence of Different Types of Domestic Violence 11

2.1 Prevalence of Domestic Violence and Violence by Anyone 11

2.2 Different Forms of Spousal Violence 14
2.3 Health Consequences of Violence 21
2.4 Initiation and Frequency of Spousal Violence 22
2.5 Help Seeking 23

Risk Factors for the Experience of Domestic Violence 27

3.1 Woman’s Characteristics 27

3.2 Husband’s/partner’s Characteristics 33
3.3 Union Characteristics 36
3.4 Household Characteristics 38
3.5 Intergenerational Effects 42
3.6 Multivariate Regression Analysis: Identifying the Major Risk
Factors for Domestic Violence 43

Domestic Violence and Women’s Empowerment 53

4.1 Domestic Violence and Women’s Participation in

Decisionmaking 54
4.2 Violence and Norms that Reinforce Inequality
in Marital Relationships 63
4.3 Controlling Behaviors by Husbands and Domestic Violence 68

Domestic Violence and Demographic and Health Outcomes 73

5.1 Measures of Women’s Nutritional Status 74

5.2 Indicators of Women’s Reproductive Health and Spousal
Violence 76
5.3 Access to Antenatal and Delivery Care and Spousal Violence 88

Contents iii
 5.4 Children’s Mortality and Health and Mother’s Experience of
Spousal Violence 90

References 97

Appendix A DHS Domestic Module with Recommendations for

Implementation and Other Domestic Violence Questions 105

Appendix B Distribution of Ever-married Women and Currently

Married Women by Variables Used in the Multivariate
Logistic Regression 119

iv Contents
 Tables and Figures

Table 1.1 Description of the Demographic and Health Surveys (DHS)

included in this report and associated violence modules 3

Table 1.2 Percent distribution of women in the DHS domestic violence

samples, by country and background characteristics 4

Table 2.1 Percentage of women age 15-49 who have experienced any
violence by anyone since the age of 15 years (or since first
marriage in Egypt), percentages of ever-married women
age 15-49 who have experienced violence by a husband/partner
ever and in the 12 months preceding the survey, and the types
of questions used to estimate violence, by country 12

Table 2.2 Among women who have ever been pregnant, percentage who
have ever experienced violence during pregnancy by anyone 14

Table 2.3 Percentage of ever-married women age 15-49 who have

experienced specific acts of violence included in the modified
CTS ever or in the 12 months preceding the survey 15

Table 2.4 Percentage of ever-married women age 15-49 who have

experienced specific marital sexual acts of violence included
in the modified CTS ever or in the past 12 months 16

Table 2.5 Percentage of ever-married women age 15-49 who report

experiencing specific behaviors by their husbands that
constitute emotional violence 16

Table 2.6 Percentage of ever-married women age 15-49 who have ever
experienced different combinations of spousal violence

Table 2.7 Percentage of ever-married women age 15-49 who report

hitting or physically mistreating their husbands ever and in
the 12 months preceding the survey, by whether they
themselves have experienced any violence by their husbands 19

Table 2.8 Percentage of ever-married women age 15-49 who report

various types of health outcomes as a consequence of acts
carried out by their husbands or partners, by whether the
respondent ever experienced spousal violence 21

Table 2.9 Percent distribution of ever-married women age 15-49 who

have ever experienced spousal violence by when in their
marriage/partnership the violence first began, according to
marital duration 22

Tables and Figures v

 Table 2.10 Among ever-married women age 15-49 who report ever
experiencing violence by a spouse or partner, the frequency
of violence in the 12 months preceding the survey 23

Table 2.11 Among women who have ever experienced violence by anyone,
percentage who never sought help from anyone and percentage
who sought help from specific sources, by source(s) from which
help was sought 24

Table 2.12 Percent distribution of women who experienced violence by

anyone and did not seek help by reason for not seeking help 25

Figure 2.1 Percentage of all women who have experienced any violence
by anyone 13

Figure 2.2 Percent distribution of ever-married women who have

experienced spousal violence (emotional, physical, or sexual),
by type of violence experienced 18

Figure 2.3 Percentages of ever-married women who have experienced

spousal violence and not physically abused their husband,
percentage who have experienced spousal violence and have
physically abused their husbands, and percentage who have
physically abused their husbands but have not themselves
experienced physical violence 20

Table 3.1.1 Percentage of ever-married women age 15-49 who have ever
experienced spousal violence, by background characteristics 28

Table 3.1.2 Percentage of ever-married women age 15-49 who experienced

spousal violence in the past 12 months, by background
characteristics 29

Table 3.2.1 Percentage of ever-married women age 15-49 who have ever
experienced spousal violence, by husband’s characteristics 33

Table 3.2.2 Percentage of ever-married women age 15-49 who experienced

spousal violence in the past 12 months, by husband’s
characteristics 34

Table 3.3.1 Percentage of ever-married women age 15-49 who have ever
experienced spousal violence, by characteristics of the union 36

Table 3.3.2 Percentage of ever-married women age 15-49 who experienced

spousal violence in the past 12 months, by characteristics of
the union 37

Table 3.4.1 Percentage of ever-married women age 15-49 who have ever
experienced spousal violence, by household characteristics 39

vi Tables and Figures

Table 3.4.2 Percentage of ever-married women age 15-49 who experienced
spousal violence in the past 12 months, by household
characteristics 39

Table 3.5 Percentages of ever-married women age 15-49 who experienced

spousal violence ever and in the past 12 months, by whether their
mother was ever beaten by their father 43

Table 3.6.1 Correlates of a currently married woman's likelihood of having

ever experienced spousal violence: adjusted odds ratios estimated
using logistic regression 44

Table 3.6.2 Correlates of a currently married woman's likelihood of

having experienced spousal violence in the past 12 months:
adjusted odds ratios estimated using logistic regression 45

Table 3.7 Adjusted odds ratios of the likelihood of experiencing spousal

violence ever and in the 12 months preceding the survey for
respondents whose father beat their mother compared with
respondents whose fathers did not beat their mothers (including
those who do not know if their father beat their mother):
logistic regression results 46

Figure 3.1 Percentage of ever-married women age 15-49 who have ever
experienced spousal violence, by number of children ever born 31

Figure 3.2 Percentage of ever-married women age 15-49 who have

experienced violence in the preceding 12 months, by relative
household wealth status 42

Figure 3.3 Adjusted odds ratios: odds of experiencing spousal violence ever
and in the past 12 months, by frequency with which a husband
comes home drunk, Nicaragua 1998 49

Table 4.1 Percent distribution of currently married women age 15-49 by

who in their household makes different household decisions:
the women alone, the women jointly with their husbands, their
husbands alone, or someone else (alone or jointly with others in
the household), according to whether they have experienced
violence by their husband ever, in the past 12 months, or never 56

Table 4.2 Percentage of currently married women age 15-49 who

experienced violence in the past 12 months, according to whether
they make different decisions on their own or jointly with their
husband, or whether their husband alone makes the decision, by
type of decision 62

Tables and Figures vii

 Table 4.3 Percentage of ever-married women age 15-49 by whether they
have ever experienced violence by their husband, have experienced
violence by their husband in the last one year, or never experienced
violence by their husband, according to whether they agree or
disagree with different reasons justifying wife-beating and reason 64

Table 4.4 Percentage of ever-married women age 15-49 who agree with
each different reason justifying wife-beating, by reason, among
women who have experienced violence by their husband, ever,
in the last one year, or never 66

Table 4.5 Percentage of ever-married women age 15-49 who agree and
who disagree with different reasons for which a wife is justified
in refusing to have sex with her husband, by whether they have
ever experienced spousal violence, experienced spousal violence
in the past one year, or never experienced spousal violence,
according to reason 67

Table 4.6 Percentage of ever-married women age 15-49 who have

experienced violence by their husband ever, in the past one year,
or never, who agree with different reasons for which a wife is
justified in refusing to have sex with her husband, by reason 68

Table 4.7 Percentage of ever-married women age 15-49 who have

ever experienced violence by their husband, who have
experienced violence by their husband in the past one year,
and who have never experienced violence by their husband,
according to whether their husband shows different
controlling behaviors 69

Table 4.8 Percentage of ever-married women age 15-49 who have

experienced violence by their husband ever, in the past 12
months, or never, according to the number of marital control
behaviors shown by their husbands 70

Figure 4.1 Percentage of currently married women who have experienced

violence in the 12 months preceding the survey, according to the
person who has the final say in decisions about making large
household purchases 63

Figure 4.2 Percentage of women who have ever experienced spousal

violence, according to the number of controlling behaviors
shown by their husbands 71

Table 5.1 Percentage of ever-married women 15-49 who are underweight

(BMI<18.5), overweight (BMI 25.0-29.9), and obese (BMI>30.0)
and percentage who are anemic, by whether women have
experienced violence by their husband ever, in the past 12
months, or never 75

viii Tables and Figures

Table 5.2 Mean number of children ever born to ever-married women
age 15-49 by age in years, according to whether they have ever
experienced violence by their husband or not 77

Table 5.3 Percent distribution of births during the five years (three years
in India) preceding the survey and current pregnancies by
fertility planning status, according to whether the mother has
ever experienced violence by her husband or not 78

Table 5.4 Cumulative percent distribution of births during the five years
(three years in India) preceding the survey by the interval since
the previous birth and the median number of months since
previous birth according to whether the mother has ever
experienced violence by her husband or not 80

Table 5.5 Percent distribution of currently married women age 15-49

by their use of contraception, according to whether they have
experienced violence by their husband ever, in the past 12
months, or never 82

Table 5.6 Percentage of currently married women with need for family
planning, percentage with unmet need for family planning, and
percentage of total need satisfied by whether they have
experienced violence by their husbands ever, in the past 12
months, or never 84

Table 5.7 Among ever-married women who have ever had a live birth,
the percentage of ever-married women who have ever had a
nonlive birth or had a terminated pregnancy (miscarriage,
abortion, or stillbirth), according to whether they have ever
experienced violence by their husband or not 86
Table 5.8 Percentage of ever-married women who report having an STI
in the 12 months preceding the survey, according to whether
they have experienced violence by their husband ever, in
the past 12 months or never 87

Table 5.9 Percentage of births in the five years (three years in India)
preceding the survey, by whether antenatal care was received
and timing of such care and whether the delivery was assisted
by a medical professional, according to whether the mother has
ever experienced violence by her husband or not 89

Table 5.10 Infant and child mortality rates for the five years before the
survey, according to whether the mother has ever experienced
violence by her husband or not 91

Tables and Figures ix

 Table 5.11 Percentage of children age 12-35 months who received specific
vaccinations at any time before the interview, according to
whether the mother has ever experienced violence by her
husband or not 93

Table 5.12 Percentage of children age 0-59 months who are undernourished,
and percentage of children age 6-59 months who are anemic, by
whether the mother has experienced violence by her husband
ever, in the past 12 months, or never 94

Figure 5.1 Percentages of all births and all births not wanted at all born to
women who have ever experienced violence 79

Figure 5.2 Under-five child mortality rates by mother’s experience of

spousal violence 92

x Tables and Figures

 Preface

One of the most significant contributions of the MEASURE DHS+ program is the
creation of an internationally comparable body of data on the demographic and
health characteristics of populations in developing countries. The DHS Analytical
Studies series and the DHS Comparative Reports series examine these data, focusing
on specific topics. The principal objectives of both series are: to provide information
for policy formulation at the international level, and to examine individual country
results in an international context. Whereas Comparative Reports are primarily de-
scriptive, Analytical Studies take a more analytical approach.
The Analytical Studies series comprises in-depth, focused studies on a variety of
substantive topics. The studies are based on a variable number of data sets, depending
on the topic under study. A range of methodologies is used, including multivariate
statistical techniques. The topics covered are selected by MEASURE DHS+ staff in
conjunction with the MEASURE DHS+ Scientific Advisory Committee and
USAID.
It is anticipated that the Analytical Studies will enhance the understanding of sig-
nificant issues in the fields of international population and health for analysts and
policymakers.

Martin Vaessen
Project Director

Preface xi
 Acknowledgments

The authors wish to thank Dr. Mary Ellsberg of PATH and Ms. Anne Cross of
DHS for their insightful review of this manuscript. We also give heartfelt thanks to
Ms. Ladys Ortiz for all her patience and assistance in taming and tabulating data.

Acknowledgments xiii
 Executive Summary

This study uses household and individual-level data from the Demographic and
Health Surveys (DHS) program to examine the prevalence and correlates of domestic
violence and the health consequences of domestic violence for women and their chil-
dren. Nationally representative data from nine countries—Cambodia (2000), Colom-
bia (2000), the Dominican Republic (2002), Egypt (1995), Haiti (2000), India
(1998-1999), Nicaragua (1998), Peru (2000), and Zambia (2001-2002)—are ana-
lyzed within a comparative framework to provide a multifaceted analysis of the phe-
nomenon of domestic violence.
Scientific investigation of the problem of domestic violence is a relatively recent Scientific investigation
endeavor. It is only within the past 30 years that violence against women has been of the problem of
acknowledged internationally as a threat to the health and rights of women as well as domestic violence is a
relatively recent
to national development. With the recognition of violence against women as a global
endeavor. It is only
problem came the need for the development of methodologies to collect data on vio- within the past 30
lence ethically and in a manner that maximizes the validity and reliability of the data. years that violence
To this end, the DHS program began to collect information on the prevalence of against women has
domestic violence against women within the context of the household in the early been acknowledged
1990s. However, it was not until the late 1990s that the DHS program developed a internationally as a
standard module of questions in consultation with experts on domestic violence threat to the health
and rights of women
measurement, gender, and survey research. The module and its implementation con- as well as to national
form to the recommendations of the World Health Organization for ethical collec- development.
tion of data on domestic violence.
The proportions of ever-married women reporting spousal/intimate partner vio-
lence vary across countries. They are highest at 48 percent in Zambia, 44 percent in
Colombia, and 42 percent in Peru, and lowest at 18 percent in Cambodia, 19 percent
in India, and 22 percent in the Dominican Republic. In Egypt and Nicaragua, about
one in three ever-married women reports the experience of domestic violence.
Women who had ever been pregnant were asked about their experience of violence
during pregnancy. The proportions of women who reported spousal abuse during
pregnancy were highest in Colombia and Nicaragua at 11 percent, and lowest in
Cambodia at 1 percent, with Haiti and the Dominican Republic in the middle at 5
percent each.
In five of the nine countries included in this report, data on domestic violence
were collected by asking about several discrete acts of violence that can be categorized
as primarily physical, emotional, or sexual in nature. In all five of these countries, the
most frequently reported acts of physical violence were being pushed, shaken, slapped
or targeted with a thrown object or having one’s arm twisted. In all countries, more
than one in six women report having experienced at least one of these acts at some
time. The percentage of women reporting an act of sexual violence by their spouse
ranges from 4 percent in Cambodia to 17 percent in Haiti. At least one in ten
women in each country has been emotionally abused (threatened or publicly humili-
ated) by her husband at some time.
While the majority of this report is concerned with violence by husbands against
their wives, in some countries, data were also collected on whether women had ever
been physically violent against their husbands when their husbands were not being
violent toward them. The proportion of women reporting being violent against their

Executive Summary xv
 husbands is higher among women who have ever experienced spousal violence than
among women who have never experienced spousal violence. However, in all coun-
tries except the Dominican Republic, women who have physically abused their hus-
bands remain a small fraction of the proportion who have been abused by their hus-
bands.
Women who report having experienced violence from their husbands also fre-
quently report several immediate physical consequences. In Colombia, more than
half of the women who experienced violence reported that they had bruises and
aches. Between five and 13 percent of women who had experienced violence in the
six countries for which data are available report having an injury or broken bone. De-
spite the injurious outcomes that are often associated with the experience of domestic
violence, many women do not seek help for the violence; those not seeking help
range from 41 percent in Nicaragua to 78 percent in Cambodia. Most of those who
do seek help turn to their own families, friends or neighbors.
Examining the characteristics of the women who experience violence and the con-
texts in which they live helps to identify some of the common risk factors, if any, for
violence. In all countries, women who have been married more than once or who are
divorced or separated report higher rates of violence than women who are currently
married and have been married only once. This is not surprising since domestic vio-
lence can be an important reason for marriage dissolution. Women who married at a
young age and those who have multiple children are also more likely to report having
experienced violence. In most countries, women who are older than their husbands
are more likely to report having experienced violence. In five of the nine countries
analyzed, women living in urban households are more likely to report violence than
those living in rural households. The wealth of a household has an inconsistent and
often nonlinear relationship with the experience of violence. Women whose husbands
frequently return home drunk are several times more likely to report having experi-
enced violence than are women whose husbands do not come home drunk. Having a
family history of domestic violence between one’s parents significantly increases the
likelihood of experiencing violence oneself. In all countries where these data are avail-
able, the last two factors are consistently and positively associated with a woman’s
likelihood of experiencing violence. These relationships, identified in the bivariate
analyses, largely hold true in the multivariate analyses as well.
In most countries, Gender relations and roles may affect or be affected by the prevalence of violence
women who are older against women in a given society. Among the indicators of gender relations and roles
than their husbands considered here are currently married women’s participation in various types of
are more likely to
household decisions, their acceptance of wife-beating by husbands, attitudes toward a
report having
experienced violence. woman’s right to refuse to have sex with her husband, and controlling behaviors by
husbands that could strongly circumscribe women’s lives. The findings demonstrate
that rates of domestic violence tend to be lower for couples who share responsibility
for household decisions than for couples in which the husband or the wife makes
household decisions alone. With regard to attitudes about gender rights and roles, in
every country studied, women who agree, for example, that there are circumstances
under which it is acceptable for a husband to hit his wife are more likely to report
having ever experienced violence. However, there is no consistent relationship be-
tween a woman’s experience of violence and the degree to which she feels that a
woman has the right to refuse sex to her husband. As for the relationship of control-
ling behaviors exhibited by a husband and the respondent’s experience of violence,
the data indicate that for each of the six controlling behaviors for which information

xvi Executive Summary

was collected, rates of violence are much higher for women who say that their hus-
band manifests the behavior than for women who say he does not. For example, in
Colombia, the prevalence of violence among women whose husbands frequently ac-
cuse them of being unfaithful is 76 percent, compared with 34 percent among women
whose husbands do not accuse them of infidelity. This research also shows that the
likelihood of experiencing violence increases with the number of controlling behav-
iors exhibited by the husband. In the Dominican Republic, 10 percent of women
whose husbands do not manifest any of the controlling behaviors report having ever
experienced violence, compared with 17 percent of women whose husbands exhibit
one or two controlling behaviors and 76 percent of women whose husbands manifest
five or six of the given controlling behaviors.
Domestic violence not only poses a direct threat to women’s health, but also has
adverse consequences for other aspects of women’s health and well-being and for the
survival and well-being of children. This study examines the bivariate relationships of
domestic violence with a number of demographic and health outcomes, including
women’s and children’s nutritional status, women’s fertility, the intendedness of a
woman’s most recent birth, birth spacing, unmet need and contraceptive use, the
likelihood of having a non-live birth, the prevalence of sexually transmitted infections
(STIs), access to antenatal and delivery care, infant and child mortality, and vaccina-
tion coverage for children age 12-35 months.
The analyses find that the experience of domestic violence does indeed have a sig- Domestic violence not
nificant relationship with many health indicators, such that experience of violence only poses a direct
results in negative outcomes for the health of women and children. While fertility in threat to women’s
health, but also has
most countries is higher among women who have experienced violence than among
adverse
women who have not, the relationship tends to be weak. However, in all countries consequences for
except Haiti, women who have ever experienced violence are less likely to have had a other aspects of
birth that was wanted at the time of conception than women who have never experi- women’s health and
enced violence. For example, in Colombia, 58 percent of births in the past five years well-being and for the
to women who had ever experienced violence were unwanted, compared with 45 per- survival and well-
cent of births to women who had never experienced violence. In most countries, the being of children.
experience of violence is associated with slightly higher rates of ever use of contracep-
tion. However, women who have ever experienced violence are also more likely than
other women to not be currently using contraception, suggesting higher rates of con-
traceptive discontinuation among women who have experienced violence. Unmet
need is higher in the majority of the countries analyzed in this report for women who
have ever experienced violence than for those who have not. Self-reported prevalence
of STIs is also higher among women who have experienced violence than among
women who have not.
Women are not the only ones to suffer health-related repercussions from domestic
violence. Starting from conception, children of mothers who have experienced vio-
lence are at a disproportionate risk for poor health outcomes. In all countries except
Cambodia and Haiti, mothers are less likely to receive antenatal care in the first tri-
mester of their pregnancy if they have experienced violence than if they have not.
Furthermore, the likelihood of having had a non-live birth is higher by 33 to 72 per-
cent in eight of the nine countries among women who have ever experienced vio-
lence, than among women who have never experienced violence, and in seven of the
nine countries included here, under-five mortality rates are higher for mothers who
have experienced violence than for mothers who have not. With regard to vaccination
against childhood diseases, in Colombia, Egypt, Nicaragua, and Peru, the proportion

Executive Summary xvii

 of children age 12-35 months who are fully vaccinated is higher by five to 10 percent
among mothers who have not experienced violence than among mothers who have.
The proportion is higher by 38 percent in India and by 49 percent in the Dominican
Republic. As with maternal nutritional status, the association between a mother’s
Women are not the experience of domestic violence and child nutritional status is variable and inconsis-
only ones to suffer tent.
health-related
repercussions from
domestic violence. This report provides a comprehensive analysis of the phenomenon of domestic
Starting from violence from an international perspective, covering both the prevalence of violence
conception, children and its contexts and correlates. It is hoped that the report will be a useful tool for
of mothers who have raising awareness about this problem of wide-ranging significance as well as for in-
experienced violence forming the work of policymakers and program planners.
are at a
disproportionate risk
for poor health
outcomes.

xviii Executive Summary

 1
Introduction

1.1 Background

O ver the past 30 years, in the wake of such global events as the United Nations’
conferences on population and development and on women, the international
community has become increasingly aware of the importance of women’s gendered
social and health status in relation to key demographic and health outcomes.
Violence against women became a key issue in this regard, and early research on
the relationship between violence against women and reproductive health in the de-
veloping world (Heise et al., 1995; Heise, 1993) contributed to a deeper awareness of
the problem and the adverse health outcomes associated with it. Acceptance of gen-
der-based violence as a threat to women’s health and human rights was formalized
when 189 governments signed on to the Platform for Action of the 1995 United Na-
tions’ Beijing World Conference on Women. This platform explicitly recognizes that
violence against women creates an obstacle to the achievement of the objectives of
equality, development, and peace at the national level and violates the human rights
of women at the individual level. It further recognizes that the lack of data and statis-
tics on the incidence of violence against women makes the elaboration of programs
and monitoring of changes difficult (United Nations, 1995a).
Violence against women takes many forms. The 1993 Declaration on the Elimina-
tion of Violence Against Women of the United Nations General Assembly defined
such violence 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
private life.” This definition includes all forms of violence against women over the
entire life cycle. While some forms of violence tend to be specific to a life-cycle stage,
such as female feticide through sex-selective abortion, female infanticide, and female
genital cutting, other forms of violence cut across all ages. Violence can be in the
form of physical violence, sexual abuse, emotional or psychological abuse, verbal
abuse, and specific acts of violence during pregnancy. Women are also harmed by
limiting their access to food and medical care, carrying out dowry deaths and honor
killings, and coercing them to have sex through rape and/or sexual harassment. Men
who hurt women can be intimate partners, family members, or other men. The sub-
set of violence by intimate partners is usually referred to as “domestic violence,” al-
though the term is not always clearly defined.
It is within this context of increasing global awareness of the problem of violence
against women, along with the association of such violence with adverse demographic
and health outcomes, and the lack of representative information about the phenome-
non, that the Demographic and Health Surveys (DHS) program collects data on the
prevalence of domestic and other forms of violence against women within the house-
hold. Since its inception, the primary objective of the DHS program has been to pro-

1 • Introduction 1
 vide a comparable body of data on the demographic and health characteristics of
populations in developing countries. Traditionally, these data have included nation-
ally representative information on fertility, family planning, infant and child mortal-
ity, reproductive health, child health, and the nutritional status of women and chil-
dren. Since domestic violence is a health hazard in itself and plays a critical role in
women’s ability to attain other important demographic and health goals, domestic
violence data provide an important complement to the traditional focus areas of the
DHS program. Nonetheless, as of September 2003, nationally representative data on
domestic violence have been collected only in 11 countries that have implemented
DHS surveys. This report provides a summary of findings on domestic violence for
nine of these countries. Throughout this report, the term “domestic violence” is used
interchangeably with “spousal violence” or “intimate partner violence,” unless other-
wise specified. The terms “spouse” and “intimate partner” include any partners with
whom the respondent is living or has lived with as if married. It follows that terms
such as “currently married” or “ever-married” include “currently partnered” and “ever-
partnered” women.
The DHS survey is an The DHS survey is an ideal vehicle for studying not only the linkages between
ideal vehicle for domestic violence and health and demographic outcomes, but also the context in
studying not only the which violence takes place. The DHS Household Questionnaire collects data on sex,
linkages between
age, education, household headship, relationship to the household head for all
domestic violence and
health and household members, household possessions, and household access to various ameni-
demographic ties such as toilet facilities, water, and electricity. The DHS Women’s Questionnaire
outcomes, but also collects data for women age 15 to 49 years on a variety of characteristics, including
the context in which age, marital status, parity, contraceptive use, education, employment, and empower-
violence takes place. ment status, as well as their husband’s education, occupation, and alcohol consump-
tion. Women’s attributes combined with the reported attributes of their husbands
provide the characteristics of marital unions. With this information, it is possible to
describe the household context of violence, discuss the characteristics of women who
have experienced spousal abuse (as well as the characteristics of the abuser), and iden-
tify risk factors stemming from individual, union, and household-level conditions.
Accordingly, this report presents the prevalence of various types of violence: it de-
scribes the characteristics of the women who experience violence, as well as the char-
acteristics of their partners, marriages, and households and explores the relationship
between violence and indicators of women’s empowerment, demographic outcomes,
and women’s and children’s health and nutrition. The main purpose of this document
is to shed light on the phenomenon of gender-based violence, which has been sub-
jected to little close empirical examination, yet is theorized to have important link-
ages to the physical and psychological health of significant proportions of women and
children around the world. Specifically, Chapter 2 of this report discusses the cross-
national prevalence of violence by anyone against women; various forms of spousal
violence, including emotional, physical, and sexual violence; and violence by women
against their intimate partners. In Chapter 3, prevalence of spousal violence accord-
ing to individual, spousal, marital, and household characteristics is examined to better
understand some of the risk factors associated with violence. Chapter 4 discusses the
linkages between domestic violence and other indicators of women’s empowerment,
including their participation in household decisionmaking and their beliefs about
gender roles. Chapter 5 examines the bivariate relationship between selected demo-
graphic and health indicators for women and children and women’s experience of
spousal violence.

2 Introduction • 1
 Table 1.1 lists the countries included in this report, with the dates of fieldwork
and household and individual sample sizes, as well as the sample size of women ad-
ministered the domestic violence questions. The differences in the DHS sample size
and the sample size for the domestic violence data for any given country arise from
one or more sources. The largest source of difference is due to the fact that in about
half of the countries, the domestic violence module of questions was implemented in
only a subsample of the households selected for the DHS sample. Differences also
arise as a result of two security and ethical precautions increasingly mandated by the
DHS program (see below) for the collection of data on domestic violence. The first
requires that the interviewer does not continue with the questions on domestic vio-
lence if privacy cannot be ensured; the second requires that, in sample households
where more than one woman is eligible for the DHS survey, the domestic violence
module be administered to only one, randomly selected woman. A final source of
difference is that the domestic violence questions in some countries were only admin-
istered to ever-married women, even though the DHS sample included all women
age 15-49. Not all of the countries in Table 1.1 have information on all variables ex-
amined in this report; for example, in some countries, women were only asked about
ever-experience of spousal violence, while in others they were asked both about the
ever-experience of violence and the experience of violence in the past year. Conse-
quently, some countries may be excluded from some tables. Table 1.2 shows the sam-
ple of women who were interviewed about domestic violence by background charac-
teristics.

Table 1.1 Description of the Demographic and Health Surveys (DHS) included in this report and associated violence modules

Explicit
instruction
Number Eligiblity to
Number of of criterion discontinue
households women for Number interview if
Dates of Implementing inter- inter- women’s of Eligibility criterion for privacy not
Country fieldwork organization viewed viewed interview women domestic violence module possible
One randomly selected Yes
National Institute of All ever-married woman in
2/1/2000- Statistics/Ministry of women household,
Cambodia 6/1/2000 Health 12,236 15,351 15-49 2,403 age 15-49
All
3/1/2000- women
Colombia 7/1/2000 PROFAMILIA 10,907 11,585 15-49 11,536 All women 15-49 Yes
All One randomly selected
Dominican 7/4/2002- women woman in household,
Republic 12/10/2002 CESDEM 27,135 23,384 15-49 8,746 age 15-49 Yes
Ever-
married
11/1/1995- National Population women Ever-married women
Egypt 1/1/1996 Council 15,567 14,779 15-49 7,123 15-49 No
All One randomly selected
3/1/2000- Institut Haïtien de women woman in household,
Haiti 7/1/2000 l'Enfance 9,595 10,159 15-49 3,389 age 15-49 Yes
Ever-
married
11/1/1998- International Institute for women Ever-married women,
India 7/1/2000 Population Sciences 92,486 90,303 15-49 90,303 age 15-49 No
All One randomly selected
12/1/1997- Instituto Nacional de women ever-married woman in
Nicaragua 5/1/1998 Estadísticas y Censos 11,528 13,634 15-49 8,507 household, age 15-49 Yes
All
7/1/2000- Instituto Nacional de women
Peru 11/1/2000 Estadística e Informática 28,900 27,843 15-49 27,259 All women 15-49 Yes
All One randomly selected
11/1/2001- women woman in household,
Zambia 5/1/2002 Central Statistical Office 7,126 7,658 15-49 5,029 age 15-49 Yes

1 • Introduction 3
Table 1.2 Percent distribution of women in the DHS domestic violence samples, by country and background characteristics
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Background Dominican
characteristic Cambodia Colombia Republic Egypt Haiti India Nicaragua Peru Zambia
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Age
15-19 4.1 19.6 19.3 4.6 25.7 9.2 8.7 20.5 23.3
20-24 10.9 17.2 16.6 14.5 15.6 18.4 16.1 17.2 21.6
25-29 14.6 14.9 15.7 18.3 14.3 19.9 19.2 15.3 17.8
30-34 20.5 13.9 14.3 17.8 12.0 16.9 19.0 14.6 12.8
35-39 19.1 13.5 14.1 18.1 11.3 14.7 16.6 12.4 9.9
40-44 16.1 11.5 10.4 13.7 10.7 11.8 12.0 11.2 8.3
45-49 14.7 9.3 9.5 13.0 10.4 9.2 8.4 9.0 6.3

Residence
Urban 16.2 77.5 68.4 46.5 46.0 26.2 61.9 70.1 40.6
Rural 83.8 22.5 31.6 53.5 54.0 73.8 38.1 29.9 59.4

Education
No education 31.0 3.3 4.2 44.3 30.3 53.4 18.7 5.0 12.1
Primary 56.3 31.8 45.6 25.2 43.9 16.9 43.8 28.4 58.2
Secondary 12.5 50.1 33.7 23.8 24.7 21.8 32.0 44.8 26.3
Secondary+ 0.2 14.7 16.6 6.8 1.1 7.9 5.5 21.8 3.4

Marital status
Never married u 34.1 22.2 u 30.8 u 0.0 36.3 24.6
Married 86.5 24.9 18.3 92.6 57.1 93.8 36.7 31.1 60.0
Living together u 26.2 41.9 u 1.8 u 43.2 24.6 0.8
Widowed 9.1 2.1 0.6 5.1 2.0 4.2 1.3 1.4 4.9
Divorced/separated 4.5 12.7 17.0 2.2 8.3 2.0 18.9 6.6 9.7
Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0
Number of women 2,403 11,536 8,746 7,123 3,389 90,303 8,507 27,259 5,029
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
u = Unknown (not available)

Descriptions of indicators and any theoretical assumptions are provided separately

in each chapter. However, some general guidelines are applicable throughout this
report: 1) Respondents with missing values are excluded from each table, unless oth-
erwise noted, and as long as missing cases comprise 2 percent or less of the relevant
population, they are not reported; 2) calculations performed on exceptionally small
samples (26 to 49 cases) are noted in the respective table, and when the effective
sample size is 25 or fewer, the statistic is suppressed.

1.2 Domestic Violence Measurement in the DHS Program

The measurement of domestic violence within the DHS program has been evolv-
ing in keeping with the research on how to increase the validity of prevalence meas-
urement and in response to higher ethical standards in the collection of sensitive data
(Ellsberg et al., 2001; World Health Organization, 2001). In this section, we discuss
the steps taken in the DHS program to respond to these changing standards for re-
search in domestic violence and the extent to which the information in this report
reflects these concerns.

1.2.1 Increasing the Validity of DHS Violence Indicators

The first time domestic violence data were collected as part of a DHS was in Co-
lombia in 1990. In 1995, questions on domestic violence were fielded in Egypt as
part of a module of questions investigating the status of women in the country, and in
the same year, violence was again measured in Colombia. All of these initial attempts
at measuring domestic violence were isolated and did not use standardized questions.
Realizing this, in 1998-99 the DHS program set about developing a more standard-
ized approach to the measurement of domestic violence with the most valid measures

4 Introduction • 1
available. After consultation with experts on domestic violence measurement, gender,
and survey research, the DHS domestic violence module was developed. To design
this module, the DHS program built on the set of questions first implemented as part
of the 1998 Nicaragua DHS survey. The current DHS domestic violence module is
accompanied by guidelines on its ethical implementation. These guidelines were
adapted from corresponding World Health Organization guidelines (World Health
Organization, 2001). The complete module and the guidelines can be found in Ap-
pendix A.
Not all countries for which domestic violence data were collected by DHS surveys
have used the module. While data on violence were collected in Egypt long before
the development of the module, some of the countries where domestic violence data
were collected after the development of the module chose not to use it. In general,
however, the different approaches used to measure prevalence of domestic violence in
the nine countries included in this report fall into two categories. The first is a single-
question threshold approach, and the second is one—embodied in the DHS domes-
tic violence module—that combines the first approach with the use of a modified
Conflict Tactics Scale (CTS) to measure spousal violence.
The single-question threshold approach: This approach was used in Egypt, India,
Peru, and Zambia. The respondent is asked a single question to determine whether
she has ever experienced violence. Women who give a positive response are then
asked more questions, such as who the perpetrator was/is (including the husband),
and in Egypt, India, and Zambia, they are asked about the frequency of the violence.
No followup questions are asked of women who say “no” to the initial question.
Thus, the woman is given only one chance to disclose the occurrence of violence.
The modified CTS approach, as embodied in the domestic violence module: This ap-
proach involves implementing a modified version of the CTS to get information on
spousal violence and then a series of single questions to get at violence experienced at
the hands of someone other than a husband or partner, as well as violence during
pregnancy. The original CTS, developed by sociologist Murray Straus in the 1970s,
consists of a series of individual questions regarding specific acts of violence, such as
slapping, punching, and kicking. The original scale had 19 items (Straus, 1979,
1990). The modified list used by the DHS program includes only about 15 acts of
physical and sexual violence (see Appendix A). If the respondent affirms that any one
of the specified acts or outcomes has taken place, she is considered to have experi-
enced violence. The modified CTS approach was used in Cambodia, Colombia
(2000), Dominican Republic, Haiti, and Nicaragua.
The modified CTS approach has several advantages over a single-question thresh-
old approach, particularly in the context of cross-cultural research. By asking sepa-
rately about specific acts of violence, the violence measure is not affected by different
understandings between women of what constitutes violence. A woman has to say
whether she has, for example, ever been “slapped,” not whether she has ever experi-
enced “violence” or even “beatings” or “physical mistreatment.” All women would
probably agree what constitutes a slap, but what constitutes a violent act or what is
understood as violence may vary among women and across cultures.

1 • Introduction 5
 Nonetheless, not everyone agrees that measuring violence through discrete acts is
the most meaningful approach to measurement. For example, Smith, Tessaro, and
Earp (1995) have argued that surveys that measure discrete violent behaviors are in-
capable of capturing the “chronic vulnerability and gendered nature of battered
women’s experiences.” However, the purpose of asking questions about domestic vio-
lence in a national-level survey is to get the best estimates of the prevalence of the
phenomenon. For valid cross-national comparisons, it is important that the questions
have the same meaning in all cultural contexts. In this regard, questions about dis-
crete behaviors travel most easily across cultural and linguistic borders.
Another advantage of the modified CTS approach is that it gives respondents
multiple opportunities to disclose the experience of violence. The level of comfort in
All women would disclosing such experiences to anyone, let alone to an interviewer, is likely to vary
probably agree what
constitutes a slap, but
among cultures as well as among women sharing the same culture. The level of com-
what constitutes a fort in disclosing such experiences to anyone, let alone to an interviewer, is likely to
violent act or what is vary among cultures as well as among women sharing the same culture. Some women
understood as may not be immediately willing to disclose their experience of violence the very first
violence may vary time they are asked, and hence an approach that uses a single gatekeeping question
among women and would yield a lower prevalence. Also, a single question is much less likely than multi-
across cultures.
ple questions are to capture women’s varied experiences of violence. Thus, an ap-
proach that asks about violence from many different angles using separate questions,
is likely to encourage disclosure because it gives women some time to think about
their experiences and permits them to disclose when they are ready and/or when they
are asked a question describing an experience with which they identify.
The modified CTS approach corrects several inadequacies of the original CTS.
Although it is the most commonly used quantitative measure of domestic violence,
the original CTS has also been criticized on several grounds (c.f. DeKeseredy and
Schwartz, 1998), including: 1) it situated abuse in the context of disputes, disagree-
ments, or differences, rather than allowing for the possibility that abuse can occur
even without any other form of conflict; 2) it did not include sexual violence, which is
often a complement of other forms of physical violence; and 3) it grouped acts of vio-
lence into categories that suggest that the act determines severity, rather than its con-
sequences. Most of these shortcomings of the original CTS do not apply, however, to
the modified CTS recommended by the DHS program. The modified CTS incorpo-
rates questions on sexual violation along with questions on physical violence. Further,
the DHS program implements the CTS in a way that does not assume that violence
takes place only in circumstances characterized by conflict. The module also contains
questions that investigate the consequences of violence: one set of questions asks
about physical outcomes of the violence, such as bruises or broken bones. Notably,
however, there is no further probing into possible motives for the violence that took
place, and there is no investigation into the meaning for the woman of a given act of
violence. In this report, no attempt was made to rank the severity of abuse.
On the basis of one of these two approaches to the reporting of spousal violence,
two indicators of the prevalence of spousal or intimate partner violence are defined
and used throughout this report, namely, having ever experienced spousal violence
and having experienced spousal violence in the 12 months preceding the survey.
While the former measure reflects lifetime experience, the latter identifies women
who are currently at risk. Spousal violence measures (unless otherwise indicated) ex-
plicitly include both physical and sexual violence perpetrated by husbands (including
current or past husband/partner) in Cambodia, Colombia, the Dominican Republic,

6 Introduction • 1
Haiti, Nicaragua, and Zambia. In the remaining countries, the questions used did not
separately ask about acts of sexual violence, and hence such violence would be in-
cluded only if respondents themselves see their experience of sexual violence as an
experience of physical violence, beatings, or mistreatment.
The advantages of the CTS approach, compared with the single-question thresh-
old approach, suggest that violence data collected with the latter approach may un-
derestimate prevalence. The extent to which this is true, however, is likely to differ
across countries and within countries by culture and region. The extent of underesti-
mation may also depend on how acceptable the reporting of violence is and the very
prevalence of violence that is being measured. Consequently, it is important that
comparison of the prevalence of violence across countries be carried out with caution.

1.2.2 Ensuring the Ethical Collection of Violence Data

Much of the information typically collected in a DHS survey is very personal and
sensitive in nature (e.g., information on sexual behavior and condom use). Conse-
quently, the DHS program already has strict procedures that meet international re-
quirements of informed consent and privacy of information. Precautions include the
requirement that names of respondents are never disclosed and are excluded from all
data sets. In addition to these precautions, several other safety and ethical procedures
and guidelines are recommended when a country considers collecting domestic vio-
lence data as a part of the planned DHS. These guidelines, in keeping with World
Health Organization (2001) ethical and safety recommendations for research on do-
mestic violence, include:
• An instruction, built into the domestic violence module, that requires the in-
terviewer to continue the interview only if privacy is ensured. If privacy cannot be
obtained, the interviewer must skip the module and enter an explanation of what
happened.
• At the start of the module, each respondent is read a statement to inform her
that the next set of questions are very personal in nature and will explore different
aspects of a woman’s life. The statement also assures the respondent that her answers
are completely confidential and that no one else will be told her answers. This state-
ment is in addition to the informed consent obtained at the start of the DHS inter-
view.
• Special training is provided for interviewers and supervisors to sensitize them
to the problem of domestic violence and to the specific challenges involved in collect-
ing data on violence. The need to develop a rapport with the respondent and ensure
privacy is emphasized both during the training and practice sessions.
• Only one eligible woman in each selected household is to be administered
the module questions. In households with more than one woman eligible for the
DHS survey, the woman administered the module is to be randomly selected through
a specially designed sample-selection procedure. By interviewing only one woman in
each household for the domestic violence questions, possible security breaches due to
other persons in the household knowing that information on domestic violence was
given are minimized.
• Information on organizations that provide services or referrals to victims of
domestic violence is made available to any respondent who asks the interviewer for
help.
• If men are interviewed, they are not asked domestic violence questions.

1 • Introduction 7
 It is also recommended that translators not be used to administer the domestic
violence questions. The use of translators is minimized in the survey because DHS
guidelines require that questionnaires be translated into the major languages of the
country. Accordingly, in Cambodia, the questionnaire was translated into Khmer; in
Colombia, the Dominican Republic, and Peru, it was translated into Spanish; in
Haiti, it was translated into French and Creole; in Zambia, it was translated into
seven languages, and in India, it was translated into 17 languages. To minimize any
changes in meaning through translation, the DHS program routinely performs back-
translation to check the accuracy of the translated questionnaire.
Most of these recommendations have been followed in countries where the DHS
domestic violence module has been implemented.

1.2.3 Attempts to Minimize Underreporting of Violence

There is often a culture of silence around the topic of domestic violence, which
makes the collection of data on this sensitive topic particularly challenging. Even
women who want to speak about their experience with domestic violence may find it
difficult because of feelings of shame or fear.
There is often a DeKeseredy and Schwartz (1998), for example, note that while all victims’ surveys
culture of silence exhibit a certain amount of underreporting, it is assumed that surveys that incorpo-
around the topic of rate questions on intimate violence are particularly susceptible to this shortcoming.
1

domestic violence,
Building rapport with the respondent, ensuring privacy, providing the respondent
which makes the
collection of data on with multiple opportunities for disclosure, and asking longer, more probing questions
this sensitive topic following the simple measures embodied in the CTS have all been identified as pos-
particularly sible ways to encourage the reporting of violence (c.f. DeKeseredy and Schwartz,
challenging. Even 1998; Ellsberg et al., 2001). There are several ways in which the DHS program has
women who want to attempted to encourage disclosure. The new module, as discussed above, provides
speak about their respondents with multiple opportunities for disclosure, not only by asking them many
experience with
domestic violence
different times about any experience of violence, but also by asking them about many
may find it difficult different forms of violence. The module is generally located in the latter part of the
because of feelings of DHS questionnaire; therefore by the time the respondent is asked about her experi-
shame or fear. ence of violence, the interviewer and respondent are fairly well acquainted. Several of
the ethical and safety guidelines described above also contribute directly to promoting
disclosure of any experience of violence. For example, the special training focuses on
asking about violence in nonjudgmental tones. Also, the option of discontinuing the
interview if complete privacy cannot be obtained increases the likelihood that vio-
lence questions are asked only when the respondent feels secure.
Despite these precautions, concern remains about possible underestimation of vio-
lence. However, in at least one country, Cambodia, there is independent corrobora-
tion of the DHS spousal violence estimate. The Cambodia DHS estimate is almost
identical to the corresponding estimate from the Household Survey of Domestic Vio-
lence in Cambodia (Ministry of Women’s Affairs and Project Against Violence,
1996). When interpreting differentials in prevalence among subgroups in a given
country, caution should always be exercised. While a large part of any substantial dif-
ferences in violence between subgroups undoubtedly reflects actual differences in

1
The assumption that shame is associated with domestic violence, with underreporting being
a consequence of such shame, might be a cultural artifact (associated with the researcher). To
the authors’ knowledge, there have been no studies that attempt to discern whether or not
domestic violence is a shameful or embarrassing topic in all cultural contexts.

8 Introduction • 1
prevalence, differential underreporting by women in the different subgroups can also
contribute to the exaggeration or narrowing of differences in prevalence. Caution is
also advised when comparing the overall prevalence of violence among countries, es-
pecially among countries that have used different approaches to measure prevalence.

1 • Introduction 9
 2
Prevalence of Different Types of Domestic Violence

T his chapter documents the prevalence of different forms of violence, estimated

from responses given by women age 15-49 who were asked about their experi-
ence of violence. The types of violence discussed include physical and sexual violence
by anyone; physical, sexual, and emotional violence by a spouse/partner; violence dur-
ing pregnancy; and violence by wives against their husbands. In Section 2.1, the
prevalence of violence by anyone, prevalence of spousal violence, and violence during
pregnancy are discussed. In Section 2.2, an in-depth examination is conducted of
various forms of spousal violence, including types of physical, sexual, and emotional
violence, based on information obtained from women’s responses to the Conflict
Tactics Scale (CTS) questions. Also discussed is violence by wives against their hus-
band/partner. In Section 2.3, the relationship between injuries and the experience of
violence is examined. This discussion is followed by a discussion in Section 2.4 of the
timing of the initiation of spousal violence and its frequency. Finally, in Section 2.5,
help-seeking behaviors of women who have ever experienced violence by anyone are
discussed. Details of questions used in each country are provided in Appendix A.

2.1 Prevalence of Domestic Violence and Violence by Anyone

Table 2.1 shows the prevalence of any violence since age 15 (since first marriage in
Egypt) by anyone for all women age 15-49 and the percentages of ever-married
women who have experienced violence ever and in the 12 months preceding the sur-
vey. Women are said to have experienced violence if they say “yes” to any one of the
relevant questions summarized in the last column of the table. Countries differen-
tially include sexual violence: in Egypt, India, and Peru, no explicit questions were
asked about sexual violence; in Cambodia, Colombia, the Dominican Republic,
Haiti, and Nicaragua, women were asked about sexual violence by current or last
husband/partner; and in Zambia, all women were asked about sexual violence by any-
one, including their husband.
Table 2.1 shows that the proportion of women reporting violence by anyone since
the age of 15 (or since first marriage in Egypt) is high in all countries: in Zambia,
more than half of women report having experienced violence; in Peru and Colombia,
more than two out of five women have experienced violence; in Egypt, Haiti, and
Nicaragua, the proportion is one in three; and in the Dominican Republic, Cambo-
dia, and India, it is about one in four (see Figure 2.1).
Rates of spousal/intimate partner violence among ever-married women vary simi-
larly across countries, with the rates being highest at 48 percent in Zambia, 44 per-
cent in Colombia, and 42 percent in Peru, and lowest at 18 percent in Cambodia.
With the exception of Colombia, rates of spousal violence alone are much lower
(Cambodia, Haiti, Peru, and Zambia) or somewhat lower (Dominican Republic,
Egypt, India, and Nicaragua) than the rates of any violence.

2 • Prevalence of Different Types of Domestic Violence 11

Table 2.1 Percentage of women age 15-49 who have experienced any violence by anyone since the age of 15 years (or since first
marriage in Egypt), percentages of ever-married women age 15-49 who have experienced violence by a husband/partner ever and in
the 12 months preceding the survey, and the types of questions used to estimate violence, by country
Ever-married women
Percentage
Percentage of beaten by a
women ever Percentage ever spouse/
beaten by beaten by a partner in the
Country anyone spouse/partner past 12 months Definition of having experienced violence: a “yes” on one or more
Cambodia1 23.4 17.5 15.4 Items on the modified CTS and questions on being hit, slapped,
(n=2,403) (n=2,403) (n=2,403) kicked, or physically hurt by someone ever and/or during
pregnancy

Colombia 41.0 44.1 Items on the modified CTS and questions on being hit, slapped,
(n=11,536) (n=7,602) u kicked, or physically hurt by someone ever and/or during
pregnancy

Dominican 23.9 22.3 11.0 Items on the modified CTS and questions on being hit, slapped,
Republic (n=8,746) (n=6,807) (n=6,807) kicked, or physically hurt by someone ever and/or during
pregnancy
1
Egypt 35.0 34.4 12.5 Questions on having ever been beaten since first married and
(n=7,123) (n=7,123) (n=7,123) during any pregnancy

Haiti 35.2 28.8 21.0 Items on the modified CTS and questions on being hit, slapped,
(n=3,389) (n=2,347) (n=2,347) kicked, or physically hurt by someone ever and/or during
pregnancy
1
India 21.0 18.9 10.3 Question on having been beaten or mistreated physically since
(n=90,303) (n=90,303) (n=90,303) age 15

Nicaragua1 32.6 30.2 13.2 Items on the modified CTS and questions on being hit, slapped,
(n=8,507) (n=8,507) (n=8,507) kicked, or physically hurt by someone ever and/or during
pregnancy

Peru 47.4 42.4 u Questions on being pushed, hit, attacked physically by

(n=27,259) (n=17,369) spouse/partner and/or hit, slapped, kicked or hurt physically by
anyone

Zambia 58.7 48.4 26.5 Questions on having been beaten by husband, beaten by
(n=5,029) (n=3,792) (n=3,792) anyone, forced to have sex by anyone including the husband, or
forced to have sex with a third party
1
Sample includes only ever-married women
u = Unknown (not available)

Rates of violence in the year prior to the survey among all ever-married women are
necessarily similar to or lower than the rates of violence ever reported by the same
women, and they measure the extent to which women are currently at risk of vio-
lence. Current violence rates will be more similar to rates of ever-experience of vio-
lence in countries where the status of women does not allow them to renegotiate the
terms of their relationships with their partners and where women cannot easily leave
violent relationships (through formal divorce, for example). Information on the ex-
perience of recent violence was not obtained in Colombia and Peru, but for the other
countries, the proportion of women who have experienced spousal violence in the

12 Prevalence of Different Types of Domestic Violence•2

year preceding the survey ranges from a high of 27 percent in Zambia to a low of 10
percent in India. In Cambodia, most of the women (88 percent) who report ever be-
ing abused by a husband or partner also report being abused in the past 12 months.
This proportion is also very high in Haiti, at 73 percent. In the Dominican Republic,
India, and Zambia, about half of the women who report ever experiencing spousal
violence also report experiencing violence in the past year. Only in Egypt and Nicara-
gua does the proportion of women reporting recent violence fall below half to 36 and
44 percent, respectively.

Figure 2.1
Percentage of all women who have experienced any
violence by anyone

Cambodia 23

Colombia 41

Dominican Republic 24

Egypt 35

Haiti 35

India 21

Nicaragua 33

Peru 47

Zambia 59

0 20 40 60 80
Percent

Note: Cambodia, Egypt, India, and Nicaragua are ever-married women.

Women age 15-49 who were ever pregnant (including those currently pregnant)
were asked whether they had ever experienced violence or physical mistreatment by
anyone during any pregnancy. Table 2.2 shows that 13 percent of women in Colom-
bia have experienced violence by someone during pregnancy, compared with 6 to 7
percent in the Dominican Republic and Haiti. In these countries, for most women
reporting violence during pregnancy, the perpetrator was a husband. Violence during
pregnancy is least prevalent in Cambodia, where only 3 percent of women report
beatings by anyone during pregnancy and only 1 percent report violence by a hus-
band. In Egypt, 11 percent of women report being beaten during pregnancy by
someone, and in Nicaragua, 11 percent of women report being beaten during preg-
nancy by a husband. The true estimates in both Egypt and Nicaragua may be some-
what higher, however, since the reported estimates do not include the experience of
women who have been beaten only during pregnancy (and not before and after) and
who did not say “yes” to the questions on having ever experienced violence.

2 • Prevalence of Different Types of Domestic Violence 13

 Table 2.2 Among women who have ever been pregnant, per-
centage who have ever experienced violence during pregnancy by
anyone
––––––––––––––––––––––––––––––––––––––––––––––––––––––
Percentage of women
who ever experienced
violence during pregnancy
––––––––––––––––––––––––– Number of
By By a women ever
Country anyone husband pregnant
––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia1 2.5 1.3 2,288
Colombia 12.8 10.6 7,286
Dominican Republic 6.2 5.1 6,467
2
Egypt 11.1 u 6,652
Haiti 6.5 5.4 2,226
Nicaragua2 u 11.1 8,142
––––––––––––––––––––––––––––––––––––––––––––––––––––––
1
Question on whether woman was beaten during pregnancy did not
include “husband” as a pre-coded response category
2
In Egypt and Nicaragua, only women who had reported ever
experiencing violence were asked the question on whether they had
experienced violence during pregnancy. Hence the reported
proportions do not include women who experienced violence only
during pregnancy but not before and after and did not say “yes” to
the questions on ever experience of violence.
u = Unknown (not available)

2.2 Different Forms of Spousal Violence

The different forms of violence discussed in this section include specific acts of
physical, sexual, and emotional violence perpetrated by the husband/partner against
2
his wife and any violence perpetrated by the wife against the husband/partner.
Physical spousal violence. For the five countries where the modified CTS was
used, Table 2.3 shows the percentages of women who report the different types of
violence included in the scale to measure spousal physical violence. Small variations
in the actual wording of the different “acts” of violence included are noted in the
footnotes to the table.
In all five countries, acts described in items (a) and (b) in the table (“push you,
shake you, or throw something at you” and “slap you or twist your arm”) tend to be
the ones most commonly reported by women. In all countries, more than one in six
women report having experienced at least one of these acts. Acts (a) to (e) are, in
general, more commonly reported than acts (f) to (h). In Colombia, 40 percent of
women report at least one act from (a) to (e); in the remaining countries, this propor-
tion is 16 to 27 percent. By contrast, acts (f) to (h) are reported by 4 to 11 percent of

2
While these data are rarely of interest on their own, they are discussed here to provide
insight into the pattern of responses given by women across countries where virtually identical
questions were fielded.

14 Prevalence of Different Types of Domestic Violence•2

Table 2.3 Percentage of ever-married women age 15-49 who have experienced specific acts of violence included in the
modified CTS ever or in the 12 months preceding the survey
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Dominican
Cambodia Colombia Republic Haiti Nicaragua
(n=2,403) (n=7,602) (n=6,807) (n=2,347) (N=8,507)
–––––––––––– –––––––––––– –––––––––––– –––––––––––– ––––––––––––
Past Past Past Past Past
Item Ever year Ever year Ever year Ever year Ever year
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
(a) Push you, shake you, or throw
1
something at you 10.0 8.6 36.6 na 15.1 7.8 12.5 8.5 22.3 9.6
2
(b) Slap you or twist your arm 11.3 9.8 30.7 na 11.8 6.2 11.2 7.9 17.1 6.4
(c) Punch you with a fist or
3
something that could hurt you 6.1 5.4 8.8 na 9.6 5.1 9.3 5.9 19.4 7.3
(d) Bite you na na 3.5 na na na na na na na
(e) Kick or drag you 5.9 5.1 12.4 na 3.7 2.0 6.8 4.1 9.2 3.1

Only one type of act (a-e) 6.6 6.0 10.7 na 5.5 2.8 5.8 4.6 7.2 4.0
Any two types of acts (a-e) 3.7 3.4 13.9 na 5.2 2.6 3.7 2.5 6.3 3.2
Any three types of acts (a-e) 2.8 2.3 8.4 na 4.0 2.3 1.9 1.5 6.5 2.2
Any four types of acts (a-e) 2.7 2.3 5.2 na 3.1 1.5 5.2 3.0 7.1 2.4
All five types of acts (a-e) u u 1.5 na u u u u u u
Any act (a-e) 15.9 14.0 39.7 na 17.8 9.3 16.6 11.7 27.1 11.8

(f) Try to strangle or burn you 0.6 0.5 4.5 na 3.3 2.1 1.9 1.3 7.0 2.6
(g) Threaten you with a knife, gun,
or other type of weapon 3.1 2.6 8.4 na 4.1 2.5 2.9 2.0 8.8 2.8
(h) Attack you with a knife, gun, or
other type of weapon 1.3 1.0 3.9 na 2.5 1.5 0.9 0.8 u u
Only one type of act (f-h) 3.1 2.8 5.7 na 3.5 2.2 2.9 2.4 6.9 2.6
Any two types of acts only (f-h) 0.6 0.4 3.1 na 1.4 0.8 1.1 0.7 4.5 1.4
All three types of act (f-h) 0.2 0.2 1.6 na 1.2 0.8 0.2 0.1 u u
Any act (f-h) 3.9 3.4 10.4 na 6.1 3.8 4.2 3.2 11.4 4.0

Any physical violence: At least one

act from (a-h) 16.4 14.6 40.0 na 18.4 9.8 17.3 12.5 27.6 11.9
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
1
Only “push and shake you”
2
Hit you with his hand
3
Hit you with a hard object"
na = Not applicable; u = Unknown (not available)

women. The proportion of women reporting at least one of the listed acts (a-h) is 16
to 18 percent in Cambodia, Haiti, and the Dominican Republic, 28 percent in Nica-
ragua, and 40 percent in Colombia. Few women report experiencing all of the types
of acts listed as (a) to (e), and an even smaller proportion report all of the acts listed
as (f) to (h) in any country.
Sexual spousal violence. Table 2.4 shows the percentage of women reporting dif-
ferent acts of marital sexual violence by a current or last husband/partner. Zambia is
the only country where this information is not obtained with a CTS format. In Zam-
bia, all women were asked whether they had ever been forced to have sex by anyone,
including their husband, and whether they had ever been forced to have sex with a
third person.
The percentage of ever-married women reporting that they have experienced an
act of sexual violence by their husband or partner ranges from 17 percent in Haiti, to
10 to 11 percent in Colombia and Nicaragua, and 4 to 6 percent in the remaining
countries. Notably, with the exception of Nicaragua, in the remaining four countries
for which data are available, at least two-thirds of the women who report ever experi-
encing spousal sexual violence also report such violence for the 12 months preceding
the survey. In Nicaragua, this proportion is also high but less than half (38 percent).

2 • Prevalence of Different Types of Domestic Violence 15

Table 2.4 Percentage of ever-married women age 15-49 who have experienced specific marital sexual acts of violence included in the
modified CTS ever or in the past 12 months
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Dominican
Cambodia Colombia Republic Haiti Nicaragua Zambia
(n=2,403) (n=7,602) (n=6,807) (n=2,347) (N=8,507) (n=3,792)
–––––––––––– –––––––––––– –––––––––––– –––––––––––– –––––––––––– ––––––––––––
Past Past Past Past Past Past
Item Ever year Ever year Ever year Ever year Ever year Ever year
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
(a) Threaten you in order to have
sexual intercourse even when
you did not want to u u u u u u u u 6.6 2.4 u u
(b) Physically force you to have
sexual intercourse even when
you did not want to 3.4 2.9 11.0 u 6.0 4.0 16.7 14.4 8.7 3.0 5.1 3.9
(c) Force you to perform other
types of sexual acts you
1
did not want to 1.4 1.2 u u 3.4 2.2 6.0 5.5 5.7 2.4 0.0 0.0

Only one type of act (a-c) 2.5 2.2 11.0 u 3.5 2.3 11.4 9.7 3.4 1.6 5.1 3.9
Only two types of act (a-c) 1.1 0.9 na u 3.0 2.0 5.7 5.1 2.9 0.9 0.0 0.0
All three types of acts (a-c) na na na na na na na na 3.9 1.5 na na

At least one act (a-c) 3.6 3.2 11.0 u 6.4 4.2 17.0 14.8 10.2 3.9 5.1 3.9
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
1
In Zambia, respondents were asked whether they had ever been forced to have sex with another person.
na = Not applicable
u = Unknown (not available)

Emotional spousal violence. As part of the domestic violence module, women

were asked about several different behaviors that can be considered as constituting
emotional violence. The only two behavior questions common to the modified CTS
used across countries were “Does/did your husband ever say or do something to hu-
miliate you in front of others?” and “Does/did your husband threaten you or someone
close to you with harm?” The percentage of ever-married women reporting that their
3
husbands/partners do exhibit these behaviors are shown in Table 2.5.

Table 2.5 Percentage of ever-married women age 15-49 who report experiencing specific behaviors by their
husbands that constitute emotional violence
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Specific acts Experienced
` of emotional violence at least
––––––––––––––––––––––– Experienced one of the
Ever at least Experienced specified acts
Ever threatened one of the both of the of violence in
humiliated her or those specified acts specified acts the 12 months
her in front close to her of emotional of emotional preceding
Country of others with harm violence violence the survey
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia 7.9 9.3 13.5 3.7 12.1
Colombia 11.5 u 11.5 u u
Dominican Republic 15.1 9.9 17.7 7.2 11.3
Haiti 11.9 6.9 13.2 5.5 10.8
Nicaragua 27.7 16.5 29.0 22.3 15.9
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
u = Unknown (not available)

3
The cross-cultural validity of the items on the emotional violence scale has not yet been
unequivocally established; hence, much more care is needed in interpreting this information
compared with the information on physical and sexual violence. This is also a reason why data
on emotional violence are not included in the rates of violence reported, analyzed, and used
elsewhere in this report.

16 Prevalence of Different Types of Domestic Violence•2

 Twelve percent or more of women report that their husbands have subjected them
to at least one of these behaviors. In Nicaragua, 29 percent of women report experi-
encing at least one such behavior and 22 percent report being subjected to both types
of behaviors. In the Dominican Republic, 18 percent of women report emotional vio-
lence, and in the remaining three countries, the proportion is 12 to 14 percent. The
percentage of women experiencing emotional violence in the past year, in the four
countries for which these data are available, are fairly similar: in all four countries, 11
to 16 percent of women have experienced emotional violence at the hands of their
husbands/partners in the past 12 months.
Combinations of spousal violence. Since emotional, physical, and sexual violence
are likely to co-occur, Table 2.6 shows the percentage of ever-married women who
report different combinations of emotional, sexual, and physical violence. Figure 2.2
shows the percent distribution of women who report emotional, physical, or sexual
violence according to types of violence they have experienced.

Table 2.6 Percentage of ever-married women age 15-49 who have ever experienced different combinations of spousal violence
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Emotional Emotional Physical Emotional, Emotional, Number
and and and physical, physical, of ever-
Emotional Physical Sexual physical sexual sexual and or married
Country only only only only only only sexual sexual women
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia 5.5 7.6 0.3 5.8 0.4 0.9 2.1 22.3 2,403
Colombia 1.7 24.9 1.0 5.3 0.1 5.5 4.4 42.8 7,602
Dominican Republic 5.5 5.4 0.5 7.3 0.4 1.1 4.5 24.7 6,807
Haiti 2.6 5.5 8.4 4.2 1.0 2.3 5.4 29.3 2,347
Nicaragua 5.2 3.8 0.5 14.7 0.6 0.6 8.5 33.9 8,507

Based on the items on the CTS only, 43 percent of ever-married women in Co-
lombia, 34 percent in Nicaragua, 29 percent in Haiti, 25 percent in the Dominican
Republic, and 22 percent in Cambodia have experienced emotional, physical, or sex-
ual violence by their current or last husband (Table 2.6). In Cambodia, women are
most likely to report physical violence only, followed by emotional and physical vio-
lence only and by only emotional violence. Other types of violence on their own or in
combination are far less common (see Figure 2.2). In Colombia, few women report
either emotional violence or sexual violence alone or in combination. Women are
most likely to report only physical violence (25 percent), followed by combinations of
physical violence with sexual and/or emotional violence. In the Dominican Republic
and Nicaragua, sexual violence is least likely to be reported; however, physical and
emotional violence alone or in combination are most common. Haiti is the only
country where sexual violence alone is reported more often than any other form of
violence on its own or in combination. As shown in Figure 2.2, sexual violence ac-
counts for 29 percent of violence reported in the CTS by women in Haiti. The next
most common types of violence reported are physical only (19 percent) and all three
forms of violence (18 percent).

2 • Prevalence of Different Types of Domestic Violence 17

 Figure 2.2
Percent distribution of ever-married women who have experienced
spousal violence (emotional, physical, or sexual), by type of violence
experienced

Cambodia Colombia

10% 4%
4% 9%
24%
2% 13%

0%
26%
12%
59%
34% 2%
1%

Dominican Republic Haiti

18% 22% 9%
18%

5% 19%
2% 8%
3%
22%
14%
29% 29%
2%

Emotional only

Nicaragua
Physical only
15%
25% Sexual only
11%
Emotional and physical only
2% 2%
2%
Emotional and sexual only

43%
Physical and sexual only

Emotional, physical, and sexual

18 Prevalence of Different Types of Domestic Violence•2

 Violence by women against their husband/partner. Spousal violence by the hus-
band is not the only form of spousal violence. Women may also sometimes be the
perpetrators of violence. To measure violence by wives against their husbands, (vio-
lence that is not in self-defense alone) women were asked, “Have you ever hit,
slapped, kicked, or done anything else to physically hurt your (last) husband at times
when he was not already beating or physically hurting you?” The phrase “when he
was not already beating or physically hurting you” is included in the question to help
minimize the reporting of violence resorted to in self-defense. However, since the
question does not explicitly exclude acts committed in response to a perceived or
known threat, it remains unclear how much of the reported violence by women is
occurring because women who have already experienced abuse are acting violently in
anticipation of further abuse and how much of it is due to women initiating abuse
without any known threat. To begin to sort out some of these issues, though not con-
clusively, the proportion of women reporting violence against their husbands is also
presented by whether women have themselves been abused.
Accordingly, Table 2.7 shows the percentage of all ever-married women who say
that they have ever physically abused their husbands or physically abused their hus-
bands in the past year. Also shown in the table is the variation in these proportions
by women’s own experience of violence. Figure 2.3 shows the proportions of ever-
married women who have experienced spousal violence and have not abused their
husbands, have experienced spousal violence and abused their husbands, and have
abused their husbands but have not themselves experienced spousal violence.

Table 2.7 Percentage of ever-married women age 15-49 who report hitting or physically mistreating
their husbands ever and in the 12 months preceding the survey, by whether they themselves have
experienced any violence by their husbands
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Ever experienced Never experienced All ever-
violence by husband violence by husband married women
––––––––––––––––––– ––––––––––––––––––– ––––––––––––––––––
Beaten Beaten him Beaten Beaten him Beaten Beaten him
husband in the past husband in the past husband in the past
Country ever 12 months ever 12 months ever 12 months
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia 10.0 7.9 2.1 1.9 3.5 2.9
1
Colombia 13.4 u u u u u
Dominican Republic 29.3 16.1 8.5 3.7 13.1 6.5
Haiti 14.5 11.9 1.0 0.8 4.9 4.0
1
Nicaragua 15.1 u u u u u
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
1
The question used was “Do you remember if you have ever been the one to hit first?” and is asked
only of women who reported in the modified CTS that they had experienced violence from their
husband.
u = Unknown (not available)

There are only three countries (Cambodia, the Dominican Republic, and Haiti)
where data are available for all ever-married women on violence by women against
their husbands. Additionally, for Colombia and Nicaragua, the information is avail-
able only for women who have themselves experienced spousal violence.
In Cambodia and Haiti, 4 to 5 percent of ever-married women report having
physically abused their husband/partner at some time. In the Dominican Republic,
this proportion is higher at 13 percent. Since estimates of spousal abuse by women
obtained from the women themselves may underestimate the extent of such violence,
it is reassuring that for at least one of these countries, an independent comparator
based on men’s reports of such violence is available. The Cambodia 1996 Household

2 • Prevalence of Different Types of Domestic Violence 19

 Survey on Domestic Violence done under the aegis of the Ministry of Women’s Af-
fairs and Project Against Domestic Violence of Cambodia (Ministry of Women’s
Affairs and Project Against Violence, 1996) asked men if they had experienced vio-
lence by their wives. The proportion of men reporting such violence is 3 percent; this
estimate compares favorably with the 4 percent reported by women in the Cambodia
DHS survey.
In all countries, women who are themselves abused are more likely to report hav-
ing abused their partners. For example, in the Dominican Republic, 29 percent of the
women who have experienced spousal violence say they have beaten their spouse,
compared with 9 percent of women who have never experienced violence. In the re-
maining two countries where this comparison is possible, the differentials are simi-
larly large. In Cambodia, 10 percent of women who have themselves been abused
have abused their husbands, compared with 2 percent of women who have never been
abused; in Haiti, the corresponding proportions are 15 and 1 percent, respectively.
In Colombia and Nicaragua, data were obtained on women’s perpetration of vio-
lence only for women who were abused. About one in eight women who have them-
selves been abused have abused their husbands in these countries. Figure 2.3 high-
lights the fact that in most countries, the proportion of ever-married women who
have physically abused their husbands is only a fraction of those who have themselves
been abused. Even in the Dominican Republic, the proportion of ever-married
women who have abused their husbands is far less than those who have themselves
been abused.

Figure 2.3
Percentages of ever-married women who have experienced spousal violence
and not physically abused their husbands, percentage who have experienced
spousal violence and have physically abused their husbands, and percentage
who have physically abused their husbands but have not themselves
experienced physical violence

Cambodia

Colombia

Dominican Republic

Haiti

Nicaragua

0 10 20 30 40 50

Have experienced spousal violence and not physically abused their husband
Have experienced spousal violence and physically abused their husband
Have not expereienced spousal violence but have abused their husband
Note: In Colombia, the percentage of women who hve physically abused their
husband but have not experienced spousal violence was not assessed.

20 Prevalence of Different Types of Domestic Violence•2

 Together these data suggest that a large part of the reported violence by women
against their spouses is occurring in relationships where violence is being used by
both spouses. Whether this use of violence by either spouse is in self-defense cannot
be determined from these data, however.

2.3 Health Consequences of Violence

In addition to the CTS questions about different acts perpetrated by the hus-
band/partner, ever-married women were also asked whether they had ever experi-
enced specific health-related outcomes because of something the husband/partner
may have done. The purpose of these questions (which follow the modified CTS
questions in the module) is multifold. For women who have already provided infor-
mation on violence, they provide information on health consequences that could be
related to the violence. For women who have not yet reported spousal violence, they
provide an alternative way of reporting the experience of violence. For all women,
they provide another opportunity for disclosure. However, in Colombia and the Do-
minican Republic, these questions were asked only of women who had already dis-
closed violence. In Egypt too, where the domestic violence module was not used, the
question on possible health outcomes was directly linked to women’s reporting of
violence.
Table 2.8 shows the percentages of women reporting different types of physical
health consequences and visits to health facilities resulting from something done by
the husband, according to whether women reported any violence or not on other
items of the physical and sexual part of the CTS. The table shows that a large pro-
portion of women who experience violence have sustained some injuries, particularly

Table 2.8 Percentage of ever-married women age 15-49 who report various types of health outcomes as a consequence of acts
carried out by their husbands or partners, by whether the respondent ever experienced spousal violence
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Health outcome
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Had bruises Had injury or Had to go to a Had at
1
Violence status and aches broken bone health facility Other least one Had none
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Reported violence 36.5 6.5 6.3 u 38.0 62.0
Did not report violence 0.2 0.0 0.1 u 0.2 99.8
1,2
Colombia
Reported violence 53.3 10.2 27.5 Pregnancy aborted: 2.5 54.1 45.9
Had loss of function: 2.1
2
Dominican Republic
Reported violence 47.5 12.9 20.7 u 50.0 50.0
1,2
Egypt
Reported violence 18.0 u u Needed medical attention: 10.2 19.0 81.0

Haiti
Reported violence 15.5 7.7 9.2 u 18.9 81.1
Did not report violence 0.1 0.1 0.0 u 0.2 99.8
2
Nicaragua
Reported violence 22.7 4.9 4.9 u 23.8 76.2
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
1
In Colombia, women were asked whether, as a consequence of something their husband/partner did, the woman lost, temporarily or
permanently, an organ, a physical function, or part of the body. In Egypt, women were asked whether they were hurt during a beating
such that they needed medical attention, whether they got it or not.
2
In Colombia, Egypt, the Dominican Republic, and Nicaragua, questions on health outcomes were asked only of women who
reported experiencing one of the acts asked about in the modified CTS, rather than all women eligible for the domestic violence
module
u = Unknown (not available)

2 • Prevalence of Different Types of Domestic Violence 21

 bruises and aches. In Colombia, more than half of the women who reported experi-
encing violence reported that they had bruises and aches. This proportion is 48 per-
cent in the Dominican Republic, 37 percent in Cambodia, 23 percent in Nicaragua,
and 16 to 18 percent in Egypt and Haiti. Between 5 and 13 percent of women who
had experienced violence in the six countries report having an injury or broken bone.
In Colombia, 28 percent of women report having had to go to a health facility as a
consequence of something the husband had done to her; this proportion is 21 percent
in the Dominican Republic and 9 percent or less in the remaining countries.
There are only two countries, Cambodia and Haiti, where data on injuries are
available for both the women who had and who had not reported violence in earlier
questions. These data clearly show that the types of injuries asked about are common
only among women who have also reported experiencing spousal violence. Overall,
these data emphasize that spousal violence directly causes severe negative health con-
sequences for a significant proportion of women who suffer such abuse.

2.4 Initiation and Frequency of Spousal Violence

In order to understand the nature and causes of violence, it is important to also
document what is known about the initiation of spousal violence. Table 2.9 shows
the percent distribution of ever-married women who report spousal violence, by when
in their marriage they said the abuse began, for the four countries for which these
data are available. The data are shown according to marital duration to minimize
truncation and censoring; nonetheless, caution should be exercised in interpreting the

Table 2.9 Percent distribution of ever-married women age 15-49 who have ever experienced spousal
violence by when in their marriage/partnership the violence first began, according to marital duration
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Years after marriage
––––––––––––––––––––––––––––––––––––––– Since
Duration Before 10 years divorce/
of marriage marriage 0-2 years 3-4 years 5-9 years or more separation Total
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
0-4 (0.0) (89.2) (8.1) na na (2.7) 100.0
5-9 0.0 32.1 37.0 28.4 na 2.5 100.0
10-14 2.0 38.8 27.6 20.4 9.2 2.0 100.0
15+ 2.2 23.1 12.1 24.7 36.3 1.6 100.0
Total 1.5 34.9 20.5 21.7 19.7 1.6 100.0
Colombia
0-4 6.8 86.7 6.6 na na na 100.0
5-9 2.3 65.8 16.7 15.1 na na 100.0
10-14 1.4 59.2 17.6 17.1 4.7 na 100.0
15+ 2.1 54.6 13.9 15.3 14.0 na 100.0
Total 2.7 62.7 13.8 13.4 7.4 na 100.0
Dominican Republic
0-4 3.1 87.0 7.4 na na 2.5 100.0
5-9 0.6 70.4 19.6 9.3 na 0.0 100.0
10-14 1.1 49.0 18.0 21.8 9.2 0.8 100.0
15+ 1.3 47.0 11.9 18.9 20.4 0.6 100.0
Total 1.4 58.2 14.3 14.9 10.6 0.7 100.0
Haiti
0-4 4.4 90.4 4.4 na na 0.7 100.0
5-9 3.4 52.9 39.5 4.2 na 0.0 100.0
10-14 2.1 54.3 25.5 13.8 4.3 0.0 100.0
15+ 2.2 41.4 32.0 14.7 9.4 0.4 100.0
Total 2.8 56.1 26.2 9.8 4.7 0.3 100.0
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
na = Not applicable
u = Unknown (not available)

22 Prevalence of Different Types of Domestic Violence•2

results because these problems are not eliminated, especially for women who have
been married less than five years. Caution is also needed because these data are de-
pendent on women’s recall of the first violent event in relation to the beginning of
their marriage; such recall may vary by the length of the marriage.
Table 2.9 shows that in all countries except Cambodia, for women who had ex-
perienced violence by the time of the survey, violence is most likely to have started
within two years of the marriage at all marital durations. Further, in these countries,
the vast majority of women (70 percent or more) at most marital durations have ex-
perienced violence in the first five years of marriage. In Cambodia, the timing of the
start of violence varies greatly by marital duration; however, even here, over two-
thirds of women who have been married less than 15 years have experienced the first
episode of violence within five years of marriage.
Another aspect of violence that needs examination is the frequency with which it
takes place. Table 2.10 shows the frequency of spousal violence in the 12 months pre-
ceding the survey for ever-married women who report any spousal violence. This in-
formation is not available for Colombia and Peru.

Table 2.10 Among ever-married women age 15-49 who report ever experiencing violence by a
spouse or partner, the frequency of violence in the 12 months preceding the survey
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Times experienced violence
in the past 12 months Number of
–––––––––––––––––––––––––––––––––––––– ever-married
Frequently Not in women who have
(5 or more Sometimes the past ever experienced
Country times) (1-4 times) 12 months spousal violence
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia 36.0 50.7 12.0 420
1
Dominican Republic 42.3 5.7 52.0 1,519
2
Egypt 9.1 35.4 54.6 2,451
Haiti 41.8 29.9 27.8 676
3
India 14.4 40.1 44.8 17,102
Nicaragua 29.3 10.8 57.9 2,570
Zambia 4.3 41.8 53.9 1,836
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
1
For the Dominican Republic, frequency is non-numeric and is instead reported as “frequently,”
“sometimes,” or not at all in the past year. This question was asked only of women who reported
violence in the modified CTS.
2
For Egypt, “frequently” is defined as six or more times, and “sometimes” is defined as one to
five times in the past year.
3
For India, frequency is non-numeric and is instead reported as “many times” or “sometimes.”
Furthermore, in India, it is not possible to be certain that women are reporting on violence
experienced from the husband in the past year.

Among women who report any spousal violence, 42 percent report experiencing
frequent violence in the past one year in the Dominican Republic and Haiti, 36 per-
cent report frequent violence in Cambodia, 29 percent do so in Nicaragua, and 14
percent in India. Only in Egypt and Zambia is this proportion below 10 percent. In
addition, it is also clear from Table 2.10 that among women who have experienced
violence in the 12 months preceding the survey, frequent violence is more common
among women in the Dominican Republic, Haiti, and Nicaragua than infrequent
violence, whereas the opposite is true in the remaining countries.

2.5 Help Seeking

In this final section of the chapter, data are presented on help-seeking behavior of
women who have ever experienced any violence by anyone: whether they seek help,

2 • Prevalence of Different Types of Domestic Violence 23

 from whom they seek help, and if they do not seek help, the reasons they give for not
doing so. Seeking help is loosely defined to also include talking about the abuse to
someone. Specifically, Table 2.11 shows, for women who have ever experienced vio-
lence by anyone, the percentages who have never sought help for the problem and
who have sought help from different sources. Women who said that they did seek
help from some source could specify one or more sources: thus, the percentages will
not add to 100 percent. Table 2.12 shows the percent distribution of women who did
not seek help for the violence they experienced by the main reason for not doing so.

Table 2.11 Among women who have ever experienced violence by anyone, percentage who never sought help from anyone and percentage who
sought help from specific sources, by source(s) from which help was sought
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
1
Sought help from
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Did Women’s
not Doctor/ organi- Other Number
seek Own Friends/ Husband/ Lawyer/ health zations/ organi- of
Country help family In-laws neighbor boyfriend Police courts center NGOs zations Other women
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia 77.5 14.1 2.6 9.9 u 0.2 0.0 0.2 u 0.0 3.8 504
Colombia 62.0 25.6 4.7 10.1 1.2 15.6 3.2 0.2 u 4.5 2.5 4,710
Dominican Republic 58.8 21.7 6.5 15.7 0.5 14.3 1.6 0.0 0.3 3.4 2.2 1,922
Egypt 52.8 43.6 u 2.9 0.4 u u 0.0 u u 3.5 2,491
Haiti 68.7 19.7 3.0 8.3 0.4 1.6 1.0 1.0 u u 1.8 1,120
Nicaragua 40.5 33.6 7.5 26.3 u 12.8 2.8 6.7 3.3 u 7.2 2,822
Peru 57.8 32.8 4.7 5.6 0.9 15.0 3.8 0.6 u 3.0 2.2 12,883
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
1
Respondents could specify multiple sources of help.
2
Excludes women with information missing on help-seeking behavior.
NGO = Nongovernmental organization
u = Unknown (not available)

As is clear from Table 2.11, in most countries, the majority of women do not seek
help. Those not seeking help among women who have ever experienced violence
ranges from 41 percent in Nicaragua to 78 percent in Cambodia. Most women who
do seek help do so from their own families. Friends and/or neighbors are also an im-
portant source for help in Cambodia, Colombia, the Dominican Republic, Haiti and
Nicaragua. Institutions that generally have in their mandate the provision of assis-
tance to abused women are rarely used. The police (and similar authorities charged
with these duties across countries) are used by 13 to 16 percent of abused women
only in Colombia, the Dominican Republic, Nicaragua, and Peru.
The reasons given by abused women for not seeking help are shown in Table 2.12.
In the Dominican Republic, Egypt, and Haiti, about half the women say that they
did not seek help because it is “no use.” In Nicaragua, 41 percent of women say that it
is not necessary. In Cambodia, the most common reason given is that the respondent
was embarrassed about the abuse. This reason is also quite common in the remaining
countries, with 10 percent of women in Egypt to 21 percent in Haiti mentioning it.
Being afraid of further beatings is rarely mentioned by women in Cambodia and
Egypt, but in Nicaragua, 18 percent of women give this as the main reason for not
seeking help. This reason is also relatively common in the Dominican Republic (8
percent) and Haiti (9 percent). In Cambodia, over one in ten women who have not
sought help say that it is because they do not know where or to whom to go.

24 Prevalence of Different Types of Domestic Violence•2

 Table 2.12 Percent distribution of women who experienced violence by anyone and did not seek help by reason
for not seeking help
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Reason for Dominican
not seeking help Cambodia Republic Egypt Haiti Nicaragua
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Don't know whom to go to 12.3 9.0 8.7 7.2 2.1
1
No use 18.2 48.0 51.2 51.7 7.3
2
Part of life 4.7 4.4 7.7 7.9 40.6
Afraid of divorce 3.6 3.9 0.7 0.3 0.0
3
Afraid of further beatings 3.7 7.7 1.2 9.3 17.5
Afraid of getting person into trouble 1.5 8.1 11.2 2.8 0.0
Embarrassed 48.4 16.2 10.0 20.5 18.9
4
Other 7.5 2.7 9.2 0.1 13.5

Total 100.0 100.0 100.0 100.0 100.0

5 6
Number of women 362 1,542 1,315 667 2,330
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
1
The reason given was worded “thought wouldn't help.”
2
The reason given was worded “thought was not necessary.”
3
The reason given was worded “fear of husband.”
4
Includes “don't know.”
5
Does not include women who had missing information on this variable.
6
Includes women who may have talked about the violence with someone.

Overall, these data show that in most countries, women suffer abuse silently. They
tend not to seek help mainly because they think that the help will be of no use—they
think that it is part of life or they are embarrassed by the abuse. Further, institutional
help of any sort is rarely sought.

2 • Prevalence of Different Types of Domestic Violence 25

 3
Risk Factors for the Experience of Domestic Violence

T he factors and processes that contribute to the phenomenon of domestic violence

are not clearly understood. Further, the relationships between some background
characteristics and violence do not necessarily run in only one direction. For example,
while poverty is recognized by many to be a risk factor for domestic violence, violence
is also a risk factor for poverty since abuse can result in increased vulnerability to fal-
ling into poverty. By examining selected background characteristics of the individuals
and relationships affected by intimate partner violence, it is possible to begin to dis-
cern certain factors that are associated with an increased risk of experiencing domestic
violence.
This chapter first presents the results of bivariate analyses that show the character-
istics and context of violence in terms of women’s own characteristics, characteristics
of their husbands and the marital union, and characteristics of their household. In
addition, the intergenerational effects of violence are examined by exploring whether
women’s own risk of experiencing violence varies by their mother’s experience of
spousal violence. Two indicators of violence are examined: ever-experience of spousal
violence and experience of spousal violence in the past 12 months. Only ever-married
women are included in the analysis, where any woman who has lived with a man is
considered “married.” All data on the husband’s partner’s characteristics are obtained
through the reports of wives/partners. Finally, logistic regression is used to determine
the factors that have a consistently significant and direct effect on a married woman’s
risk of ever and current experience of violence across different countries.

3.1 Woman’s Characteristics

The variation in the percentages of women ever experiencing and recently experi-
encing spousal violence is examined for the following characteristics of women: cur-
rent marital status, current age, age at first marriage, number of children ever born,
education, and work status in the past 12 months.
Current marital status: The first panels in Tables 3.1.1 and 3.1.2 show how preva-
lence of ever-experience of violence and current experience of violence varies among
women who are currently married and have been married only once, women who are
currently married and have been married more than once, currently divorced or sepa-
rated women and currently widowed women. Given that spousal violence is a com-
mon reason for divorce, it is not surprising that in most countries, the highest rates of
the ever-experience of spousal violence are reported by women who are currently di-
vorced/separated or in a second or higher order marriage, and the lowest rates are
reported by women who are still married to their first husband/partner or who have
been widowed (Table 3.1.1).

3 • Risk Factors for the Experience of Domestic Violence 27

Table 3.1.1 Percentage of ever-married women age 15-49 who have ever experienced spousal violence, by background characteristics
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Background Dominican
characteristics Cambodia Colombia Republic Egypt Haiti India Nicaragua Peru Zambia
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Marital status
Currently married
Married only once 15.5 35.1 14.3 34.3 24.5 18.3 21.8 38.5 48.2
Married more than once 27.3 60.5 27.8 48.6 28.2 33.2 45.5 58.2 45.0
Divorced/separated 37.4 61.0 34.5 38.8 45.4 42.8 39.2 59.5 57.9
Widowed 17.5 55.0 19.6 24.1 46.3 16.2 20.4 50.7 41.2

Current age
15-19 4.0 38.5 19.6 28.7 25.9 13.0 26.6 30.9 38.4
20-24 13.7 43.4 25.7 34.1 33.2 17.1 26.6 37.3 49.3
25-29 21.4 42.9 24.5 34.4 25.2 20.6 29.1 41.3 53.2
30-34 19.1 43.8 23.2 37.1 31.4 21.5 32.4 43.1 48.5
35-39 18.3 45.3 21.7 36.3 27.4 20.5 32.5 44.8 46.4
40-44 12.7 43.3 23.3 33.2 22.0 19.4 33.2 45.2 50.0
45-49 22.1 48.0 15.7 31.7 36.2 17.1 30.0 44.3 44.0

Age at first marriage

<15 19.2 58.5 31.1 42.2 32.0 25.6 39.2 53.5 54.4
15-19 17.4 50.7 24.1 38.6 29.4 18.7 31.3 48.5 48.5
20-24 17.6 40.0 16.1 28.6 31.4 11.3 22.0 38.9 46.3
25+ 17.2 27.0 12.2 19.4 17.8 8.1 16.8 29.0 36.3

Number of children ever born

0 7.1 27.9 15.5 22.4 27.7 12.4 18.2 22.2 38.9
1-2 15.2 39.8 20.4 30.4 24.4 16.1 24.2 37.6 48.2
3-4 17.3 50.3 24.6 33.9 24.9 21.1 34.2 45.0 50.4
5+ 21.5 54.4 25.8 42.0 35.7 23.9 36.9 52.7 49.0

Education
No education 20.9 48.8 21.9 41.5 24.8 23.5 33.1 43.7 46.8
Primary 16.8 48.0 24.8 42.5 30.3 20.7 31.8 46.0 49.4
Secondary or higher 12.1 41.0 19.5 17.5 34.7 9.8 26.9 40.0 47.1

Work status
Not working 18.8 37.7 19.2 36.2 26.0 14.9 25.9 36.2 48.9
Working, paid in cash 18.2 43.4 25.7 21.0 30.5 26.7 35.4 46.2 49.5
Working, paid in kind 15.6 48.2 13.7 u * u u 45.6 (44.7)
Working, not paid 19.1 49.2 13.8 54.7 * 22.3 31.1 42.4 46.4
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Note: Figures in bold represent bivariate relationships that are not statistically significant based on the chi-square test (p>0.05). Figures in
parentheses are based on 25-49 unweighted cases. An asterisk indicates that a figure is based on fewer than 25 unweighted cases and has been
suppressed.
u = Unknown (not available)

Current spousal violence rates (experience of violence in the past 12 months) also
vary significantly with women’s marital status in all countries except Haiti. However,
the expectation that the end of a marriage will translate into an end of the risk of
spousal abuse is not borne out in every country. While widowed women in every
country, except Cambodia, have the lowest rates of current violence, currently di-
vorced women continue to have rates that are much higher than those for women in
their first marriage or widowed women. In most countries, women who are currently
in their second or higher order marriages have the highest rates of violence. Notably,
women who are in their second or higher order marriages are about 50 percent more
likely to report current violence than women in their first marriage in all countries
except Haiti and Zambia.
Age: A woman’s age is thought to affect the likelihood that she will experience
domestic violence. Researchers argue that as a woman ages, she often grows in social
status as she becomes not only a wife, but a mother, and perhaps a more economically
productive or socially influential member of her community; thus, older women are

28 Risk Factors for the Experience of Domestic Violence •3

less likely to report current experience of abuse than young women (Fernandez, 1997;
McClusky, 2001). Tables 3.1.1 and 3.1.2 show that both ever-experience and current
experience of violence do vary significantly by age in most, if not all countries; how-
ever, the pattern of variation differs substantially.
Ever-experience of violence is generally hypothesized to increase with age, since
the older an ever-married woman is, the longer has been her period of exposure to
the risk of violence. However, Table 3.1.1 does not support this expectation. Al-
though ever-experience of violence varies significantly with age, it does not increase
monotonically with age. In most countries, the rate of ever-experience of violence
fluctuates inconsistently within a narrow range with age. In Egypt and India, it first
rises, peaking for women age 30-34, and then falls. Nonetheless, in keeping with the
exposure argument, in all countries except the Dominican Republic, women in the
youngest age group have the lowest rates of ever-experience of violence.

Table 3.1.2 Percentage of ever-married women age 15-49 who experienced spousal violence in the past 12 months, by
background characteristics
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Background Dominican
characteristics Cambodia Republic Egypt Haiti India Nicaragua Zambia
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Marital status
Currently married:
Married only once 14.2 8.7 13.0 20.4 10.5 11.9 28.1
Married more than once 22.1 12.9 19.1 21.0 17.7 17.3 28.2
Divorced/separated 27.8 14.4 9.4 24.5 12.1 12.9 25.2
Widowed 14.7 2.0 0.5 16.4 2.8 4.4 7.8

Current age
15-19 4.0 15.4 21.0 25.4 10.4 18.2 33.3
20-24 12.2 16.7 18.8 31.4 11.4 15.7 35.3
25-29 19.1 13.4 14.1 19.4 12.3 13.9 29.7
30-34 16.8 11.3 12.9 26.2 11.5 13.8 24.2
35-39 16.8 9.6 12.6 22.4 9.9 10.9 19.8
40-44 10.9 5.4 8.2 13.0 7.8 11.5 16.6
45-49 18.1 5.5 4.5 12.5 5.9 6.7 15.8

Age at first marriage

< 15 16.7 17.3 13.3 28.8 13.6 16.5 29.4
15-19 15.0 11.7 13.8 20.9 10.5 14.0 26.2
20-24 15.8 7.5 11.6 22.8 5.7 9.5 25.9
25+ 15.8 4.4 8.2 12.2 4.2 6.2 21.9

Number of children ever born

0 5.9 9.9 10.5 23.3 8.0 11.1 28.0
1-2 13.7 10.0 15.4 21.7 9.4 12.0 32.9
3-4 15.4 12.2 12.2 20.3 11.0 14.6 28.4
5+ 18.6 11.3 10.9 20.3 12.1 13.7 19.4

Education
No education 18.0 9.8 14.1 18.0 13.3 13.9 27.2
Primary 15.1 13.0 15.0 23.5 9.9 13.6 26.0
Secondary or higher 10.5 8.9 8.3 21.9 5.1 12.4 27.2

Work status
Not working 17.4 10.2 13.3 21.8 8.7 12.4 29.5
Working, paid in cash 15.1 11.9 6.2 20.8 13.7 14.2 23.9
Working, paid in kind 14.0 7.8 u * u u (25.6)
Working, not paid 17.9 6.3 23.2 * 11.4 13.2 26.2
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Note: Figures in bold represent bivariate relationships that are not statistically significant based on the chi-square test
(p>0.05). Figures in parentheses are based on 25-49 unweighted cases. An asterisk indicates that a figure is based on
fewer than 25 unweighted cases and has been suppressed.
u = Unknown (not available)

3 • Risk Factors for the Experience of Domestic Violence 29

 By contrast, experience of violence in the past year is not affected by duration of
exposure (except for women who have been married for less than one year). Net of
the duration of exposure effect, the likelihood of experiencing violence tends to de-
cline with age in almost all countries (Table 3.1.2). Specifically, in the Dominican
Republic, Egypt, Nicaragua, and Zambia, current experience of violence is highest
for women in the two youngest age groups (ages 15-19 and 20-24). After age 24,
experience of violence in the past year generally declines with age in these countries,
with Haiti being the exception. In Cambodia, the relationship between age and re-
cent violence has an inverted U-shape, starting at 4 percent of women age 15-19 who
have experienced violence in the past year, then climbing to a high of 19 percent in
the 25-29 age group, followed by a decline to 11 percent among those age 40-44.
Then there is an unusual upswing at the oldest age group: 18 percent of women age
45-49 report having experienced domestic violence in the past year. In India, recent
experience of violence is relatively low only among women above age 34.
Overall, these data suggest that in most countries, younger women are more at risk
of being currently abused than older women. This is borne out not just by the current
abuse rates but also by the fact that rates of ever-experience of violence do not in-
crease consistently with age.
Overall, these data Age at first union: A woman’s young age at first union is generally thought to be
suggest that in most another risk factor for the experience of domestic violence. This expectation has both
countries, younger contextual- and individual-level explanations. At the contextual level, age at marriage
women are more at
risk of being currently
is a reflection of the status of women (Mason, 1987), a correlate of violence, with
abused than older very early marriages being more common in societies where women’s status is low. At
women. the individual level, a woman’s age at marriage is expected to be related to her risk of
experiencing violence, because when she marries at a very young age she has not been
given a chance to acquire the life skills and the maturity needed to ensure her self-
interest and security in marriage and within the spousal relationship.
The expectation that the experience of violence varies with age at marriage is sup-
ported by the data for most countries. As shown in Tables 3.1.1 and 3.1.2, in all
countries except Cambodia and, for current abuse, Zambia, the relationship between
violence and age at first marriage is significant and in the expected direction.
For virtually all countries, those who married at the youngest ages report the most
violence for both ever-experience of violence as well as that experienced in the past 12
months. Those who marry at 25 years of age or older consistently report the least
violence. For example, 42 percent of women in Egypt who married before age 15 re-
port having ever experienced violence, compared with less than half that proportion
among those who married at age 25 or older (19 percent). Similarly, in Colombia, 59
percent of women who married before age 15 report having ever experienced vio-
lence, while 27 percent of those who married at age 25 or older report having ever
experienced violence.
Number of children: Several studies indicate that the risk of experiencing violence is
positively associated with women’s number of children (e.g., Ellsberg, 2000; Martin
et al., 1999). What remains unclear is the direction of the relationship - whether in-
creased fertility leads to violence, or violence leads to increased fertility. The relation-
ship between violence and the number of children a woman has borne can be concep-
tualized such that when there are more children in a household, there is less income
per capita: insufficient resources may lead to exacerbated levels of stress for the head
of the household, which may lead to violence in some instances; hence, the more
children, the greater likelihood of violence (Martin et al., 1999). However, the rela-

30 Risk Factors for the Experience of Domestic Violence •3

tionship may work in the opposite direction. The existence of greater numbers of
children in a household is a result of, rather than a cause of, spousal violence, in that
women who are subject to partner violence may be less able to control their own
sexuality and fertility than women who are not subject to violence (Johnson, 2003).
In all countries except Haiti, women with no children have the lowest rates of
ever-experience of violence, and in most countries, women with five or more children
have the highest rate of ever-experience of violence (Table 3.1.1 and Figure 3.1).
Further, in most countries, the reported rate of violence increases fairly consistently
with the number of children. For example, in Peru, 22 percent of women who have
no children report ever experiencing violence, compared with 38 percent of women
with one or two children, 45 percent of women with three or four children, and 53
percent of women with five or more children. This pattern is similar to that in Cam-
bodia, Colombia, the Dominican Republic, Egypt, India, and Nicaragua. The bivari-
ate relationship between violence and parity is more consistent for women’s ever-
experience of violence than for their recent experience of violence (Table 3.1.2).

Figure 3.1
Percentage of ever-married women age 15-49 who have
ever experienced spousal violence, by number of children
ever born
60

50

40

30

20

10

0
Cambodia Colombia Dominican Egypt Haiti India Nicaragua Peru Zambia
Republic
0 1-2 3-4 5+

Education: Education has been shown to be a source of empowerment for women,

facilitating their ability to “gather and assimilate information, manipulate and control
the modern world, and interact effectively with modern institutions” (Kishor, 2000;
however, see Malhotra and Mather, 1997). It is hypothesized that women with more
education have greater resources to draw upon in times of need, such as when dealing
with a violent partner. Thus, it is expected that women with more education experi-
ence less violence.
In Cambodia, Colombia, India, and Nicaragua, the relationship between ever-
experience of violence and education is negative and monotonic: the more education
a woman has, the less likely she is to report having ever experienced violence. For

3 • Risk Factors for the Experience of Domestic Violence 31

 example, in Cambodia, 21 percent of women who have no education report having
ever experienced violence, compared with 17 percent of those with primary education
and 12 percent of those with secondary or higher education. In the Dominican Re-
public, Egypt, Peru, and Zambia, the highest rates of violence are found among
women with primary education and the lowest rates are found among women with
secondary or higher education. In Haiti, by contrast, education is positively related to
the ever-experience of violence: the more education a Haitian woman has, the more
likely she is to report that she has ever experienced violence.
Despite the varying patterns across countries in the relationship between education
and the ever-experience of violence, it is worth noting that the differentials by educa-
tion are never very large. In two countries, the rates for women with secondary or
higher education are either higher than those for women with no education (Haiti) or
are the same (Zambia), whereas in four countries (Colombia, the Dominican Repub-
lic, Nicaragua, and Peru), the rate for women with secondary or higher education is
at least 72 percent of the rate for women with no education. Only in Egypt and India
are the rates of ever-experience of violence among the most educated women less
than half the rate for women with no education.
In most countries, variation in rates of recent violence by education is similar to
that observed for the ever-experience of violence: in general, as the level of education
increases, the likelihood that a woman will report that she has experienced violence in
the past year decreases. Haiti again proves to have an unusual relationship between
recent violence and education, with women with the least education also reporting
the least violence, while recent violence does not vary significantly with education in
Nicaragua and Zambia.
Work status: Similarly to education, women who are engaged in paid employment
are hypothesized to have more say over financial and other household matters than
women who are not active in the labor market (Malhotra and Mather, 1997; see dis-
cussion in García, 2000). However, it is interesting to note that frequently, women
engaged in paid employment are more likely to be subjected to domestic violence
than those who are not in the labor force. However, it is interesting to note that fre-
quently, women engaged in paid employment are more likely to be subjected to do-
mestic violence than those who are not in the labor force.
In Colombia, the Dominican Republic, Haiti, India, Nicaragua, and Peru, women
who report that they are currently working and earning cash also report significantly
higher levels of ever-experience of violence than do women who are not currently
working. For example, in Peru, 46 percent of working women earning cash report
having ever experienced domestic violence, while 36 percent of nonworking women
report the same. Only in Egypt are women in paid employment significantly less
likely to have ever experienced violence than those who do not work: 36 percent of
women not working report having ever experienced violence, compared with 21 per-
cent of those who do work for cash. There is no consistent relationship across coun-
tries between violence and employment by type of payment.
The relationship between work status and violence in the past 12 months is similar
to the one noted in Table 3.1.1 for Egypt (where women working for cash experience
less violence) and India (where working women experience more violence). The rela-
tionship is no longer significant for the Dominican Republic, Haiti, and Nicaragua.
For Zambia, where ever-experience of violence was not significantly related to
women’s work status, the relationship of current violence with work is significant and

32 Risk Factors for the Experience of Domestic Violence •3

negative (working women are somewhat less likely than women who do not work to
experience current violence).

3.2 Husband’s/partner’s Characteristics

To fully understand spousal violence, the characteristics of the husband or partner
who is the alleged perpetrator of the violence also need to be examined. Accordingly,
this section discusses how women’s experience of violence, ever (Table 3.2.1) and in
the past 12 months (Table 3.2.2), varies by their partner’s education, occupation, and
4
alcohol consumption.

Table 3.2.1 Percentage of ever-married women age 15-49 who have ever experienced spousal violence, by husband’s characteristics
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Husband’s/partner’s Dominican
characteristics Cambodia Colombia Republic Egypt Haiti India Nicaragua Peru Zambia
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Education
No education 24.8 50.7 29.7 42.1 21.7 25.8 32.2 39.5 43.2
Primary 18.4 46.7 24.0 42.2 31.3 23.4 31.9 46.1 48.8
Secondary or higher 12.4 41.4 18.8 23.2 32.7 13.6 26.9 40.9 49.3
Don't know/missing 11.4 59.7 26.2 * 24.5 18.2 36.5 46.7 37.9

Occupation
Agriculture 16.7 44.7 20.7 36.7 31.0 21.5 27.0 40.7 46.7
Nonagricultural 19.2 41.9 22.6 33.4 28.6 17.2 31.7 43.1 49.9

Alcohol use
Doesn’t drink 12.6 u 13.7 u 26.5 u u u u
Never gets drunk 10.5 31.2 16.1 u 20.1 u 22.8 28.1 u
Drunk occasionally 14.1 42.2 24.1 u 35.9 u 28.9 43.0 u
Drunk frequently 48.6 70.3 54.0 u 71.3 u 47.0 78.7 u
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Note: Figures in bold represent bivariate relationships that are not statistically significant based on the chi-square test (p>0.05). Figures in
parentheses are based on 25-49 unweighted cases. An asterisk indicates that a figure is based on fewer than 25 unweighted cases and has been
suppressed.
u = Unknown (not available)

Education: Women whose husbands have secondary or higher levels of education

have lower rates of ever-experience of violence in most countries. The exceptions are
Haiti, Peru, and Zambia (Table 3.2.1). In countries such as Cambodia, Colombia,
the Dominican Republic, India, and Nicaragua, the relationship between education
and violence is negative and monotonic. For example, in Cambodia, 25 percent of
women whose husbands have no education report having ever experienced violence,
compared with 18 percent of women whose husbands have only primary education
and 12 percent of women whose husbands have secondary or higher education. Haiti
again is an unusual case, in that the relationship between men’s education and vio-
lence is positive and monotonic, echoing the same relationship indicated between
Haitian women’s education and experience of violence.
Having a husband with secondary or higher education is also less likely to be asso-
ciated with a woman’s experience of violence in the past year for most countries, the
exceptions being Haiti and Zambia (Table 3.2.2).

4
For a small proportion of women who are in second or higher order nonabusive marriages,
the reported characteristics may not be the characteristics of the husband that abused them.
This is because in the DHS, information on husband’s characteristics is available only for a
woman’s current or most recent husband.

3 • Risk Factors for the Experience of Domestic Violence 33

 Table 3.2.2 Percentage of ever-married women age 15-49 who experienced spousal violence in the past 12 months, by
husband’s characteristics
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Husband’s/partner’s Dominican
characteristics Cambodia Republic Egypt Haiti India Nicaragua Zambia
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Education
No education 22.0 18.4 13.7 17.2 14.7 13.0 24.3
Primary 16.2 12.7 15.4 22.9 12.1 14.6 28.1
Secondary or higher 10.7 8.4 9.7 22.3 7.3 11.9 25.7
Don't know/missing 11.4 9.2 * 18.9 11.8 11.2 13.6

Occupation
Agriculture 14.9 9.0 12.5 28.3 11.9 11.1 26.6
Non-agricultural 16.7 11.3 12.4 20.5 9.3 14.1 26.2

Alcohol use
Doesn't drink 11.5 5.8 u 20.3 u u u
Never gets drunk 8.3 5.1 u 15.1 u 8.1 u
Drunk occasionally 11.9 12.3 u 29.1 u 14.4 u
Drunk frequently 43.9 34.3 u 26.7 u 29.9 u
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Note: Figures in bold represent bivariate relationships that are not statistically significant based on the chi-square test
(p>0.05). Figures in parentheses are based on 25-49 unweighted cases. An asterisk indicates that a figure is based on
fewer than 25 unweighted cases and has been suppressed.
u = Unknown (not available)

Overall, in most countries, the pattern of variation in women’s experience of vio-

lence by their husband’s education is similar to that by their own education. This is
not surprising since a man’s level of education is unlikely to be independent of the
education of his wife: Becker’s (1973) theory of positive assortative mating asserts
that people tend to marry those to whom they are similar, particularly on such di-
mensions as education.
Occupation: Some of the literature indicates that in developing societies where agri-
cultural land is inherited exclusively by sons, women are more likely to be culturally
devalued (e.g., Dyson and Moore, 1983; Miller, 1981), and hence at a higher risk of
violence. Accordingly, in Tables 3.2.1 and 3.2.2, women’s experience of violence is
examined according to whether their husbands are employed in agriculture or not.
The tables show that the relationship between a husband’s occupation and domestic
violence is inconsistent across countries.
Specifically, women’s ever-experience of violence does not vary significantly by her
husband’s occupation in Cambodia, the Dominican Republic, or Haiti. In Nicaragua,
Peru, and Zambia, husbands in nonagricultural occupations were significantly more
likely to be reported as having ever been violent; however, the differentials in the
rates are not larger than about five percentage points in any country. For example, in
Zambia, 50 percent of women with husbands in a nonagricultural occupation re-
ported that they had ever experienced spousal violence, while 47 percent of women
with husbands in agricultural occupations reported that they had ever experienced
violence. Egypt, India, and Colombia are the only countries where there is some evi-
dence that women whose husbands are in agricultural occupations have experienced
higher rates of violence. Notably, however, even in these countries, the differential by
husband’s occupation is small. In Egypt, 37 percent of women whose husbands were
in agriculture had ever experienced violence, as compared to 33 percent of women
whose husbands were in nonagricultural occupations; in India, 22 percent of women
whose husbands were in agriculture had ever experienced violence, as compared to 17

34 Risk Factors for the Experience of Domestic Violence •3

percent of women whose husbands were in nonagricultural occupations; and in Co-
lombia, the corresponding figures were 45 and 42 percent, respectively.
Concerning women’s experience of violence in the past year, four countries showed
a significant relationship between violence and husband’s occupation: in Haiti and
India, women with husbands in an agricultural occupation reported violence at a sig-
nificantly higher rate than women whose husbands were not in agricultural occupa-
tions; in Nicaragua and the Dominican Republic, the reverse was true. The relation-
ship between occupation and experience of violence in the past year was not
significant in Cambodia, Egypt, or Zambia.
Alcohol consumption: Of all measurable variables hypothesized to influence the like-
lihood of domestic violence, a partner’s habitual drunkenness has one of the strogest,
most consistent relationships to the phenomenon (c.f. Coker et al., 2000; Johnson,
2003).
The results from the Demographic and Health Surveys (DHS) questionnaires re-
inforce this body of literature: the relationship between the experience of violence by
wives and the frequency of drunkenness among men who consume alcohol is positive,
monotonic, and highly significant in all countries where data on drunkenness are
available (Tables 3.2.1 and 3.2.2). Women who report that their husbands come
home drunk frequently are the most likely to report having ever experienced domestic
violence. Women who report that their husbands come home drunk frequently are
the most likely to report having ever experienced domestic violence.
For example, in Cambodia, about 11 to 13 percent of women whose husbands
never come home drunk (either because they abstain from drinking alcohol or be-
cause they do not get drunk) report having ever experienced violence, whereas almost
half (49 percent) of the women whose husbands come home drunk frequently report
ever-experience of violence. Similarly, in Colombia, 31 percent of women whose
husbands do not come home drunk report having ever experienced violence, com-
pared with 70 percent of women whose husbands come home drunk frequently. No-
tably, among the six countries for which data on drunkenness are available, violence is
between two and five times more common among women whose husbands get drunk
frequently than among women whose husbands never get drunk.
This relationship generally remains consistent and strong when considering
women’s experience of violence in the past year: the more frequently a husband
comes home drunk, the more likely a woman is to have experienced recent violence.
For example, in Nicaragua, 8 percent of women whose husbands do not get drunk
report having experienced violence in the past year, while 30 percent of women
whose husbands come home drunk frequently have experienced violence in the past
year.
There are three countries in which husbands who do not drink at all are distin-
guished from those who do drink, but never come home inebriated: Cambodia, the
Dominican Republic, and Haiti. In Haiti and Cambodia, women with husbands who
do not drink at all report levels of violence falling between those experienced by
women whose husbands drink but never come home drunk and those of women
whose husbands come home drunk on occasion. In the Dominican Republic, women
who report that their husbands do not drink at all report almost the same amount of
violence as those who report that their husband does not come home drunk.

3 • Risk Factors for the Experience of Domestic Violence 35

 3.3 Union Characteristics
Risk factors for women’s vulnerability to spousal violence include not only their
own characteristics and the characteristics of their husbands, but also how their own
characteristics compare with those of their husbands. Status inconsistency theory as
applied to the marital dyad suggests that when two people of different ascribed or
achieved statuses engage in a marital union, the result may be tensions that ultimately
result in marital dissatisfaction or dissolution (Mueller et al., 1979). The literature
suggests that where men are of higher educational status than women, thus having
both higher ascribed (on the basis of gender) and achieved (on the basis of higher
educational attainment) statuses, they are more likely to assert unequal, and even vio-
lent, power in the relationship (Hornung et al., 1981). Hornung et al. (1981) also
find that when women have greater achieved status than their husbands, there is an
increased risk of marital violence.
Thus, in Tables 3.3.1 and 3.3.2, the bivariate relationships between prevalence of
spousal violence against women, ever and in the past one year, and spousal age and
educational differences are examined. In addition, the variation in the prevalence of
violence by marital duration is also shown in these tables for women who are cur-
rently in their first marriage. Since information on duration of union is available only
for current unions, women who are not in their first union are not included in the last
panels of Tables 3.3.1 and 3.3.2.

Table 3.3.1 Percentage of ever-married women age 15-49 who have ever experienced spousal violence, by characteristics of the union
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Characteristics Dominican
of union Cambodia Colombia Republic Egypt Haiti India Nicaragua Peru Zambia
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Spousal age difference
Husband is
Younger 20.3 43.9 26.7 40.3 32.6 18.9 34.3 43.7 50.8
0-4 years older 16.0 38.2 18.5 35.7 26.8 16.8 25.6 39.6 51.9
5-9 years older 14.9 38.8 15.2 35.2 20.7 18.7 26.5 40.2 45.9
10-14 years older 12.2 40.6 19.4 31.8 27.6 20.7 32.4 37.7 43.9
15+ years older 12.5 36.1 18.0 36.9 23.6 23.3 28.2 39.8 44.2

Spousal educational difference

Husband has
Less education 22.9 42.6 28.3 33.6 28.6 16.7 31.0 46.7 42.8
Both have none 23.1 48.5 20.5 41.8 21.7 25.9 31.9 36.6 39.0
Same education 14.4 39.0 19.6 21.5 18.5 10.2 26.7 38.5 50.9
More education 16.0 45.4 22.8 34.9 34.5 17.5 29.9 43.0 49.9
1
Marital duration
0-4 years 11.3 27.3 11.0 22.7 23.5 10.1 15.8 26.5 38.2
5-9 years 17.0 35.8 16.6 38.9 28.8 18.4 22.5 36.8 53.7
10-14 years 18.6 37.4 13.5 39.2 24.1 21.4 23.4 42.2 52.9
15+ years 15.0 39.2 15.6 35.7 23.1 20.8 23.8 44.0 51.1
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Note: Figures in bold represent bivariate relationships that are not statistically significant based on the chi-square test (p>0.05).
1
Includes only currently married women who have been married only once.

Spousal age difference: Differences in spousal age, in which the husband is older
than the wife, are theorized to imply power imbalances in the relationship: because
age often confers seniority, ascribed power associated with age intersects with the
power associated with maleness in many cultures, such that a wife younger than her
husband may be at a comparative disadvantage. There is, however, little in the em-
powerment literature regarding the converse situation, when the wife is older than
the husband. Theoretically, one could argue that socially ascribed power increases
with age, regardless of gender; this may be true for most relationships but may not

36 Risk Factors for the Experience of Domestic Violence •3

apply to the relationship of a woman to her husband. In fact, it may be more likely
that because relationships in which women are older than their husbands are so con-
trary to the normative marital arrangement in most societies, they may be at greater
risk for marital discord.
As shown in Table 3.3.1, in six of the nine countries in which age difference be-
tween partners is significantly associated with the likelihood of having ever experi-
enced violence (Colombia, the Dominican Republic, Egypt, Haiti, Nicaragua, and
Peru), women who are older than their husbands are the most likely to report having
experienced violence from their partner. This finding is particularly striking in the
Dominican Republic, where 27 percent of women who are married to a younger man
report having ever experienced violence, compared to an average of 18 percent of
women who have married someone older than themselves (Table 3.3.1). Notably,
there is also no consistent relationship between ever-experience of spousal violence
and spousal age difference where the husband is older than the wife. India is the only
country where, among couples in which the husband is older, the rate of violence in-
creases steadily with spousal age difference.
Rates of recent violence also tend to be higher for women older or similar in age to
their husbands (Table 3.3.2). Overall, however, there is little variation in the recent
experience of violence by spousal age difference.

Table 3.3.2 Percentage of ever-married women age 15-49 who experienced spousal violence in the past 12 months, by
characteristics of the union
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Characteristics Dominican
of union Cambodia Republic Egypt Haiti India Nicaragua Zambia
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Spousal age difference
Husband is
Younger 18.3 13.5 13.8 22.7 12.0 13.9 33.9
0-4 yrs older 14.6 10.6 13.3 22.5 9.9 12.8 31.0
5-9 yrs older 13.5 8.4 14.1 16.8 10.8 13.1 27.7
10-14 yrs older 9.8 10.5 10.8 23.0 11.5 14.2 25.3
15+ yrs older 10.9 7.9 13.4 19.0 11.6 13.9 22.8

Spousal educational difference

Husband has
Less education 20.9 16.7 13.6 15.5 8.4 14.3 24.1
Both have none 19.6 11.4 12.9 16.8 15.0 13.4 22.0
Same education 13.5 13.0 9.6 15.3 5.2 9.7 30.7
More education 13.8 11.9 13.0 26.2 9.3 13.6 26.7
1
Marital duration
0-4 years 10.9 9.2 15.1 22.7 7.9 12.6 31.7
5-9 years 15.5 9.9 16.9 24.4 12.2 14.9 35.9
10-14 years 16.7 8.7 14.9 20.9 12.5 13.2 25.4
15+ years 13.7 7.5 9.4 16.0 10.2 8.8 20.0
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Note: Figures in bold represent bivariate relationships that are not statistically significant based on the chi-square test
(p>0.05).
1
Includes only currently married women who have been married only once.

Spousal educational difference: The acquisition of a formal education usually confers

many benefits upon those who hold it, such as functional literacy, access to improved
employment opportunities, inroads into selective collegial networks, and entrée into
more exclusive marriage markets. In short, formal education often provides opportu-
nities for both improved day-to-day life skills as well as improved social status within
the larger community. As such, education can be understood as a status indicator,
with those who have more of it deemed as more powerful or influential than those

3 • Risk Factors for the Experience of Domestic Violence 37

 who have less. The implication of status inconsistency theory is that those of equal
status are most likely to have a non-conflictual relationship.
Table 3.3.1 shows that women in relationships in which both partners are edu-
cated and have the same number of years of education are the least likely to report
ever experiencing partner violence. This is true in Cambodia, Colombia, the Domini-
can Republic, Egypt, Haiti, India, and Nicaragua. In Peru and Zambia, it is women
who share a lack of education with their husbands who are least likely to report ever
having experienced violence. Both of these findings indicate that status consistency
within the marital dyad may be associated with a decreased probability of experienc-
ing domestic violence, although this is clearly not an unqualified association, since
the data for Cambodia, Colombia, Egypt, India, and Nicaragua indicate that in these
countries, women are at greatest risk of violence when both partners are uneducated.
It is of further interest to note that in the Dominican Republic and Peru, it is the
women who have achieved a higher level of education than their husbands who are
most likely to report having ever experienced spousal violence. Results are generally
the same for experience of violence in the past year (Table 3.3.2). Notably, women
who are educated at all and have the same educational status as their partner have a
much lower prevalence of current violence in Egypt, India, and Nicaragua than other
women.
Marital duration: Marital duration is a better measure of exposure to the likelihood
of spousal violence than is woman’s age (shown in Tables 3.1.1 and 3.1.2); rates of
ever-experience of violence can be expected to rise with marital duration because a
longer marriage provides a longer period of exposure. However, marital duration can
also be considered a proxy for compatibility in a marriage, particularly in cultures
where divorce is legal and socially accepted. In such a case, the experience of violence,
both ever and current, is likely to be negatively associated with marital duration.
With regard to the relationship between marital duration and ever-experience of
violence, in every country where the relationship is significant (all countries except
Haiti [Table 3.3.1]), women in unions of the shortest duration (zero to four years)
are the least likely to report violence, while those in relationships that have lasted
longer than four years are significantly more likely to report having ever experienced
violence. For example, in Peru, 27 percent of those in a union of less than five years’
duration report having ever experienced violence, as compared to 37 percent of those
in unions five to nine years long, 42 percent in unions that have lasted 10 to 14 years,
and 44 percent in unions of 15 years or longer. This consistent relationship of ever-
experience of violence with duration of union would suggest that there is an exposure
effect.
However, the data also show a downturn in reporting of ever-experience of vio-
lence, as well as experience of violence in the past year (Table 3.3.2), among women
whose unions have lasted 15 years or longer. It is of interest to note that with regard
to experience of violence in the past year, in four of the seven countries (Egypt, Haiti,
Nicaragua, and Zambia), the women least likely to report recent violence are those
whose unions are of the longest duration, supporting the idea that the longevity of a
union reflects in part the compatibility of the couple.

3.4 Household Characteristics

An important aspect of the context of women’s lives is the characteristics of the
households in which they reside, including the location of the household (urban or

38 Risk Factors for the Experience of Domestic Violence •3

rural), the composition of the household (nuclear or not), and the wealth of the
household. A priori, urban residence and nuclear family composition are expected to
be positively associated, and wealth negatively associated, with the risk of violence.
Area of residence: The anonymity of urban living is generally believed to be associ-
ated with a higher risk of violence. Indeed, in six of the nine countries shown in Ta-
ble 3.4.1, women living in urban areas are significantly more likely to report having
ever experienced violence from their husband or partner than rural women, and only
in two countries (India and Egypt) are they significantly less likely to do so. In Cam-
bodia and Haiti, the ever-experience of violence does not vary by residence.

Table 3.4.1 Percentage of ever-married women age 15-49 who have ever experienced spousal violence, by household characteristics
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Household Dominican
characterisitcs Cambodia Colombia Republic Egypt Haiti India Nicaragua Peru Zambia
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Residence
Urban 16.2 45.7 23.3 29.4 28.7 14.4 32.3 43.4 53.4
Rural 17.7 39.2 20.3 38.8 28.9 20.6 26.9 40.6 45.4

Family structure
Nuclear 18.3 40.7 19.8 33.9 26.3 22.7 28.8 42.7 45.3
Nuclear (female headed) 26.0 67.2 30.3 28.3 32.2 23.3 44.9 60.9 50.0
Nonnuclear 14.7 43.4 23.3 35.9 32.2 16.3 29.2 40.0 50.6

Wealth quintile
Lowest 24.4 41.1 23.8 42.1 28.9 27.4 28.3 40.5 48.4
Second 18.6 46.3 24.6 43.8 26.6 24.0 31.9 46.2 42.8
Middle 15.1 51.2 25.1 40.5 35.2 20.3 33.8 49.2 45.5
Fourth 14.4 42.8 22.1 30.8 26.7 15.3 31.4 41.6 51.0
Highest 14.4 38.1 16.3 18.0 26.8 7.8 25.8 33.3 54.4
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Note: Figures in bold represent bivariate relationships that are not statistically significant based on the chi-square test (p>0.05).

Current experience of violence (Table 3.4.2) also varies as expected in the Domini-
can Republic, Nicaragua, and Zambia. In Egypt, India, and Haiti, rural women are
more likely to be currently abused than urban women. The case of Haiti is interesting
in that ever-experience of violence does not vary by residence, but current experience
does. Twenty-three percent of rural women report current spousal violence, com-
pared with only 18 percent of urban women.

Table 3.4.2 Percentage of ever-married women age 15-49 who experienced spousal violence in the past 12 months, by
household characteristics
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Household Dominican
characteristics Cambodia Republic Egypt Haiti India Nicaragua Zambia
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Residence
Urban 13.9 11.8 10.4 17.5 7.2 14.4 29.6
Rural 15.6 9.4 14.4 23.3 11.4 11.3 24.6

Family structure
Nuclear 16.4 11.6 12.8 19.2 13.1 13.7 27.3
Nuclear (female headed) 17.6 10.0 2.2 24.2 5.4 14.2 10.8
Non-nuclear 13.4 10.6 13.3 20.3 8.8 12.5 27.4

Wealth quintile
Lowest 22.2 12.7 16.9 20.3 16.4 12.4 28.3
Second 16.4 13.3 16.3 22.7 13.7 13.5 23.6
Middle 12.2 12.6 14.6 29.7 10.9 14.9 24.6
Fourth 13.4 10.1 10.7 18.9 7.4 15.1 26.9
Highest 12.5 6.7 5.6 14.0 3.4 10.0 28.8
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Note: Figures in bold represent bivariate relationships that are not statistically significant based on the chi-square test
(p>0.05).

3 • Risk Factors for the Experience of Domestic Violence 39

 Household structure: Kishor (2000) suggests that there exist variables that indicate
direct evidence of empowerment, those that indicate sources of empowerment, and
those that indicate settings in which empowerment might be expected to occur (or
not). Family structure can be considered a setting within which women are empow-
ered to act or are constrained from acting–possibly through the use of violence. Pre-
vious research has indicated that when a woman lives with her in-laws, especially in
highly patriarchal societies, she is at greater risk of subordination to her husband, as
well as other members of his family, particularly her mother-in-law. While most lit-
erature associates patrilocal extended family living arrangements with less autonomy
and empowerment for women, it may also be that women living within an extended
family receive a degree of protection from domestic violence, given the regular pres-
ence of other family members in the household.
For the purpose of this analysis, two categories of nuclear households have been
identified: those with a married couple living with or without children (nuclear) and
those with a woman heading the household living alone or with children (nuclear,
5
female headed). All other households are categorized as nonnuclear.
The results presented in Table 3.4.1 indicate that ever-experience of violence does
not vary between women living in nonnuclear and nuclear (married couple) house-
holds in Nicaragua. In Colombia, the Dominican Republic, Egypt, Haiti, and Zam-
bia, it is women who are in nonnuclear households that report higher levels of abuse
than women in nuclear households. Only in Cambodia, India, and Peru is the expec-
tation borne out that women who live in nuclear households have higher rates of vio-
lence.
Since household structure can change quickly, a more relevant analysis is with re-
cent experience of violence. Table 3.4.2, however, shows that in most countries, cur-
rent violence rates do not vary between residents of nuclear and nonnuclear house-
holds. The differential is relatively large and significant only in India; it is women in
nuclear households who have higher rates of current violence.
What is also notable is that in Cambodia, Colombia, the Dominican Republic,
Haiti, India, Nicaragua, and Peru, the ever-experience of violence is highest in nu-
clear households that are headed by a woman. What is also notable is that in Cam-
bodia, Colombia, the Dominican Republic, Haiti, India, Nicaragua, and Peru, the
ever-experience of violence is highest in nuclear households that are headed by a
woman.
However, in four countries (the Dominican Republic, Egypt, India, and Zambia),
current experience of violence is lowest among women living in female-headed nu-
clear households. This contrast suggests that women who have been in abusive rela-
tionships at some point do leave the household in which the abuse is taking place,
thus explaining why women currently on their own are more likely to report ever ex-
periencing violence but least likely to report current violence.
Relative wealth status: A common assumption in the literature on domestic vio-
lence is that women who are poor are more likely to experience violence than women
who are not poor (e.g. Jewkes, 2002; Heise, 1998). Poverty is not necessarily seen as a
causal factor, but is generally assumed to significantly increase the risk of domestic

5
Since it is not possible to unequivocally determine the relationships between household
members from the DHS data, no attempt is made here to divide nonnuclear households into
those in which the household includes the respondent’s in-laws and those in which the
respondent is co-resident with members of her natal family.

40 Risk Factors for the Experience of Domestic Violence •3

violence. However, the association between poverty and domestic violence is unlikely
to be entirely unidirectional: the perpetration and experience of such violence may
contribute to aggravating, perpetuating or even causing household poverty (Byrne,
Resnick, Kilpatrick, Best, and Saunders, 1999). Notably too, population-based em-
pirical research finds only mixed support for a consistently positive association be-
tween violence and poverty (Ellsberg, Pena et al., 1999; Johnson, 2003; Kishor and
Johnson, 2003).
Recent advances in the use of survey-based household assets data allow researchers
to reliably evaluate the distribution of poverty in populations (Filmer and Pritchett,
2001). This recently developed wealth index has been tested in a large number of
countries in relation to inequities in household income, use of health services, and
health outcomes (Rutstein, Gwatkin and Johnson, 2000). It is an indicator of wealth
that is consistent with expenditure and income measures (Rutstein, 1999). The
wealth index is constructed using household asset data (including country-specific
assets) and principal components analysis. The asset information was collected
through the DHS household questionnaire, and includes household ownership of a
number of consumer items ranging from a television to a bicycle or car, as well as
dwelling characteristics such as type of drinking water available, sanitation facilities
used, roofing and flooring.
Each asset is assigned a weight (factor score) generated through principal compo-
nents analysis, and the resulting asset scores are standardized in relation to a standard
normal distribution with a mean of zero and a standard deviation of one (Gwatkin,
Rutstein, Johnson, Pande and Wagstaff, 2000). Each household is then assigned a
score for each asset, and the scores are summed by household; individuals are ranked
according to the total score of the household in which they reside. The sample is
then divided into population quintiles: each quintile is designated a rank, from one
(poorest) to five (wealthiest).
Table 3.4.1 indicates that there are only three countries (Cambodia, Egypt and
India) of the nine included where there is a more or less steady negative relationship
of ever-experience of violence by household wealth status. In addition, in most of the
countries, women in the highest wealth quintile are also the least likely to report hav-
ing ever experienced spousal violence. However, it is not necessarily the case that
poor women are therefore most likely to experience violence; in fact, in the majority
of countries where the relationship between household wealth and ever-experience of
violence is significant, it takes the shape of an inverted U, with a peak in reporting of
violence in the third quintile; this is the case for Colombia, the Dominican Republic,
Haiti, Nicaragua and Peru. In Zambia, women in the wealthiest households are most
likely to report having ever been beaten by their husbands (Figure 3.2).
The bivariate relationship between wealth and experience of violence in the past
year is also inconsistent. Nonetheless, current violence rates are consistently lower
among women in wealthier households than poorer households in two of the seven
countries for which data on current violence are available. In addition, in Nicaragua
and the Dominican Republic, women in the wealthiest households have the lowest
rate of violence, although the relation between recent violence and wealth is non-
linear.

3 • Risk Factors for the Experience of Domestic Violence 41

 Figure 3.2
Percentage of ever-married women age 15-49 who have
experienced spousal violence in the preceding 12 months,
by relative household wealth status

35

30

25

20

15

10

0
Cambodia Dominican Egypt Haiti India Nicaragua
Republic
Lowest quintile Second quintile Middle quintile Fourth quintile Highest quintile

3.5 Intergenerational Effect

Research has documented an important negative effect of domestic violence on
children, even if the children are not themselves abused: male children who see their
mother being abused by their father are at a higher risk of becoming abusers in their
intimate relationships as adults, while female children are more likely to enter abusive
spousal relationships as adults (Kalmuss, 1984; Seltzer and Kalmuss, 1988). To ex-
amine whether this relationship is found across countries, Table 3.5 shows how
women’s own experience of violence, ever and in the past one year, varies by whether
their mother was abused by their father. The information on the mother’s experience
is reported by the respondent. Since a fairly significant proportion of respondents did
not know whether their mothers were beaten or not, the experience of violence by
women in the “don’t know” category is also reported in the table. The intergenera-
tional influence of domestic violence is reported only for six countries, since the in-
formation was not obtained in Egypt, India, and Zambia.
Table 3.5 clearly shows that in all countries, women who knew that their mothers
were abused by their fathers were much more likely to have ever experienced and to
be currently experiencing violence than women who replied “no” to the question
about the father’s abuse of the mother. The differentials are large and significant in
all countries. For example, in Cambodia, 30 percent of women whose fathers beat
their mothers have ever experienced violence and almost the same percentage (28
percent) report current violence, compared with 15 and 13 percent, respectively,
among those who said that their fathers did not beat their mothers.
In all countries, the prevalence of violence, ever and current, among women who
said that they did not know whether their father beat their mother lies between those
who said “yes” and those who said “no.” This is to be expected since for at least some
proportion of these women the answer is “yes.”

42 Risk Factors for the Experience of Domestic Violence •3

 Table 3.5 Percentages of ever-married women age 15-49 who experienced spousal violence ever and in the
past 12 months, by whether their mother was ever beaten by their father
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Dominican
Family history Cambodia Colombia Republic Haiti Nicaragua Peru
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
EVER EXPERIENCED SPOUSAL VIOLENCE
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Father beat mother
No 15.2 36.1 20.0 27.0 27.4 35.8
Yes 29.7 55.4 36.3 37.8 36.6 50.0
Don't know 20.7 46.5 27.9 32.1 35.4 46.3
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
EXPERIENCED VIOLENCE IN THE PAST 12 MONTHS
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Father beat mother
No 13.1 u 9.2 20.5 11.4 u
Yes 28.1 u 21.6 33.2 17.2 u
Don't know 17.7 u 18.2 22.8 15.5 u
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Note: All bivariate relationships are statistically significant based on the chi-square test (p<0.05).
u = Unknown (not available)

3.6 Multivariate Regression Analysis: Identifying the Major Risk

Factors for Domestic Violence
In order to identify the factors that significantly increase or decrease the risk of ex-
periencing violence, multivariate logistic analyses were conducted for all countries.
Factors considered include all the individual, husband/partner, union and household
characteristics discussed in the bivariate analysis. Two dependent variables are ana-
lyzed for each country: ever-experience of spousal violence and experience of spousal
violence in the 12 months preceding the survey. For each dependent variable, a re-
spondent is coded “1” if she has experienced violence and coded “0” otherwise. The
analysis is perforce restricted to currently married women age 15-49, because some of
the variables relevant for the multivariate analysis are available only for currently mar-
ried women (e.g., spousal age and spousal age difference) or are more appropriate for
only currently married women (duration of union). However, since the percent dis-
tributions of ever-married women and currently married women are very similar for
almost all indicators (see Appendix B), the multivariate and bivariate results remain
comparable and the multivariate results can be considered to be generally representa-
tive of the experience of ever-married women.
Tables 3.6.1 and 3.6.2 show the odds ratios calculated from the coefficients of the
logistic regressions for each of the two dependent variables, respectively. Each odds
ratio gives the increase or decrease in the odds of the event (ever-experience of vio-
lence or current experience of violence) occurring for a given value of the independent
variable as compared to the reference category. For example, an odds ratio of 1.6 in
Table 3.6.1 for the age category 20 to 24 says that the odds that a woman age 20 to
24 years has ever experienced violence are 60 percent higher than if she were only 15
to 19 years of age (the reference category). This multivariate analysis adds to the
bivariate discussion by identifying the factors that significantly affect the likelihood of
violence net of all other factors hypothesized as relevant. In addition to the regres-
sions reported in Tables 3.6.1 and 3.6.2, another set of logistic regressions was run
for the six countries for which data on mother’s abuse by the father were available.

3 • Risk Factors for the Experience of Domestic Violence 43

Table 3.6.1 Correlates of a currently married woman's likelihood of having ever experienced spousal violence: adjusted odds ratios estimated using
logistic regression
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Dominican
Characterisitcs Cambodia Colombia Republic Egypt Haiti India Nicaragua Peru Zambia
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Individual characteristics
Age (r: age group 15-19) ns ns * *** * *** ns ns ns
20-24 1.59 0.98 0.76 0.94 0.55* 1.13* 0.64** 0.94 1.17
25-29 2.06 0.81 0.67 0.85 0.34*** 1.12 0.53*** 1.03 1.31
30-34 1.78 0.77 0.50* 1.10 0.34*** 1.07 0.50*** 1.01 1.00
35-39 1.09 0.73 0.37*** 1.16 0.28*** 1.00 0.50** 0.97 0.89
40-44 0.77 0.59* 0.42** 1.25 0.22*** 0.93 0.50** 1.03 0.98
45-49 1.02 0.68 0.29*** 0.87 0.24*** 0.84 0.46** 1.07 0.78
Woman's age at marriage
(r: <15 years old) ns *** ns ns ns *** * *** ns
15-19 0.81 0.75* 0.92 0.89 1.01 0.84*** 0.91 0.84* 0.83
20-24 1.00 0.55*** 0.69* 0.83 0.97 0.65*** 0.75* 0.62*** 0.74
25+ 0.92 0.35*** 0.76 0.66* 0.90 0.53*** 0.57** 0.45*** 0.57*
Number of unions (r: one)
Two or more unions 2.04*** 1.35*** 1.90*** 1.34* 1.22 1.63*** 2.43*** 1.62*** 0.80**
Number of children ever born
(r: none) *** *** *** *** ns *** *** *** ns
1-2 1.74 1.25 1.57* 1.54*** 0.99 1.31*** 1.36 1.59*** 1.12
3-4 2.42* 1.58*** 2.11*** 1.53*** 1.25 1.39*** 1.80*** 1.82*** 1.01
5+ 4.34*** 1.77*** 1.94*** 1.66*** 1.29 1.43*** 2.11*** 2.19*** 1.02
Woman's education level (r: none) * ns ns *** * *** ns * ns
Primary 0.78 1.18 1.18 1.04 1.46* 1.05 0.93 1.22** 1.10
Secondary or higher 0.41* 1.00 1.33 0.57*** 1.18 0.74*** 0.95 1.15 0.90
Work status (r: not working) ns *** *** *** ns *** *** *** ns
Working, paid in cash 1.01 1.44*** 1.37*** 0.78* 1.10 1.57*** 1.41*** 1.32*** 1.11
Working, paid in kind 0.71 1.32 1.49 u 0.81 u u 1.17 0.84
Working, no pay 1.07 1.55*** 0.74 1.79*** 0.28 1.25*** 1.17 1.16** 1.06
Husband/partner's characteristics
Husband's education level
(r: none) ns ns ns *** ns *** *** ns ns
Primary 1.04 1.04 1.01 0.89 1.26 1.01 0.96 1.51* 0.91
Secondary or higher 0.79 0.94 0.93 0.63*** 1.25 0.83*** 0.69* 1.52* 0.87
Husband's occupation
(r: nonagriculture)
Agricultural occupation 0.69* 1.01 0.83 0.71*** 0.89 0.90*** 0.78** 0.82*** 1.09
Husband's drunkenness
(r: doesn't drink) *** *** *** u *** u *** *** u
Never comes home drunk 0.94 r 1.07 u 1.40 u r r u
Comes home drunk sometimes 1.12 1.74*** 1.69*** u 2.29*** u 1.47*** 1.95*** u
Comes home drunk frequently 6.06*** 4.94*** 6.39*** u 3.32*** u 2.79*** 8.52*** u
Union characteristics
Spousal age difference
(r: husband is younger) ns ns * ns ns ** * *** ns
Husband is 0-4 years older 0.83 0.90 0.83 1.04 0.81 0.96 0.92 0.85*** 0.81
Husband is 5-9 years older 0.84 0.81* 0.72* 0.90 0.68* 1.03 0.90 0.82*** 0.69
Husband is 10-14 years older 0.80 0.86 0.78 0.84 0.68 1.06 1.19 0.69*** 0.68
Husband is 15+ years older 0.85 0.72* 0.65** 0.87 0.60* 1.03 0.83 0.75*** 0.67
Spousal educational difference
(r: husband has less education) ns *** ns ns ns *** ns *** ns
Both have no education 0.67 1.29 0.77 0.96 1.20 0.89* 0.95 1.12 0.85
Both have same level of education 0.93 0.82* 0.95 0.96 0.86 0.78*** 0.94 0.78*** 1.08
Husband has more education 0.81 1.10 0.97 1.09 1.05 0.90* 1.00 0.82*** 1.08
Marital duration (r: 0-4 years) ns * ns *** ns *** ** *** **
5-9 years 0.64 1.38*** 1.28 1.59*** 1.08 1.50*** 1.58*** 1.41*** 1.54***
10-14 years 0.63 1.40* 1.29 1.40* 1.63 1.58*** 1.68*** 1.50*** 1.59**
15+ years 0.43 1.35 1.81* 0.96 1.45 1.61*** 1.69* 1.43*** 1.95**
Household characteristics
Residence (r: urban)
Rural 0.94 0.80* 0.91 0.96 0.62*** 0.76*** 0.89 0.77*** 0.72**
Family structure (r: nonnuclear)
Nuclear 1.32 1.22*** 1.05 1.05 1.19 1.17*** 0.97 1.07 0.98
Household wealth status
(r: poorest quintile) ns ** ns *** ns *** ns *** ns
Second quintile 0.85 1.16 0.96 1.07 1.12 0.87*** 1.05 1.08 0.77**
Middle quintile 0.66* 1.43*** 0.93 0.88 1.19 0.72*** 1.03 1.06 0.90
Fourth quintile 0.76 1.21 0.83 0.71** 0.82 0.54*** 1.02 0.87 0.93
Wealthiest quintile 0.77 1.08 0.72 0.51*** 0.86 0.30*** 0.89 0.63*** 1.11
Constant 0.23 0.80 0.18*** 0.55*** 0.39*** 0.22*** 0.42*** 0.81*** 0.67***
-2 log likelihood 1652.39 7150.02 4291.47 7761.40 2286.36 68785.08 6991.44 19409.17 4605.39
Nagelkerke R square 0.18 0.16 0.16 0.12 0.08 0.11 0.14 0.15 0.04
Number of women 2,024 5,935 4,795 6,435 2,017 84,202 6,484 15,653 3,409
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Note: Shading represents bivariate relationships that are not statistically significant based on the chi-square test (p>0.05).
r = Reference (omitted) category; ns = not significant; u = Unknown (not available)
***: p<.005; **: p<.01; *: p<.05

44 Risk Factors for the Experience of Domestic Violence •3

 Table 3.6.2 Correlates of a currently married woman's likelihood of having experienced spousal violence in the past 12
months: adjusted odds ratios estimated using logistic regression
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Dominican
Characterisitcs Cambodia Republic Egypt Haiti India Nicaragua Zambia
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Individual characteristics
Age (r: age group 15-19) ns ns *** *** *** *** ns
20-24 1.40 0.71 0.72 0.53* 1.02 0.53*** 1.03
25-29 1.84 0.55* 0.50*** 0.31*** 0.99 0.37*** 1.04
30-34 1.59 0.48* 0.56* 0.31*** 0.89 0.31*** 0.98
35-39 1.03 0.34** 0.57 0.25*** 0.76** 0.21*** 0.91
40-44 0.75 0.32** 0.42* 0.15*** 0.60*** 0.22*** 0.80
45-49 0.90 0.28*** 0.24*** 0.17*** 0.46*** 0.15*** 0.61
Woman's age at marriage
(r: <15 yrs old) ns ns ns ns *** ns *
15-19 0.81 0.93 1.01 1.00 0.90*** 1.03 0.72**
20-24 1.02 0.81 1.19 1.03 0.69*** 1.01 0.64**
25+ 0.96 0.64 1.04 0.84 0.66*** 0.89 0.47**
Number of unions (r: one)
Two or more unions 1.66* 1.43*** 1.39 1.13 1.38*** 1.40*** 1.14
Number of children ever born
(r: none) *** ** * ns *** *** ns
1-2 1.77 1.60* 1.56*** 0.95 1.23*** 1.37 1.04
3-4 2.74* 2.27*** 1.52* 1.24 1.29*** 1.78* 0.93
5+ 5.08*** 2.14** 1.67** 1.15 1.40*** 2.47*** 0.76
Woman's education level (r: none) ns ns *** ns *** ns ns
Primary 0.75 1.26 0.92 1.37 0.94 0.76 1.05
Secondary or higher 0.42* 1.29 0.53*** 1.13 0.70*** 0.69 0.97
Work status (r: not working) ns ** *** ns *** * ns
Working, paid in cash 0.96 1.41*** 0.62** 1.11 1.40*** 1.28** 0.96
Working, paid in kind 0.73 1.54 u 1.01 u u 0.94
Working, no pay 1.14 0.73 1.76*** 0.34 1.09* 1.06 1.08
Husband/partner's characteristics
Husband's education level (r: none) ns ns ns ns *** ns ns
Primary 0.93 1.21 0.77 0.98 0.91 1.04 0.86
Secondary or higher 0.72 1.03 0.69 1.02 0.79*** 0.81 0.71
Husband's occupation
(r: non-agriculture)
Agricultural occupation 0.68* 0.78 0.74*** 0.96 0.91*** 0.72*** 1.13
Husband's drunkenness
(r: doesn't drink) *** *** u *** u *** u
Never comes home drunk 0.82 0.76 u 1.43 u r u
Comes home drunk sometimes 1.07 1.62*** u 2.37*** u 2.14*** u
Comes home drunk frequently 6.07*** 7.39*** u 4.08*** u 4.80*** u
Union characteristics
Spousal age difference
(r: husband is younger) ns ns ns * * ns ns
Husband is 0-4 years older 0.87 0.87 1.02 0.77 0.82* 0.90 0.64
Husband is 5-9 years older 0.88 0.74 0.91 0.62* 0.87 0.86 0.60
Husband is 10-14 years older 0.78 0.85 0.78 0.59* 0.88 1.02 0.57*
Husband is 15+ years older 0.77 0.60* 0.91 0.51*** 0.82* 0.77 0.56
Spousal educational difference
(r: husband has less education) ns ns ns ns ** ns ns
Both have no education 0.55 1.53 0.65* 1.05 0.88 0.88 0.89
Both have same level of education 0.98 1.01 0.93 0.94 0.81*** 0.74* 1.15
Husband has more education 0.83 0.85 0.85 1.22 0.93 0.99 1.12
Marital duration (r: 0-4 years) ns ns ** ns *** ns ns
5-9 years 0.58 1.08 1.34 1.18 1.31*** 1.47* 1.05
10-14 years 0.53 1.05 1.24 1.65 1.24*** 1.42 0.77
15+ years 0.36* 0.95 0.82 1.63 1.24* 1.25 0.83
Household characteristics
Residence (r: urban)
Rural 0.92 0.87 0.82 0.60*** 0.76*** 0.81* 0.58***
Family structure (r: non-nuclear)
Nuclear 1.34 1.38** 0.93 1.21 1.21*** 0.90 1.01
Household wealth status
(r: poorest quintile) ns ns *** ns *** ns **
Second quintile 0.84 0.93 0.93 1.14 0.86*** 0.89 0.76**
Middle quintile 0.57** 0.88 0.70*** 1.20 0.68*** 0.95 0.83
Fourth quintile 0.85 0.84 0.58*** 0.81 0.49*** 1.13 0.69**
Wealthiest quintile 0.82 0.86 0.41*** 0.80 0.26*** 0.85 0.97
Constant 0.18*** 0.10*** 0.15*** 0.36** 0.10*** 0.14*** 0.38***
-2 Log Likelihood 1550.57 3063.37 4699.36 2113.40 48571.00 4497.69 3867.82
Nagelkerke R square 0.18 0.16 0.08 0.09 0.09 0.11 0.04
Number of women 2,024 4,795 6,435 2,017 84,202 6,484 3,409
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Note: Figures in bold represent bivariate relationships that are not statistically significant based on the chi-square test
(p>0.05).
r = Reference (omitted) category; ns = not significant; u = Unknown (not available)
***: p<.005; **: p<.01; *: p<.05

3 • Risk Factors for the Experience of Domestic Violence 45

 Table 3.7 reports the odds ratios of violence for women whose fathers beat their
mothers compared with those whose fathers did not or who did not know whether
their father beat their mother. Although these regressions also include all the vari-
ables shown in Tables 3.6.1 and 3.6.2, so as not to make the discussion repetitive,
only the odds ratios for the intergenerational effect variable are shown.

Table 3.7 Adjusted odds ratios of the likelihood of experiencing spousal violence ever and in the 12 months
preceding the survey for respondents whose father beat their mother compared with respondents whose fathers
did not beat their mothers (including those who do not know if their father beat their mother): logistic regression
results
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Dominican
Family history Cambodia Colombia Republic Haiti Nicaragua Peru
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
EVER EXPERIENCED SPOUSAL VIOLENCE
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Father beat mother
No/don't know 1.0 1.0 1.0 1.0 1.0 1.0
Yes 2.26*** 2.16*** 2.00*** 1.96*** 1.61*** 1.63***
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
EXPERIENCED VIOLENCE IN THE PAST 12 MONTHS
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Father beat mother
No/don't know 1.0 u 1.0 1.0 1.0 u
Yes 2.47*** u 2.02*** 1.96*** 1.47*** u
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Note: The regressions from which these odds ratios are obtained include all the variables in Tables 3.6.1 and
3.6.2 as controls.
u = Unknown (not available)

Risk factors for having ever experienced and for currently experiencing
spousal violence
Age: The experience of violence significantly varies between age groups only in
about half the countries: the Dominican Republic, Egypt, Haiti, India, and Nicara-
gua. In the Dominican Republic, Haiti, and Nicaragua, the risks of ever experiencing
violence and currently experiencing it are consistently lower for older women com-
pared with women age 15 to 19, with the risks diminishing more or less steadily with
age. By contrast, in Egypt and India, the odds of having ever experienced violence do
not vary consistently or significantly with age; however, the odds of currently experi-
encing violence are all lower for older women than for women age 15-19 and are par-
ticularly low at the oldest ages. Overall, this analysis suggests that, controlling for all
other characteristics, a currently married woman’s age does not affect her risk of ex-
periencing violence; where it does, it is the youngest women who are more at risk
than older women. The suggested negative association of ever-experience of violence
and age obtained here for some countries contrasts with the bivariate association dis-
cussed earlier (see Table 3.1.1). However, the similarity in age distributions of the
samples of currently married and ever-married women implies that the negative asso-
ciation is in fact the net effect of age and is not due to the restriction of the multivari-
ate analysis to currently married women.
Woman’s age at marriage: As in the case of age, a woman’s age at marriage is associ-
ated with her risk of experiencing violence only in about half of the countries. Unlike
the association with age, however, age at marriage is more consistently related to the
risk of ever-experience of violence than current experience of violence. In Colombia,
India, and Peru, the odds of ever experiencing violence fall consistently with age at
marriage, so that women who were first married at ages younger than 15 have the

46 Risk Factors for the Experience of Domestic Violence •3

highest odds of ever having experienced violence and those married beyond age 24
have the lowest. This is also true for Nicaragua, except that in Nicaragua, the odds
are significantly lower of ever experiencing violence only for women who first married
at ages older than 19. In the Dominican Republic and Egypt, age at marriage is not a
significant factor affecting the risk of violence; nonetheless, women who married be-
tween ages 20 and 24 in the Dominican Republic and women who married at an age
older than 24 in Egypt do have significantly lower odds of having ever experienced
violence. Zambia is the only country where age at first marriage is more significantly
related to current experience of violence than ever-experience of violence. Here too,
however, the risk of violence declines as age at first marriage increases. These data
suggest that across almost all countries, the risk of violence declines with increased
age at marriage and is much lower among women who have married at older ages
than among those married at younger ages, particularly ages younger than 15.
Number of unions: In all countries except Zambia, the ever-experience of violence is
much higher for women who are in a marriage of second or higher order than for
women in their first marriage. In all countries except Zambia, the ever-experience of
violence is much higher for women who are in a marriage of second or higher order
than for women in their first marriage.
This is to be expected, since a second marriage can occur only if the first has
ended, and for first marriages ending in a divorce, one reason for termination can be
violence in the relationship. However, Table 3.6.2 shows that in all seven countries
for which data on current experience of violence are available, women who are in sec-
ond or higher order marriages are also at a higher risk of current violence than
women in their first marriages, and this difference is significant in four of the seven
countries.
Number of children: When compared with women with no children, the odds of
ever experiencing violence as well as of currently experiencing violence increase
sharply with women’s number of children in all countries except Haiti and Zambia.
This consistent and strong positive association with violence persists despite controls
for various exposure variables, age, and wealth of the household. Furthermore, the
relationship is equally valid in most countries for the ever-experience and current ex-
perience of violence.
Own education level: Ever-experience of violence and current experience of violence
(where data are available) vary with women’s education in Cambodia, Egypt, Haiti,
India, and Peru. In Cambodia, Egypt, and India, the odds of experiencing violence
are significantly lower only among women who have a secondary or higher level of
education, whereas in Haiti and Peru, odds do not differ between women with no
education and women with secondary or higher education, but they are significantly
higher for women with only primary education. Thus, there is not a consistent one-
to-one relationship between a woman’s level of education and her risk of experienc-
ing violence.
Work status: With the exception of Cambodia, Haiti, and Zambia, in all other
countries, women’s likelihood of experiencing violence varies with their work status.
However, the pattern of variation is not consistent across countries. In Colombia, the
Dominican Republic, India, Nicaragua, and Peru, women who earn cash are signifi-
cantly more likely to have ever experienced, as well as to currently experience, vio-
lence than women who are not currently working at all. In Egypt, by contrast, they
are significantly less likely to have done so. Women who work but are not paid are

3 • Risk Factors for the Experience of Domestic Violence 47

 either no different from women who do not work at all, or they have significantly
higher odds of experiencing violence in Colombia, India, Peru, and also Egypt. No-
tably, Egypt is the one country in which the odds of violence are lower if women
work with pay but higher if they work without it, compared with the reference cate-
gory.
Husband’s education level: Since the focus of the analysis here is spousal violence,
the education of the husband may be the more relevant factor for a woman’s risk of
violence. However, only in Egypt, India, and Nicaragua does the risk of women ever
experiencing violence vary significantly with the husband’s level of education. In all
three countries, the odds of having ever experienced violence are lower only for
women whose husbands have at least a secondary level of education. In Peru, in con-
trast to all other countries, the odds of violence are higher if the husband has any
education (primary or secondary education) than if he has none. Current experience
of violence is similarly associated with the husband’s level of education only in India.
Husband’s occupation: The odds that a woman will experience violence are consis-
tently lower for women whose husbands are engaged in agricultural occupations, as
compared with women whose husbands are engaged in nonagricultural ones, in each
of the five countries where the differential is significant: Cambodia, Egypt, India,
Nicaragua, and Peru. Thus, women’s risk of violence either does not vary by whether
the husband is in an agricultural or nonagricultural occupation or, where it does, it is
always higher for women whose husbands are in nonagricultural occupations. Factors
such as area of residence and husband’s education, which could be thought to explain
this effect, are already controlled for in the equation.
Husband’s drunkenness: In all countries for which data are available, women whose
husbands frequently come home drunk have much higher odds of experiencing vio-
lence than women whose husbands do not drink or never come home drunk. This
relationship is strong and significant for both ever-experience and current experience
of violence. Figure 3.3 illustrates this relationship in Nicaragua.

48 Risk Factors for the Experience of Domestic Violence •3

 Figure 3.3
Adjusted odds ratios: odds of experiencing spousal violence
ever and in the past 12 months, by frequency with which a
husband comes home drunk, Nicaragua 1998
6.00

4.80

4.00

2.79

2.14
2.00
1.47
1.00 1.00

0.00
Ever experienced violence Experienced violence in the past 12 months

Does not drink Drunk sometimes Drunk frequently

Spousal age difference: The odds of experiencing violence do not vary by spousal age
difference in Cambodia, Egypt, and Zambia. In Peru, all women who are younger
than their husbands have a lower risk of experiencing violence than do women who
are older than their husbands, whereas in Colombia, the Dominican Republic, and
Haiti, only women who are either five to nine years or 15 or more years younger than
their husbands have odds of ever experiencing violence that are significantly lower
than the odds for women who are older than their husbands. The pattern for current
experience of violence is similar, with women with larger spousal age differences hav-
ing lower odds of experiencing current violence than do women who are older than
their husbands.
Spousal educational difference: Spousal educational difference has a significant effect
on women’s ever-experience of violence only in Colombia, India, and Peru. In all of
these countries, couples with the same amount of education have odds of having ever
experienced violence that are significantly lower than the odds for women whose
husbands have less education than they have. In addition, in India and Peru, the odds
are also lower for women whose husbands have more education than them. Notably,
with regard to current experience of violence, in India and Nicaragua, it is only
women whose husbands have the same amount of education who are at a lower risk
of experiencing violence than women in other categories of spousal educational dif-
ferentials. Thus, if there is a difference in risks of violence by spousal educational dif-
ferences, the category of women who appear to have consistently significantly lower
odds are women who have the same level of education as their husbands.
Marital duration: Marital duration measures exposure to the risk of spousal vio-
lence. Nonetheless, the odds of having ever experienced violence are consistently and
positively associated with duration only in the Dominican Republic, India, Nicara-
gua, and Zambia. By contrast, in Colombia, Egypt, and Peru, the odds of having ever

3 • Risk Factors for the Experience of Domestic Violence 49

 experienced violence are higher for marital durations of five to nine and 10 to 14
years than for longer or shorter durations, and in Cambodia and Haiti, the risk of
ever experiencing violence does not vary significantly with marital duration. Marital
duration also does not have a consistent relationship across countries with a woman’s
current risk of violence. In the Dominican Republic, Haiti, and Zambia, the odds of
current experience of violence do not vary by marital duration. In Cambodia, they are
significantly lower only at durations of 15 or more years, while in India and Nicara-
gua, they are higher at durations above zero to four years but are highest at durations
of five to nine years.
Place of residence: In the majority of countries, there is a significant relationship be-
tween residence and the odds of experiencing violence. In each of these cases, the
odds of having currently or ever experienced violence are lower for rural than for ur-
ban women.
6
Family structure: In Colombia and India, the odds of women ever experiencing
violence are significantly higher if women live in nuclear rather than nonnuclear
households. Similarly, in the Dominican Republic and India, the odds of a woman
currently experiencing violence are also higher if she lives in a nuclear household than
if she does not.
Household wealth status: Poverty has typically been considered a significant risk fac-
tor for violence. However, as is evident in Tables 3.6.1 and 3.6.2, there is no consis-
tent relationship between the risk of violence and the poverty-wealth status of the
households in which women reside. India is the only country where the odds of ex-
periencing violence are lower for women living in households other than the poorest
ones and where the odds decline steadily with increasing wealth. In Egypt and Peru,
women in the wealthiest households also have significantly lower odds of having ever
experienced violence than women in the poorest households. However, in Cambodia,
it is women living in households in the middle quintile who are the only ones that
have significantly lower odds of experiencing violence, as compared to women in the
poorest quintile. In Colombia, too, women in the middle quintile are different in
terms of their odds of experiencing violence; however, unlike women in Cambodia,
the odds of experiencing violence are higher for these women than for women from
poorer or richer households. In Zambia, by contrast, women in households belonging
to the middle, fourth, and wealthiest quintiles are no different from those in the
poorest quintile; however women in the second quintile have significantly lower odds
of experiencing any violence. Household wealth has a similarly inconsistent relation-
ship across countries with the likelihood of women currently experiencing violence.
Father beat mother: Table 3.7 shows that despite controls for all the variables dis-
cussed, women whose mothers experienced spousal violence are consistently much
more likely to experience violence than women who said that their father did not beat
their mother or who did not know whether their father beat their mother. Women
whose mothers experienced spousal violence are consistently much more likely to ex-
perience violence than women who said that their father did not beat their mother or
who did not know whether their father beat their mother.
In four of the six countries, women who report that their mothers were abused are
twice as likely to have ever experienced violence, and in three of the four countries for
which information on current violence is available, they are also about twice as likely

6
There is no separate category for female-headed nuclear households since the number of
currently married women living in such households is extremely limited.

50 Risk Factors for the Experience of Domestic Violence •3

to report current violence. In Nicaragua and Peru, daughters of abused mothers are at
least 60 percent more likely to report having ever experienced spousal violence than
daughters who either do not know or whose mothers were not abused.
In conclusion, this analysis has shown that there are almost no individual, marital,
or household characteristics that emerge as factors that identify women at an in-
creased risk of violence in all countries. In fact, there are only three variables that are
consistently related to a currently married woman’s risk of violence across all coun-
tries: their number of children, their husband’s drunkeness, and being able to report
spousal abuse of the mother. Notably, only one of these three variables, a woman’s
number of children, was considered as one of the respondent’s own characteristics.
This is also one of the variables with which the direction of causality remains particu-
larly unclear. Are women who have more children more likely to experience violence,
or are women who experience violence less likely to be able to control their fertility?
This issue is not resolved by the finding that women’s ever-experience of violence and
their current experience of violence vary similarly with number of children.
Neither of the other two variables that have a cross-nationally consistent, strong
and positive influence on a woman’s risk of violence, namely, the husband’s drunken-
ness and the experience of spousal violence by the respondent’s mother are directly
related to women themselves, but are aspects of the circumstances of women’s lives.
While a husband’s excessive drinking cannot be assumed to itself be the cause of the
abuse, it does appear to be an enabling circumstance that is strongly related to a
woman’s risk of violence. The only one of the significant variables where the direc-
tion of causality is clear is the effect of the mother’s experience of spousal violence.
This analysis reinforces and extends to six more countries the literature that empha-
sizes the intergenerational effects of violence.
It is also notable that other factors usually assumed to be negatively linked with the
risk of violence, such as household wealth and education of the husband and wife, are
not always significant nor necessarily negative in their effects. Factors such as early
age at marriage and large spousal age and educational differences, often used as indi-
cators of women’s disempowerment, are only weakly related to the risk of violence
and not always in the predicted direction. Overall, this analysis shows that while
women’s risk of violence does vary in some countries by selected individual and
household characteristics, there is little consistency in the nature and significance of
the variation across countries. This is in stark contrast to the cross-cultural relevance
of factors such as husband’s drunkenness or a family history of violence.

3 • Risk Factors for the Experience of Domestic Violence 51

 4
Domestic Violence and Women’s Empowerment

T his chapter examines the association between domestic violence and some hy-
pothesized correlates of violence, namely, selected indicators of women’s em-
powerment and the evidence of controlling behaviors by husbands. The latter can be
seen as factors that directly inhibit women’s empowerment.
The prevalence rate of violence against women is almost by definition an indicator
of women’s status at a societal level. It could be argued that in societies where women
routinely experience violence, women are clearly devalued. However, research does
not necessarily indicate a consistent negative relationship between violence and
women’s status. Early research, for example, found a U-shaped relationship between
the status of women in different states of the United States and the experience of vio-
lence (Yllo, 1983). Yllo explained these results by arguing the following: violence
rates were high where women’s status was low because the low status resulted in lim-
ited options for women; they were relatively high where women’s status was high be-
cause women’s high status constituted a threat to the dominance of men. The latter
explanation is consistent with status inconsistency explanations for family violence
(Yick, 2001). Status inconsistency theories see violence as resulting from resource
imbalance among members of a family, where resources include both material and
nonmaterial (such as education and prestige, etc.) assets. Patriarchal norms typically
imply that men will have more resources than women, and the empowerment of
women can upset this balance. Women can experience violence when patriarchal
norms are threatened by resource imbalance in favor of the woman, which over time
can generate stressors within the family (Gelles, 1993).
Status inconsistency theories assume that a relationship exists between empower-
ment and violence, and they implicitly suggest that violence will result when women’s
empowerment is inconsistent with patriarchal norms. At the individual level, how-
ever, the direction of causality is likely to be unclear. Although an individual woman’s
expression of empowerment can result in violence when such behavior is perceived as
violating normative gender roles (as claimed by the status inconsistency theories) and
evidenced in efforts to empower women (Goetz, 1997 and Sen Gupta, 1996), the
violence itself can cause individual women to be disempowered. To the extent that
the latter is true, causality may run not from empowerment to violence, but from vio-
lence to disempowerment. In fact, low empowerment and spousal violence together
may form a vicious circle, making it difficult to discern the direction of causality. This
is likely to be particularly true in societies where women cannot or may not leave vio-
lent marriages.
In cross-sectional data of the type available from the Demographic and Health
Surveys (DHS) questionnaires, the direction of causality is difficult to disentangle.
While the experience of spousal violence could have occurred in the 12 months be-
fore the interview or at any time before that, women’s empowerment is evaluated at

4 • Domestic Violence and Women’s Empowerment 53

 the time of interview Consequently, this chapter explores the extent to which
women’s empowerment levels vary by whether they have ever experienced spousal
violence, experienced spousal violence in the last one year, or never experienced vio-
lence, as well as whether women’s experience of violence varies by their empower-
ment level.
Three different sets of indicators of women’s empowerment are used in this chap-
ter. The first set purports to measure women’s degree of control over their environ-
ment by measuring their participation in household and children-related decision-
making. The second set derives from the assumption that a fundamental element of
empowerment is the rejection of seemingly immutable and essentially unequal rights
and privileges on the basis of the sex of an individual. Two examples of such “rights”
normatively ascribed to men include the right of husbands to regulate and control
“their” women’s behavior, through force if necessary, and the right to expect wives to
be submissive to husbands’ sexual demands. Acceptance by women of this norma-
tively prescribed power of men over women reflects an acceptance of unequal gender
roles on the one hand and a lack of conscientization about women’s entitlement to
bodily security and integrity, on the other (Correa and Petchesky, 1994; Sen and Bat-
liwala, 2000; United Nations, 1995b). For example, acceptance of the beating of
wives by husbands in a society is indicative of low status for women, absolutely and
relative to men. At the level of the individual woman, too, acceptance by her of the
right of men to beat their wives is indicative of her acceptance of women’s lower
status relative to men. While such attitudes do not necessarily signify approval of
these rights for men, they do signify women’s acceptance of norms that give men
these rights.

4.1 Domestic Violence and Women’s Participation in

Decisionmaking
Participation in decisions about one’s own needs, household needs, and the needs
of children is an indicator of women’s engagement with and control over their imme-
diate household environments. There are many reasons to expect that women’s par-
ticipation in decisionmaking will vary by their domestic violence status. Domestic
violence lowers women’s self-esteem and erodes their mental health (Astbury, 1999;
Ellsberg et al., 1999; Fikree and Bhatti, 1999), thereby affecting women’s capacity, as
well as willingness, to participate. Women may also be actively prevented from par-
ticipating in the control of what happens in their households through the controlling
behavior of their partners. Yet another direction of association derives from the status
inconsistency theory discussed earlier. Violence may result from women’s attempts to
control some of the decisions that are not normatively perceived to be in the realm of
women’s control, such as economic decisions. In other words, behaviors that are per-
ceived to be violating gender roles may fuel violence. Research, in fact, identifies male
control of household decisionmaking as a predictor of partner violence (Levinson,
1989; Oropesa, 1997). However, whether male control of decisionmaking is a conse-
quence of abused women’s inability to participate or is itself a correlate of the violence
preventing women from participating remains unclear.

54 Domestic Violence and Women’s Empowerment • 4

 In light of the possibilities discussed above, the different directions of causality be-
tween women’s decisionmaking and the experience of violence are explored in Tables
4.1 and 4.2. Table 4.1 shows women’s participation in personal, household, and
child-related decisions separately for women who have ever experienced spousal vio-
lence, experienced such violence in the past one year, or never experienced it. Table
4.2 shows how the likelihood of experiencing spousal violence in the recent past (past
one year) varies by whether women make different decisions alone or jointly with
their husbands or whether the husband decides alone. In this latter table, only the
likelihood of violence in the past one year is examined to better understand whether
women’s participation in decisionmaking is giving rise to violence. Both tables are
restricted to currently married women in order to focus on the variation between vio-
lence and women’s participation in decisionmaking when husbands are potentially a
force. For the small minority of currently married women who have been married
more than once, the husband who is presumed to be present to make decisions or Violence may result
participate in them may not be the one who was violent towards the respondent. This from women’s
latter possibility is further minimized if attention is focused on the experience of vio- attempts to control
lence only in the past one year. some of the decisions
In the presentation of the data on decisionmaking, women who make decisions on that are not
their own are listed separately from those who make decisions jointly with their hus- normatively perceived
to be in the realm of
band. However, the literature on women’s status and empowerment does not really
women’s control, such
clarify the nature of women’s control over decisions that would be considered reflec- as economic
tive of the empowerment of women. Does control require that women make deci- decisions.
sions by themselves, or can joint decisionmaking be an equally valid measure of con-
trol and hence empowerment? The lack of conceptual clarity on the issue of control
should be kept in mind when interpreting the results below.
In the DHS survey, women’s participation is measured for several different deci-
sions. With some variation across the different countries in the specific wording of
the question, women were asked who usually makes the specified decisions in their
household. There is also some variation across the countries in the specific decisions
asked about. In Tables 4.1 and 4.2, the data related to the decisions included in most
of the countries are shown. These results are also typical of the decisions not specifi-
cally shown in the tables.

Does participation in decisionmaking vary by women’s experience of

violence?
Since gender-role incompatibility is one explanation for the expected variation in
women’s decisionmaking by their violence status, the results are discussed separately
for the different types of decisions. The expectation is that husbands are much more
likely to be the main decisionmakers among couples where the wife is abused.
Decisions about what food to cook: In almost all societies, women are the ones typi-
cally expected to make decisions about food and what to cook. The presumption is
that participation in or even control of such decisions is unlikely to represent any
gender-role incompatibility. Thus, it is no surprise that in all of the seven countries
for which information is available for these types of decisions, the majority of cur-
rently married women report making such decisions on their own. Further, the pat-
tern of variation by violence status in who makes these types of decisions is quite con-
sistent across countries. Husbands alone rarely make these decisions, but in most

4 • Domestic Violence and Women’s Empowerment 55

 Table 4.1 Percent distribution of currently married women age 15-49 by who in their household makes different household
decisions: the women alone, the women jointly with their husbands, their husbands alone, or someone else (alone or jointly
with others in the household), according to whether they have experienced violence by their husband ever, in the past 12
months, or never
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Don’t know/
Type of decision and Jointly missing/
whether experienced Woman with Husband decision
violence by husband alone husband alone Other not made Total
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
DECISIONS ABOUT WHAT FOOD TO COOK EACH DAY
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Colombia
Ever experienced violence 74.8 12.7 3.5 8.9 0.0 100.0
Experienced violence in the past 12 months u u u u u -
Never experienced violence 73.2 14.2 2.5 9.8 0.3 100.0
Egypt
Ever experienced violence 68.0 14.9 6.6 10.1 0.4 100.0
Experienced violence in the past 12 months 67.9 12.9 8.0 10.9 0.3 100.0
Never experienced violence 68.8 19.2 4.5 7.1 0.4 100.0
Haiti
Ever experienced violence 82.7 5.0 1.9 9.6 0.8 100.0
Experienced violence in the past 12 months 81.0 5.7 1.8 10.5 1.0 100.0
Never experienced violence
76.4 9.8 3.7 10.0 0.2 100.0
India
Ever experienced violence 76.7 4.0 4.1 15.3 0.0 100.0
Experienced violence in the past 12 months 74.9 4.5 4.8 15.8 0.0 100.0
Never experienced violence 69.9 4.5 3.5 22.1 0.0 100.0
Nicaragua
Ever experienced violence 62.9 24.5 9.5 2.2 0.9 100.0
Experienced violence in the past 12 months 63.0 22.7 10.9 2.2 1.3 100.0
Never experienced violence 54.2 34.6 7.6 2.2 1.5 100.0
Peru
Ever experienced violence 76.9 12.4 2.9 7.8 0.1 100.0
Experienced violence in the past 12 months u u u u u -
Never experienced violence 73.8 13.8 2.4 9.9 0.2 100.0
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
DECISIONS ABOUT RESPONDENT'S OWN HEALTH CARE
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced violence 37.2 49.4 10.4 2.7 0.3 100.0
Experienced violence in the past 12 months 35.7 50.6 10.7 3.0 0.0 100.0
Never experienced violence 37.6 52.4 7.7 2.2 0.2 100.0
Colombia
Ever experienced violence 69.1 16.4 10.8 3.5 0.1 100.0
Experienced violence in the past 12 months u u u u u -
Never experienced violence 63.0 23.8 10.7 2.5 0.1 100.0
Haiti
Ever experienced violence 47.3 21.4 26.4 4.5 0.5 100.0
Experienced violence in the past 12 months 44.8 20.6 28.9 5.1 0.6 100.0
Never experienced violence 40.8 33.7 20.3 4.9 0.3 100.0
India
Ever experienced violence 30.0 14.8 43.1 12.1 0.0 100.0
Experienced violence in the past 12 months 25.7 15.1 46.9 12.3 0.0 100.0
Never experienced violence 27.6 17.5 38.4 16.5 0.0 100.0
Peru
Ever experienced violence 61.9 20.5 15.6 1.9 0.1 100.0
Experienced violence in the past 12 months u u u u u -
Never experienced violence 58.3 24.2 15.6 1.7 0.1 100.0
Zambia
Ever experienced violence 28.0 10.3 51.1 10.3100.0 0.3
Experienced violence in the past 12 months 27.4 9.5 51.8 11.1100.0 0.3
Never experienced violence 30.8 11.5 44.2 13.3100.0 0.2
Continued…
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
u = Unknown (not available)

56 Domestic Violence and Women’s Empowerment • 4

 Table 4.1—Continued
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Don’t know/
Type of decision and Jointly missing/
whether experienced Woman with Husband decision
violence by husband alone husband alone Other not made Total
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
DECISIONS REGARDING CHILDREN’S ILLNESS
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced violence 23.9 66.9 4.5 0.9 3.7 100.0
Experienced violence in the past 12 months 24.2 67.0 4.0 1.1 3.8 100.0
Never experienced violence 19.0 69.1 2.1 1.0 8.8 100.0
Dominican Republic
Ever experienced violence 40.2 41.7 8.6 3.8 5.8 100.0
Experienced violence in the past 12 months 38.7 43.2 8.4 3.0 6.7 100.0
Never experienced violence 24.1 57.2 6.4 2.4 9.9 100.0
Egypt
Ever experienced violence 27.3 48.0 19.7 1.5 3.5 100.0
Experienced violence in the past 12 months 25.0 48.2 20.2 2.4 4.1 100.0
Never experienced violence 25.4 51.2 14.8 1.1 7.5 100.0
Haiti
Ever experienced violence 26.9 39.8 16.4 3.5 13.3 100.0
Experienced violence in the past 12 months 24.7 40.0 18.5 3.9 12.9 100.0
Never experienced violence 21.5 48.4 12.5 3.6 14.0 100.0
Nicaragua
Ever experienced violence 42.9 38.0 13.5 0.3 5.3 100.0
Experienced violence in the past 12 months 44.7 34.6 14.9 0.4 5.5 100.0
Never experienced violence 27.8 54.2 11.9 0.5 5.6 100.0
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
DECISIONS ABOUT VISITS TO FAMILY, FRIENDS, OR RELATIVES
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced violence 20.9 72.5 3.4 1.7 1.4 100.0
Experienced violence in the past 12 months 21.4 71.9 3.2 1.9 1.6 100.0
Never experienced violence 15.4 78.5 2.5 2.5 1.1 100.0
Colombia
Ever experienced violence 36.5 43.6 12.5 3.4 4.0 100.0
Experienced violence in the past 12 months u u u u u -
Never experienced violence 28.3 56.5 9.3 3.2 2.7 100.0
Egypt
Ever experienced violence 8.2 20.3 67.6 3.0 0.9 100.0
Experienced violence in the past 12 months 5.5 18.5 71.7 3.0 1.3 100.0
Never experienced violence 7.5 38.3 52.4 1.4 0.4 100.0
Haiti
Ever experienced violence 56.6 30.9 7.2 4.5 0.9 100.0
Experienced violence in the past 12 months 53.3 33.0 8.0 4.7 1.0 100.0
Never experienced violence 49.2 37.3 8.2 4.2 1.2 100.0
India
Ever experienced violence 16.1 22.6 45.9 15.3 0.1 100.0
Experienced violence in the past 12 months 12.1 22.3 49.4 16.2 0.0 100.0
Never experienced violence 13.4 25.6 38.5 22.4 0.0 100.0
Nicaragua
Ever experienced violence 21.1 49.1 27.8 0.9 1.1 100.0
Experienced violence in the past 12 months 23.0 43.1 31.9 0.9 1.2 100.0
Never experienced violence 12.9 65.4 19.6 0.5 1.6 100.0
Peru
Ever experienced violence 30.2 53.2 13.6 2.0 1.0 100.0
Experienced violence in the past 12 months u u u u u 100.0
Never experienced violence 25.0 59.6 11.5 3.1 0.9 100.0
Zambia
Ever experienced violence 18.1 23.7 56.5 1.4 0.3 100.0
Experienced violence in the past 12 months 18.2 22.3 57.6 1.6 0.3 100.0
Never experienced violence 16.3 26.3 55.8 1.4 0.2 100.0
Continued…
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
u = Unknown (not available)

4 • Domestic Violence and Women’s Empowerment 57

 Table 4.1—Continued
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Don’t know/
Type of decision and Jointly missing/
whether experienced Woman with Husband decision
violence by husband alone husband alone Other not made Total
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
DECISIONS ABOUT MAKING LARGE HOUSEHOLD PURCHASES
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced violence 37.2 48.4 12.5 1.6 0.3 100.0
Experienced violence in the past 12 months 37.5 47.6 12.8 1.7 0.3 100.0
Never experienced violence 25.7 59.3 9.4 5.5 0.1 100.0
Colombia
Ever experienced violence 27.1 41.2 25.0 6.1 0.5 100.0
Experienced violence in the past 12 months u u u u u -
Never experienced violence 18.9 53.6 20.8 6.1 0.6 100.0
Egypt
Ever experienced violence 14.2 28.2 47.0 10.6 0.1 100.0
Experienced violence in the past 12 months 12.0 27.0 50.4 10.6 0.0 100.0
Never experienced violence 14.0 42.1 36.4 7.4 0.1 100.0
Haiti
Ever experienced violence 40.2 23.3 25.1 10.4 0.9 100.0
Experienced violence in the past 12 months 35.8 22.6 29.0 11.6 1.1 100.0
Never experienced violence 31.2 40.4 18.0 10.0 0.4 100.0
India
Ever experienced violence 13.4 28.9 39.4 18.2 0.0 100.0
Experienced violence in the past 12 months 10.2 27.7 43.6 18.6 0.0 100.0
Never experienced violence 10.0 31.3 32.6 26.1 0.0 100.0
Nicaragua
Ever experienced violence 17.3 44.3 32.4 1.3 4.7 100.0
Experienced violence in the past 12 months 17.1 41.1 35.9 1.4 4.5 100.0
Never experienced violence 9.6 60.3 25.4 0.5 4.2 100.0
Peru
Ever experienced violence 23.7 50.2 22.1 3.6 0.3 100.0
Experienced violence in the past 12 months u u u u u -
Never experienced violence 18.2 56.0 20.3 5.2 0.3 100.0
Zambia
Ever experienced violence 11.7 24.0 62.4 1.6 0.2 100.0
Experienced violence in the past 12 months 11.3 21.9 64.2 2.3 0.4 100.0
Never experienced violence 11.3 25.1 61.0 2.4 0.1 100.0
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
DECISIONS ABOUT HAVING ANOTHER CHILD
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced violence 12.8 68.9 3.7 0.7 14.0 100.0
Experienced violence in the past 12 months 12.2 68.4 3.5 0.7 15.2 100.0
Never experienced violence 8.8 72.5 1.9 0.4 16.5 100.0
Dominican Republic
Ever experienced violence 33.3 38.4 9.5 1.3 17.5 100.0
Experienced violence in the past 12 months 32.4 38.1 9.1 1.0 19.4 100.0
Never experienced violence 22.4 52.5 5.1 1.4 18.6 100.0
Egypt
Ever experienced violence 4.8 67.6 21.5 0.4 5.7 100.0
Experienced violence in the past 12 months 4.8 65.7 24.6 0.4 4.6 100.0
Never experienced violence 3.0 77.0 14.7 0.3 5.0 100.0
Haiti
Ever experienced violence 24.1 40.2 8.5 0.7 26.6 100.0
Experienced violence in the past 12 months 24.0 39.3 9.5 0.8 26.4 100.0
Never experienced violence 16.8 43.9 8.0 0.4 30.9 100.0
Zambia
Ever experienced violence 9.9 33.7 52.3 1.2 2.9 100.0
Experienced violence in the past 12 months 9.6 32.6 53.8 1.4 2.6 100.0
Never experienced violence 9.0 36.7 48.9 0.8 4.6 100.0
Continued…
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
u = Unknown (not available)

58 Domestic Violence and Women’s Empowerment • 4

 Table 4.1—Continued
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Don’t know/
Type of decision and Jointly missing/
whether experienced Woman with Husband decision
violence by husband alone husband alone Other not made Total
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
DECISIONS ABOUT CONTRACEPTION
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced violence 19.9 61.6 1.9 1.2 15.4 100.0
Experienced violence in the last one year 20.6 61.6 1.8 1.3 14.8 100.0
Never experienced violence 15.1 67.9 1.8 0.6 14.6 100.0
Egypt
Ever experienced violence 14.4 64.2 9.4 0.2 11.8 100.0
Experienced violence in the last one year 14.3 63.4 11.6 0.1 10.5 100.0
Never experienced violence 12.7 65.9 6.7 0.3 14.4 100.0
Haiti
Ever experienced violence 24.8 33.4 5.6 0.3 36.1 100.0
Experienced violence in the last one year 22.8 36.4 6.2 0.3 34.3 100.0
Never experienced violence 16.2 30.4 2.7 0.6 50.0 100.0
Nicaragua
Ever experienced violence 36.5 34.7 10.4 1.0 17.4 100.0
Experienced violence in the last one year 42.1 32.5 12.5 0.5 12.5 100.0
Never experienced violence 22.3 52.2 9.1 1.1 15.2 100.0
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
u = Unknown (not available)

countries, the proportion of husbands making these decisions alone is somewhat

higher in couples where the woman has experienced violence. Women who have
never experienced violence are more likely than those who have to make decisions
about food jointly with their husbands. The variation by violence status is greatest for
women who make these decisions alone: women who have experienced violence are
more likely than those who have not to make decisions about food on their own. This
pattern is most marked in Nicaragua.
Decisions about respondents’ own health care: Among all of the types of decisions
asked about, perhaps those with the most strategic importance for the self-interest of
women are decisions about their own health care. For this type of decision, which
concerns women’s own health needs, more than for any of the others considered here,
it can be argued that women should be final arbiters. Nonetheless, women are not
necessarily participating in these decisions. In only three of the six countries for
which this information is available, namely, Colombia, Haiti, and Peru, are these de-
cisions most likely to be made by women alone. In India and Zambia, these decisions
are most likely to be made by husbands alone, and in Cambodia, they are most likely
to be made jointly by women and their husbands.
There is also no consistent pattern of variation in decisionmaking with regard to
the respondent’s health care across countries by the violence status of women. While
most countries do fall into the pattern seen for other decisions, in Haiti, India, and
Zambia, husbands’ control over these decisions is significantly more common among
women who have experienced violence than among those who have never experi-
enced violence. The likelihood of joint decisionmaking about women’s health care is
greater among women who have never experienced violence in all countries for which
data are available. Also, in Colombia, Haiti, India, and Peru, making such decisions
by themselves is more common among women who have experienced violence than
among those who have not, whereas in Zambia, this pattern is reversed.

4 • Domestic Violence and Women’s Empowerment 59

 Decisions regarding children’s illness: Women are traditionally the caretakers of chil-
dren; but as with their own health care, they are not necessarily the final arbiters of
decisions related to what to do when a child falls ill. However, it is encouraging that
in all countries, the large majority of women are participating alone or jointly in mak-
ing these decisions. Notably, joint decisionmaking is more common among women
who have not experienced violence than among those who have. Women are more
likely to be making these decisions on their own if they have experienced violence. In
fact, in Nicaragua, women who have experienced violence are more likely to make
these decisions alone than jointly with their husbands, whereas women who have
never experienced violence are about twice as likely to make these decisions jointly
with their husbands than alone. In all countries except Cambodia, a nonnegligible
proportion of husbands make these decisions on their own, and this proportion is
consistently higher for women who have experienced violence, compared with
women who have not. For example, in Haiti, husbands of 19 percent of women who
have experienced violence recently make these decisions alone, compared with hus-
bands of 13 percent of women who have never experienced violence.
Decisions about visits to family and friends: Women’s participation in such decisions
has a bearing on their freedom of movement. In patriarchal societies and where
women are cloistered, it is less likely that women will be free to make these decisions
alone. This expectation is borne out in the data. Haiti is the only country where these
decisions are most likely to be made by women alone. In Egypt, India, and Zambia,
they are most likely to be made by husbands alone, and in the remaining countries,
women are most likely to make these decisions jointly with their husbands.
The variation in women’s participation by violence is similar across countries. In
every country, joint decisionmaking is most common among women who have never
experienced violence, and in all countries but Haiti, these decisions are more likely to
be made by husbands alone if women have experienced violence, particularly if
women have experienced violence in the past one year. Even in Egypt and India,
where these decisions are most likely to be made by husbands, husbands are much
more likely to be making these decisions if the woman has experienced violence than
if she has not. In Egypt, for example, among women who have not experienced vio-
lence, 52 percent report that their husbands make these decisions alone, whereas
among women who have experienced violence in the past one year, this proportion is
72 percent. Notably too, in all countries except Egypt and India, women who have
experienced violence ever or recently are more likely to make these decisions alone,
compared with women who have never experienced violence.
Decisions about large household purchases: In most cultures, decisions about major
household purchases are not typically within the normatively prescribed purview of
women. It can be hypothesized that gender-role incompatibility is most likely if
women try to dominate these decisions. Nonetheless, the data show that in about half
of the countries, these decisions are most often made jointly by husbands and wives.
In Egypt, India, and Zambia, husbands alone are most likely to make these decisions.
In Haiti, the person most likely to make these decisions varies by whether women
have experienced violence or not. Women who have experienced violence are most
likely to make these decisions themselves, whereas women who have never experi-
enced violence are most likely to make these decisions jointly with their husbands. In
every country except Zambia, joint decisionmaking is much more common among
women who have never experienced violence than among women who have. By con-

60 Domestic Violence and Women’s Empowerment • 4

trast, these decisions are more commonly made by women themselves and by their
husbands only among women who have experienced violence, compared with women
who have not. The likelihood that a husband makes the decision alone is higher
among women who have recently experienced violence than among those who have
ever experienced violence or never experienced violence. This is true for all countries
for which short- and long-term data on violence are available.
Decisions about having another child and contraception: In most countries, women
report making these decisions jointly with their husbands. The only exceptions are
Nicaragua and Zambia. In Nicaragua, women are most likely to decide about the use
of contraception by themselves if they have ever experienced violence but jointly with
their husbands if they have never experienced violence. In Zambia, the decision to
have another child is most often made by husbands alone. In two of the three coun-
tries, namely Cambodia and Egypt, where information is available for both types of
decisions, women are much more likely to make the decision regarding contraception
by themselves than the decision about having another birth. The pattern of deci-
sionmaking varies fairly consistently across all countries. Compared with women who
have experienced violence, women who have never experienced violence are more
likely to make these decisions jointly with their husbands and are much less likely to
make these decisions on their own or to have husbands who decide alone.

Does the likelihood of experiencing violence vary by women’s

participation in household decisionmaking?
A show of empowerment, for example, making decisions that are traditionally ex-
pected to be controlled by men, can be hypothesized as “invoking” violence. How-
ever, the results shown in Table 4.2 do not fully bear out this hypothesis. Table 4.2
shows the percentages of currently married women who have experienced violence in
the 12 months preceding the survey, according to whether they alone make different
decisions, make decisions jointly with their husbands, or have husbands who make
the different decisions alone.
As can be seen from Table 4.2 and Figure 4.1, the highest rates of violence are not
consistently experienced by women who make different decisions by themselves. In
fact, for most countries, the highest rates of recent violence are generally among
women whose husbands make decisions alone, irrespective of the decision. In fact, for
most countries, the highest rates of recent violence are generally among women
whose husbands make decisions alone, irrespective of the decision.
In Egypt and India, this is true for all decisions, and in the remaining countries for
which data on recent violence are available, this is true for the majority of decisions.
Nonetheless, married women who make decisions by themselves do have much
higher rates of violence than women who make decisions jointly with their husbands.
In fact, the most consistent result in Table 4.2 is that the rates of recent violence are,
with very few exceptions, the lowest in all countries for women who make decisions
jointly with their husbands. Even in the case of decisions that are traditionally ex-
pected to be made by women, such as what food to cook each day, violence is more
common among women who make this decision alone than among those who make
this decision jointly with their husband.

4 • Domestic Violence and Women’s Empowerment 61

 Table 4.2 Percentage of currently married women age 15-49 who experienced violence in the past 12 months, according to whether they
make different decisions on their own or jointly with their husband, or whether their husband alone makes the decision, by type of decision
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Women who Women who make Women whose
make decision decision jointly husbands make the
by themselves with their husband decision alone Other
–––––––––––––––––– –––––––––––––––––– –––––––––––––––––– –––––––––––––––––––
Percentage Percentage Percentage Percentage
who who who who
experienced experienced experienced experienced
violence Number violence Number violence Number violence Number
in the past of in the past of in the past of in the past of
Type of decision 12 months women 12 months women 12 months women 12 months women
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Own health care 14.1 780 14.4 1,078 19.4 169 19.5 47
Children's illness 18.0 413 14.4 1,429 24.0 52 16.3 20
Visits to family, friends, or relatives 19.4 339 13.7 1,610 17.7 55 11.9 50
Making large household purchases 20.1 573 12.3 1,196 19.2 206 5.3 101
Contraceptive use 19.2 331 13.6 1,389 14.7 38 26.8 15
Having another child 19.1 196 14.1 1,494 23.8 46 26.2 9
Dominican Republic
Children's illness 14.1 1,431 7.9 2,857 12.2 359 11.3 141
Have another child 13.1 1,289 7.6 2,625 15.3 311 7.2 75
Egypt
What food to cook each day 18.4 4,523 13.6 1,169 28.5 344 24.7 540
Children's illness 17.9 1,718 17.9 3,304 22.8 1,090 34.2 85
Visits to family, friends, or relatives 13.2 512 10.7 2,111 23.1 3,809 29.4 127
Making large household purchases 15.8 929 13.5 2,458 23.4 2,647 23.2 562
Contraceptive use 20.1 876 18.1 4,309 28.2 505 6.0 17
Having another child 24.5 239 16.6 4,862 26.7 1,130 22.2 20
Haiti
What food to cook each day 21.4 1,555 13.4 172 11.2 64 22.1 198
Own health care 21.8 846 13.9 614 27.2 433 21.8 96
Children's illness 22.3 455 17.8 924 28.2 269 22.2 71
Visits to family, friends, or relatives 21.5 1,019 19.0 712 20.8 157 23.2 84
Making large household purchases 22.1 667 12.8 724 30.2 392 23.8 202
Contraceptive use 25.7 366 24.1 621 37.3 67 11.5 11
Having another child 26.8 370 18.8 857 24.0 162 34.0 10
India
What food to cook each day 10.9 64,255 10.6 3,963 13.6 3,290 7.8 18,783
Own health care 9.4 25,379 9.2 15,313 12.3 35,462 8.1 14,135
Visits to family, friends, or relatives 9.0 12,560 9.2 22,619 12.8 36,038 7.9 19,036
Making large household purchases 9.8 9,653 9.3 27,869 13.3 30,571 7.8 22,193
Nicaragua
What food to cook each day 14.9 3,847 9.6 2,154 17.9 553 13.3 148
Children's illness 18.6 2,179 9.3 3,370 16.2 837 11.7 32
Visits to family, friends, or relatives 20.2 1,033 9.5 4,126 19.5 1,489 18.3 43
Making large household purchases 19.4 803 9.9 3,788 17.6 1,857 26.0 49
Contraceptive use 21.4 1,787 9.2 3,212 17.7 643 5.6 74
Zambia
Own health care 25.3 900 23.6 335 29.7 1,452 25.4 364
Visits to family, friends, or relatives 28.8 525 24.2 767 28.0 1,716 32.6 42
Making large household purchases 26.8 352 24.3 751 28.4 1,887 29.8 63
Having another child 27.8 288 25.2 1,079 29.0 1,544 40.0 30

The results of Tables 4.1 and 4.2 thus do not support the hypothesis that the ex-
perience of violence by women is higher when there are gender-role violations. What
the data do uphold is the expectation that violent relationships are characterized by
husbands’ somewhat greater control of decisionmaking, whereas nonviolent spousal
relationships are characterized by joint decisionmaking. What is a surprise, however,
is that violent relationships are also characterized by women making decisions alone.
In fact, from Table 4.1, it is clear that decisionmaking by women alone, as well as by
men alone, is much more common for women who have experienced violence than
for women who have never done so. This suggests that for some women, the experi-
ence of violence separates them from the control of decisions relevant to their well-
being, while for other women, it enhances their control of decisions. Further research
is needed to help identify the factors that would explain this difference.

62 Domestic Violence and Women’s Empowerment • 4

 Figure 4.1
Percentage of currently married women who have
experienced violence in the 12 months preceding the
survey, according to the person who has the final say
in decisions about making large household purchases

30
Percentage

20

10

0
Cambodia Egypt Haiti India Nicaragua Zambia

Woman alone Jointly with husband Husband alone

4.2 Domestic Violence and Norms that Reinforce Inequality in

Marital Relationships
The DHS asks women two questions that tap into their attitudes about the roles
of wives relative to their husbands. The first asks women whether they agree that a
husband is justified in beating his wife for different specified reasons. These reasons,
which range from those that involve suspicions about a wife’s moral character to
those that may be considered more trivial, such as not cooking properly, were chosen
to provide variation in the perceived seriousness of behavioral-norm violation. The
second question asks women whether they think that a wife is justified in refusing to
have sex with her husband under several specified circumstances. Both of these ques-
tions explore women’s acceptance of norms that accord men power over women and
subordinate the rights of women to those of men. As mentioned above, seeing as jus-
tified the power of men over women reflects an acceptance of unequal gender roles as
well as a lack of conscientization about women’s entitlement to bodily security and
integrity. Thus, women who agree with the right of men to physically and sexually
dominate women are hypothesized to be less empowered than women who reject
these rights.
The direction of association between women’s acceptance of men’s rights over
women and domestic violence is not clear. It could be hypothesized that women who
accept the subordinate roles of wives relative to husbands will be less subject to vio-
lence because they are likely to conform to traditional role expectations in other ways
as well. In other words, their behavior is less likely to create status incompatibility.
However, this hypothesis would be valid if violence is indeed a result of gender-role
violations by women. It could also be hypothesized that the experience of violence
“teaches” women to accept these norms. Since the direction of causality is unclear,
the violence-attitude relationship is explored by examining, in turn, both violence and

4 • Domestic Violence and Women’s Empowerment 63

 attitudes as dependent variables. Accordingly, Tables 4.3 and 4.5 compare the preva-
lence of violence among women who agree and women who disagree with each atti-
tude question, and Tables 4.4 and 4.6 examine women’s extent of agreement with
each attitude question, according to women’s experience of violence.
Acceptance of wife-beating: The question on women’s attitudes toward wife-beating
was asked in eight of the nine countries included in this report. The specific reasons
that respondents were asked to consider vary among countries. Nonetheless, as can be
seen from Table 4.3, there is no evidence that women who agree that husbands are
justified in beating their wives experience lower rates of violence. In every country

Table 4.3 Percentage of ever-married women age 15-49 by whether they have ever experienced violence by their husband, have experienced
violence by their husband in the last one year, or never experienced violence by their husband, according to whether they agree or disagree with
different reasons justifying wife-beating and reason
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Whether woman agrees with reason justifying a husband beating his wife
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Dominican
Cambodia Republic Egypt Haiti India Nicaragua Zambia
Reason justifying a husband beating his ––––––––– ––––––––– ––––––––– ––––––––– ––––––––– ––––––––– –––––––––
wife/experience of violence by husband Yes No Yes No Yes No Yes No Yes No Yes No Yes No
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
If she goes out without telling him
Ever experienced violence 19.5 16.9 25.2 22.3 u u 31.3 26.0 25.3 15.3 31.3 30.3 50.3 40.3
Experienced violence in the past
12 months 17.3 14.7 14.8 10.6 u u 27.8 18.0 13.9 8.3 15.4 12.9 26.3 21.0
Never experienced violence 80.5 83.1 74.8 77.7 u u 68.7 74.0 74.7 84.7 68.7 69.7 49.7 59.7
Number of women 658 1,710 250 6,539 u u 750 1,557 32,956 56,883 1,302 7,050 3,119 642
If she neglects the children
Ever experienced violence 19.5 16.7 26.0 22.2 44.2 25.3 32.2 25.7 25.1 14.9 31.3 30.3 50.6 44.9
Experienced violence in the past
12 months 17.8 14.3 15.4 10.4 23.4 13.7 27.8 18.5 13.6 8.2 15.5 12.7 26.7 23.0
Never experienced violence 80.5 83.3 74.0 77.8 55.8 74.7 67.8 74.3 74.9 85.1 68.7 69.7 49.4 55.1
Number of women 720 1,629 519 6,258 3,356 3,167 656 1,667 36,089 53,773 1,585 6,800 2,432 1,329
If she argues with him/answers back
Ever experienced violence 19.9 16.9 23.6 22.4 42.4 17.9 28.3 27.5 u u u u 50.0 46.7
Experienced violence in the past
12 months 18.6 14.5 15.4 10.7 23.0 9.0 22.5 20.9 u u u u 27.1 23.3
Never experienced violence 80.1 83.1 76.4 77.6 57.6 82.1 71.7 72.5 u u u u 50.0 53.3
Number of women 473 1,877 94 6,697 4,568 1,956 254 2,042 u u u u 2,108 1,651
If she refuses to have sex with him
Ever experienced violence 20.1 17.1 14.3 22.5 40.3 22.1 29.8 26.6 u u 30.2 30.4 51.3 45.7
Experienced violence in the past
12 months 17.6 15.0 11.5 10.8 21.3 12.4 25.7 19.5 u u 15.1 13.2 27.4 23.3
Never experienced violence 79.9 82.9 85.7 77.5 59.7 77.9 70.2 73.4 u u 69.8 69.6 48.7 54.3
Number of women 224 2,089 71 6,719 4,578 1,819 427 1,861 u u 576 7,759 1,990 1,735
If she does not prepare the food
Ever experienced violence 22.9 16.8 23.1 22.5 47.9 30.1 27.5 27.5 26.6 16.4 u u 51.1 46.1
Experienced violence in the past
12 months 21.2 14.6 15.0 10.6 27.7 15.2 22.9 20.7 14.9 8.8 u u 27.9 23.0
Never experienced violence 77.1 83.2 76.9 77.5 52.1 69.9 72.5 72.5 73.4 83.6 u u 48.9 53.9
Number of women 290 2,046 191 6,596 1,760 4,784 274 2,052 22,182 67,692 u u 1,807 1,962
If she talks to other men/he suspects
she is unfaithful
Ever experienced violence u u u u 39.7 26.6 u u 24.1 16.4 31.7 30.0 u u
Experienced violence in the past
12 months u u u u 20.2 16.1 u u 14.0 8.5 16.1 12.3 u u
Never experienced violence u u u u 60.3 73.4 u u 75.9 83.6 68.3 70.0 u u
Number of women u u u u 4,241 2,266 u u 29,548 59,953 2,176 6,136 u u
Agrees with any reason
Ever experienced violence 20.1 15.7 26.4 22.0 39.8 12.0 30.9 24.8 23.8 13.0 32.0 29.3 49.8 38.4
Experienced violence in the past
12 months 18.2 13.5 15.3 10.3 21.3 6.0 26.7 16.6 13.2 6.8 15.9 11.9 26.4 17.5
Never experienced violence 79.9 84.3 73.6 78.0 60.2 88.0 69.1 75.2 76.2 87.0 68.0 70.7 50.2 61.6
Number of women 969 1,434 638 6,168 5,433 1,166 1,000 1,345 49,689 40,614 2,756 5,752 3,337 455
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
u = Unknown (not available)

64 Domestic Violence and Women’s Empowerment • 4

and for almost all reasons, rates of violence are either similar or much higher among
women who agree that a husband is justified in beating his wife.
The last row in Table 4.3 shows a summary measure of the variation seen by spe-
cific reason. It shows how rates of violence vary between women who agree with at
least one reason, compared to those who agree with no reason. In every country, rates
of violence are higher for women who agree at all with wife beating than among
women who do not agree with any reason. Compared with rates among women who
do not agree with any reason, rates for the prevalence of ever-experience of violence
among women who agree with wife-beating are higher by about 9 percent in Nicara-
gua; 20 to 30 percent in Cambodia, the Dominican Republic, Haiti, and Zambia; 80
percent in India; and more than 200 percent in Egypt. Notably, these differentials are
even higher in the case of recent violence. While these results may in part be ex-
plained by possible rationalization by women of their own experience of violence, the
results are also compatible with the idea that agreement with norms about gender
roles does not protect women against spousal violence.
Table 4.4 examines whether agreement with the different reasons justifying a hus-
band beating his wife varies by women’s experience of violence. With few exceptions,
agreement is higher among women who have experienced violence than among those
who have never experienced violence, although this variation is small for several rea-
sons in many countries, particularly, the Dominican Republic, Haiti, and Nicaragua.
The higher levels of agreement among women who have experienced violence is quite
marked when agreement is measured not for each specific reason, but for any reason
at all (the last row of the table). The proportion of women who agree with at least
one reason is higher among women who have ever experienced violence than among
women who have never experienced violence by about 5 to 10 percent in Nicaragua
and Zambia and by 18 percent or more in the other countries. Further, in every
country, the differential in agreement is even higher between women who have re-
cently experienced violence and those who have never experienced violence.
A wife’s right to refuse sex to her husband: Women were asked if they thought a wife
is justified in refusing sex to her husband if: 1) she has recently given birth, 2) she
knows her husband has sex with other women; 3) she knows her husband has a sexu-
ally transmitted infection (STI); and 4) she is tired or not in the mood. Agreement
with one or more of these reasons can be understood as challenging gender norms,
thereby suggesting the likelihood that rates of violence will be higher among women
who agree with these reasons than among women who do not agree.

4 • Domestic Violence and Women’s Empowerment 65

Table 4.4 Percentage of ever-married women age 15-49 who agree with each different reason justifying wife-beating, by reason, among women who
have experienced violence by their husband, ever, in the last one year, or never
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Reason justifying a husband beating his wife/ Dominican
experience of violence by husband Cambodia Republic Egypt Haiti India Nicaragua Zambia
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
If she goes out without telling him
Percentage who agree among women who have
Ever experienced violence 30.6 4.1 u 36.5 48.8 15.8 85.4
Experienced violence in the last one year 30.9 5.0 u 42.4 49.0 17.9 85.5
Never experienced violence 26.7 3.5 u 30.2 33.6 15.1 79.3
If she neglects the children
Percentage who agree among women who have
Ever experienced violence 33.5 8.8 64.5 32.8 52.9 19.3 67.0
Experienced violence in the last one year 34.7 10.9 63.8 37.1 52.6 22.0 67.6
Never experienced violence 29.2 7.3 43.6 26.1 37.0 18.3 61.4
If she argues with him
Percentage who agree among women who have
Ever experienced violence 22.4 1.5 84.2 11.2 u u 57.4
Experienced violence in the last one year 23.8 2.0 85.5 11.6 u u 59.4
Never experienced violence 19.1 1.4 61.2 10.7 u u 53.9
If she refuses to have sex with him
Percentage who agree among women who have
Ever experienced violence 10.7 0.7 80.2 19.8 u 6.8 55.6
Experienced violence in the last one year 10.7 1.1 79.4 22.3 u 7.8 56.7
Never experienced violence 9.0 1.1 63.6 17.6 u 6.8 49.6
If she does not prepare the food
properly/on time/ burns the food
Percentage who agree among women who have:
Ever experienced violence 15.8 2.9 36.6 11.7 34.5 u 50.2
Experienced violence in the last one year 16.7 3.9 39.8 12.8 35.5 u 52.4
Never experienced violence 11.3 2.8 21.3 11.7 22.2 u 45.2
If she talks to other men/he suspects
she is unfaithful
Percentage who agree among women who have:
Ever experienced violence u u 73.2 u 41.6 26.8 u
Experienced violence in the last one year u u 69.7 u 44.5 31.2 u
Never experienced violence u u 59.5 u 30.6 25.0 u
Percentage who agree with at least one
reason among women who have
Ever experienced violence 46.3 11.0 93.9 48.2 69.1 34.4 90.5
Experienced violence in the last one year 47.7 13.4 94.3 54.4 70.3 39.0 91.7
Never experienced violence 39.1 8.9 76.1 40.6 51.7 31.5 85.7
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
u = Unknown (not available)

Table 4.5 shows mixed results with regard to the expectation of higher rates of
violence among women who agree than among women who do not agree that women
are justified in refusing sex to their husband. In most countries and for most reasons,
there is very little difference in violence rates by women’s agreement, and in several
countries, women who do not agree with a wife’s right to refuse sex have higher rates
of violence. These differences are also obvious when women who agree with all rea-
sons are compared with women who disagree with one or more reasons (last row in
Table 4.5). Rates of ever-experience of violence among women who agree with all
reasons are higher in Colombia but lower in Haiti than for women who disagree with
one or more reasons. In the rest of the countries, there is little or no variation at all.
Rates of recent experience of violence do not vary by agreement at all in Cambodia;
they are marginally higher among women who agree, in the Dominican Republic;
and they are lower in the remaining countries for which data are available.

66 Domestic Violence and Women’s Empowerment • 4

Table 4.5 Percentage of ever-married women age 15-49 who agree and who disagree with different reasons for which a wife is justified in refusing
to have sex with her husband, by whether they have ever experienced spousal violence, experienced spousal violence in the past one year, or
never experienced spousal violence, according to reason
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Whether woman agrees with each reason justifying a wife refusing sex to her husband
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Dominican
Reason for which a wife is justified in Cambodia Colombia Republic Haiti Nicaragua Peru Zambia
refusing her husband sex/ ––––––––– ––––––––– ––––––––– ––––––––– ––––––––– ––––––––– –––––––––
experience of violence by husband Yes No Yes No Yes No Yes No Yes No Yes No Yes No
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
She knows her husband has an STI
Ever experienced violence 18.6 16.0 44.5 34.7 22.6 18.1 27.8 24.7 u u 42.9 39.3 48.5 48.4
Experienced violence in the past
12 months 15.8 15.6 u u 10.9 7.5 21.0 20.7 u u u u 25.3 25.9
Never experienced violence 81.4 84.0 55.5 65.3 77.4 81.9 72.2 75.3 u u 57.1 60.7 51.5 51.6

Number of women 1,649 674 7,322 155 6,548 230 2,075 223 u u 15,575 400 3,336 397

She knows her husband has sex

with other women
Ever experienced violence 18.3 16.5 44.2 43.7 22.7 18.6 25.4 35.6 30.4 30.7 42.4 44.3 48.3 48.8
Experienced violence in the past
12 months 15.8 15.0 u u 10.8 10.1 20.6 23.4 13.2 14.8 u u 24.4 28.1
Never experienced violence 81.7 83.5 55.8 56.3 77.3 81.4 74.6 64.4 69.6 69.3 57.6 55.7 51.7 51.2

Number of women 1,627 687 6,776 605 6,188 547 1,786 490 7,917 437 16,021 695 2,821 887

She has recently given birth

Ever experienced violence 18.2 16.5 44.1 40.4 22.5 21.6 27.2 27.0 30.6 23.9 42.7 40.7 48.7 45.7
Experienced violence in the past
12 months 15.4 16.1 u u 10.7 10.5 20.2 24.7 13.3 11.7 u u 25.3 26.9
Never experienced violence 81.8 83.5 55.9 59.6 77.5 78.4 72.8 73.0 69.4 76.1 57.3 59.3 51.3 54.3

Number of women 1,705 646 7,180 319 6,508 266 2,044 269 8,214 178 16,368 479 3,494 273

She is tired or not in the mood

Ever experienced violence 17.9 16.8 44.8 41.1 22.6 21.7 26.3 30.6 30.2 33.9 42.6 43.6 48.3 49.8
Experienced violence in the past
12 months 15.0 16.4 u u 11.3 8.9 19.3 25.4 13.1 15.7 u u 25.2 25.9
Never experienced violence 82.1 83.2 55.2 58.9 77.4 78.3 73.7 69.4 69.8 66.1 57.4 56.4 51.7 50.2

Number of women 1,646 653 6,322 1,080 5,190 1,502 1,779 476 7,800 513 14,403 1,978 2,617 1,084

She agrees with all reasons

Ever experienced violence 18.0 16.6 45.3 40.4 22.4 22.5 23.6 34.3 30.0 31.5 42.4 42.5 48.0 48.9
Experienced violence in the past
12 months 15.4 15.4 u u 11.2 9.7 18.5 25.4 13.1 14.0 u u 23.7 27.3
Never experienced violence 82.0 83.4 54.7 59.6 77.6 77.5 76.4 65.7 70.0 68.5 57.6 57.5 52.0 51.1

Number of women 1,503 900 5,547 2,056 4,776 2,030 1,506 839 7,508 999 12,927 4,442 2,043 1,749
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
u = Unknown (not available)

There is also no clear pattern evident in the rates of agreement by experience of

violence (Table 4.6) across the different countries. In some countries, rates of agree-
ment are higher for women who have experienced violence, and in others, they are
either no different or lower. Additionally, the variation tends to be small in almost all
countries and for all reasons.
Overall, a review of Tables 4.5 and 4.6 shows that women’s views about wives be-
ing able to refuse sex with their husbands do not appear to be consistently related to
violence, either as a possible outcome or a possible cause.

4 • Domestic Violence and Women’s Empowerment 67

Table 4.6 Percentage of ever-married women age 15-49 who have experienced violence by their husband ever, in the past one year, or never, who
agree with different reasons for which a wife is justified in refusing to have sex with her husband, by reason
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Reason for which a wife is justified
in refusing sex to her husband/ Dominican
Experience of violence by husband Cambodia Colombia Republic Haiti Nicaragua Peru Zambia
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
She knows her husband has an STI
Percentage who agree among women who have
Ever experienced violence 73.1 97.3 96.8 89.8 u 90.6 88.0
Experienced violence in the last one year 70.3 u 97.3 88.5 u u 87.7
Never experienced violence 67.7 95.5 96.0 88.0 u 89.0 87.9

She knows her husband has sex with other women

Percentage who agree among women who have
Ever experienced violence 70.9 89.5 92.2 70.7 93.6 92.3 74.2
Experienced violence in the last one year 69.7 u 91.4 74.7 93.4 u 71.6
Never experienced violence 67.0 88.9 90.6 78.2 92.8 92.2 74.6

She has recently given birth

Percentage who agree among women who have
Ever experienced violence 74.0 94.7 95.7 86.7 97.6 94.9 92.7
Experienced violence in the last one year 71.0 u 95.2 83.9 97.2 u 91.9
Never experienced violence 70.3 94.2 95.6 87.3 96.1 93.8 91.6

She is tired or not in the mood

Percentage who agree among women who have
Ever experienced violence 70.3 84.6 76.7 72.9 91.6 83.2 92.7
Experienced violence in the last one year 67.0 u 80.5 69.7 91.3 u 91.9
Never experienced violence 68.1 82.0 76.1 77.0 91.7 82.8 91.6

Percentage who agree with all reasons among

women who have
Ever experienced violence 64.4 75.2 70.1 55.2 87.8 74.4 53.5
Experienced violence in the last one year 62.6 u 73.1 56.7 87.5 u 50.3
Never experienced violence 62.1 71.2 70.2 67.6 88.5 74.5 54.3
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
u = Unknown (not available)

4.3 Controlling Behaviors by Husbands and Domestic Violence

Male dominance over women can be manifested in many different ways, including
control over household decisionmaking, which was discussed earlier. In fact, some
male behaviors have been identified in the literature as risk factors for violence
(Campbell et al., 2003), and the World Health Organization includes coercive and/or
controlling behaviors in its definition of gender-based violence (World Health Or-
ganization, 2004). The DHS questionnaire sought information on different combi-
nations of six such behaviors, namely: whether the respondent’s husband is jealous or
angry if she talks to other men; he frequently accuses her of being unfaithful; he does
not permit her to meet her girlfriends; he limits her contacts with her family; he in-
sists on knowing where she is all the time; and he does not trust her with money. In
Peru, information was obtained for only one of these behaviors; in Colombia and
Haiti, it was sought only for five of the six behaviors; and in the remaining three
countries, information is available for all six behaviors. In the questionnaire, each of
these behaviors was described in a phrase, and women were asked whether the phrase

68 Domestic Violence and Women’s Empowerment • 4

Table 4.7 Percentage of ever-married women age 15-49 who have ever experienced violence by their husband, who have
experienced violence by their husband in the past one year, and who have never experienced violence by their husband, according to
whether their husband shows different controlling behaviors
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Respondent’s husband’s behavior
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Jealous/ Frequently Does not Insists on
angry if accuses her permit her Limits her knowing Does not
she talks to of being to meet contact with where she is trust her
other men unfaithful girlfriends her family all the time with money
––––––––– ––––––––– ––––––––– ––––––––– ––––––––– –––––––––
Experience of violence by husband Yes No Yes No Yes No Yes No Yes No Yes No
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced violence 38.1 12.7 50.2 12.9 41.6 15.7 44.9 15.9 40.7 14.1 39.7 14.3
Experienced in the past 12 months 32.1 11.5 43.5 11.4 38.5 13.7 39.3 14.0 35.3 12.5 35.2 12.5
Never experienced violence 61.9 87.3 49.8 87.1 58.4 84.3 55.1 84.1 59.3 85.9 60.3 85.7

Number of women 439 1,945 289 2,097 150 2,224 112 2,261 298 2,068 290 2,093

Colombia
Ever experienced violence u u 75.5 33.5 70.4 32.7 77.8 36.0 59.6 30.5 66.2 37.0
Experienced in the past 12 months u u u u u u u u u u u u
Never experienced violence u u 24.5 66.5 29.6 67.3 22.2 64.0 40.4 69.5 33.8 63.0

Number of women u u 1,909 5,693 2,280 5,322 1,458 6,141 3,521 4,079 1,820 5,782

Dominican Republic
Ever experienced violence 34.1 12.9 58.2 15.5 50.0 16.4 52.6 18.9 30.2 13.9 54.2 19.7
Experienced in the past 12 months 17.8 5.0 36.4 5.8 27.8 7.0 30.4 8.4 15.6 5.4 33.4 8.8
Never experienced violence 65.9 87.1 41.8 84.5 50.0 83.6 47.4 81.1 69.8 86.1 45.8 80.3

Number of women 3,047 3,710 1,112 5,681 1,237 5,543 732 6,059 3,571 3,224 545 6,226

Haiti
Ever experienced violence 36.1 15.1 46.3 19.9 37.3 22.3 u u 31.5 19.6 26.9 28.9
Experienced in the past 12 months 27.3 12.1 36.4 14.7 29.7 16.4 u u 25.5 12.1 22.3 20.8
Never experienced violence 63.9 84.9 53.7 80.1 62.7 77.7 u u 68.5 80.4 73.1 71.1

Number of women 1,398 854 683 1,637 815 1,512 u u 1,551 775 929 1,313

Nicaragua
Ever experienced violence 41.8 20.0 58.3 23.9 53.3 24.7 54.4 25.5 37.8 22.3 52.3 23.9
Experienced in the past 12 months 20.2 7.1 33.3 8.6 28.3 9.5 30.8 9.7 17.5 8.7 26.7 9.3
Never experienced violence 58.2 80.0 41.7 76.1 46.7 75.3 45.6 74.5 62.2 77.7 47.7 76.1

Number of women 3,959 4,391 1,600 6,798 1,682 6,717 1,417 6,978 4,417 3,983 1,938 6,424

Peru
Ever experienced violence u u u u u u 77.1 39.5 u u u u
Experienced in the past 12 months u u u u u u u u u u u u
Never experienced violence u u u u u u 22.9 60.5 u u u u

Number of women u u u u u u 1,359 16,010 u u u u

––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
u = Unknown (not available)

applied to their relationship with their husband. Table 4.7 shows how rates of vio-
lence vary between women whose husbands manifest each of these behaviors and
women whose husbands do not do so. Table 4.8 and Figure 4.2 summarize this in-
formation by examining how rates of violence vary with the number of controlling
behaviors manifested, rather than by any specific behavior. The expectation is that
rates of violence will be higher for women whose husbands show controlling behav-
iors and that they will rise with the number of controlling behaviors manifested.
Table 4.7 clearly shows that for each type of behavior listed, rates of violence are
much higher for women who say that their husband manifests the behavior than for
women who say he does not. In fact, in the case of most behaviors, the violence rates

4 • Domestic Violence and Women’s Empowerment 69

 Table 4.8 Percentage of ever-married women age 15-49 who have experienced violence by their
husband ever, in the past 12 months, or never, according to the number of marital control
behaviors shown by their husbands
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Number of marital control behaviors
shown by the respondent’s husband
––––––––––––––––––––––––––––––––––
Experience of violence by husband None 1-2 3-4 5-6
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced violence 10.2 26.9 49.0 46.2
Experienced violence in the last one year 9.2 23.3 41.2 43.2
Never experienced violence 89.8 73.1 51.0 53.8

Number of women 1,693 447 200 63

Colombia
Ever experienced violence 22.5 44.9 74.1 89.1
Experienced violence in the last one year u u u u
Never experienced violence 77.5 55.1 25.9 10.9

Number of women 3,040 2,687 1,435 440

Dominican Republic
Ever experienced violence 10.0 17.0 42.5 75.9
Experienced violence in the last one year 3.2 6.9 23.1 46.3
Never experienced violence 90.0 83.0 57.5 24.1

Number of women 2,215 3,050 1,143 399

Haiti
Ever experienced violence 11.8 19.8 39.9 41.2
Experienced violence in the last one year 8.8 13.4 32.8 31.3
Never experienced violence 88.2 80.2 60.1 58.8

Number of women 367 959 804 215

Nicaragua
Ever experienced violence 16.6 24.1 46.7 65.2
Experienced violence in the last one year 5.3 8.6 21.9 39.0
Never experienced violence 83.4 75.9 53.3 34.8

Number of women 2,651 3,402 1,577 878

–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
u = Unknown (not available)

are at least twice as high when a given behavior is manifested than when it is not, in
all countries. Further, the rates of violence experienced by women whose husbands
manifest any given behavior are remarkably high. For example, the prevalence of vio-
lence among women whose husbands frequently accuse them of being unfaithful
ranges from 46 percent in Haiti to 76 percent in Colombia. In the case of respon-
dents whose husbands limit their contact with their families, these rates vary from 45
percent in Cambodia to 78 and 77 percent in Colombia and Peru, respectively.
Table 4.8 shows further that the likelihood of violence increases with the number
of controlling behaviors manifested by a husband. For example, in the Dominican
Republic, the ever-experience of violence is only 10 percent among women whose
husbands do not manifest any of these behaviors, 17 percent among those who mani-
fest one or two of the behaviors, but then rises to 76 percent among women whose
husbands manifest all five or six of these behaviors. In Colombia, the progression in
violence rates by numbers of behaviors is from 23 percent among women whose hus-
bands show none of these behaviors to 89 percent among those whose husbands
show five or six of the behaviors.

70 Domestic Violence and Women’s Empowerment • 4

 Figure 4.2
Percentage of women who have ever experienced spousal
violence, according to the number of controlling behaviors
shown by their husbands
100

90 +
80
+ (
70
#
60

50 &
+ # &
40 (* *
30
#&
+
(*
20
#
10 &*
(
0
0 1-2 3-4 5-6
& Cambodia * Colombia + Dominican Republic # Haiti ( Nicaragua

Thus, the DHS data show clearly that the relationship of controlling behaviors
and the risk of violence is valid in all countries. Further, the likelihood of violence
escalates rapidly with increases in the number of such behaviors, so that the manifes-
tation of any given behavior is often associated with at least a doubling of violence
rates.
In conclusion, controlling behaviors manifested by husbands appear to be far more
important in increasing women’s risk of experiencing violence than their lack of em-
powerment, as measured by indicators of decisionmaking and acceptance of gender-
role norms.

4 • Domestic Violence and Women’s Empowerment 71

 5
Domestic Violence and Demographic and Health Outcomes

D omestic violence poses a direct threat to women’s health (Heise et al., 1994)
and also has adverse consequences for other indicators of women’s and chil-
dren’s health and well-being. Several reviews of the relevant literature (e.g., Heise et
al., 1999; Campbell, 2002) emphasize the linkages between the experience of domes-
tic violence and both fatal and nonfatal outcomes for women and their children. Fatal
outcomes related to domestic violence for women can result directly through homi-
cide or indirectly through suicide and maternal or AIDS-related mortality. Nonfatal
outcomes include manifestations of adverse mental, physical, and reproductive health
outcomes and negative health behaviors (Heise et al., 1999).
Included among the mental health problems found to occur more frequently
among abused women than among those who are not abused are higher rates of de-
pression, posttraumatic stress, and eating disorders. Poor physical health among
abused women manifests as chronic conditions including chronic pain, injuries, gas-
trointestinal disorders, and generally poor health status among others. Abused
women’s reproductive health is also compromised through much higher rates of gy-
necological problems, HIV and sexually transmitted infections (STIs), miscarriages,
abortions, unwanted pregnancy, and low birth weight (Campbell, 2002). Negative
health behaviors include overeating, alcohol and drug abuse, and sexual risk-taking.
Although the pathways from maternal experience of violence to the health and sur-
vival of children are not well understood, research provides evidence of increased
mortality and undernutrition among children of abused mothers (Jejeebhoy, 1998;
Ganatra et al., 1998; Asling-Monemi et al., 2003).
The Demographic and Health Surveys does not collect information on mental and
physical health outcomes identified as sequelae of the experience of violence. How-
ever, data on the nutritional status of women can provide a summary measure of
women’s general health. Compromised nutritional status, especially in the form of
being extremely underweight or obese, is not only a risk factor for diseases such as
hypertension and diabetes, but is also hypothesized as one of the many adverse health
outcomes of having experienced violence (Heise et al., 1999). The DHS survey also
provides extensive information on women’s reproductive health and the health and
survival of children. Thus, this chapter examines how several different demographic
and health outcomes for women and children vary by women’s experience of spousal
violence (physical and sexual). Specifically, Section 5.1 examines women’s physical
health through two measures of nutritional status: body mass index (BMI) and ane-
mia status. Women’s reproductive health is discussed in Section 5.2. Indicators of
women’s reproductive health include measures of women’s fertility and their ability to
have only the children they want and when they want them, the occurrence of
nonlive births, and the self-reported prevalence of STIs. Women’s access to maternal
health care is discussed in Section 5.3. Section 5.4 presents data on infant and child

5 • Domestic Violence and Demographic and Health Outcomes 73

 mortality and on the nutritional and immunization status of children. The DHS sur-
vey typically does not have any measures of mental health for women or children.

5.1 Measures of Women’s Nutritional Status

The BMI is an important measure of women’s nutritional status and is based on a
woman’s weight in relation to her height (Shetty and James, 1994). The BMI is de-
2 7
fined as weight in kilograms divided by the height in meters squared (kg/m ). Three
categories of malnourishment are defined: a BMI of less than 18.5 indicates chronic
energy deficiency; a BMI between 25.0 and 29.9 indicates overweight; and a BMI of
30.0 or higher indicates obesity. While chronic energy deficiency is associated with
higher mortality and morbidity, overweight and obesity are known risk factors for
many health conditions, including diabetes, heart disease, hypertension, gall bladder
disease, and some forms of cancer. Research suggests that the risk of severe obesity is
higher among abused women, particularly women who have experienced sexual or
nonsexual abuse in childhood (Felitti, 1993; Williamson et al., 2002). Table 5.1
Notably, in most
shows the percentages of women who fall into the different malnourishment catego-
countries, obesity is
unrelated to the ries, according to whether they have experienced violence ever, in the past 12
recent experience of months, or never. The BMI measures exclude women who were pregnant at the time
violence, even when it of the survey or women who gave birth during the two months preceding the survey.
is higher among Table 5.1 does not provide support for a consistent or significant positive relation-
women who have ship between the experience of violence by women and their nutritional status. With
ever experienced the exception of India, in no other country does the prevalence of underweight, over-
violence.
weight, or obesity vary substantially by women’s violence status. In India, women
who have experienced violence are more likely (41 percent) than those who have
never experienced violence (34 percent) to be underweight. However, compared with
women who have experienced violence, women who have never experienced violence
are more likely to be overweight or obese. Obesity is somewhat more common among
women who have ever experienced violence in Egypt, Haiti, Nicaragua, and Peru, but
the differentials tend to be small. Notably, in most countries, obesity is unrelated to
the recent experience of violence, even when it is higher among women who have
ever experienced violence.
For five of the nine countries, Table 5.1 also shows how the prevalence of anemia
varies by violence status of women. Anemia is characterized by a low level of hemo-
globin in the blood. Anemia usually results from a nutritional deficiency of iron,
8
folate, vitamin B12, or certain other nutrients. This type of anemia is commonly
referred to as iron-deficiency anemia and provides another important measure of
malnourishment. Anemia may have detrimental effects on the health of women and
children and can become an underlying cause of maternal mortality and perinatal
mortality. In Table 5.1, nonpregnant women with a hemoglobin level below 11.9
grams/deciliter (g/dl) and pregnant women with a hemoglobin level below 10.9 g/dl
are considered anemic (Centers for Disease Control and Prevention, 1998).

7
In pounds and inches, the formula for BMI is
BMI = 730 × [(weight in pounds)/(height in inches)]2
8
Anemia can also be caused by malaria or worm infestation.

74 Domestic Violence and Demographic and Health Outcomes • 5

Table 5.1 Percentage of ever-married women 15-49 who are underweight (BMI<18.5), overweight (BMI 25.0-
29.9), and obese (BMI>30.0) and percentage who are anemic, by whether women have experienced violence
by their husband ever, in the past 12 months, or never
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Body mass index (BMI) Anemia
––––––––––––––––––––––––––––––––––––– ––––––––––––––––––
Number Percentage Number
Experience of violence 25.0- of with of
by husband <18.5 29.9 30.0+ women anemia women
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced 20.2 7.2 0.3 415 60.3 413
Experienced in the last year 21.1 5.7 0.1 365 60.3 364
Never experienced 18.7 7.6 1.5 1,958 58.5 1942

Colombia
Ever experienced 2.6 33.5 11.0 1,322 u u
Experienced in the last year u u u u u u
Never experienced 2.6 32.9 12.0 1,906 u u

Egypt
Ever experienced 1.4 29.0 21.0 1,340 u u
Experienced in the last year 1.4 28.4 19.4 776 u u
Never experienced 0.9 34.7 18.9 2,191 u u

Haiti
Ever experienced 6.0 18.2 13.7 637 55.3 605
Experienced in the last year 6.2 22.0 5.1 489 53.0 472
Never experienced 8.0 22.9 9.7 1,641 53.5 1,598

India
Ever experienced 41.2 5.4 1.1 15,929 55.9 15,504
Experienced in the last year 43.7 4.2 0.8 8,587 57.3 8,381
Never experienced 33.8 8.7 2.4 66,945 50.7 65,189

Nicaragua
Ever experienced 2.4 32.8 17.8 2,493 u u
Experienced in the last year 2.7 30.2 15.9 1,094 u u
Never experienced 2.7 33.2 15.5 5,713 u u

Peru
Ever experienced 0.0 40.7 18.3 7,130 32.5 1,768
Experienced in the last year u u u u u u
Never experienced 0.0 38.8 18.0 9,605 32.8 2,170

Zambia
Ever experienced 10.4 11.6 3.7 1,813 u u
Experienced in the last year 10.1 10.7 2.6 949 u u
Never experienced 13.4 10.6 3.8 1,917 u u
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Note: Pregnant women and women with a birth in the preceding two months are excluded from the data on
BMI.
u = Unknown (not available)

As in the case of the BMI-based malnourishment measures, there is no significant

association between anemia and the experience of violence. Although in three of the
four countries for which data are available, women who have ever experienced vio-
lence are more likely to be anemic than women who have not experienced violence,
the differentials tend to be small. Only in India is the differential in the prevalence of
anemia more substantial, particularly between women who have recently experienced
violence (57 percent are anemic) and women who have never experienced violence
(51 percent are anemic).
A possible reason why no or only a weak association is found between malnour-
ishment and violence is that the measure of violence being used does not appropri-
ately capture the types of violence that are best known to be associated with obesity.
For example, Felitti (1993) found obesity to be positively associated with childhood

5 • Domestic Violence and Demographic and Health Outcomes 75

 abuse. However, the measure of violence shown in Table 5.1 captures only the ex-
perience of spousal violence. Any violence by anyone other than the spouse is ex-
cluded, thereby effectively excluding childhood abuse of any kind.

5.2 Indicators of Women’s Reproductive Health and Spousal

Violence
An extensive literature suggests that the experience of physical and sexual abuse
increases the risk of unwanted pregnancies and STIs. The hypothesized pathways are
both direct and indirect (Heise et al., 1999). Lack of sexual autonomy and control in
the face of actual or threatened violence and fear of repercussion if contraception is
used or condom use is requested are direct pathways to unwanted pregnancy and in-
creased risk of HIV/STIs. Any secret use of contraception by women can mediate
these effects (Population Council, 1998). Indirect pathways include high-risk sexual
behavior, such as multiple partners and unprotected sex, low self-esteem, and com-
promised mental health. Gynecological problems are also more common among
women who have experienced spousal abuse (Campbell, 2002), including those re-
lated to STIs. Adverse pregnancy outcomes have also been related to abuse (Je-
jeebhoy, 1998; Pearlman et al., 1990). A metastudy of the literature, however, finds
only mixed support for consistently adverse pregnancy outcomes (Petersen et al.,
1997). In addition, the experience of violence can lead to pregnancy loss through de-
liberate termination of pregnancy (Glander et al., 1998).
This section examines whether women who have experienced violence differ from
those who have not in terms of their fertility, wantedness of their births, timing of
births, and contraceptive use. This is followed by a discussion of the variation in
women’s lifetime experience of nonlive births as a proxy for abortions and miscar-
riages. Women’s access to reproductive health care before and during delivery is ex-
amined next. The section ends with a discussion of whether the reported prevalence
of STIs varies by women’s experience of violence.

5.2.1 Fertility-related indicators

Lack of sexual autonomy associated with the experience of domestic violence can
have several different fertility-related outcomes, including a large number of births,
births that are unwanted, short intervals between births, and low contraceptive use,
especially in relation to expressed need for fertility control. Each of these outcomes is
examined in relation to women’s violence status in the discussion below.
Do women who experience violence have more births than women who have not
experienced violence?
To compare fertility between women who have experienced violence and women
who have not, Table 5.2 shows the mean number of children ever born to ever-
married women by age group and the ever-experience of violence. Research reveals an
association of violence and higher fertility, although the direction of causality remains
unclear (Ellsberg et al., 1999).
Table 5.2 shows that ever-married women age 15-49 who have ever experienced
violence have a higher number of children ever born in all nine countries than women
who have never experienced violence. In six of these countries, women who have ex-
perienced violence have, on average, at least half a child more than women who have
not experienced violence.

76 Domestic Violence and Demographic and Health Outcomes • 5

 Table 5.2 Mean number of children ever born to ever-married women age 15-49 by age in years,
according to whether they have ever experienced violence by their husband or not
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Age groups Ever-married
Experience of violence ––––––––––––––––––––––––––––––––––––– women
by husband 15-19 20-29 30-39 40-49 15-49
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced * 2.5 4.2 6.2 4.3
Never experienced 0.5 1.9 4.0 5.6 3.8

Colombia
Ever experienced 0.9 2.0 2.9 4.0 2.9
Never experienced violence 0.7 1.6 2.5 3.4 2.4

Dominican Republic
Ever experienced 1.0 2.2 3.5 4.1 3.0
Never experienced violence 0.8 1.9 3.0 3.8 2.7

Egypt
Ever experienced 0.9 2.5 4.3 6.1 4.0
Never experienced 0.5 1.9 3.9 5.2 3.4

Haiti
Ever experienced (0.6) 2.0 4.8 5.7 3.9
Never experienced violence 0.8 2.0 4.3 5.8 3.8

India
Ever experienced 0.9 2.5 4.0 4.8 3.4
Never experienced violence 0.6 2.1 3.6 4.4 2.9

Nicaragua
Ever experienced 1.0 2.6 4.5 5.9 3.9
Never experienced violence 0.9 2.2 3.9 5.3 3.3

Peru
Ever experienced 0.9 2.0 3.6 5.3 3.6
Never experienced violence 0.7 1.7 3.1 4.2 2.9

Zambia
Ever experienced 1.0 2.6 4.9 7.3 4.0
Never experienced violence 0.7 2.3 5.0 7.1 3.9
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Note: Figures in parentheses are based on 25-49 unweighted cases. An asterisk indicates that a
figure is based on fewer than 25 unweighted cases and has been suppressed.

Since the likelihood of having ever experienced violence varies by age and age dis-
tributions vary among countries, Table 5.2 also shows the mean number of births by
age group. As expected, the average number of children ever born increases with
women’s age as women proceed through their family-building process. What is nota-
ble, however, is that the mean number of births in most age groups and countries
tends to be higher for women who have experienced violence than for women who
have not. Specifically, fertility in all age groups (with a sufficient number of cases) is
higher for women who have ever experienced violence, compared with women who
have never experienced violence in Cambodia, Colombia, the Dominican Republic,
Egypt, India, Nicaragua, and Peru. Notably, in Egypt and Peru, women age 40-49
(ages by which most women have completed their family-building process) have
about one child more if they have ever experienced violence than if they have never
experienced violence. In Haiti and Zambia, ever-experience of violence is associated
with higher fertility in at least two of the four age groups.
Overall, these data show unequivocally that fertility for women who have experi-
enced violence is higher than that for women who have not.

5 • Domestic Violence and Demographic and Health Outcomes 77

 Are abused women more likely to have mistimed or unwanted births?
Unwanted fertility can be investigated by examining women’s responses to the
question that directly asks, for all births in the past five years (three years in India),
whether at the time of conception, the birth was wanted then, wanted later, or not
wanted at all. These data are shown in Table 5.3 by women’s experience of violence.

Table 5.3 Percent distribution of births during the five years (three years in India) preceding the survey and
current pregnancies by fertility planning status, according to whether the mother has ever experienced
violence by her husband or not
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Not Number
Mother’s experience Wanted Wanted wanted of
of violence by husband then later at all Missing Total births
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced 59.7 10.9 29.2 0.2 100.0 414
Never experienced 70.5 7.9 21.1 0.4 100.0 1,654

Colombia
Ever experienced 41.7 26.6 31.5 0.1 100.0 1,908
Never experienced 54.6 27.5 17.8 0.1 100.0 2,631

Dominican Republic
Ever experienced 45.8 30.4 22.0 1.8 100.0 1,020
Never experienced 59.1 28.2 11.5 1.1 100.0 3,450

Egypt
Ever experienced 63.6 10.8 25.0 0.6 100.0 2,350
Never experienced 70.8 10.5 18.3 0.4 100.0 3,835

Haiti
Ever experienced 47.6 25.4 27.0 0.0 100.0 603
Never experienced 42.0 25.9 31.9 0.2 100.0 1,719

India
Ever experienced 72.4 13.0 14.2 0.4 100.0 7,341
Never experienced 79.8 11.7 8.3 0.2 100.0 31,769

Nicaragua
Ever experienced 59.3 15.3 23.9 1.5 100.0 2,145
Never experienced 68.5 15.8 14.6 1.1 100.0 5,098

Peru
Ever experienced 38.7 23.5 37.6 0.1 100.0 4,736
Never experienced 48.4 24.5 26.9 0.3 100.0 7,446

Zambia
Ever experienced 58.9 22.1 18.8 0.3 100.0 2,198
Never experienced 63.5 18.4 17.9 0.2 100.0 2,248

In all countries except Haiti, women who have experienced violence are consis-
tently less likely to say that their birth was wanted when it was conceived, compared
with women who have never experienced violence. The absolute differences in these
proportions are relatively large (9 percentage points or more) in five of the nine coun-
tries. For example, in Cambodia, 60 percent of women who have ever experienced
violence say that the birth was wanted at the time it was conceived, compared with 71
percent of women who have never experienced violence. The differential in Colombia
is even larger, with 42 percent of women who had experienced violence saying that
the birth was wanted then, compared with 55 percent among women who had never
experienced violence. In Haiti, by contrast, women who have experienced violence
are somewhat more likely than women who have never experienced violence to say
that the birth was wanted at the time of conception.

78 Domestic Violence and Demographic and Health Outcomes • 5

 The hypothesized lack of sexual autonomy among abused women suggests that
abused women should be at a greater risk of having a mistimed as well as an un-
wanted birth. Contrary to expectations, however, the likelihood of a birth being mis-
timed varies little or not at all by violence status of the mother, although the likeli-
hood of having a birth that is not wanted at all is consistently higher among women
who have experienced violence than among those who have not, in all countries ex-
cept Haiti. In fact, in this bivariate analysis, the likelihood of a woman having a birth
that was not wanted at all is 37 to 40 percent higher in Cambodia, Egypt, and Peru if
she has experienced violence than if she has not. This differential is even higher in all
the remaining countries except Haiti and Zambia. In the Dominican Republic,
women who have experienced violence are almost twice as likely as those who have
not to have a birth that was not wanted at all. This difference is only 5 percent in
Zambia. In Haiti, by contrast, women who have never experienced violence are about
18 percent more likely to have a birth that was not wanted at all, compared with
women who have experienced violence. Further, in all countries except Haiti, births
not wanted at all are disproportionately born to women who have experienced vio-
lence. For example, in Colombia, only 42 percent of all births born in the five years
preceding the survey were born to women who have experienced violence (see Figure
5.1), but 56 percent of those not wanted at all.

Figure 5.1
Percentages of all births and all births not wanted at all
born to women who have ever experienced violence
60

50

40

30

20

10

0
Cambodia Colombia Dominican Egypt Haiti India Nicaragua Peru
Republic
All births All births not wanted
at all

Are women who experience violence less successful than other women
in being able to space their births?
In order to examine this question, Table 5.4 shows the cumulative percent distri-
bution of births during the five years (three years in India) before the survey, by the
interval since the last birth, separately for births to women who have and have not
experienced violence. Also shown is the median number of months since previous
birth. First births are excluded since they do not have a preceding birth.

5 • Domestic Violence and Demographic and Health Outcomes 79

 Table 5.4 Cumulative percent distribution of births during the five years (three years in India) preceding the survey by
the interval since the previous birth and the median number of months since previous birth according to whether the
mother has ever experienced violence by her husband or not
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Months since previous birth Median Number
Mother’s experience of –––––––––––––––––––––––––––––––––––––––––––– months since of
violence by husband <12 12-17 18-23 24-35 36-47 48+ previous birth births
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced 0.7 10.0 24.2 61.0 82.0 100.0 30.2 327
Never experienced 1.6 8.4 21.1 54.2 74.9 100.0 34.3 1,158

Colombia
Ever experienced 0.9 11.3 27.8 52.2 65.3 100.0 34.9 1,247
Never experienced 1.1 13.0 25.8 46.1 60.3 100.0 38.8 1,455

Dominican Republic
Ever experienced 1.4 16.6 30.1 52.4 68.0 100.0 34.3 711
Never experienced 1.3 13.0 27.4 54.9 69.7 100.0 33.2 2,055

Egypt
Ever experienced 1.7 12.9 27.4 61.7 79.7 100.0 31.4 1,723
Never experienced 1.6 12.8 26.5 57.4 75.5 100.0 32.2 2,416

Haiti
Ever experienced 0.9 10.9 27.3 65.5 83.0 100.0 30.1 424
Never experienced 2.0 11.6 27.1 64.8 82.7 100.0 30.5 1,175

India
Ever experienced 1.5 8.9 21.6 56.7 79.5 100.0 33.1 5,023
Never experienced 1.5 9.6 23.8 56.8 78.6 100.0 33.3 18,246

Nicaragua
Ever experienced 2.5 16.8 32.7 60.5 75.9 100.0 30.0 1,604
Never experienced 2.0 15.8 31.9 59.9 75.2 100.0 30.6 3,459

Peru
Ever experienced 0.8 7.0 18.7 47.8 64.0 100.0 37.3 3,361
Never experienced 0.6 7.8 21.9 49.1 65.4 100.0 36.4 4,640

Zambia
Ever experienced 1.2 5.9 15.4 60.5 81.7 100.0 32.7 1,624
Never experienced 0.8 5.9 16.7 59.4 82.2 100.0 33.4 1,582

The table only weakly supports (and in only a few countries) the hypothesis that
women who have experienced violence are less likely to be able to space their births
than other women. In all countries except Cambodia and Colombia, the median
birth interval for second or higher order births is very similar for women who have
and have not experienced violence. In Cambodia and Colombia, the median birth
interval is shorter for births to women who have experienced violence than for births
to other women, but the difference is only about four months.
An examination of the cumulative distribution of births by birth interval shows
that in all countries except India, Peru, and Zambia, the proportion of births born
after a short birth interval (less than two years) was higher for women who have ex-
perienced violence than for those who have not. However, in most countries the dif-
ference is very small. This difference becomes somewhat wider if the proportion of
births born within 36 months is examined. Overall, the data suggest that even in
countries where birth intervals are generally shorter for women who experience vio-
lence compared with other women, the differences tend to be quite small.

80 Domestic Violence and Demographic and Health Outcomes • 5

5.2.2 Contraceptive use and contraceptive need
The high level of unwanted pregnancies associated with violence hypothesized in
the research on violence implies a low level of contraceptive use by women coupled
with a high level of need for fertility control. Lower contraceptive use among women
who have experienced violence is also suggested by the higher fertility and unwanted-
ness of the last birth, as seen earlier. Further, research on contraceptive use suggests
that women are often reluctant to raise the issue of contraception with their partners
for fear of reprimand or violence (Bawah et al., 1999; Blanc et al., 1996). However,
the fact that birth intervals do not vary significantly by violence status of the mother
argues against large contraceptive use differences. Evidence of clandestine contracep-
tive use by women (Biddlecom and Fapohunda, 1998; Population Council, 1998)
suggests that at least some women do not depend on their partner’s approval when
they want to control their fertility. That women may, in fact, be quite resourceful,
even in the face of violence, is also suggested by the much higher rates of women In all countries except
making contraception-related decisions on their own among ever-abused, as also re- India, women who
cently abused women, compared with women who have never experienced violence have never
experienced violence
(Table 4.1). In light of this ambiguity, women’s contraceptive use is examined in
are more likely than
some detail in this section. other women to have
Table 5.5 shows the percent distribution of currently married women who have never used
ever experienced spousal violence, experienced violence in the past one year, and contraception. In
never experienced spousal violence, according to whether they have ever used contra- other words, in most
ception. The table shows women according to whether they are currently using con- countries, violence is
traception (a traditional or a modern method), have ever used but are not currently associated with
slightly higher rates of
using, or have never used any contraception at all. In all countries except India, ever-use of
women who have never experienced violence are more likely than other women to contraception.
have never used contraception. In other words, in most countries, violence is associ-
ated with slightly higher rates of ever-use of contraception.
A higher rate of ever-use of contraception does not necessarily imply higher rates
of current use, however. In fact, in Cambodia, the Dominican Republic, Egypt, In-
dia, and Zambia, current use of contraception does not vary much by women’s ever-
experience of violence. The only countries where there is a somewhat larger variation
in current contraceptive use by women’s ever-experience of violence are Nicaragua
and Peru, and in these countries, women who have experienced violence are more
likely (not less likely) to be currently using contraception than women who have not
experienced violence. Further, in about half of the countries, women who have ever
experienced violence are more likely to be using a modern contraceptive method,
compared with women who have never experienced violence. In addition, in all coun-
tries except India, there is virtually no difference in current contraceptive use rates
between women who have ever experienced violence and women who have recently
experienced violence. In India and Zambia, current contraceptive use is lower among
women who have recently experienced violence, compared with all other women.
In some countries, the differences in ever-use of contraception by violence status
of women are explained by the fact that women who have experienced violence are
more likely to have discontinued use of contraception. They had used contraception
in the past but were not using at the time of the survey. Notably, the rate of discon-
tinuation is perforce likely to be low in countries where sterilization constitutes a high

5 • Domestic Violence and Demographic and Health Outcomes 81

 Table 5.5 Percent distribution of currently married women age 15-49 by their use of contraception, according to whether
they have experienced violence by their husband ever, in the past 12 months, or never
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Currently using contraception
––––––––––––––––––––––––– Ever used Never
Using a Using a but not used Number
Experience of violence modern traditional currently contra- of
by husband method method Total using ception Total women
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced 19.9 1.9 21.8 21.9 56.3 100.0 341
Experienced in the past year 20.0 1.5 21.5 23.1 55.4 100.0 308
Never experienced 18.0 5.9 24.0 12.7 63.3 100.0 1,737

Colombia
Ever experienced 65.9 12.7 78.6 18.6 2.8 100.0 2,320
Experienced in the past year u u u u u 100.0 u
Never experienced 62.9 13.1 76.0 18.1 5.9 100.0 3,577

Dominican Republic
Ever experienced 66.7 4.6 71.2 21.4 7.4 100.0 1,004
Experienced in the past year 66.2 6.8 72.9 20.3 6.8 100.0 522
Never experienced 66.1 4.3 70.4 18.3 11.3 100.0 4,263

Egypt
Ever experienced 44.8 1.6 46.4 25.5 28.2 100.0 2,301
Experienced in the past year 44.3 1.2 45.6 24.7 29.7 100.0 1,228
Never experienced 45.7 2.3 47.9 21.4 30.7 100.0 4,298

Haiti
Ever experienced 22.9 4.8 27.7 36.6 35.7 100.0 484
Experienced in the past year 23.3 4.7 28.0 37.8 34.1 100.0 411
Never experienced 19.1 5.9 25.0 24.6 50.4 100.0 1,512

India
Ever experienced 43.3 4.2 47.6 6.1 46.3 100.0 15,706
Experienced in the past year 38.5 4.0 42.4 6.3 51.3 100.0 9,002
Never experienced 42.7 5.6 48.4 7.1 44.5 100.0 68,976

Nicaragua
Ever experienced 61.5 2.8 64.3 22.2 13.5 100.0 1,918
Experienced in the past year 62.4 2.6 65.0 22.2 12.7 100.0 909
Never experienced 57.2 3.1 60.3 19.1 20.6 100.0 4,873

Peru
Ever experienced 53.4 17.9 71.4 20.8 7.9 100.0 6,101
Experienced in the past year u u u u u 100.0 u
Never experienced 48.9 18.6 67.5 20.9 11.6 100.0 9,077

Zambia
Ever experienced 25.9 9.2 35.2 38.0 26.9 100.0 1,452
Experienced in the past year 24.9 8.2 33.1 37.5 29.4 100.0 834
Never experienced 25.0 8.6 33.6 33.5 33.0 100.0 1,605
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
u = Unknown (not available)

proportion of all contraceptive use, such as India, where sterilization accounts for
about two-thirds of all contraceptive use, and to some extent Colombia and Nicara-
gua, where it accounts for about one-third of contraceptive use. Higher rates of dis-
continuation defined in this crude way are most evident, in Cambodia, Haiti, and
Zambia, for women who have experienced violence, compared with women who have
not. For example, in Cambodia, 22 percent of women who have ever experienced
violence have ever used contraception but are not currently using it; among women
who have never experienced violence, this proportion is only 13 percent. In Zambia,
the corresponding proportions are 38 percent among abused women and 34 percent
among those who have not experienced violence. This difference is also evident in the

82 Domestic Violence and Demographic and Health Outcomes • 5

Dominican Republic, Egypt, and Nicaragua, but in these countries, the differences
are much smaller. Multivariate analyses of these data show that even after all relevant
controls are introduced into the equations, the experience of violence increases the
likelihood of having ever used, but not currently using, contraception among ever-
married women by 77 percent in Cambodia, 44 percent in Haiti, and 14 percent in
the Dominican Republic (Kishor and Johnson, 2003.)
Overall, Table 5.5 gives a mixed picture of how contraceptive use varies by vio-
lence status of women. Perhaps the only conclusion that is consistently apparent for
most countries is that violence is not associated with lower rates of ever or current
contraceptive use. Further, for some countries, the data strongly suggest that women
who have experienced violence are not only more likely to have tried contraception,
but are also more likely to have discontinued it. Perhaps in these countries, violence
interferes not with women’s ability to use contraception, but to use it consistently and
continuously to meet their fertility desires.
Little or no variation in current contraceptive use by violence status does not
eliminate the possibility that unmet need for contraception will vary by violence
status of women. If need for contraception is higher among women who have experi-
enced violence than among those who have not, a similar rate of current contracep-
tive use among these two groups would imply that more of the contraceptive needs of
women who have never experienced violence have been met, compared with those of
women who have experienced violence. Accordingly, Table 5.6 shows the total need
for family planning, unmet need (for spacing and for limiting), and the percentage of
the total need that is being satisfied, according to women’s violence status. Currently
married women who are not using any method of contraception but who do not want
any more children or want to wait two or more years before having another child are
defined as having an unmet need for family planning. Current contraceptive users are
said to have met the need for family planning. The total demand for family planning
is the sum of the met need and the unmet need. The footnotes in the table provide
detailed definitions of these concepts.
As hypothesized, women who have experienced violence tend to have higher total
need for family planning than women who have not experienced violence. This is
true in all countries except India. Despite the somewhat higher current contraceptive
use rates in some countries among women who have experienced violence, the higher
need is also manifested in higher unmet need in most countries. More specifically, in
seven countries, total unmet need is higher among women who have experienced vio-
lence than among women who have not experienced violence, and in all countries,
women who have ever experienced violence have higher unmet need for limiting
births than women who have never experienced violence. Since both the total need
and met need (current contraceptive use) can vary with violence, the percentage of
need that is satisfied is given in Table 5.6. Overall, the percentage of need satisfied is
about the same by violence status of women in Colombia, Haiti, India, Nicaragua,
and Zambia; in the remaining countries, it is lower for women who have ever experi-
enced violence than for those who have never experienced violence. Notably, how-
ever, the percentage of need that is satisfied tends to be lowest among women who
have recently experienced violence. This is true in all countries for which information
on recent violence is available, except Haiti.

5 • Domestic Violence and Demographic and Health Outcomes 83

 Table 5.6 Percentage of currently married women with need for family planning, percentage with unmet need
for family planning, and percentage of total need satisfied by whether they have experienced violence by their
husbands ever, in the past 12 months, or never
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
2
Total Unmet need
need for ––––––––––––––––– Percentage Number
Experience of violence family For For of need of
1
by husband planning spacing limiting Total satisfied women
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced 64.8 17.1 25.9 43.0 33.7 341
Experienced in the last year 65.1 17.5 26.1 43.7 33.0 308
Never experienced 55.1 16.9 14.2 31.1 43.5 1,737

Colombia
Ever experienced 84.3 1.9 3.8 5.8 93.2 2,320
Experienced in the last year u u u u u u
Never experienced 82.6 3.2 3.4 6.6 92.0 3,577

Dominican Republic
Ever experienced 84.6 7.5 5.9 13.4 84.2 1,004
Experienced in the last year 89.7 9.4 7.4 16.8 81.3 522
Never experienced 80.9 6.2 4.4 10.5 87.0 4,263

Egypt
Ever experienced 67.7 6.3 15.0 21.3 68.5 2,301
Experienced in the last year 67.7 8.4 13.7 22.1 67.3 1,228
Never experienced 64.2 6.9 9.3 16.3 74.7 4,298

Haiti
Ever experienced 70.5 13.5 29.3 42.8 39.3 484
Experienced in the last year 70.6 15.6 27.0 42.5 39.7 411
Never experienced 64.9 15.1 24.8 39.9 38.6 1,512

India
Ever experienced 64.4 7.3 9.6 16.8 73.9 15,706
Experienced in the last year 61.5 9.1 10.0 19.1 69.0 9,002
Never experienced 64.1 8.4 7.3 15.8 75.4 68,976

Nicaragua
Ever experienced 80.0 5.7 10.0 15.7 80.4 1,918
Experienced in the last year 84.0 8.0 11.0 19.0 77.4 909
Never experienced 75.0 6.4 8.3 14.7 80.4 4,873

Peru
Ever experienced 81.4 2.9 7.2 10.1 87.6 6,101
Experienced in the last year u u u u u u
Never experienced 78.1 3.7 6.9 10.6 86.4 9,077

Zambia
Ever experienced 64.6 14.8 14.6 29.5 54.4 1,452
Experienced in the last year 63.0 16.6 13.3 29.9 52.5 834
Never experienced 61.3 14.2 13.5 27.7 54.8 1,605
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
1
Total need includes of the percentages of women with met need and unmet need. Met need is the percentage
of women who are currently using contraception for spacing or for limiting. Nonusers who are pregnant or
amenorrheic and women whose pregnancy was the result of a contraceptive failure are not included in the
category of unmet need but are included in total need for contraception (since they would have been using had
their method not failed).
2
Unmet need for spacing includes pregnant women whose pregnancy was mistimed, amenorrheic women who
are not using family planning and whose last birth was mistimed, and fecund women who are neither pregnant
nor amenorrheic and who are not using any method of family planning and say they want to wait two or more
years for their next birth. Also included in unmet need for spacing are fecund women who are not using any
method of family planning and say they are unsure whether they want another child or who want another child
but are unsure when to have the birth unless they say it would not be a problem if they discovered they were
pregnant in the next few weeks. Unmet need for limiting refers to pregnant women whose pregnancy was
unwanted, amenorrheic women whose last child was unwanted, and fecund women who are neither pregnant
nor amenorrheic, who are not using any method of family planning, and who want no more children. Excluded
from the unmet need category are pregnant and amenorrheic women who became pregnant while using a
method (these women are in need of a better method of contraception).
u = Unknown (not available)

84 Domestic Violence and Demographic and Health Outcomes • 5

 This discussion of contraceptive use and need for family planning does suggest
that the experience of violence is associated with greater need for family planning.
However, the greater need among women who have experienced violence, compared
with that among women who have not, does not always imply that these women have
consistently higher unmet need or lower contraceptive use than women who have
never experienced violence. In fact, countries are about evenly distributed in terms of
differentials by violence status of women in how their need is divided between unmet
need and current contraceptive use (comparing Tables 5.5 and 5.6). For example, in
Cambodia, Egypt, and India, women who have experienced violence have higher
unmet need and have similar or lower contraceptive use rates than women who have
not experienced violence. In the Dominican Republic, Haiti, Nicaragua, and Zambia,
by contrast, both unmet need and contraceptive use are somewhat higher among
women who have experienced violence than among those who have not. In the re-
maining two countries (Colombia and Peru), contraceptive use, but not unmet need,
is higher for women who have experienced violence. Thus, in the majority of coun-
tries, higher need associated with the experience of violence is met by higher use of
contraception, even if sometimes it also means higher rates of unmet need.

5.2.3 Nonlive births

Research suggests that violence is positively associated with adverse pregnancy
outcomes such as abortions and miscarriages (Janssen et al., 2003; Jejeebhoy, 1998).
There are many posited causal routes, including lower weight gain during pregnancy
among women who have experienced violence (Berenson et al., 1997; Curry et al.,
1998), higher rates of STIs, and delay in accessing antenatal care (see below).
In the DHS survey, all women are asked if they have ever had a pregnancy that did
not end in a live birth (i.e., ended in miscarriage or abortion or was a stillbirth).
Based on this question, the proportion of women who have ever had a nonlive birth is
defined here only for women who have ever been pregnant, that is, ever had a birth,
are currently pregnant for the first time, or ever had a terminated pregnancy. Women
who have never been pregnant are excluded from the denominator for this analysis
since experience of violence cannot affect women’s risk of having a nonlive birth if
they have never been pregnant. In defining this variable, there is no differentiation
among miscarriage, abortion, and stillbirths, since it is not clear to what extent
women themselves will have the language to correctly differentiate among these three
outcomes and, more importantly, all three outcomes can be expected to be positively
associated with domestic violence. Accordingly, Table 5.7 shows how the percentage
of women who have ever had a nonlive birth among women who have ever been
pregnant varies by their experience of violence. Because the timing of the nonlive
births is not known, data are not presented separately for women who have recently
experienced violence.
Table 5.7 shows that in every country, women who have experienced violence are
more likely to have had a terminated pregnancy, compared with women who have
never experienced violence. The differential by violence tends to be large. In most
countries, the likelihood of having had a nonlive birth is higher by 33 to 72 percent
among women who have ever experienced violence than among women who have
never experienced violence. These results have to be interpreted with care, however,
since the exposure to pregnancy and, hence, to having a nonlive birth, will vary by age
and number of pregnancies, neither of which is being controlled for. However, in a

5 • Domestic Violence and Demographic and Health Outcomes 85

 multivariate analysis conducted for Cambodia, the Dominican Republic and Haiti, a
highly significant effect of violence was found on the likelihood of having had a
nonlive birth (Kishor and Johnson 2003). In this analysis, even after controlling for
age and the number of children ever born (as well as other relevant factors), the ex-
perience of violence was associated with a 91 percent increase in the likelihood of a
nonlive birth in the Dominican Republic, a 53 percent increase in Cambodia, and a
29 percent increase in Haiti.
Table 5.7 Among ever-married women who have ever had a live
birth, the percentage of ever-married women who have ever had a
nonlive birth or had a terminated pregnancy (miscarriage, abortion,
or stillbirth), according to whether they have ever experienced
violence by their husband or not
–––––––––––––––––––––––––––––––––––––––––––––––––––––
Percentage
who ever had a Number
nonlive birth/ of women
Experience of violence terminated who have ever
1
by husband pregnancy had a birth
–––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced 26.1 411
Never experienced 19.1 1,865

Colombia
Ever experienced 30.8 3,228
Never experienced 21.6 3,966

Dominican Republic
Ever experienced 39.7 1,450
Never experienced 23.1 4,921

Egypt
Ever experienced 37.9 2,321
Never experienced 28.0 4,332

Haiti
Ever experienced 24.1 600
Never experienced 15.7 1,602

India
Ever experienced 25.3 16,019
Never experienced 19.0 66,367

Nicaragua
Ever experienced 25.8 2,487
Never experienced 17.4 5,611

Peru
Ever experienced 26.9 7,217
Never experienced 16.7 9,537

Zambia
Ever experienced 23.8 1,759
Never experienced 19.9 1,849
–––––––––––––––––––––––––––––––––––––––––––––––––––––
1
Includes women who are currently pregnant

5.2.4 Prevalence of sexually transmitted infections

The hypothesized positive association of STIs and violence is based on research
that finds high rates of forced sex among women who are abused by their male part-
ners, a lesser ability to negotiate and use condoms and access counseling and testing,
and a higher prevalence of risky sexual behaviors and drug use among adolescents and
adults who were abused as children (Campbell, 2002; Campbell and Alford, 1989;
Cohen et al., 2000; Heise et al., 1999). From the DHS surveys, it is possible to ex-

86 Domestic Violence and Demographic and Health Outcomes • 5

amine the association between the prevalence of STIs and violence through self-
reports of STIs and STI symptoms. The DHS survey asks all women who have ever
had sexual intercourse whether they had an STI in the last 12 months and whether
they had a genital sore or ulcer in the last 12 months. Additionally, in Haiti, women
were also asked if they had a malodorous vaginal discharge. Women who said “yes” to
one or all questions are said to have had an STI in the 12 months preceding the sur-
vey. Table 5.8 shows how this self-reported prevalence of STIs varies by the violence
status of women. Data on self-reported STIs are not available for India, Egypt, and
Nicaragua.

Table 5.8 Percentage of ever-married women who report having an

STI in the 12 months preceding the survey, according to whether
they have experienced violence by their husband ever, in the past
12 months or never
–––––––––––––––––––––––––––––––––––––––––––––––––––––
Percentage
who had
an STI Number
Experience of violence in the past of
by husband 12 months women
–––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced 7.0 420
Experienced in the last year 7.3 369
Never experienced 2.8 1,983
Colombia
Ever experienced 3.1 3,345
Experienced in the last year u u
Never experienced 1.2 4,257

Dominican Republic
Ever experienced 3.7 1,527
Experienced in the last year 4.6 731
Never experienced 1.0 5,279

Haiti
Ever experienced 18.4 643
Experienced in the last year 21.4 492
Never experienced 10.3 1,703
Peru
Ever experienced 5.4 7,370
Experienced in the last year u u
Never experienced 3.7 9,998

Zambia
Ever experienced 7.7 1,836
Experienced in the last year 8.6 961
Never experienced 3.3 1,955
–––––––––––––––––––––––––––––––––––––––––––––––––––––
u = Unknown (not available)

The proportion of women reporting an STI is low in all countries, but particularly
in Colombia. Despite this, in all countries, the prevalence of STIs among women
who have experienced violence is higher than that among women who have not. Fur-
ther, prevalence is even higher if the violence has been recent. In most countries, the
self-reported prevalence of STIs among women who have experienced violence is at
least twice that among women who have never experienced violence. Although Table
5.8 shows only bivariate results, a multivariate analysis of the data for three of these
countries shows that even after controlling for relevant socioeconomic factors and
other behaviors, violence is significantly and positively associated with the likelihood
of reporting an STI or STI symptom (Kishor and Johnson, 2003).

5 • Domestic Violence and Demographic and Health Outcomes 87

 5.3 Access to Antenatal and Delivery Care and Spousal Violence
Maternal health and birth outcomes partly depend on the care received by the
mother during pregnancy and delivery. The Safe Motherhood Initiative proclaims
that all pregnant women must receive basic, professional antenatal care (Harrison,
1990). Ideally, antenatal care should monitor a pregnancy for signs of complications,
detect and treat preexisting and concurrent problems of pregnancy, and provide ad-
vice and counseling on preventive care, diet during pregnancy, delivery care, postnatal
care, and related issues. The number of antenatal checkups and the timing of the first
checkup are important for the health of the mother and the outcome of the preg-
nancy. The conventional recommendation for normal pregnancies is that once preg-
nancy is confirmed, antenatal checkups should be scheduled at four-week intervals
during the first seven months, then every two weeks until the last month, and weekly
thereafter (MacDonald and Pritchard, 1980). Four antenatal checkups—one each
during the third, sixth, eighth, and ninth months of pregnancy—have been recom-
mended as the minimum necessary (Park and Park, 1989). Studies on the timing of
the initial antenatal checkup, however, show that even when antenatal care is initi-
ated as late as the third trimester, there is a substantial reduction in perinatal mortal-
ity (Ramachandran, 1992). Another important thrust of Safe Motherhood programs
is to encourage deliveries under proper hygienic conditions and under the supervision
of trained health professionals.
This section examines whether women’s access to and timing of antenatal care
(ANC) varies by their experience of violence. Accordingly, Table 5.9 shows the per-
cent distribution of births in the five years (three years in India) preceding the survey
to ever-married women by whether the mother received antenatal care and, if she
did, the timing of the first ANC visit, according to the mother’s experience of vio-
lence. The table also includes information on the percentages of births that were de-
livered with the assistance of a medical professional. While little is known about the
relationship between proper delivery care and the experience of violence, some re-
search in developed countries suggests that women who have experienced violence are
more likely than other women to delay seeking antenatal care (Dietz et al., 1997).
Table 5.9 shows that although countries vary greatly in women’s access to ANC,
this access does not vary substantially by women’s experience of violence. This is par-
ticularly true in Colombia, the Dominican Republic, Haiti, Nicaragua, and Zambia,
where access to ANC is virtually universal. Among the remaining countries, where
ANC is received for two-thirds or less of births, the pattern of variation by violence
status of the mother varies by country. In Egypt and India, the experience of violence
has a negative effect on the likelihood of receiving ANC. In Egypt, mothers who
were abused received ANC for only 32 percent of births, compared with 41 percent
of births for mothers who were not abused. The corresponding proportions for India
were 58 and 67 percent, respectively. By contrast, ANC does not vary by violence
status of the mother in Peru, and in Cambodia, ANC is lower for births to women
who have never experienced violence than to women who have. In general, this sug-
gests that the experience of violence is not a hurdle in countries where ANC is nearly
universal, but it is a significant hurdle in some of the countries where ANC is not
universal.

88 Domestic Violence and Demographic and Health Outcomes • 5

Table 5.9 Percentage of births in the five years (three years in India) preceding the survey, by whether antenatal care was received
and timing of such care and whether the delivery was assisted by a medical professional, according to whether the mother has ever
experienced violence by her husband or not
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Timing of first ANC visit (months)
among those who received
––––––––––––––––––––––––––––––––––––– Delivery
Don’t assisted by Number
Mother’s experience of Percentage know/ a medical of
violence by husband with ANC <4 4-5 6-7 8+ missing Total professional births
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced 48.2 24.5 19.1 30.5 17.0 9.1 100.0 30.2 250
Never experienced 44.2 22.2 31.4 24.9 17.4 4.1 100.0 34.3 1,039

Colombia
Ever experienced 89.4 73.6 18.9 6.3 1.0 0.2 100.0 88.2 1,282
Never experienced 92.5 78.2 15.9 4.9 0.9 0.2 100.0 88.4 1,836

Dominican Republic
Ever experienced 98.2 77.4 16.9 4.5 0.3 0.9 100.0 98.0 657
Never experienced 99.1 84.4 12.1 2.7 0.4 0.4 100.0 98.7 2,243

Egypt
Ever experienced 32.3 71.8 15.5 8.4 2.8 1.5 100.0 36.0 2,121
Never experienced 40.9 79.2 13.4 4.6 1.7 1.0 100.0 48.8 3,291

Haiti
Ever experienced 80.4 62.3 20.8 13.1 2.1 1.9 100.0 64.0 359
Never experienced 82.5 62.1 22.5 12.2 2.3 0.8 100.0 61.2 995

India
Ever experienced 58.4 41.6 35.3 17.6 5.5 0.2 100.0 32.4 6,274
Never experienced 67.4 51.9 30.0 14.8 3.1 0.0 100.0 44.7 26,492
Nicaragua
Ever experienced 86.0 69.0 19.3 7.9 2.3 1.5 100.0 72.3 1,335
Never experienced 86.5 72.4 16.9 7.6 2.0 1.2 100.0 67.9 3,236

Peru
Ever experienced 65.5 66.9 20.6 9.3 3.1 0.2 100.0 57.5 4,372
Never experienced 65.8 72.2 16.9 8.7 2.1 0.3 100.0 57.1 6,751

Zambia
Ever experienced 95.3 13.1 55.2 28.6 2.5 0.5 100.0 45.6 1,267
Never experienced 95.7 16.7 54.8 26.6 1.6 0.2 100.0 42.9 1,319

Table 5.9 also shows that the timing of the first visit among mothers who did re-
ceive any ANC varies by the violence status of the mother. In all countries except
Cambodia and Haiti, mothers are much less likely to receive ANC in the first trimes-
ter of their pregnancy if they have experienced violence than if they have not. This
differential becomes much smaller by the fifth month in all countries except Cambo-
dia, Egypt, India, and Zambia. Cambodia is particularly interesting. In this country,
the likelihood that a mother received ANC for her birth in the first trimester of the
pregnancy is slightly lower (22 percent) if she never experienced violence than if she
did (25 percent), but the cumulated likelihood that she received an ANC checkup by
month five is much higher if she has never experienced violence (54 percent) than if
she has (44 percent). Thus, as in most other countries (in Cambodia too), the data
suggest that the experience of violence is associated with a delay in receiving ANC,
even among women who do receive any ANC.
As in the case of ANC, countries vary greatly in women’s access to medical assis-
tance during delivery. In Cambodia, Egypt, and India, where such access is particu-
larly limited, births to mothers who have experienced violence are somewhat less
likely to have been delivered with the assistance of a health professional, compared

5 • Domestic Violence and Demographic and Health Outcomes 89

 with births to mothers who have never experienced violence. In Zambia, where access
to medical care at delivery is also limited, by contrast, delivery with the assistance of a
medical professional is more likely if the mother has experienced violence than if she
has not.
Overall, the data show that access to ANC and proper delivery care is, in the ma-
jority of countries, lower for women who have experienced violence than for other
women, but the relationship tends to be weak and inconsistent. What is more robust
is the relationship between the experience of violence and delay in the timing of
ANC. Women who have experienced violence, in most countries, access ANC later
than women who have not experienced violence.

5.4 Children’s Mortality and Health and Mother’s Experience of

Spousal Violence
The negative effects on children of witnessing frequent marital violence are well
documented (Edelson, 1999; Jouriles et al., 1989; McCloskey et al., 1995). They in-
clude emotional, behavioral, and physical health problems that become more evident
as the child grows. Since the DHS does not have information on the health of chil-
dren older than five years, it is not possible to examine the health consequences of
mother’s experience of abuse for children much beyond infancy. This section explores
whether a mother’s experience of violence has negative effects on the health and sur-
vival of her young children. Some evidence already exists that infant and child mor-
tality rates for abused mothers are higher than those for mothers who have not been
abused (Asling-Monemi et al., 2003; Binka et al., 1995; Jejeebhoy, 1998). While the
pathways through which the survival and health of children are put at risk by a
mother’s experience of violence are unclear, there is some limited evidence that chil-
dren of abused mothers are more likely to be malnourished and less likely to be im-
munized than other children. Accordingly, in this section, the linkages between
mothers’ experience of violence and infant and child mortality rates, immunization
rates, and the nutritional status of children are explored.

5.4.1 Are infant and child mortality rates higher for women who have
ever experienced violence?
Table 5.10 shows the following five different mortality rates for children born to
ever-married mothers in the five years preceding the survey according to whether the
mother has ever experienced violence:

Neonatal mortality: The probability of dying in the first month of life

Postneonatal mortality: The probability of dying after the first month of life
but before the first birthday
Infant mortality (1q0): The probability of dying before the first birthday
Child mortality (4q1):The probability of dying between the first and fifth
birthdays
Under-five mortality (4q0): The probability of dying before the fifth birthday.

Table 5.10 shows evidence of higher rates of infant and child mortality among
women who have ever experienced violence, compared with those among women
who have not. In Cambodia, Egypt, India, Nicaragua, and Zambia, all five of the

90 Domestic Violence and Demographic and Health Outcomes • 5

 Table 5.10 Infant and child mortality rates for the five years before the survey, according to whether the
mother has ever experienced violence by her husband or not
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Infant Child Under-five
Mother’s xperience Neonatal Postneonatal mortality mortality mortality
of violence by husband mortality mortality (1q0) (4q1) (5q0)
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced 53 52 105 42 142
Never experienced 41 50 91 30 118
Colombia
Ever experienced 20 9 29 4 33
Never experienced 13 7 20 4 24
Dominican Republic
Ever experienced 21 14 35 8 43
Never experienced 24 12 35 9 44
Egypt
Ever experienced 44 37 81 30 109
Never experienced 35 36 72 23 93
Haiti
Ever experienced 30 60 90 62 147
Never experienced 33 76 109 47 150
India
Ever experienced 54 32 86 38 121
Never experienced 46 23 69 28 96
Nicaragua
Ever experienced 22 29 51 12 63
Never experienced 20 22 42 10 52
Peru
Ever experienced 23 20 43 21 63
Never experienced 22 21 43 16 58
Zambia
Ever experienced 32 63 95 85 171
Never experienced 31 60 92 79 163

different infant and child mortality rates are higher for women who have experienced
violence than for women who have not; in Colombia, all rates except the child mor-
tality rate are higher; and in Peru, three of the five rates (namely, the neonatal mor-
tality rate, the child mortality rate, and the under-five mortality rate) are higher.
There is less consistency in the Dominican Republic and Haiti, where only one of the
five rates is higher.
The differentials in infant and child mortality rates by mother’s violence status
evident for most countries in Table 5.10, are not necessarily large (see Figure 5.2);
however, the consistency with which the differentials occur suggests that the experi-
ence of violence by mothers could be putting the survival of their young children at
risk. In fact, a study that controlled for other factors affecting infant and child mor-
tality using data from León, Nicaragua, found that one third of all child deaths were
attributable to the experience of spousal violence by the mother (Asling-Monemi et
al., 2003).

5 • Domestic Violence and Demographic and Health Outcomes 91

 Figure 5.2
Under-five child mortality rates by mother’s
experience of spousal violence

Cambodia

Colombia

Dominican Republic

Egypt

Haiti

India

Nicaragua

Peru

Zambia

0 20 40 60 80 100 120 140 160 180

Experienced violence Did not experience violence

5.4.2 Child immunization and the experience of violence by the mother

The vaccination of children against six serious, but preventable, diseases (tubercu-
losis, diphtheria, pertussis, tetanus, poliomyelitis, and measles) has been a corner-
stone of child health care throughout the world. According to World Health Organi-
zation guidelines, a child should have received one BCG vaccination for tuberculosis,
three doses each of the polio and the DPT (diphtheria, pertussis, and tetanus) vac-
cines, and one dose of the measles vaccine before his or her first birthday. Incomplete
vaccinations can put the child’s health at risk. Table 5.11 shows the percentage of
children age 12-35 months who have received the required combination of vaccina-
tions, according to the mother’s experience of violence.
In Cambodia, the Dominican Republic, Egypt and India, children of mothers who
have experienced violence are much less likely than other children to have received
the required numbers of the different vaccinations. Haiti is the only country where
children of mothers who have experienced violence are more likely to receive each of
the required vaccinations, as compared with children of mothers who have not ex-
perienced violence. However, if the likelihood of receiving all of the required vaccina-
tions is examined, in six of the nine countries, children of mothers who have experi-
enced violence are less likely to be fully immunized. Specifically, the proportion of
fully vaccinated children age 12-35 months among mothers who have not experi-
enced violence, compared with children of mothers who have experienced violence, is
higher by at least 5 to 10 percent in Colombia, Egypt, Nicaragua, and Peru; 38 per-
cent in India; and 49 percent in the Dominican Republic. Also, in five countries,
children of mothers who have experienced violence are more likely to have received
none of the required vaccinations than children of mothers who have not experienced

92 Domestic Violence and Demographic and Health Outcomes • 5

any violence. In Cambodia, for example, 19 percent of children of mothers who have
not experienced violence have received none of the required vaccinations, but this
proportion is 26 percent among children of mothers who have ever experienced vio-
lence. In all other countries, except India, the differential, though evident, is very
small.
Overall, these data provide evidence that in most countries, children of mothers
who have experienced violence are disadvantaged in their access to life-saving routine
immunizations.

Table 5.11 Percentage of children age 12-35 months who received specific vaccinations at any time before
the interview, according to whether the mother has ever experienced violence by her husband or not
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Vaccinations received
–––––––––––––––––––––––––––––––––––––––––––––
Three Three Number
Mother’s experience of doses doses of
violence by husband BCG of polio of DPT Measles All None children
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced 61.0 46.8 45.9 51.6 40.4 25.6 128
Never experienced 71.7 53.6 48.4 58.1 38.8 18.7 457
Colombia
Ever experienced 94.1 69.5 76.0 39.8 23.3 1.6 662
Never experienced 92.9 71.6 79.5 39.5 24.4 1.3 930

Dominican Republic
Ever experienced 88.5 40.4 41.7 83.0 24.8 4.4 393
Never experienced 94.2 44.6 62.7 91.2 37.0 2.2 1,222

Egypt
Ever experienced 92.6 85.4 84.1 88.0 78.5 1.7 776
Never experienced 95.1 89.9 89.6 91.8 84.8 2.3 1,160

Haiti
Ever experienced 79.1 46.6 45.9 56.9 38.4 14.3 206
Never experienced 69.2 41.6 45.2 52.9 32.5 18.2 543

India
Ever experienced 62.3 53.2 44.7 39.7 30.2 18.9 1,966
Never experienced 73.8 61.5 58.0 53.3 41.6 13.3 8,227

Nicaragua
Ever experienced 94.5 85.6 82.5 89.8 74.5 2.0 721
Never experienced 96.0 86.3 84.9 90.3 79.6 1.7 1,704
Peru
Ever experienced 96.3 74.7 83.6 77.6 60.1 1.7 1,647
Never experienced 96.5 76.2 84.3 82.0 63.8 1.7 2,592

Zambia
Ever experienced 91.8 82.3 81.3 84.3 72.2 4.4 727
Never experienced 91.3 82.3 79.8 83.9 71.7 4.4 427

5.4.3 Nutritional status of children and mother’s experience of

violence

Table 5.12 shows the variation in child nutritional status, according to the
mother’s experience of spousal violence ever, in the past one year, or never. Two indi-
cators of nutritional status are shown. The first indicator, percentage of children who
are undernourished, is derived from the three commonly used nutrition indicators:
weight-for-age, height-for-age and weight-for-height. Each of these indicators of
nutritional status is typically expressed in standard deviation units (Z scores) from the
median for the international reference population. For this analysis, children who are

5 • Domestic Violence and Demographic and Health Outcomes 93

 more than two standard deviations below the reference median on any one of the
three indices are counted as undernourished. The second indicator of nutritional
status included in Table 5.12 is the percentage of children who are anemic. A child is
considered anemic if he or she has a hemoglobin level below 11.0 g/dl.
In Colombia, Egypt, India, Nicaragua, and Peru, children of mothers who have
ever experienced spousal violence, as well as those who have recently experienced
spousal violence, are more likely to be undernourished than are children of mothers
who have never experienced spousal violence. The differences are relatively large only
in India, however.

Table 5.12 Percentage of children age 0-59 months who are undernourished, and
percentage of children age 6-59 months who are anemic, by whether the mother has
experienced violence by her husband ever, in the past 12 months, or never
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Children 0-59 months Children 6-59 months
––––––––––––––––––––– –––––––––––––––––––––
Percentage Number Number
Mother’s experience of under- of Percentage of
violence by husband nourished children anemic children
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cambodia
Ever experienced 58.6 303 71.8 276
Experienced in the last year 59.7 272 71.2 250
Never experienced 59.5 1,170 61.9 1,030

Colombia
Ever experienced 17.3 1,518 u u
Experienced in the last year u u u u
Never experienced 13.6 2,122 u u

Dominican Republic
Ever experienced 9.8 751 u u
Experienced in the last year 9.8 430 u u
Never experienced 11.1 2,434 u u
Egypt
Ever experienced 35.3 1,805 u u
Experienced in the last year 37.2 1,045 u u
Never experienced 33.5 2,802 u u

Haiti
Ever experienced 28.1 470 70.3 402
Experienced in the last year 27.5 424 71.9 360
Never experienced 30.8 1,171 63.8 1,052

India
Ever experienced 67.0 4,652 76.8 3,944
Experienced in the last year 68.3 2,959 77.3 2,532
Never experienced 57.6 19,769 73.3 16,314

Nicaragua
Ever experienced 31.5 1,598 u u
Experienced in the last year 29.2 809 u u
Never experienced 27.1 3,799 u u

Peru
Ever experienced 28.4 3,781 49.5 833
Experienced in the last year u u u u
Never experienced 26.0 5,761 49.2 1,136

Zambia
Ever experienced 53.2 1,512 u u
Experienced in the last year 50.8 865 u u
Never experienced 53.6 1,577 u u
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
u = Unknown (not available)

94 Domestic Violence and Demographic and Health Outcomes • 5

 The analysis of anemia among children shows similarly mixed results. In all four
countries for which these data are available, children of mothers who have experi-
enced violence are more likely to be anemic than children whose mothers have never
experienced violence; however, the differences by violence status are relatively large,
at 8 to 10 percentage points, only in Cambodia and Haiti. Interestingly, the differen-
tial by violence status in anemia rates is higher for children than for women them-
selves in both Cambodia and Haiti, whereas in India the differential is smaller.
Overall, in seven of the nine countries, children of mothers who have experienced
violence tend to perform more poorly on either one or both of the given nutritional
status measures, than children of mothers who have never experienced violence. For
most countries except India, the differentials are small on at least one of the two indi-
cators. Thus, although a negative relationship between the nutritional status of chil-
dren and a mother’s violence status is evident in most countries, it is either weak or
varies in strength by the indicator of undernutrition.

5 • Domestic Violence and Demographic and Health Outcomes 95

 References

Asling-Monemi, K., R. Pena, M.C. Ellsberg, and L.A. Persson. 2003. Violence
against women increases the risk of infant and child mortality: A case-referent study
in Nicaragua. Bulletin of the World Health Organization 81(1): 10-18.

Astbury, J. 1999. Promoting women’s mental health. Geneva: World Health Organi-
zation.

Bawah, A.A., P. Akweongo, R. Simmons, and J.F. Phillips. 1999. Women’s fears
and men’s anxieties: The impact of family planning on gender relations in northern
Ghana. Studies in Family Planning 30(1): 54-66.

Becker, G.S. 1973. A theory of marriage: Part I. Journal of Political Economy 81(4):
813-846.

Berenson, A.B., C.M. Wiemann, T.F. Rowe, and V.I. Rickert. 1997. Inadequate
weight gain among pregnant adolescents: Risk factors and relationship to infant birth
weight. American Journal of Obstetrics and Gynecology 176(6): 1220-1224.

Biddlecom, A.E., and B.M. Fapohunda. 1998. Covert contraceptive use: Prevalence,
motivations, and consequences. Studies in Family Planning 29(4): 360-372.

Binka, F.N., G.H. Maude, M. Gyapong, D.A. Ross, and P.G. Smith. 1995. Risk
factors for child mortality in northern Ghana: A case-control study. International
Journal of Epidemiology 24(1): 127-135.

Blanc, A.K., B. Wolff, A.J. Gage, A.C. Ezeh, S. Neema, and J. Ssekamatte-
Ssebuliba. 1996. Negotiating reproductive outcomes in Uganda. Calverton, Maryland:
Macro International Inc. and Institute of Statistics and Applied Economics
[Uganda].

Byrne, C.A., H.S. Resnick, D.G. Kilpatrick, C.L. Best, and B.E. Saunders. 1999.
The socioeconomic impact of interpersonal violence on women. Journal of Consulting
and Clinical Psychology 67(3): 362-366.

Campbell, J.C. 2002. Health consequences of intimate partner violence. Lancet 359:
1331-1336.

Campbell, J.C., and P. Alford. 1989. The dark consequences of marital rape. Ameri-
can Journal of Nursing 89: 946-949.

Campbell, J.C., and L.A. Lewandowski. 1997. Mental and physical health effects of
intimate partner violence on women and children. Psychiatric Clinics of North America
20(2): 353-374.

References 97
 Campbell, J.C., D. Webster, J. Koziol-McLain, C. Block, D. Campbell, M.A.
Curry, F. Gary, N. Glass, J. McFarlane, C. Sachs, P. Sharps, Y. Ulrich, S.A. Wilt, J.
Manganello, Z. Zu, J. Schollenberger, V. Frye, and K. Laughon. 2003. Risk factors
for femicide in abusive relationships: Results from a multisite case-control study.
American Journal of Public Health 93(7): 1089-1097.

Centers for Disease Control and Prevention (CDC). 1998. Recommendations to

prevent and control iron deficiency in the United States. Morbidity and Mortality
Weekly Report 47(RR-3): 1-29.

Cohen, M., C. Deamant, S. Barkan, J. Richardson, M. Young, S. Holman, K. Anas-

tos, J. Cohen, and S. Melnick. 2000. Domestic violence and childhood sexual abuse
in HIV-infected women and women at risk for HIV. American Journal of Public
Health 90(4): 560-565.

Coker, A.L., P.H. Smith, R.E. McKeown, and M.J. King. 2000. Frequency and cor-
relates of intimate partner violence by type: Physical, sexual, and psychological batter-
ing. American Journal of Public Health 90: 553-559.

Correa, S. and R. Petchesky. 1994. Reproductive and sexual rights: A feminist per-
spective. In G. Sen, A. Germain, and L.C. Chen (eds.), Population policies reconsid-
ered: Health, empowerment and rights. Boston, Massachusetts: Harvard School of Pub-
lic Health.

Curry, M.A., N. Perrin, and E. Wall. 1998. Effects of abuse on maternal complica-
tions and birth weight in adult and adolescent women. Obstetrics and Gynecology
92(4): 530-534.

DeKeseredy, W.S., and M.D. Schwartz. 1998. Measuring the extent of woman abuse
in intimate heterosexual relationships: A critique of the Conflict Tactics Scales. U.S.
Department of Justice, Violence Against Women Online Resources. Available at
http://www.vaw.umn.edu/documents/vawnet/ctscritique/ctscritique.html.

Dietz, P.M., J.A. Gazmararian, M.M. Goodwin, F.C. Bruce, C.H. Johnson, and
R.W. Rochat. 1997. Delayed entry into prenatal care: Effect of physical violence.
Obstetrics and Gynecology 90(2): 221-224.

Dyson, T., and M. Moore. 1983. On kinship structure, female autonomy and demo-
graphic behavior in India. Population and Development Review 9(1): 35-60.

Edelson, J.L. 1999. Children witnessing of adult domestic violence. Journal of Inter-
personal Violence 14(8): 839-870.

Ellsberg, M.C. 2000. Candies in hell: Research and action on domestic violence against
women in Nicaragua. Umeå, Sweden: Umeå University.

98 References • 1
Ellsberg, M.C., L. Heise, R. Pena, S. Agurto, and A. Winkvist. 2001. Researching
domestic violence against women: Methodological and ethical considerations. Studies
in Family Planning 32(1): 1-16.

Ellsberg, M.C., R. Peña, A. Herrera, J. Liljestrand, and A. Winkvist. 1999. Wife

abuse among women of childbearing age in Nicaragua. American Journal of Public
Health 89(2): 241-244.

Felitti V.J. 1993. Childhood sexual abuse, depression and family dysfunction in adult
OBE patients: A case control study. Southern Medical Journal 86(7): 732-736.

Fernandez, M. 1997. Domestic violence by extended family members in India. Jour-

nal of Interpersonal Violence 12(3): 433-455.

Fikree, F.F. and L.I. Bhatti. 1999. Domestic violence and health of Pakistani
women. International Journal of Gynaecology and Obstetrics 65(2): 195-201.

Filmer, D., and L. Pritchett. 2001. Estimating wealth effects without expenditure
data—or tears: An application to educational enrollments in states of India. Demo-
graphy 38(1): 115-132.

Ganatra, B.R., K.J. Coyaji, and V.N. Rao. 1998. Too far, too little, too late: A com-
munity-based case-control study of maternal mortality in rural west Maharashtra,
India. Bulletin of the World Health Organization 76(6): 591-598.

García, B. (ed.). 2000. Women, poverty and demographic change. Liege: Oxford Uni-
versity Press for IUSSP.

Gelles, R.J. 1993. Through a sociological lens: Social structure and family violence.
In R.J. Gelles and D.R. Loseke (eds.), Current controversies on family violence. New-
bury Park, California: Sage Publications.

Glander, S.S., M.L. Moore, R. Michielutte, and L.H. Parsons. 1998. The prevalence of
domestic violence among women seeking abortion. Obstetrics and Gynecology 91:
1002-1006.

Goetz, A.M. 1997. Managing organisational change: The “gendered” organisation of

space and time. Gender and Development 5(1): 17-27.

Goetz, A.M., and R. Sen Gupta. 1996. Who takes the credit? Gender, power and
control over loan use in rural credit programs in Bangladesh. World Development
24(1): 45-63.

Gwatkin, D.R., S. Rutstein, K. Johnson, R.P. Pande, and A. Wagstaff. 2000. Socio-
economic differences in health, nutrition and poverty. HNP/Poverty Thematic Group of
The World Bank. Washington, D.C.: The World Bank.

References 99
 Harrison, K.A. 1990. The political challenge of maternal mortality in the Third
World. Maternal mortality and morbidity—A call to women for action. Special issue,
May 28, 1990.

Heise, L.L. 1993. Reproductive freedom and violence against women: Where are the
intersections? Journal of Law, Medicine and Ethics 21(2): 206-216.

Heise, L.L. 1998. Violence against women: An integrated, ecological framework.

Violence Against Women 4(3): 262-290.

Heise, L., M. Ellsberg, and M. Gottemoeller. 1999. Ending violence against women.
Population Reports, Series L, No. 11. Baltimore, Maryland: Johns Hopkins Univer-
sity School of Public Health, Population Information Program.

Heise, L., K. Moore, and N. Toubia. 1995. Sexual coercion and reproductive health: A
focus on research. New York: The Population Council.

Heise, L., J. Pitanguy, and A. Germain. 1994. Violence against women: The hidden
health burden. World Bank Discussion Paper #225. Washington D.C.: The World
Bank.

Hornung, C.A., B.C. McCullough, and T. Sugimoto. 1981. Status relationships in

marriage: Risk factors in spouse abuse. Journal of Marriage and the Family 43: 675-
692.

Janssen, P.A., V.L. Holt, N.K. Sugg, I. Emanuel, C.M. Critchlow, and A.D. Hen-
derson. 2003. Intimate partner violence and adverse pregnancy outcomes: A popula-
tion-based study. American Journal of Obstetrics and Gynecology 188(5): 1341-1347.

Jejeebhoy, S.J. 1998. Associations between wife-beating and fetal and infant death:
Impressions from a survey in rural India. Studies in Family Planning 29(3): 300-308.

Jewkes, R. 2002. Intimate partner violence: Causes and prevention. Lancet

359(9315): 1423-1429.

Johnson, K. 2003. Dialectics of power and violence in the home: A comparative

analysis of women’s experience of domestic violence in Haïti and Nicaragua. Disser-
tation, University of Maryland.

Jouriles, E.N., C.M. Murphy, and K.D. O’Leary. 1989. Interspousal aggression,
marital discord, and child problems. Journal of Consulting and Clinical Psychology
57(3): 453-455.

Kalmuss, D. 1984. The intergenerational transfer of marital aggression. Journal of

Marriage and the Family 46: 11-19.

100 References • 1
Kishor, S. 2000. Empowerment of women in Egypt and links to the survival and
health of their infants. In H.B. Presser, and G. Sen (eds.), Women’s empowerment and
demographic processes: Moving beyond Cairo. New York: Oxford University Press.

Kishor, S., and K. Johnson. 2003. Women’s health at the nexus of poverty and do-
mestic violence: Evidence from the developing world. Paper presented at the 2003
Annual Meeting of the Population Association of America, Minneapolis, Minnesota.

Levinson, D. 1989. Violence in cross-cultural perspective. Newbury Park, California:

Sage Publications.

MacDonald, P.C., and J.A. Pritchard. 1980. Williams obstetrics. 16th ed. New York:
Appleton-Century-Crofts.

Malhotra, A., and M. Mather. 1997. Do schooling and work empower women in
developing countries? The case of Sri Lanka. Sociological Forum 12(4): 599-630.

Martin, S.L., A.O. Tsui, K. Maitra, and R. Marinshaw. 1999. Domestic violence in
northern India. American Journal of Epidemiology 150(4): 417-426.

Mason, K.O. 1987. The impact of women’s social position on fertility in developing
countries. Sociological Forum 2(4): 718-745.

McCloskey, L.A., A.J. Figueredo, and M.P. Koss. 1995. The effects of systemic fam-
ily violence on children’s mental health. Child Development 66(5): 1239-1261.

McClusky, L. 2001. Here, our culture is hard: Stories of domestic violence from a Mayan
community in Belize. Austin, Texas: University of Texas Press.

Miller, B.D. 1981. The endangered sex: Neglect of female children in rural North India.
Ithaca, New York: Cornell University Press.

Ministry of Women’s Affairs and Project Against Violence. 1996. Household survey
on domestic violence in Cambodia. Phnom Penh, Cambodia.

Mueller, C.W., T.L. Parcel, and F.C. Pampel. 1979. The effect of marital dyad
status inconsistency on women’s support for equal rights. Journal of Marriage and the
Family 56: 1121-1139.

Oropesa, R.S. 1997. Development and marital power in Mexico. Social Forces 75(4):
1291-1317.

Park, J.E., and K. Park. 1989. Textbook of preventive and social medicine. 12th ed. Ja-
balpur, India: M/S Banarsidas Bhanot Publishers.

Pearlman, M.D., J.E. Tintinalli, and R.P. Lorenz. 1990. Blunt trauma during preg-
nancy. New England Journal of Medicine 323: 1609-1613.

References 101
 Petersen, R., J.A. Gazmarian, A.M. Spitz, D.L. Rowley, M.M. Goodwin, L.E. Saltz-
man, and J.S. Mark. 1997. Violence and adverse pregnancy outcomes: A review of the
literature and directions for further research. American Journal of Preventive Medicine
13(5): 366-373.

Population Council. 1998. Secrecy and silence: Why women hide contraceptive use.
Population Briefs 4(3): 3.

Ramachandran, P. 1992. Need of organization of antenatal and intrapartum care in

India. Demography India 21(2): 179-193.

Rutstein, S., D. Gwatkin, and K. Johnson. 1999. Wealth versus expenditure: Com-
parison between the DHS wealth index and household expenditures in four depart-
ments of Guatemala. Calverton, Maryland: ORC Macro (Unpublished).

Rutstein, S., D. Gwatkin, and K. Johnson. 2000. Poverty, health inequality, and its
health and demographic effects. Paper presented at the 2000 Annual Meeting of the
Population Association of America, Los Angeles, California.

Seltzer, J.A., and D. Kalmuss. 1988. Socialization and stress explanations for spousal
abuse. Social Forces 67(2): 473-491.

Sen, G. and S. Batliwala. 2000. Empowering women for reproductive rights. In

H.B. Presser and G. Sen (eds.), Women’s empowerment and demographic processes:
Moving beyond Cairo. New York: Oxford University Press.

Shetty, P.S., and W.P.T. James. 1994. Body mass index—a measure of chronic
energy deficiency in adults. Rome: FAO.

Smith, P.H., I. Tessaro, and J.A.L. Earp. 1995. Women’s experiences with battering:
A conceptualization from qualitative research. Women’s Health Issues 5(4): 173-182.

Straus, M.A. 1979. Measuring intrafamily conflict and violence: The conflict tactics
(CT) scales. Journal of Marriage and the Family 41(1): 75-88.

Straus, M.A. 1990. Measuring intrafamily conflict and violence: The conflict tactic
(CT) scales. In M.A. Straus and R.J. Gelles (eds.), Physical violence in American fami-
lies: Risk factors and adaptations to violence in 8,145 families. New Brunswick, New
Jersey: Transaction Publishers, pp. 29-47.

United Nations General Assembly. 1993. Declaration on the elimination of violence

th
against women. 85 Plenary Meeting, A/RES/48/104. Geneva: United Nations Gen-
eral Assembly.

United Nations. 1995a. Report of the Fourth World Conference on Women, Beijing 4-15
September 1995. New York: United Nations.

102 References • 1
United Nations. 1995b. Population and development: Programme of action adopted at the
International Conference on Population and Development: Cairo 5-13 September 1994.
New York: Department for Economic and Social Information and Policy Analysis,
United Nations.

Williamson, D.F., T.J. Thomson, R.F. Anda, W.H. Dietz, and V. Felitti. 2002. Body
weight and obesity in adults and self-reported abuse in childhood. International Journal
of Obesity and Related Metabolic Disorders 26(8): 1075-1082.

World Health Organization. 2001. Putting women first: Ethical and safety recommen-
dations for research on domestic violence against women. Geneva, Switzerland: Depart-
ment of Gender and Women’s Health.

World Health Organization. 2004. Gender-based violence Web site:

http://www.who.int/gender/violence/en/, February.

Yick, A.G. 2001. Feminist theory and status inconsistency theory: Application to
domestic violence in Chinese immigrant families. Violence Against Women 7(5):
545-562.

Yllo, K. 1983. Sexual inequality and violence against wives in American states. Jour-
nal of Comparative Family Studies 14: 67-86.

References 103
 Appendix A

DHS Domestic Module with

Recommendations for Implementation
and Other Domestic Violence Questions
Recommendations for enhancing the quality and safety of research on domestic
violence

(Adapted from recommendations made by Lori Heise and Mary Ellsberg,

CHANGE, and the WHO Multi-Country Study of Women’ Health and
Domestic Violence, Core Protocol)

The following measures should be adopted in all countries where the domestic
violence module is being implemented. These measures are aimed at ensuring
women’s safety and the ethical implementation of the domestic violence module, and
at maximizing the disclosure of actual violence.

♦ Selection of eligible women. Only one woman per household should receive the
domestic violence module. The subsample may be composed of all women of
reproductive age or ever-married women, although in countries where the age of
marriage is late, using all reproductive age women may mean that a large group
of women will never have been exposed to the risk of spousal violence. Specific
methods should be used to ensure that respondents are randomly selected within
the household.

♦ Training. All members of the staff should receive special training. This includes
the administrative and technical personnel as well as both male and female field
staff. It is important to ensure that all staff understand the purpose of the
module and why special measures are being used. Field staff need to receive
additional training in how to administer the module using the safety procedures
established by the survey, how to deal with crisis situations, and how to prepare
themselves emotionally for the work. It is usually recommended that at least part
of the training be carried out with men and women separately, since their
experiences and reactions to the training are likely to be quite different. It is
recommended that local women’s groups who work in the field of violence
prevention be invited to participate in the training.

♦ Informed consent. The introductory sentence in the violence module should be

treated as an additional informed consent procedure. The respondent should be
reassured about the confidentiality of the information. If more than one woman
in the household is being interviewed with the core questionnaire, the interviewer
should informally explain that no one else in the household is being asked the
questions that the respondent is going to be asked and that no one else will know
what has been discussed.

♦ Privacy. The need for absolute privacy must be stressed with interviewers. In
addition to using a range of techniques for ensuring privacy, they should be free
to reschedule the interview to another time in order to carry out the interview in
private. If any other adult comes into the room while the module is being
implemented, the interviewer must immediately stop and, if necessary, change
the subject. She must not resume until the adult has left and is out of hearing
distance.

Appendix B 107
 ♦ Referrals and additional information. An information sheet must be prepared for
distribution in the appropriate language, listing the options and services available
for women experiencing domestic violence, including, if possible, any legal help
and available services. At a minimum, the sheet should contain an address where
women in need can get information. The information sheet should be small
enough to be easily hidden. Countries should decide whether all women who
participate should be given this information or only women who specifically
request it. Alternatively, in order to minimize its visibility, the information on
domestic violence can be combined with other health referral information being
distributed to all women. The respondent should be asked in advance whether it
is safe for her to receive the information.

♦ Support for field workers. Emotional support to fieldworkers is essential, both in

helping interviewers withstand the demands of the fieldwork and for contributing
to the quality of the data collection process. This support can take the form of
regular debriefing meetings, to help interviewers “unload.” Supervisors also need
to be trained to give support to interviewers as needed and to identify and help
fieldworkers who are experiencing problems.

♦ Translation. The use of translators should be avoided in collecting data on

domestic violence, both because it is likely to reduce the quality of the
information given and because it violates the confidentiality of the interview,
particularly when translators are from the same community.

♦ Quality control. Quality assurance procedures need to be developed for the

domestic violence module in line with those used for the rest of the survey.
These include having field check tables on the proportion of violence modules
not being completed (DV01 coded as “2”) by each interviewer/team and the
proportion of women reporting violence by interviewer/team. The idea is to
identify individuals or supervisors who are producing data figures that are
significantly higher or lower than the rest of the fieldworkers. Additional
monitoring should be done to ensure that all procedures for implementing the
module are being followed correctly and that the data are of the highest quality.
Supervisors must identify and discipline both types of interviewers: those who are
not implementing the module in privacy and those using the need for privacy to
avoid implementing the module.

♦ Collaboration with local women’s group. Women’s groups should be involved

from the start. In some countries, it may be possible to provide referrals to local
women’s groups for respondents with problems related to violence, and they
could help in obtaining support for fieldworkers. The involvement of women’s
groups will also increase ownership of the data.

♦ Substudies on men. If a male questionnaire is also being implemented, then

questions on the prevalence of violence must not be included in the male
questionnaire (the core male questionnaire does not contain such questions).
When it is necessary to obtain such information from men, the sample of men
should not be selected from the same households as the women who receive the
domestic violence module, to avoid arousing the suspicion of husbands as to the
content of the study.

108 Appendix A
 THE DHS DOMESTIC VIOLENCE MODULE
USED WITH SMALL VARIATIONS IN CAMBODIA (2000), THE DOMINICAN REPUBLIC (2002) AND
HAITI (2000)

NO. QUESTIONS AND FILTERS CODING CATEGORIES SKIP

DV01 CHECK FOR PRESENCE OF OTHERS:

DO NOT CONTINUE UNTIL EFFECTIVE PRIVACY IS ENSURED.

PRIVACY PRIVACY
OBTAINED……..1 NOT POSSIBLE…….2────────────────────────────────────────── ─<DV28
│
?

READ TO ALL RESPONDENTS:

Now I would like to ask you questions about some other important aspects of a woman's life. I know that some of
these questions are very personal. However, your answers are crucial for helping to understand the condition of
women in (COUNTRY). Let me assure you that your answers are completely confidential and will not be told to
anyone.

DV02 CHECK 501, 502, AND 504:

CURRENTLY SEPARATED/ WIDOWED/

MARRIED/ ┌─┐ DIVORCED ┌─┐ NEVER MARRIED/ ┌─┐
LIVING ├─┘ ├─┘ NEVER LIVED └─┴───────────────── ─<DV14
WITH A MAN ? (READ IN PAST TENSE) ? WITH A MAN

DV03 When two people marry or live together, they share both good and bad
moments. In your relationship with your (last) husband/partner do (did) FRE- SOME-
the following happen frequently, only sometimes, or never? QUENTLY TIMES NEVER

a) He usually (spends/spent) his free time with you? FREE TIME .............1 2 3
b) He (consults/consulted) you on different household matters? CONSULTS ............1 2 3
c) He (is/was) affectionate with you? AFFECTIONATE.....1 2 3
d) He (respects/respected) you and your wishes? RESPECTS.............1 2 3

DV04 Now I am going to ask you about some situations which happen to
some women. Please tell me if these apply to your relationship with
your (last) husband/partner? YES NO DK

a) He (is/was) jealous or angry if you (talk/talked) to other men? JEALOUS............................ 1 2 8

b) He frequently (accuses/accused) you of being unfaithful? ACCUSES........................... 1 2 8
c) He (does/did) not permit you to meet your girl friends? NOT MEET FRIENDS......... 1 2 8
d) He (tries/tried) to limit your contact with your family? NO FAMILY......................... 1 2 8
e) He (insists/insisted) on knowing where you (are/were) at all time? WHERE YOU ARE ............. 1 2 8
f) He (does/did) not trust you with any money? MONEY ............................... 1 2 8

DV05 Now if you will permit me, I need to ask some more questions about
your relationship with your (last) husband/partner.
5B. How many times did this happen
5A. (Does/did) your (last) husband/partner ever: during the last 12 months?

a) say or do something to humiliate you in front of YES 1 ─< ┌──┬──┐

others? NO 2┐ TIMES IN LAST 12 MONTHS .... │░░│░░│
? └──┴──┘
b) Threaten you or someone close to you with YES 1 ─< ┌──┬──┐
harm? NO 2┐ TIMES IN LAST 12 MONTHS .... │░░│░░│
? └──┴──┘

Appendix A 109
 NO. QUESTIONS AND FILTERS CODING CATEGORIES SKIP

DV06 6A. (Does/did) your (last) husband/partner ever: 6B. How many times did this happen
during the last 12 months?

a) push you, shake you, or throw something at YES 1 ─< ┌──┬──┐

you? NO 2┐ TIMES IN LAST 12 MONTHS .... │░░│░░│
? └──┴──┘
b) slap you or twist your arm? YES 1 ─< ┌──┬──┐
NO 2┐ TIMES IN LAST 12 MONTHS .... │░░│░░│
? └──┴──┘
c) punch you with his fist or with something that YES 1 ─<
could hurt you? ┌──┬──┐
NO 2┐ TIMES IN LAST 12 MONTHS .... │░░│░░│
? └──┴──┘
YES 1 ─<
d) kick you or drag you? ┌──┬──┐
NO 2┐
TIMES IN LAST 12 MONTHS .... │░░│░░│
? YES 1 └──┴──┘
e) try to strangle you or burn you? ─<
NO 2┐ ┌──┬──┐
? TIMES IN LAST 12 MONTHS .... │░░│░░│
└──┴──┘
f) threaten you with a knife, gun, or other type of
weapon? YES 1 ─< ┌──┬──┐
NO 2┐ TIMES IN LAST 12 MONTHS .... │░░│░░│
g) attack you with a knife, gun, or other type of ? └──┴──┘
weapon? YES 1 ─< ┌──┬──┐
NO 2┐ TIMES IN LAST 12 MONTHS .... │░░│░░│
h) physically force you to have sexual intercourse ? └──┴──┘
with him even when you did not want to? YES 1 ─< ┌──┬──┐
NO 2┐ TIMES IN LAST 12 MONTHS .... │░░│░░│
i) force you to perform other sexual acts you did ? └──┴──┘
not want to? YES 1 ─<
┌──┬──┐
NO 2┐
TIMES IN LAST 12 MONTHS .... │░░│░░│
? └──┴──┘

DV07 CHECK DV06:

AT LEAST ONE ┌─┐ NOT A SINGLE ┌─┐

'YES' ├─┘ 'YES' └─┴────────────────────────────────────── ─<DV09
?

DV08 How long after you first got married to/started living with your (last) ┌──┬──┐
husband/partner did (this/any of these things) first happen? NUMBER OF YEARS ................. │░░│░░│
└──┴──┘
IF LESS THAN ONE YEAR, RECORD '00'. BEFORE MARRIAGE/BEFORE
LIVING TOGETHER ......................... 95
AFTER SEPARATION/DIVORCE.......... 96

DV09 Did the following ever happen because of something your (last) DV8B. How many times did this happen
husband/partner did to you: during the last 12 months?
┌──┬──┐
a) You had bruises and aches? YES 1 ─< TIMES IN LAST 12 MONTHS .... │░░│░░│
NO 2┐ └──┴──┘
?
b) You had an injury or a broken bone? ┌──┬──┐
YES 1 ─<
TIMES IN LAST 12 MONTHS .... │░░│░░│
NO 2┐
└──┴──┘
?
c) You went to the doctor or health center as a YES 1 ─< ┌──┬──┐
result of something your husband/partner did TIMES IN LAST 12 MONTHS .... │░░│░░│
NO 2┐
to you?
? └──┴──┘

DV10 Have you ever hit, slapped, kicked or done anything else to physically
hurt your (last) husband/partner at times when he was not already YES .......................................................... 1
beating or physically hurting you? NO............................................................ 2 ─<DV12

DV11 In the last 12 months, how many times have you hit, slapped, kicked or ┌──┬──┐
done something to physically hurt your (last) husband/partner at a time NUMBER OF TIMES .................. │░░│░░│
when he was not already beating or physically hurting you? └──┴──┘
1
DV12 Does (did) your husband/partner drink (alcohol)? YES .......................................................... 1
NO............................................................ 2 ─<DV14
1
Other intoxicants can be substituted/added as relevant

110 Appendix A
 NO. QUESTIONS AND FILTERS CODING CATEGORIES SKIP

DV13 How often does (did) he get drunk: very often, only sometimes, or VERY OFTEN .......................................... 1
never? SOMETIMES ........................................... 2
NEVER..................................................... 3

DV14 CHECK 501, 502 & 504:

MARRIED/LIVING WITH ┌─┐ WIDOWED/ ┌─┐

A MAN/SEPARATED/ ├─┘ NEVER MARRIED/NEVER ├─┘
DIVORCED │ LIVED WITH A MAN │
? ?
From the time you were 15 years From the time you were 15 years YES .......................................................... 1
old has anyone other than your old has anyone ever hit, slapped, NO............................................................ 2 ┐
(current/last) husband/partner hit, kicked, or done anything else to NO ANSWER ........................................... 6 ┴<DV19
slapped, kicked, or done anything hurt you physically?
else to hurt you physically?

DV15 Who has physically hurt you in this way? MOTHER ................................................. A
FATHER................................................... B
STEP-MOTHER ....................................... C
STEP-FATHER ........................................ D
SISTER .................................................... E
Anyone else? BROTHER ............................................... F
DAUGHTER .............................................G
SON ......................................................... H
RECORD ALL MENTIONED. LATE/EX-HUSBAND/EX-PARTNER ........ I
CURRENT BOYFRIEND ..........................J
FORMER BOYFRIEND ........................... K
MOTHER-IN-LAW.................................... L
FATHER-IN-LAW .................................... M
OTHER FEMALE RELATIVE/IN-LAW ..... N
OTHER MALE RELATIVE/ IN-LAW.........O
FEMALE FRIEND/ACQUAINTANCE....... P
MALE FRIEND/ACQUAINTANCE ...........Q
TEACHER ................................................ R
EMPLOYER ............................................. S
STRANGER ............................................. T

OTHER X
(SPECIFY)

DV16 CHECK DV15:

MORE THAN ┌─┐ ONLY ONE ┌─┐

ONE PERSON ├─┘ PERSON └─┴────────────────────────────────────── ─<DV18
MENTIONED ? MENTIONED

DV17 Who has hit, slapped, kicked, or done something to physically hurt you MOTHER ............................................... 01
most often? FATHER................................................. 02
STEP-MOTHER ..................................... 03
STEP-FATHER ...................................... 04
SISTER .................................................. 05
BROTHER ............................................. 06
DAUGHTER ........................................... 07
SON ....................................................... 08
LATE/EX-HUSBAND/EX-PARTNER ..... 09
CURRENT BOYFRIEND ....................... 10
FORMER BOYFRIEND ......................... 11
MOTHER-IN-LAW.................................. 12
FATHER-IN-LAW ................................... 13
OTHER FEMALE RELATIVE/IN-LAW ... 14
OTHER MALE RELATIVE/IN-LAW........ 15
FEMALE FRIEND/ACQUAINTANCE..... 16
MALE FRIEND/ACQUAINTANCE ......... 17
TEACHER .............................................. 18
EMPLOYER ........................................... 19
STRANGER ........................................... 20

OTHER 96
(SPECIFY)

Appendix A 111
 NO. QUESTIONS AND FILTERS CODING CATEGORIES SKIP

DV18 In the last 12 months, how many times has this person hit, slapped, ┌──┬──┐
kicked, or done anything else to physically hurt you? NUMBER OF TIMES .................. │░░│░░│
└──┴──┘

DV19 CHECK 201, 225, AND 226:

HAS ONE OR MORE LIVE NO LIVE BIRTHS,

OR NON-LIVE BIRTHS ┌─┐ NO NON-LIVE BIRTHS, ┌─┐
OR IS CURRENTLY ├─┘ AND IS NOT CURRENTLY └─┴─────────────────────────────── ─<DV21
PREGNANT ? PREGNANT

DV20 Has any one ever hit, slapped, kicked, or done anything else to hurt you YES .......................................................... 1
physically while you were pregnant? NO............................................................ 2 ─<DV22

DV21 Who has done any of these things to physically hurt you while you were CURRENT HUSBAND/PARTNER........... A
pregnant? MOTHER ................................................. B
FATHER................................................... C
STEP-MOTHER ....................................... D
STEP-FATHER ........................................ E
Anyone else? SISTER .................................................... F
BROTHER ...............................................G
RECORD ALL MENTIONED. DAUGHTER ............................................. H
SON .......................................................... I
LATE/LAST/EX-HUSBAND/PARTNER ....J
CURRENT BOYFRIEND ......................... K
FORMER BOYFRIEND ........................... L
MOTHER-IN-LAW................................... M
FATHER-IN-LAW ..................................... N
OTHER FEMALE RELATIVE/IN-LAW .....O
OTHER MALE RELATIVE/IN-LAW.......... P
FEMALE FRIEND/ACQUAINTANCE.......Q
MALE FRIEND/ACQUAINTANCE ........... R
TEACHER ................................................ S
EMPLOYER ............................................. T
STRANGER ............................................. U

OTHER X
(SPECIFY)

DV22 CHECK DV06, DV09, DV14, AND DV20:

AT LEAST ONE ┌─┐ NOT A SINGLE ┌─┐

'YES' ├─┘ 'YES' └─┴────────────────────────────────────── ─<DV26
?

DV23 Have you ever tried to get help to prevent or stop (this person/these YES .......................................................... 1
persons) from physically hurting you? NO............................................................ 2 ─<DV25

DV24 From whom have you sought help? MOTHER ................................................. A ┐

FATHER................................................... B │
SISTER .................................................... C │
BROTHER ............................................... D │
CURRENT/LAST/LATE HUSBAND/ │
Anyone else? PARTNER ........................................ E │
CURRENT/FORMER BOYFRIEND......... F │
RECORD ALL MENTIONED MOTHER-IN-LAW....................................G │
FATHER-IN-LAW ..................................... H │
OTHER FEMALE RELATIVE/IN-LAW ...... I │
OTHER MALE RELATIVE/ IN-LAW..........J │
FRIEND.................................................... K │
NEIGHBOR .............................................. L ├<DV26
TEACHER ............................................... M │
EMPLOYER ............................................. N │
RELIGIOUS LEADER ..............................O │
DOCTOR/MEDICAL PERSONNEL ......... P │
POLICE ....................................................Q │
LAWYER .................................................. R │
│
OTHER X │
(SPECIFY) │
│
┘

112 Appendix A
 NO. QUESTIONS AND FILTERS CODING CATEGORIES SKIP

DV25 What is the main reason you have never sought help? DON'T KNOW WHO TO GO TO............ 01
NO USE ................................................. 02
PART OF LIFE ....................................... 03
AFRAID OF DIVORCE/DESERTION .... 04
AFRAID OF FURTHER BEATINGS ...... 05
AFRAID OF GETTING PERSON
BEATING HER INTO TROUBLE ...... 06
EMBARRASSED ................................... 07
DON'T WANT TO DISGRACE
FAMILY ............................................ 08

OTHER 96
(SPECIFY)

DV26 As far as you know, did your father ever beat your mother? YES .......................................................... 1
NO............................................................ 2
DON'T KNOW .......................................... 8

THANK THE RESPONDENT FOR HER COOPERATION AND REASSURE HER ABOUT THE CONFIDENTIALITY OF HER
ANSWERS. FILL OUT THE QUESTIONS BELOW WITH REFERENCE TO THE DOMESTIC VIOLENCE MODULE ONLY.

DV27 DID YOU HAVE TO INTERRUPT THE YES YES, MORE

INTERVIEW BECAUSE SOME ADULT WAS ONCE THAN ONCE NO
TRYING TO LISTEN, OR CAME INTO THE HUSBAND ..................................... 1 2 3
ROOM, OR INTERFERED IN ANY OTHER OTHER MALE ADULT .................. 1 2 3
WAY? FEMALE ADULT ........................... 1 2 3

DV28 INTERVIEWER'S COMMENTS / EXPLANATION FOR NOT COMPLETING THE DOMESTIC VIOLENCE MODULE

Appendix A 113
Variants of the Recommended DHS Domestic Violence Module

Colombia (2000)
1. Please tell me if your spouse/partner speaks to you in such terms as “You are useless,” “You
never do anything good,” “You are a bruta,” “My mother did those things better for me.”

2. Do these situations happen in public or private?

3. Has your spouse threatened you frequently, sometimes, or never with

- Abandoning you for another woman?
- Leaving his children?
- Ceasing economic support?

4. Has anyone hit, slapped, kicked, or injured you when you were pregnant (in any of your
pregnancies)?

5. Who?

6. Does your husband frequently, sometimes, or never

a) push you or shake you?

b) hit you with his hand?
c) hit you with a hard object?
d) bite you?
e) kick or drag you?
f) threaten you with a knife, gun, or other type of weapon?
g) attack you with a knife, gun or other type of weapon?
h) try to strangle or burn you?
i) physically force you to have sexual intercourse or perform types of other sexual acts
even when you did not want to?

7. How long after you first got married to your (last) husband did (this/any of these things) first
happen?

8. Did the following ever happen because of something your (last) husband did to you?

a) you had bruises and aches

b) you had an injury or a broken bone
c) you had a pregnancy that ended in abortion or loss
d) you lost temporarily or permanently an organ, a physical function, or part of the
body.

9. Did you have to go to a doctor or medical center as a result of what your husband/partner
did?

10. When you were attacked by your (last) husband/partner, did you defend yourself with
punches or physical aggression? IF YES: how often?

114 Appendix A
11. Have you ever hit or physically attacked your husband at times when he was not already
beating or physically hurting you?

EVER-MARRIED
12. Has anyone other than your current/(last) husband hit, slapped, kicked or done anything else
to hurt you physically?

NEVER-MARRIED
Has anyone hit, slapped, kicked or done anything else to hurt you physically?

12A. Who has physically hurt you this way?

13. Has anyone hit, slapped, kicked, or injured you when you were pregnant? (in any of your
pregnancies)

14. Who?

Nicaragua (1998)
1. From the time you were 15 years old has anyone hit you or done anything else to hurt you
physically?

2. Who has physically hurt you this way? How frequently? Anyone else?

3. Now if you will permit me, I need to ask some more questions about your relationship with
your (last) husband. Does your husband ever

a) say or do something to humiliate you in front of others?

b) threaten you or someone close to you with harm?
c) push you, shake you or throw something at you?
d) slap you or twist your arm?
e) punch you with his fist or something that could hurt you?
f) kick or drag you?
g) try to strangle or burn you?
h) threaten you with a machete, gun, or other type of weapon?
i) attack you with a knife, gun or other type of weapon?
j) physically force you to have sexual intercourse even when you did not want to?
k) threaten you in order to have sexual intercourse even when you did not want to?
l) force you to perform types of other sexual acts you did not want to?

3A. How many times did this happen during the last 12 months?

4. How many times in the past 12 months (how many time previously) did the following ever
happen because of something your (last) husband did to you?

a) you had bruises and aches

b) you had an injury or a broken bone
c) you went to a health facility as a result of something your husband had done to you

Appendix A 115
5. Generally, have your children been present or within hearing distance during the time that he
was beating you?

6. During the times that you have been beaten, did you ever hit your husband/partner in self-
defense?

7. Do you remember if you have ever been the one to hit first?

8. Has your husband/partner ever hit you while you were pregnant?

9. During how many pregnancies?

Questions on Domestic Violence Used in Other Countries

Egypt
1. From the time you were married has anyone ever beaten you?

2. Can you tell me who has done this to you since you were married?
Anyone else?

3. Who is the person who beats you most often?

4. Is this person always, sometimes, or never “on something” (drugs or alcohol) when he/she
beats you?

5. Approximately, how many times were you beaten in the past one year?

6. What do you generally do when you are being beaten?

7. What is the most common reason for which you are beaten?

8. Generally, are you hurt as a result of a beating?

PROBE: Any bruises, aches, or pains?

9. Have you ever been beaten when you were pregnant?

10. Were you beaten more often or less often when you were pregnant, as compared to when you
were not pregnant?

11. Since you became pregnant, have you ever been beaten?

12. Are you beaten more often or less often now that you are pregnant as compared to when you
were not pregnant?

13. Have you ever been so seriously hurt during a beating that you needed medical attention even
if you did not see a doctor?

14. How often has this happened?

116 Appendix A
India 1998-99
1. Since you completed 15 years of age, have you been beaten or mistreated physically by any
person?

2. Who has beaten you or mistreated you physically?

Anyone else?

3. How often have you been beaten or mistreated physically in the last 12 months: once, a few
times, many times, or not at all?

Peru 2000
1. Has your spouse or partner ever pushed you, hit you, or attacked you physically?

2. Has your spouse/partner attacked you physically frequently, sometimes, or never?

FOR EVER-MARRIED WOMEN

3. Has anyone other than your current/(last) husband hit, slapped, kicked or done anything else
to hurt you physically? Who?

FOR NEVER-MARRIED WOMEN

Has anyone hit, slapped, kicked or done anything else to hurt you physically? Who?

Appendix A 117
 Appendix B

Distribution of Ever-married Women and Currently Married Women by

Variables Used in the Multivariate Logistic Regression

Table B.1 Percent distribution of ever-married women (EMW) and currently married women (CMW) by variables used in the multivariate logistic regression
(unweighted)
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Dominican
Cambodia Colombia Republic Egypt Haiti India Nicaragua Peru Zambia
–––––––––– –––––––––– –––––––––– –––––––––– –––––––––– –––––––––– –––––––––– –––––––––– ––––––––––
Variable EMW CMW EMW CMW EMW CMW EMW CMW EMW CMW EMW CMW EMW CMW EMW CMW EMW CMW
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Age
15-19 3.8 4.2 5.5 5.8 8.1 8.2 5.3 5.7 5.1 5.2 7.8 8.1 8.8 8.7 4.2 4.4 8.4 9.3
20-24 9.8 10.7 13.3 13.9 15.7 15.8 15.1 15.8 13.5 14.1 17.7 18.3 16.5 16.9 13.7 14.0 22.9 24.5
25-29 16.9 17.8 16.9 17.7 18.7 19.2 18.8 19.8 18.5 19.5 20.1 20.6 19.2 20.0 17.5 18.0 23.1 23.8
30-34 21.5 22.4 17.7 18.3 18.9 19.8 17.4 17.9 18.2 18.3 17.5 17.7 18.9 19.3 19.4 19.7 16.5 16.3
35-39 19.8 19.4 18.3 18.2 16.7 16.7 17.2 17.1 17.8 17.6 15.2 15.1 16.1 15.8 17.6 17.5 12.4 11.7
40-44 15.1 14.1 15.4 14.5 12.3 11.7 13.2 12.4 13.9 13.2 12.2 11.6 12.0 11.5 15.3 15.1 9.2 8.4
45-49 13.1 11.4 12.9 11.6 9.7 8.6 13.1 11.3 13.0 12.1 9.5 8.6 8.5 7.8 12.2 11.3 7.5 6.0

Age at marriage
<15 3.8 3.9 7.2 6.9 17.6 17.6 15.2 15.0 7.6 7.5 18.5 18.1 19.9 19.9 7.3 7.4 14.1 13.6
15-19 55.3 56.0 46.2 46.1 53.6 54.0 50.7 50.7 46.8 46.9 58.7 58.8 56.0 56.9 48.8 49.2 66.2 67.0
20-24 31.2 31.0 31.5 31.8 21.1 21.2 25.7 25.9 32.1 32.0 18.9 19.2 18.6 18.2 30.4 30.1 16.4 16.4
25+ 9.7 9.1 15.1 15.3 7.6 7.3 8.4 8.3 13.6 13.6 3.9 3.9 5.5 4.9 13.4 13.3 3.3 3.0

Number of unions
1 91.7 92.3 81.8 82.9 63.3 65.2 95.5 95.9 62.9 64.7 98.2 98.3 71.3 73.3 89.7 90.4 77.4 78.3
2+ 8.3 7.7 18.2 17.1 36.7 34.8 4.5 4.1 37.1 35.3 1.8 1.7 28.7 26.7 10.3 9.6 22.6 21.7

Number of children
ever born
0 6.4 6.1 7.0 7.1 8.0 7.7 10.4 10.4 7.4 7.8 10.4 10.3 6.1 6.2 4.7 4.7 6.3 6.5
1-2 27.5 26.4 48.8 47.7 38.3 36.4 25.4 25.3 29.6 28.4 36.0 36.0 35.4 33.2 40.2 39.2 31.4 31.4
3-4 30.2 30.4 30.9 31.7 37.1 38.7 27.1 27.3 25.7 25.7 33.1 33.3 28.7 29.4 28.7 29.1 26.5 26.7
5+ 35.9 37.1 13.3 13.5 16.6 17.2 37.1 37.0 37.2 38.2 20.5 20.4 29.9 31.2 26.4 27.0 35.8 35.4

Education level
No education 34.9 34.3 4.4 4.7 6.5 6.8 50.8 49.8 43.9 44.5 49.8 49.2 21.4 21.6 8.5 8.5 16.0 15.8
Primary 53.9 54.3 40.1 40.9 54.8 55.2 23.3 23.1 40.7 40.3 17.0 16.9 45.8 46.5 40.5 41.2 63.0 63.6
Secondary or higher 11.2 11.3 55.4 54.4 38.7 38.0 25.8 27.1 15.4 15.2 33.1 33.9 32.8 31.9 51.1 50.2 20.9 20.6

Work status
Not working 15.9 17.1 39.0 44.7 52.4 56.6 84.0 84.2 36.5 37.4 63.0 64.6 57.2 62.5 31.4 33.8 33.8 35.2
Working, not paid 14.2 14.7 3.3 3.9 1.5 1.6 2.7 2.7 0.6 0.7 13.0 13.3 3.5 3.4 17.6 19.0 28.4 28.5
Paid cash, in whole
or part 37.1 35.4 56.9 50.6 45.2 40.8 13.3 13.1 61.9 61.1 24.0 22.1 39.3 34.1 49.3 45.6 36.7 35.1
Paid in kind only 32.8 32.8 0.8 0.8 0.9 1.0 u u 0.9 0.9 u u u u 1.7 1.6 1.2 1.2

Husband's education
level
No education 17.9 17.5 5.1 5.2 7.6 7.7 33.3 31.5 29.7 30.6 26.1 25.2 21.4 22.3 2.2 2.0 8.2 7.6
Primary 53.6 54.2 39.8 42.0 49.1 52.2 28.9 29.5 38.0 39.1 18.1 18.2 42.7 44.6 33.3 34.3 49.0 52.3
Secondary or higher 27.0 27.8 54.2 52.6 35.9 34.1 37.6 38.8 24.6 23.7 55.5 56.4 32.8 31.4 63.8 63.3 41.2 39.5
Don't know/missing 1.5 0.5 0.8 0.2 7.5 6.0 0.2 0.2 7.7 6.6 0.2 0.2 3.1 1.7 0.8 0.4 1.6 0.7

Husband's occupation
Nonagricultural 30.4 29.1 74.6 72.2 76.7 75.4 68.8 69.6 95.7 96.0 64.3 64.4 60.5 58.2 61.1 59.4 43.8 41.0
Agricultural 69.6 70.9 25.4 27.8 23.3 24.6 31.2 30.4 4.3 4.0 35.7 35.6 39.5 41.8 38.9 40.6 56.2 59.0

Husband's
drunkenness
Doesn't drink 31.9 31.3 u u 30.3 32.2 u u 78.6 78.7 u u u u u u u u
Never gets drunk 4.4 4.6 32.1 34.9 23.4 24.1 u u 7.2 7.6 u u 46.5 46.5 22.0 23.0 u u
Occasionally drunk 50.5 52.5 49.0 52.1 35.2 34.7 u u 10.3 10.2 u u 41.9 41.9 68.4 69.8 u u
Frequently drunk 13.2 11.6 19.0 13.0 11.2 8.9 u u 3.9 3.5 u u 11.7 11.7 9.6 7.1 u u

Marital duration
0-4 years 12.7 13.7 20.2 21.6 17.1 17.5 19.1 20.2 19.3 19.8 19.4 20.2 16.4 16.5 18.4 18.9 23.2 25.2
5-9 years 19.9 20.5 20.5 21.3 21.2 21.2 18.1 18.9 19.8 20.5 19.1 19.7 22.3 22.7 20.0 20.1 22.8 23.3
10-14 years 21.4 22.0 18.2 18.0 20.5 21.4 17.7 18.2 19.3 19.9 17.8 18.2 18.6 19.1 19.4 19.7 19.2 19.7
15+ years 46.1 43.8 41.1 39.1 41.2 40.0 45.1 42.7 41.6 39.8 43.6 42.0 42.7 41.8 42.2 41.3 34.8 31.8

Continued…

Appendix B 119
Table B.1—Continued
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Dominican
Cambodia Colombia Republic Egypt Haiti India Nicaragua Peru Zambia
–––––––––– –––––––––– –––––––––– –––––––––– –––––––––– –––––––––– –––––––––– –––––––––– ––––––––––
Variable EMW CMW EMW CMW EMW CMW EMW CMW EMW CMW EMW CMW EMW CMW EMW CMW EMW CMW
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Spousal age
difference
Husband is younger u 18.9 u 16.6 u 14.4 u 2.9 u 13.1 u 2.7 u 16.8 u 14.4 u 2.1
Wife 0-4 years
younger u 51.4 u 38.8 u 33.7 u 27.4 u 35.5 u 38.8 u 40.4 u 45.6 u 34.0
Wife 5-9 years
younger u 21.0 u 26.6 u 26.8 u 38.5 u 28.0 u 40.7 u 24.0 u 26.1 u 42.7
Wife 10-14 years
younger u 5.8 u 11.3 u 13.1 u 20.7 u 12.9 u 13.2 u 10.9 u 9.5 u 13.1
Wife 15+ years
younger u 2.9 u 6.7 u 12.0 u 10.4 u 10.5 u 4.6 u 8.0 u 4.4 u 8.1

Spousal educational
difference
Wife has more
education 15.4 15.2 21.1 20.7 10.6 10.3 14.6 14.8 17.7 17.7 12.1 12.1 36.2 36.2 19.3 18.6 16.9 17.1
Both have no
education 12.8 12.6 1.3 1.5 3.3 3.5 27.4 25.9 26.6 27.0 23.3 22.6 11.3 11.8 1.2 1.2 4.1 3.9
Both have same
education 14.2 14.5 14.3 15.3 3.1 3.2 12.2 12.7 6.9 7.2 9.8 10.1 14.1 14.8 25.8 26.4 12.6 13.0
Husband has
more education 57.6 57.7 63.3 62.4 83.0 83.0 45.7 46.7 48.9 48.1 54.8 55.3 38.4 37.2 53.7 53.8 66.4 66.0

Residence
Urban 15.2 15.0 73.6 70.5 59.6 57.6 34.8 35.2 34.1 31.9 31.1 31.0 54.4 51.4 57.0 55.3 28.7 27.8
Rural 84.8 85.0 26.4 29.5 40.4 42.4 65.2 64.8 65.9 68.1 68.9 69.0 45.6 48.6 43.0 44.7 71.3 72.2

Family structure
Nonnuclear
(extended) 32.6 31.3 45.6 39.9 33.1 28.1 44.4 43.3 51.0 48.8 57.7 57.2 47.3 57.9 40.6 37.6 43.8 40.1
Nuclear 67.4 68.7 54.4 60.1 66.9 71.9 55.6 56.7 49.0 51.2 42.3 42.8 52.7 42.1 59.4 62.4 56.2 59.9

Wealth quintile
Lowest (poorest) 24.1 22.2 17.2 19.0 26.2 27.4 26.3 25.6 24.3 25.4 15.9 15.6 22.9 24.5 23.5 24.2 23.4 22.5
Second 21.7 21.9 21.1 21.6 25.4 24.9 22.2 22.1 20.1 20.9 16.9 16.8 21.4 21.2 23.8 24.1 22.2 22.3
Middle 20.3 20.7 22.3 21.6 20.2 20.1 19.0 19.0 21.2 21.6 19.7 19.6 20.1 20.0 22.4 22.1 23.4 24.2
Fourth 16.9 17.8 20.1 19.5 16.7 16.3 15.9 16.2 22.1 20.7 22.7 22.7 18.9 17.9 17.9 17.6 18.4 18.2
Highest (wealthiest) 17.1 17.3 19.3 18.3 11.5 11.4 16.6 17.0 12.3 11.4 24.9 25.3 16.7 16.4 12.4 12.0 12.7 12.8

Number of women 2,403 2,108 7,716 5,996 7,435 6,042 7,123 6,594 2,592 2,266 90,303 84,862 8,508 6,824 18,196 15,995 4,151 3,492
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
u = Unknown (not available)

120 Appendix B