Gender, women and health

WHO Multi-country Study on Women's Health and Domestic Violence against Women


The Study was coordinated by WHO with a core research team of international experts from the London School of Hygiene and Tropical Medicine, the Program for Appropriate Technology in Health and WHO. WHO also established an expert steering committee of researchers and advocates from different regions of the world to provide technical and scientific oversight.

Within each participating country, a collaborative research team was established to implement the WHO Study. This generally consisted of representatives from research organizations experienced in survey work, and a women’s organization providing services to abused women. Each country research team also established a consultative committee to support the study and ensure the dissemination of the results. To ensure that comparability between countries was maintained, and that the same issues and concepts were explored and analysed in the same way in each country, members of the core research team visited each country at key points, for example during interviewer training and pilot-testing.

Choice of countries and settings

Participating countries were chosen according to specified criteria, including the presence of local anti-violence groups able to use the data for advocacy and policy reform, absence of existing data, and a political environment receptive to tackling the issue.

"On the one hand it made me feel good, because it was something that I had never told anyone before. Now I've told someone."
-Woman interviewed in Brazil

In each country, the Study consisted of a cross-sectional, population-based household survey conducted in one or two settings. In half of the countries (Bangladesh, Brazil, Peru, Thailand, and the United Republic of Tanzania), surveys were conducted in (a) the capital or a large city and (b) one province or region, usually containing both urban and rural populations (see also Note 1.2). A single rural setting was used in Ethiopia, and a single large city was used in Japan, Namibia, and Serbia and Montenegro. In Samoa, the whole country was sampled. Figure 1 shows the countries participating in the WHO Study.

Definitions and measurement tools

One of the main challenges facing international research on violence against women is to develop clear definitions of different types of violence, that permit meaningful comparisons among diverse settings. Given that the way people think about violence differs between individuals and communities, the Study used conservative definitions of violence.2 The results are therefore more likely to underestimate than overestimate the true prevalence of violence.

The Study questionnaire resulted from a long process of discussion and consultation. The first draft of the core questionnaire was developed by the core research team, and was reviewed first by the expert steering committee and other experts, and then by the country teams. The questionnaire was then translated and pretested in six countries (Bangladesh, Brazil, Namibia, Samoa, Thailand, and the United Republic of Tanzania) and retested in the remaining participating countries.

The use of a single methodology across countries3 greatly reduced the difficulties that have plagued earlier research, particularly the degree to which differences in sampling, definitions, questions used, and so on, account for differences in results. While cultural biases that affect disclosure will always remain, the methodological consistency of the WHO Study ensured that the variations found in prevalence between and within countries represent, for the most part, real differences.

Since the initiation of the WHO Study, other international research initiatives have also used population-based surveys to estimate the prevalence of violence against women across countries and cultures.4 These studies provide useful comparisons with the WHO Study and, taken together, now begin to give a more comprehensive picture of violence against women around the world.

Interviewer selection and training

Women’s willingness to disclose violence is known to be influenced by characteristics of the interviewers such as sex, age, marital status, attitudes, and interpersonal skills (8). The WHO Study used female interviewers and supervisors who were selected according to criteria such as emotional maturity, ability to engage with people of different backgrounds in an empathetic and non-judgemental manner, and skills in dealing with sensitive issues. All were trained using a standardized three-week curriculum designed especially for the WHO Study (9).

Ethical and safety guidelines

The WHO Study developed a set of ethical guidelines for its work, entitled Putting women first: ethical and safety recommendations for research on domestic violence against women (10). These guidelines were adhered to in each country. Ethical permission for the study was obtained from WHO’s own ethical review group, from each of the local institutions’ ethical review boards and, in countries where it was required, from national review boards and ministries of health. In all countries, the overwhelming impression was that women were not only willing to talk about their experiences of violence, but were often deeply grateful for the opportunity to tell their stories to a non-judgemental, empathetic person (9).

"I learned a lot, from the beginning of the training till the end of the survey...The respondents really needed and enjoyed this experience, because they could talk to somebody. My career path [has] changed since the beginning of the training because I could do something which can make a difference and mean something for my country."
-Interviewer in Namibia

National dissemination and follow-up

In each country, the national results were used to produce a country report. These country reports have been disseminated at the local and national level in coordination with the country research teams, advisory groups, WHO country offices, and relevant ministries. The findings are being used in ongoing advocacy and in shaping policies, laws and programmes.

For future analysis

This report presents a summary of the initial results of the WHO Study, and reflects only a small part of the overall analysis. The database has the potential to address other important questions of relevance to public health. For example, the study investigated a number of factors that may put women at risk of partner violence, or conversely protect her. It further addressed the broader consequences of partner violence against women, including how violence affects an individual’s ability to provide for her family, maintain a job, keep her income, stay in contact with her relatives, and be an active member of groups or associations. The study also explored a range of consequences for women’s children, and the extent to which children have witnessed physical violence against their mother. Among other questions, women were asked about the birth weight of their most recent live birth in the past 5 years, their children’s attendance at school, behavioural problems, and whether any children had run away from home.

More in-depth analysis of these and other research questions will be explored in future reports and in papers to be published in the peer-reviewed literature.

Note 1.2 Certain conventions about terminology have been adopted in this summary report to reduce redundancy and make it more readable. The principal one is to permit a description of respondents as “having experienced violence” or “having sought help” when in fact the data are based on self-reporting rather than observation. Second, the term “abuse” is used frequently as a synonym for violence; thus, “ever-abused” means ever having experienced violence (or, bearing in mind the previous point, having reported ever experiencing violence). Third, in countries where research was conducted in two settings, “urban” is used to describe the capital or other large city setting and “provincial” to describe the second setting, which may have been rural or a mixture of both rural and urban. Finally, when both settings in one country produced similar results, the country name only may be used; for example, the statement “family size was larger in Bangladesh and provincial Peru”, means that family size was larger in both settings in Bangladesh but only the provincial setting in Peru. In the figures the terms “city” and “province” are used.

2The operational definitions for the different forms of violence are given in Chapter 2 of this report.

3In Japan, important modifications were made to the methodology.

4These include the World Surveys of Abuse in Family Environments (Worldsafe) and the International Violence Against Women Survey (IVAWS). In addition, the Demographic and Health Surveys (DHS) and the CDC-supported International Reporductive Health Surveys (IRHS) increasingly contain questions on violence against women as part of larger household surveys on a range of health issues.