Understanding women's attitudes towards wife beating in Zimbabwe
Comprendre l'attitude des femmes envers les brutalités conjugales au Zimbabwe
Análisis de la actitud de las mujeres ante el maltrato a las esposas en Zimbabwe
Michelle J. Hindin
Assistant Professor, Department of Population and Family Health Sciences, Bloomberg School of Public Health, The Johns Hopkins University, 615 N. Wolfe Street, E413, Baltimore, MD 21205-2179, USA (email: firstname.lastname@example.org)
OBJECTIVE: To investigate the factors associated with attitudes towards wife beating among women in partnerships in Zimbabwe in order to assist public health practitioners in preventing intimate partner violence (IPV).
METHODS: A nationally representative survey of 5907 women of reproductive age (1549 years) was conducted in Zimbabwe. Women were asked about their attitudes towards wife beating in five situations. The survey included sociodemographic characteristics, partnership characteristics, and household decision-making.
FINDINGS: Over half of all women in Zimbabwe (53%) believed that wife beating was justified in at least one of the five situations. Respondents were most likely to find wife beating justified if a wife argued with her spouse (36%), neglected her children (33%), or went out without telling her spouse (30%). Among women in partnerships (n = 3077), younger age, living in rural areas, lower household wealth, schooling at a lower level than secondary, and lower occupational status were associated with women reporting that wife beating is justified. Women who reported that they make household decisions jointly with their partners were less likely to say that wife beating is justified.
CONCLUSIONS: Zimbabwe has a long way to go in preventing IPV, particularly because the younger generation of women is significantly more likely to believe that wife beating is justified compared with older women. Given the current social and political climate in Zimbabwe, finding means to negotiate rather than settle conflict through violence is essential from the household level to the national level.
Keywords: Spouse abuse/psychology/ethnology; Women; Attitude/ethnology; Gender identity; Interpersonal relations; Social environment; Social class; Odds ratio; Regression analysis; Zimbabwe (source: MeSH, NLM).
OBJECTIF: Etudier les facteurs associés aux attitudes envers les brutalités conjugales chez les femmes mariées ou vivant maritalement au Zimbabwe afin d'aider les acteurs de la santé publique à prévenir les violences conjugales.
MÉTHODES: Une enquête a été réalisée au Zimbabwe sur un échantillon national représentatif de 5907 femmes en âge de procréer (15-49 ans). Les femmes ont été interrogées sur leur attitude envers les brutalités conjugales dans cinq situations hypothétiques. Lors de l'enquête ont également été recueillies des données sur les caractéristiques socio-démographiques du ménage, le statut de la femme par rapport à son mari/partenaire et la prise de décision au sein du ménage.
RÉSULTATS: Au Zimbabwe, plus de la moitié de l'ensemble des femmes (53 %) pensaient qu'il était justifié que la femme soit battue dans au moins une des cinq situations présentées. Les femmes enquêtées étaient davantage susceptibles de trouver justifié que la femme soit battue si elle se disputait avec son mari/ partenaire (36 %), négligeait ses enfants (33 %) ou sortait sans avertir son mari/partenaire (30 %). Parmi les femmes mariées ou vivant maritalement (n = 3077), les facteurs tels que l'âge (femmes plus jeunes), la résidence en zone rurale, le faible niveau économique du ménage, le niveau d'études ne dépassant pas le primaire et la situation professionnelle inférieure étaient associés au fait de considérer le fait d'être battue comme justifié. Les femmes qui prenaient les décisions d'ordre domestique en commun avec leur mari/partenaire trouvaient moins souvent une justification aux brutalités conjugales.
CONCLUSIÓN: Le Zimbabwe a encore un long chemin à parcourir en ce qui concerne la prévention des brutalités conjugales, d'autant plus que les jeunes générations de femmes sont beaucoup plus enclines que leurs aînées à trouver justifié que la femme soit battue par son mari/partenaire. Etant donné l'actuel climat social et politique qui règne dans le pays, il est indispensable de trouver à tous les niveaux, dans les ménages comme au niveau national, des moyens de régler les conflits par la négociation plutôt que par la violence.
Mots clés: Epouse maltraitée/psychologie/éthnologie; Femmes; Attitude/éthnologie; Identité masculin féminin; Relation interpersonnelle; Environnement social; Classe sociale; Odds ratio; Analyse régression; Zimbabwe (source: MeSH, INSERM).
OBJETIVO: Investigar los factores asociados a la actitud demostrada por las mujeres con pareja de Zimbabwe ante los casos de maltrato de la esposa, a fin de ayudar a los profesionales de la salud pública a prevenir la violencia de pareja.
MÉTODOS: Se llevó a cabo en Zimbabwe una encuesta nacionalmente representativa entre 5907 mujeres en edad fecunda (1549 años), a las que se preguntó su opinión sobre el maltrato de las esposas en cinco situaciones. La encuesta incluía datos sociodemográficos e información sobre las características de la relación y sobre la adopción de decisiones en el hogar.
RESULTADOS: Globalmente, más de la mitad de las mujeres de Zimbabwe (53%) creían que el maltrato a las esposas estaba justificado por lo menos en una de las cinco situaciones planteadas. Las entrevistadas tendían sobre todo a considerar justificado el maltrato cuando la esposa hubiese discutido con su cónyuge (36%), desatendido a sus hijos (33%) o salido de casa sin decírselo al marido (30%). Entre las mujeres con pareja (n=3077), la menor edad, el hecho de vivir en zonas rurales, la menor riqueza del hogar, un nivel de escolarización inferior al secundario y un menor estatus laboral se asociaban a una mayor tendencia a justificar el maltrato de las esposas. Las mujeres que declararon que tomaban las decisiones domésticas conjuntamente con su pareja demostraron una menor tendencia a justificar el maltrato conyugal.
CONCLUSIÓN: En Zimbabwe queda aún mucho por hacer para prevenir la violencia de pareja. En particular porque, en comparación con las mujeres de más edad, la generación más joven está significativamente más inclinada a considerar justificado el maltrato de la esposa. Dado el actual clima social y político reinante en Zimbabwe, el hallazgo de fórmulas de negociación que reemplacen el recurso a la violencia como medio de dirimir los conflictos es esencial a todos los niveles, desde el doméstico al nacional.
Palabras clave: Maltrato conyugal/psicología/etnología; Mujeres; Actitud/etnología; Identidad sexual; Relaciones interpersonales; Medio social; Clase social; Razón de diferencia; Análisis de regresión; Zimbabwe (fuente: DeCS, BIREME).
International estimates suggest that among people 1544 years of age, intimate partner violence (IPV) is the third leading cause of death (1). Yet, the factors that help to explain why, for example, one in five women has been physically or sexually assaulted by a man at some time in their lives (2) have not been well explored. Currently, WHO is undertaking a multi-country study to "explore the dimensions, health consequences and risk factors of violence" (3), and in October 2002, WHO released The world report on violence and health (4). A recent article suggested that in order to understand the causes of IPV, one needs to understand the "social context" in which it occurs, including the status of women, gender norms, socioeconomic status, and relationship dynamics (5). The present paper explores the gender dynamics associated with women's attitudes towards wife beating in a social context where there has been a long-held tradition of male dominance.
Attitudes towards IPV
Although most of the research on IPV has been conducted in developed nations, a growing body of research from the developing world has explored people's attitudes towards IPV. Most of the prevalence data has come from Latin America, where between 8% and 32% of men and women report that wife beating is justified in cases of suspected infidelity (6). Among the top reasons for society accepting wife abuse are refusing to have sex and neglecting the children. Evidence taken from different settings for example, among South African male and female public health care nurses (7), women in a community-based study in Uganda (8), and in a study of women from two regions in India (9) also indicates that IPV is considered to be a normal part of marital relationships. IPV was found to be justified by 4190% of the female respondents in at least one situation.
Relationship dynamics, women's status, and experiencing IPV
Domestic violence, according to Garciá-Moreno, is "supported or reinforced by gender norms and values that put women in a subordinate position to men" (10). Although evidence on gender dynamics in marriage and experience of IPV is limited, women might be at greater risk of IPV when their husbands have a lower relative status than they do, in terms of education or occupational prestige (5). Women are at higher risk of abuse if they live in a society that promotes male dominance, where men have conservative ideas about women, or where women have liberal ideas about the position of women. Women with very poor or very good education are least likely to be abused; however, women who have enough education to challenge existing gender norms are at the greatest risk of abuse (5). Decision-making autonomy in marriage is also associated with higher levels of abuse in the Philippines, where autonomous decision-making by either partner is associated with a higher risk of abuse, whereas decisions made jointly are protective (11).
Violence and gender in Zimbabwe
A report summarizing reproductive health in Zimbabwe (12) estimated that one in six women had been sexually abused by a man (13). Much of the wife beating in Zimbabwe is likely to stem from gender roles and social expectations. Violence against women in Zimbabwe is common and women are vulnerable because "of their low status and lack of power in the family, and because violence within marriage is widely tolerated" (14). Because violence is seen as an acceptable way to settle disputes and exercise control within marriage, women are thought to perceive that violence is acceptable and within the social norms (14).
Women lack substantial legal recourse if they experience abuse, which is a further indication that acceptance of violence within relationships is normal in Zimbabwe. For example, there are no laws that assign penalties for domestic abuse. Women could apply for a "binding-over" against anyone "who has conducted himself violently or threatened injury" or has used language or behaved in a manner that is likely to "provoke a breach of the peace or assault" (12). The binding-over can require the abuser to pay up to 200 Zimbabwean dollars (US$ 20) and if the binding-over is violated, the perpetrator forfeits the money and could be put in jail. Legally, victims of domestic violence can file for civil damages from injury (12). Although women in traditional (i.e. customary) marriages can use domestic violence as grounds for separation or divorce, police often treat domestic violence as a "domestic" problem rather than a criminal problem (13) and few women are aware even of their limited legal rights (15).
On the basis of the literature review above it is hypothesized in this paper that a high proportion of women will believe that wife beating is justified. The other main hypotheses being tested are: (1) women whose relative status is lower than their partners' status will be more likely to believe that wife beating is justified; and (2) Couples who make joint decisions in household matters will be least likely to believe that wife beating is justified, whereas male or female sole decision-makers will be more likely to believe that wife beating is justified.
The data for this study were taken from the 1999 Zimbabwe Demographic and Health Survey (ZDHS) (16). The Central Statistical Office of the Government of Zimbabwe collected the data with technical assistance from Macro International Inc., as part of the MEASURE (Monitoring and Evaluation to ASsess and Use REsults) Demographic and Health Surveys project. The MEASURE/DHS+ project is designed to "collect, analyse, and disseminate data on fertility, family planning, maternal and child health, and HIV/AIDS" (16). Macro International has been conducting the DHS surveys since 1986, and funding has been provided primarily by USAID (United States Agency of International Development).
The survey in the present paper was based on a two-stage sampling design. A household survey was undertaken in which information was collected on each member of the household, including age, sex, education, and the person's relationship to the head of the household. This information was used to determine the eligible women that is, those aged 1549 in the household, and an attempt was made to interview all of them. Questions about household decision-making, as well as attitudes towards domestic violence, were included in this national survey (16). Women were presented with a set of hypothetical circumstances and given several options for each one, from which they had to choose their answer. The dataset was designed to be nationally representative, but to obtain the correct estimates and standard errors, weights, provided by MEASURE/DHS+ with the ZDHS were used in the data analysis.
Two different groups of women were included in the analyses. First, to obtain a sense of the prevalence of attitudes towards IPV, data on all women were analysed (n = 5907). Second, to gain a better understanding of the relationship between couple dynamics and attitudes towards IPV, the focus was shifted to the sample of currently married or cohabiting women (n = 3553) for which data on household decision- making were available (n = 3077).
Circumstances of wife beating
To gain a better idea of what factors might be associated with attitudes towards wife beating, each hypothetical circumstance was examined separately. Women were asked "Sometimes a husband is annoyed or angered by things which his wife does. In your opinion, is a husband justified in beating his wife in the following situations": (16)
If she goes out without telling him?
If she neglects the children?
If she argues with him?
If she refuses to have sex with him?
If she burns the food?
From these dichotomous variables (yes/no), a single dichotomous variable was created on the basis of whether women think it is justifiable for a husband to beat his wife, under any of the circumstances. For each of the circumstances, logistic regression models were used with attitudes towards wife abuse as the outcome.
Measures of decision-making
Five questions about who makes the decisions in the household were used in this study. The questions asked in response to "Who in your family usually has the final say on the following decisions?" (16) were:
Your own health?
Large household purchases?
Daily household purchases?
Visits to family, friends, or relatives?
What food should be cooked each day?
For each of these questions, the women were given the following response options: (1) themselves (respondent), (2) husband/partner, (3) respondent and husband/partner jointly, (4) someone else, and (5) respondent and someone else jointly. Households in which people other than the husband/partner or wife contributed to the decision-making were relatively rare (67% of the responses, reflecting several different possible decision makers for example, mother-in-law). Therefore, these cases were excluded from the analysis so that the dynamics between husbands/partners and wives would remain the focus.
A set of five dichotomous variables was created for each of the five decision-making domains (or questions) above to reflect patterns of decision-making. For each domain, the variable was coded 1 if the wife had the final say over that decision alone and 0 if the wife did not have the final say. A similar set of dichotomous variables was created for each domain on the basis of whether or not the male partner had the final say in the decision. A final set of dichotomous variables was created for each domain on the basis of whether or not the couple made decisions jointly. Each set of dichotomous variables was combined as a scale (from 0 to 5) that counted the number of domains in which each (husband/partner, wife, and couple) had the final say.
Women's status, relative status, and sociodemographic characteristics
Several measures of women's status were included in the analysis to predict women's attitudes towards wife abuse. These include women's current work status (unemployed; agricultural, manual and skilled labour; and professional, sales, clerical), women's highest level of education obtained (none, primary or secondary or more), and her age. Measures of a woman's relative status were also included in the analyses. These included the difference in the wife's and husband/partner's age (wife older, husband and wife the same age (± 4 years), husband older), the difference in the wife's and husband/partner's school level (wife completed more school levels, husband/partner and wife completed the same school levels, and husband completed more school levels), and the difference in their occupational status.
Additionally, key sociodemographic factors such as residence, household wealth, and ethnicity were examined. Household wealth was calculated as a weighted sum of whether or not the household had the following items: electricity, radio, television, car, refrigerator, bicycle, and an oxcart. The weights were calculated as the inverse of the proportion of households in the sample that had these items. To evaluate the differences between the two main tribes in Zimbabwe, the Shona and the Ndebele, a variable based on the language used for the interview was included in the analysis as a proxy of ethnicity.
Women who believe that wife beating is justified
Fig. 1 shows the relationship between women's marital status, age, and their attitudes towards violence. For this analysis all the women in the survey were included. A high proportion of women believed that it was justifiable for husbands to beat their wives, although fewer older women held this belief compared with younger women. The patterns of results from Fig. 1 are still evident after statistical adjustment. In a logistic regression adjusted for the sociodemographic characteristics (data not shown), women aged 1519 years and 2024 years were over two and a half times more likely to believe that wife beating was justified (odds ratio (OR) = 4.47, P£0.001 and OR = 2.71, P£0.001, respectively) compared with women aged 4549. Among single women, the odds ratio fell to around 2 for these youngest women (OR = 1.77, P£0.001, age 1519; OR = 2.11, P£0.01, age 2024). Married women aged 1519 were over four times as likely to believe that wife beating was justified (OR = 4.15, P£0.001) compared with the oldest age group. Although younger women were significantly more likely to report that wife beating was justified, over half the women in the oldest group (4549 years) reported that wife beating was justified. These results provide strong evidence that acceptance of wife beating is the norm in Zimbabwean society.
Table 1, using cross-tabulations and c2 tests, provides weighted frequency distributions of variables that could be related to attitudes towards IPV. Nearly all of the characteristics analysed are strongly related to believing that IPV is justified. A total of 61% (3022/ 4976) of women living in rural areas believed that wife beating was justified compared with 36% (220/616) of women living in urban areas (weighted percents). Women who spoke a language other than Shona or Ndebele, women from more wealthy households, women with secondary level education or higher, and older women were less likely to accept wife beating. Women who have husbands/partners with secondary or more education, who are older, or have a professional job were least likely to approve of wife beating. Women with the same level of education as their spouses were less likely to believe that beating was justified compared with women who had either a higher or lower level of education than their spouse. The age and occupational differences between spouses were not associated with different responses regarding whether wife beating is justified.
Among the partnership characteristics, married (vs cohabiting) women, women with a partner in the household, and women in shorter marriages were more likely to believe that wife beating was justified. In households where male partners have more than the average decision-making power, women were more likely to believe that wife beating was justified, whereas women in households where decisions are made more jointly were less likely to believe that wife beating was justified. Fig. 2 shows the percentage of women who believe that wife beating is justified in each of the scenarios. Women perceived wife beating to be most justified if a wife argues with her spouse (39%) (2271/5786), if she neglects her children (37%) (2184/5831), or if she goes out without telling him (33%) (1928/5815) (weighted percents).
Finding wife beating justifiable: predictors of attitudes from multivariate models
Table 2 presents the results of six logistic regression models that incorporate attitudes towards circumstances under which wife beating is justified as outcomes. Many factors show similar associations across different circumstances, but there are some notable exceptions. For example, the models consistently show that urban women, women who come from households with higher levels of wealth, and older women are less likely to believe that wife beating is justified.
The educational level and occupational status of women are related to their attitudes towards wife beating. Compared with women with no formal schooling, women with at least some primary schooling are less likely to believe that wife beating is justified if the wife burns the food or argues with her spouse. Women who had secondary or more schooling were less likely to believe that wife beating was justified for burning the food, arguing with a spouse, and refusing to have sexual relations, compared with women with no formal schooling. Occupational status is related to attitudes towards wife beating in five of the six circumstances, with women who have a higher occupational status also believing that wife beating was not justified, compared with unemployed women. Women with higher status jobs (clerical, professional, sales) were significantly less likely to believe that wife beating was justified in any of the domains (OR = 0.52, P<0.001) compared with women who were unemployed. Interestingly, it is not simply employment but the type of job that seems to matter most. Unlike findings from previous studies (e.g. (5)), none of the measures of status difference were significant independent predictors of attitudes towards IPV after multivariate adjustment.
The relationship between attitudes towards wife beating and partnership characteristics was mixed. The attitudes to wife beating were not related to being in a cohabiting or married relationship, household headship, and whether or not the relationship was polygynous (data not shown). Women were more likely to believe that wife beating was justified for burning the food, if they were in a longer marriage (OR = 1.06, P<0.001) and if their partner lived in the household (OR = 1.37, P<0.05). Women in longer marriages were also more likely to believe that wife beating was justified if she refused to have sex (OR = 1.04, P<0.001) or for any of the five domains (OR = 1.02, P<0.05). If the partner lived in the household, women were more likely to say that wife beating was justified if the women neglected their children (OR = 1.29, P<0.01).
Women's reports of household decision-making were consistently associated with attitudes towards wife beating. Across all circumstances, and after adjustment for sociodemographic and partnership characteristics, women were more likely to say that wife beating was justified when husbands/partners had the final say in more household decisions than she did. Women who reported having the final say in more household decisions were significantly more likely to believe that husbands/partners were justified in beating their wives in all but two circumstances if the wife goes out without telling her spouse or if she refuses to have sex. By contrast, when women reported more decisions being made jointly than individually, they were significantly less likely to say that wife beating was justified in all circumstances except if she burned the food. The adjusted odds ratio indicates that for every decision made jointly (0 to 5 decisions) the prevalence of accepting violence decreases by 7%. The full difference between no decisions made jointly and all five domains of decisions made jointly was associated with more than a twofold lower acceptance of IPV after taking into account all the other demographic and household characteristics measured (adjusted OR of 0.69 = 0.935). The addition of the decision-making variables to the models adds significant explanatory power to all the models (except when refusing sex is the outcome) based on log-likelihood ratio tests. This set of results conforms to what was anticipated in hypothesis 2.
In assessing which decision-making domains within the scale were most associated with attitudes towards wife beating (data not shown), several patterns emerged. Women were most likely to approve of wife beating if they had the final say over large household purchases and over what food she would cook. Women were significantly less likely to approve of wife beating if decisions about large household purchases (OR = 0.78, P<0.001), day-to-day purchases (OR = 0.84, P<0.05), and what food she would cook (OR=0.65, P<0.01) were made jointly (data not shown).
More than half of all Zimbabwean women believe that wife beating is justified in at least some circumstances, although the proportion of women who approve of wife beating varies by scenario. Gender norms and expectations in Zimbabwe warrant the concern of public health researchers and practitioners. Even following the International Conference of Population and Development (Cairo, 1994), the Fourth World Conference on Women (Beijing, 1995), and Zimbabwe's ratification of the Convention on the Elimination of All Forms of Discrimination Against Women (May 1991), current levels of approval of wife beating in Zimbabwe are alarming.
Younger women are more likely to believe that wife beating is justified. Although data from only one point in time are presented, comparing the trend in attitudes by age (which shows that younger women are more likely than older women to believe that wife beating is justified) indicates that the next generation of women is just as likely, if not more so, to believe that wife beating is acceptable behaviour. It is unclear whether these more accepting attitudes of wife beating will persist as this younger cohort of women ages future DHS surveys or longitudinal data might be better able to answer this question.
Some of the sociodemographic characteristics associated with believing that wife beating is justified, such as age, residence, and household wealth, are unlikely to change as a result of implementing economic or social development programmes, but current efforts under way to improve women's educational status and occupational opportunities might, in fact, help to alter women's perceptions of wife beating as acceptable. The data presented in this study indicate that women with the highest levels of education and the highest status jobs are the least likely to believe that wife beating is justified. However, the women who have more liberal attitudes, as well as higher levels of education and occupational status, are a select group. It seems somewhat unlikely that education alone, for example, would cause women to believe that wife beating is not acceptable. In fact, evidence from studies of middle-school children in Zimbabwe suggest that although girls can imagine "equality in marriage" they are unable to break with the social norms that dictate men's more powerful position in marriage, particularly in terms of decision-making (17). Further work needs to be done, particularly longitudinal studies, which sort out what types of programme inputs alter beliefs and attitudes, and what programmes help to empower individuals to work against harmful social norms.
Theoretical and empirical work is needed to test the concept that altering the perception that wife beating is a social norm should lead to changes in the practice of wife beating. Changing attitudes towards wife beating at the individual level is probably not sufficient for changing practice, particularly in Zimbabwe. Although women's attitudes (i.e. believing that wife beating is justified) are less likely to be directly linked to their own experience of wife beating compared with men's attitudes, women's attitudes, as they reflect their own expectations and experience, are clearly important. Women's attitudes serve as a marker for social acceptability of wife beating. Perhaps women's expectations will change if they are presented with alternatives to living with an abusive spouse. Most women in Zimbabwe marry; there are few alternatives to marriage, and few women know of potential options to gain both financial and emotional independence from an abusive spouse. However, recent decades have seen a limited improvement in national level regulations for women to end their marriages and maintain control over their lives (18). At the same time, evidence indicates that, over the past few years, Zimbabwe has become a more violent society due to issues surrounding the recent presidential election (19, 20). In addition, it has been recently plagued with drought and some of the highest levels of human immunodeficiency virus (HIV) in the world one in three people are believed to be HIV-positive (21) with the result that many children have become orphans and many women have become household heads (22). In this type of setting, it has been suggested that violence is more likely to be accepted as a social norm (5). Women might not be empowered enough to refuse sex or negotiate for condom use, or they might be unwilling to be tested for HIV for fear of abuse (23).
The empowerment of women, either through decision-making or economic contributions, has been a key focal point for both global attempts at legislation and for several recent studies concerning health outcomes, including IPV. However, it seems that imbalances of power (either husbands/partners or wives having more or relatively more power) are associated with adverse consequences, as shown by this study and others (14, 24). Interventions that promote joint decision-making might be a promising strategy for increasing women's views towards equality in marriage while promoting men's views that household disputes should be settled with negotiation, not violence.
Future research could focus on prospective studies and interventions that test the causal nature of the associations observed in this paper. In Zimbabwe, high economic stability, violence related to land redistribution, and political tensions complicate matters and might limit the ability of the government to focus on other issues such as IPV. It is unclear whether the current climate in Zimbabwe makes further work on joint decision-making and non-violent approaches to conflict resolution both within the household and outside impossible or more needed than ever. Both individual and wider societal-level acceptance of violence needs to be addressed. In a world that is increasingly interconnected, local or national acceptance of violence might contribute to global acceptance of violence.
This work is funded, in part, by a grant to the author from the National Institute of Child Health and Human Development (NICHD), K01-HD042540.
Conflicts of interest: none declared.
1. Injury: a leading cause of the global burden of disease. Krug E, editor. Geneva: World Health Organization; 2002. WHO document WHO/HSC/PVI/99.11. 2002.
2. Venis S, Horton R. Violence against women: a global burden. Lancet 2002;359:1172.
3. World Health Organization. WHO multi-country study. Available from: URL: http://www5.who.int/violence_injury_prevention/main.cfm?p=0000000332 (accessed on 11 March 2003).
4. World report on violence and health. Krug EG, Dahlberg LL, Mercy JA, Zvi AB, Lozano R, editors. Geneva: World Health Organization; 2002.
5. Jewkes R. Intimate partner violence: causes and prevention. Lancet 2002;359:1423-9.
6. Heise L, Ellsberg M, Gottmoeller M. Ending violence against women. Population Reports Series L. Baltimore (MD): Johns Hopkins University School of Public Health, Population Information Program; 1999. Available from: URL: http://www.jhuccp.org/pr/l11edsum.shtml
7. Kim J, Motsei M. "Women enjoy punishment": attitudes and experiences of gender-based violence among PHC nurses in rural South Africa. Social Science & Medicine 2002;54:1243-54.
8. Koenig M, Lutalo T, Wabwire-Mangen F, Kiwanuka N, Wagman J, Zhao F, et al. Domestic violence in Rakai, Uganda: evidence from a community-based study. Bulletin of the World Health Organization 2003;81:53-60.
9. Jejeebhoy S. Wife-beating in rural India: a husband's right? Economic and Political Weekly 1998;33:855-62.
10. Garciá-Moreno C. Violence against women: consolidating a public health agenda. In: Sen G, George A, Ostlin P, editors. Engendering international health: the challenge of equity. Cambridge (MA): MIT Press; 2002. p. 111- 41.
11. Hindin M, Adair L. Who's at risk? Factors associated with intimate partner violence in the Philippines. Social Science & Medicine 2002;55:1385.
12. Women of the world: anglophone Africa progress report. New York: The Center for Reproductive Law and Policy; 2001: p. 131-50.
13. Women's reproductive rights in Zimbabwe: a shadow report. Prepared for the 18th session of the Committee on the Elimination of All Forms of Discrimination Against Women. Women in Law and Development in Africa, editors. New York: The Center for Reproductive Law and Policy; 1997.
14. Watts C, Keogh E, Ndlovu M, Kwaramba R. Withholding of sex and forced sex: dimensions of violence against Zimbabwean women. Reproductive Health Matters 1998;6:57-65.
15. Grant M. Moving and coping: women tenants in Gewru, Zimbabwe. In: Sheldon K, editor. Courtyards, markets, city streets: urban women in Africa. Boulder (CO): Westview Press; 1996:169-89.
16. Central Statistical Office [Zimbabwe], Macro International Inc. Zimbabwe demographic and health survey. Calverton (MD): Central Statistical Office and Macro International Inc.; 1999.
17. Ansell N. "Of course we must be equal, but ...": imagining gendered futures in two rural southern African secondary schools. Geoforum 2002;33:179-94.
18. Hindin M. For better or for worse? Women's autonomy and marital status in Zimbabwe. Population Research and Policy Review 2000;19:255-82.
19. Health-care ethics in Zimbabwe. Lancet 2002;359:791-2.
20. Focusing on health in the crisis in Zimbabwe. Lancet 2002;359:455.
21. UNAIDS/WHO Working Group on Global HIV/AIDS and STI Surveillance. Zimbabwe, 2002 update. Epidemiological fact sheets on HIV/AIDS and sexually transmitted infections. Available from: URL: http://www.who.int/emc-hiv/ fact_sheets/pdfs/Zimbabwe_EN.pdf
22. Bicego G, Rutstein S, Johnson K. Dimensions of the emerging orphan crisis in sub-Saharan Africa. Social Science & Medicine 2003;56:1235-47.
23. Laver SM. Voluntary testing and counselling for HIV. "Are adults in rural communities ready to test?" A descriptive survey. The Central African Journal of Medicine 2001;47:92-7.
24. Hindin MJ. Women's power and anthropometric status in Zimbabwe. Social Science & Medicine 2000;51:1517-28.