Domestic violence in rural Uganda: evidence from a community–based study

Michael A. Koenig,¹ Tom Lutalo,² Feng Zhao¹, Fred Nalugoda,² Fred Wabwire–Mangen,³ Noah Kiwanuka,² Jennifer Wagman,⁴ David Serwadda,³ Maria Wawer,⁴ & Ron Gray¹

ABSTRACT: Although domestic violence is an increasing public health concern in developing countries, evidence from representative, community–based studies is limited. In a survey of 5109 women of reproductive age in the Rakai District of Uganda, 30% of women had experienced physical threats or physical abuse from their current partner — 20% during the year before the survey. Three of five women who reported recent physical threats or abuse reported three or more specific acts of violence during the preceding year, and just under a half reported injuries as a result. Analysis of risk factors highlights the pivotal roles of the male partner's alcohol consumption and his perceived human immunodeficiency virus (HIV) risk in increasing the risk of male against female domestic violence. Most respondents — 70% of men and 90% of women — viewed beating of the wife or female partner as justifiable in some circumstances, posing a central challenge to preventing violence in such settings.

Keywords Spouse abuse; Women; Risk factors; Alcohol drinking/adverse effects; HIV infections; Sex behavior; Knowledge, attitudes, practice; Regression analysis; Uganda (source: MeSH, NLM).

Mots clés Epouse maltraitée; Femmes; Facteur risque; Consommation alcool/effets indésirables; HIV, Infection; Comportement sexuel; Connaissance, attitude, pratique; Analyse régression; Ouganda (source: MeSH, INSERM).

Palabras clave Maltrato conyugal; Mujeres; Factores de riesgo; Consumo de bebidas alcohólicas/efectos adversos; Infecciones por VIH; Conducta sexual; Conocimientos, actitudes y práctica; Análisis de regresión; Uganda (fuente: DeCS, BIREME).

Introduction

Over the past decade, recognition of the scope and significance of domestic violence globally has increased. Domestic violence has been defined as "the range of sexually, psychologically and physically coercive acts used against adult and adolescent women by current or former male intimate partners" (1). A growing body of evidence is highlighting the magnitude of the problem of domestic violence in developing countries (2–11). In sub–Saharan Africa, empirical evidence on the prevalence of domestic violence is limited and confined to a small number of population–based (12–15) or special–population studies (16). Recognition of the links between domestic violence and a range of adverse reproductive health outcomes — including non–use of contraception and unintended pregnancy (17, 18), poor outcomes of pregnancy and birth (19–24), gynaecological morbidity (25) and sexually transmitted diseases and human immunodeficiency virus (HIV) (18, 26, 27) — is also growing.

Our understanding of the underlying determinants of domestic violence in developing countries remains limited. A number of studies have found strong associations between socioeconomic status and domestic violence, with indicators of household wealth or education of the male partner significantly inversely associated with the risk of violence (6, 7, 11, 28, 29). Demographic characteristics are also significant risk factors for domestic violence, with several studies finding that higher age (6, 9) and higher numbers of children (9, 30) are associated with a reduced risk of violence. Other studies have found that women with a high status — as measured by their educational attainment, degree of autonomy or control over resources — are more protected from the risk of domestic violence. One consistent finding is an inverse association between women's educational attainment and the risk of domestic violence (3, 5, 9, 31). Studies have also reported that women with greater autonomy and control over resources are more protected from violence (7, 32, 33). However, some evidence shows that this association may be context–specific and that, in more conservative settings, women with high autonomy may actually be at increased risk of violence (33, 34).

Several studies in developing countries have also found a strong association between consumption of alcohol or drugs and the risk of violence (14, 15, 30, 31). A potential link between HIV status and domestic violence has also been recognized (26), with studies from Africa showing an increased risk of violence when the man is HIV positive (14) or when the woman perceives herself to be at high risk of acquiring HIV from the man (16). Finally, evidence highlights the role of intergenerational transmission of domestic violence; studies have shown that children who witness family violence are more likely to become perpetrators or victims of violence in adulthood (8, 31, 35). Thus, although some evidence does exist, the issue of domestic violence and its underlying determinants in developing countries remain inadequately understood.

Data from the Rakai Project in rural Uganda provide a unique opportunity to explore the issue of domestic violence from a community–based perspective. In the 2000–01 round of data collection, a special module of questions was fielded to assess the prevalence of domestic violence and its potential contribution to transmission of HIV in this population.

The primary study population reported in this paper consisted of 5109 sexually active women of reproductive age who lived in the 46 communities under surveillance in the Rakai Project at the time of the 2000–01 survey. This study investigates the prevalence of lifetime and recent domestic violence in this population, as well as the frequency of specific violent behaviours. We examined the specific sociodemographic and behavioural risk factors associated with recent domestic violence and the attitudes of male and female respondents toward the circumstances under which such violence is justifiable.

Setting and data

The Rakai Project was initiated in 1987 as a collaborative intervention research initiative to understand and reduce transmission of HIV/AIDS in rural Uganda. Rakai district is a rural area in south–western Uganda that borders the United Republic of Tanzania, and Lake Victoria. Education levels are low, especially for women: only 23% of women in the study population had completed >8 years of education. The economy of the study area is dominated by agriculture, with a significant proportion of both men and women engaged in agricultural production or commerce. The predominant ethnic group in Rakai is Baganda; in this group, marriage patterns are patrilineal, with women marrying into and residing with their husband's clan. Polygamous unions are common: 21% of women respondents in our study were part of such unions. Cultural norms tolerate and even condone multiple sexual partners among men (36): 42% of men reported multiple sexual partners during the previous year compared with only 4% of women. These practices, combined with limited information about HIV/AIDS and resistance to adopting safe sex behaviours, have placed the Rakai region at the centre of the HIV epidemic in Uganda; the estimated prevalence of HIV in Rakai was 16% during 1994–98 (37).

In 1994, 56 communities located on secondary roads in the Rakai District were selected and aggregated into 10 clusters. The clusters were randomly assigned to an intervention arm that received mass treatment for sexually transmitted diseases or a control arm (37). All consenting adults (men and women) between the ages of 15 and 59 years were eligible for enrolment in the study. Respondents were interviewed at their homes at regular 10–month intervals; at each visit, they were given a detailed survey that collected data on sociodemographic characteristics, health status, and sexual behaviours and partnerships. Respondents were also asked to provide blood or urine samples, or both, for detection of HIV and selected sexually transmitted diseases. All participants were given health education on HIV, sexually transmitted diseases and family planning; condoms were provided free of cost and HIV test results and counselling upon request (38).

In 1999, the Rakai surveillance area was modified to include 46 communities in the community HIV epidemiological research (CHER) survey. In the second such survey (CHER–2), which was conducted between March 2000 and February 2001, respondents were asked a series of detailed questions (adapted from the original Conflicts Tactics Scale) about their experience with domestic violence (39). Male and female respondents were asked about the occurrence of both lifetime and recent (occurring in the 12 months before the survey) male against female violence involving their current partner (Box 1). Respondents who reported violence during the preceding 12 months were asked about the specific acts of violence that took place, the precipitating factors for such violence and the nature of any resulting injuries. Respondents were also asked about the occurrence of specific female against male violent behaviour during the preceding 12 months. Finally, both male and female respondents were asked under which circumstances a man would be justified in beating his wife or female partner.

Following procedures carefully established over the previous decade for the collection of sensitive information in the Rakai Project, a number of safeguards were put in place to protect the confidentiality of information provided by respondents and to minimize potential risks to respondents associated with participating in the study. Consent to participate was obtained from all respondents upon study enrolment and at each subsequent follow–up contact (40). Interviews were conducted by highly trained interviewers of the same sex and in complete privacy; all information from the survey remained undisclosed to other family members. Completed questionnaires were maintained in secure facilities, and interview schedules were coded with the participants' study identification numbers. Where personal identifiers were used, they were separated from interview schedules containing potentially sensitive information and kept separately and securely. At time of the present survey, limited referral facilities accessible to the Rakai population had been identified. However, these referral services were available in subsequent survey rounds.

For this paper, our primary study population consisted of the 5109 women who participated in the CHER–2 survey and who were married or in a sexual partnership and aged 15–49 years at the time of CHER–2. This population resulted from a response rate of over 90% of eligible respondents for CHER–2. Although the primary focus of this study was the women's responses, we also looked at information on attitudes toward domestic violence from the corresponding sample of 3881 men aged 15–59 years who were interviewed in CHER–2 and were married or in active partnerships at the time of the study.

Results

Male against female domestic violence

Overall, 40.1% of women had ever experienced verbal abuse — verbal threats, shouting or yelling — from their current male partner and 30.4% of women had ever experienced physical threats or violence (Table 1). Physical threats and physical violence were both common during the previous 12 months.

The most common form of lifetime violence was the male partner pushing, slapping or holding down the female respondent (23.1%) (Table 2). Smaller but significant proportions of respondents reported physical threats or violence during the previous 12 months than lifetime physical threats or violence.

Fig. 1 shows the frequency of specific acts of violence among women who reported experiencing recent physical threats or violence. For almost all specific acts of violence, most women reported multiple incidences of violence during the preceding 12 months. The percentage of women reporting three or more incidents of specific physical threats or violence during the preceding 12 months ranged from 53.2% for threatening gestures to 20% for women who reported having been burned or scalded. A significant proportion of women who experienced violence reported six or more occurrences during the preceding 12 months — this ranged from 13.4% for being pushed, slapped or held down to 30.4% for threatening gestures. Although the overall number of episodes of violence cannot be estimated from our data,^a it is notable that 60% of women who reported violence during the preceding year reported three or more specific violent acts during this period.

The most commonly cited reason for physical assault by the male partner during the past year was the wife's neglect of household chores; this was cited by 28.8% of women reporting recent physical threats or violence (Table 3). Other commonly cited reasons were the woman disobeying the husband or family elders (24%), the woman's refusal to have sex (17%), arguments over money (14%) and suspected infidelity by the woman (13%). The woman's insistence on the use of condoms, use of contraception without permission or suspected HIV–positive status were less frequently cited reasons for violence (each

Overall, 44% of women who reported domestic violence (either threats or physical violence) during the previous 12 months suffered related injuries (Table 4). The most common injury was pain lasting more than one day (39.9%); this was followed by sprains, bruises or cuts (18.4%). Other injuries (such as broken bones and loss of consciousness) were much less common, but they did occur. In total, 21.5% of women required medical attention for their injuries at least once during the 12 months before the survey.

Female against male domestic violence

Respondents were also questioned about the occurrence of female against male domestic violence during the past 12 months. Overall, 19.8% of women reported verbally abusing, physically threatening or physically abusing their current male partner during the previous year (Table 5). Such behaviour primarily involved verbal threats, yelling or shouting (18.7%), but 5.4% of women reported actions constituting physical threats or violence, and 3% reported actual physical violence. It is noteworthy that roughly four out of five women who reported recent female against male domestic violence also reported recent male against female violence.

Risk factors for male against female domestic violence

The main aim of our study was to gain a better understanding of the potential risk factors associated with domestic violence in this population. To investigate this issue, we used a multivariate logistic regression analysis; our dichotomous dependent variable was defined as 0 = no episodes of physical threats or violence from the current male partner during the preceding 12 months and 1 = one or more episodes of physical threats or violence during this period. Coefficients are expressed as odds ratios relative to the omitted reference category.

Most sociodemographic variables had a limited explanatory contribution towards predicting the risk of recent domestic violence (Table 6). Age of woman, pregnancy status, use of modern contraception, religion and occupation of the male partner all failed to attain statistical significance as predictors of violence. An exception was women's education: women with secondary schooling experience (>8 years) had significantly lower risks of violence (odds ratio 0.66) than those in the reference category (no education). Women with a large number of living children (>6) also had significantly lower risks of violence (odds ratio 0.64) than the reference group.

The second major constellation of risk factors pertains to aspects of the current sexual partnership. Women in consensual unions — that is, not legally or formally married — faced significantly higher risks of violence (odds ratio 1.32) than those in the reference group (married women) and women whose partners were classified as boyfriends or other relationships (e.g., casual friends, fellow students or employees, visitors, strangers, relatives) faced significantly lower risks of violence (odds ratio 0.41). The length of the current partnership was inversely related with violence: relationships of shorter (

The final constellation of variables considered was selected risk behaviours by women or their male partners (as reported by women). The strength of the association between alcohol consumption and domestic violence was particularly noteworthy. Women whose partner frequently or always consumed alcohol before sex faced risks of domestic violence almost five times higher than those whose partners never drank before sex (odds ratio 4.62). The risk of violence to women whose partners "sometimes" consumed alcohol before sex was also significantly higher than those with partners in the non–drinking reference group (odds ratio 1.62). Women's own consumption of alcohol before sex was also modestly, but significantly, related to the risk of violence (odds ratio 1.22).

The association between the woman's perception of her male partner's HIV risk and domestic violence was also considered. Women who perceived their partner as "somewhat likely" to have been exposed to HIV were significantly more likely to report domestic violence than those in the reference category ("not at all or unlikely") (odds ratio 1.84). Particularly striking was the subgroup of women who believed that it was "very likely" that their partner was at risk of acquiring HIV. These women experienced risks of violence almost four times higher than those in the perceived low–risk reference group (odds ratio 3.72). The risk of domestic violence for women with no knowledge of their partner's risk of HIV was not significantly different for those in the reference group.

Women's age at first intercourse was also significantly associated with the risk of violence: women who became sexually active very early (

Attitudes toward domestic violence

Two main findings emerged from the analysis of data on men's and women's views about circumstances in which beating of the female partner would be justified (Fig. 2). Firstly, the proportion of respondents who believed that such a beating was justified varied considerably and depended on the specific behaviour in question.

Secondly, for almost all behaviours cited, a higher percentage of women than men believed beating to be justifiable: 16% of men and 28% of women believed beatings to be justified when a woman refused to have sex with her partner, and 22% of men and 27% of women when a woman adopted contraception without the permission of her partner. Attitudes toward infidelity by the female partner were at the other extreme: 60% of men and a striking 87% of women believed that beating was justified if the woman was unfaithful. For all situations considered, 70% of men and 90% of women believed that beating was justifiable for one or more of the situations cited (54% of men and 65% of women when infidelity by the female partner was excluded). A somewhat– surprising finding was that attitudes condoning domestic violence were more common among younger men and women. For example, 66% of men aged >35 years viewed beating of the female partner as justifiable for one or more reasons (other than infidelity); comparable figures for younger and older women were 73% and 55%, respectively (data not shown).

Discussion

Our study had two potential limitations. Firstly, the prevalence of domestic violence may have been underestimated. This is partly because of the omission of questions on sexual violence in the survey, although other studies suggest significant overlap between the occurrence of physical and sexual violence (41). A more fundamental concern is the possible under–reporting of violence by respondents, given the culturally sensitive nature of this behaviour and the possible reluctance of many respondents to acknowledge its occurrence. However, several features of the Rakai Project are likely to have increased the validity of such reporting, including the close and long–standing interaction and rapport between interviewers and respondents, the demonstrated expertise of interviewers in eliciting sensitive information and the extensive efforts undertaken to ensure privacy and protect the confidentiality of responses.

A second limitation of our study concerns gaps in the set of variables considered as potential risk factors for domestic violence. Important explanatory variables such as intergenerational exposure to violence, poverty, and gender roles and attitudes were not collected in the Rakai survey. Attitudes toward wife beating, although collected, were not included as potential risk factors in our analysis because of uncertainty over the direction of causality between domestic violence attitudes and behaviour.^b

These limitations notwithstanding, our results provide some of the most comprehensive information to date on the prevalence and nature of domestic violence among a large, representative sample of women in sub–Saharan Africa. Our results underscore the magnitude of the problem of domestic violence in this rural Ugandan setting and show that episodes of violence are neither infrequent nor isolated events.

Comparisons with other studies

Although differing somewhat in terms of timeframes and definitions, it is interesting to compare our results with those from other population–based studies of domestic violence in Africa. Two studies reported somewhat higher levels of violence — 43% of women in Zimbabwe reported experiencing physical violence in adulthood (15) and 40% of women in another Ugandan study experienced physical harm from their husband or partner (13). In contrast, somewhat lower rates of domestic violence were reported in two other studies — 25% of South African women reported physical violence from a current or past husband or partner (10% during the past year) (12) and 20% of Rwandan women reported having ever been beaten by their current partner (14).

Risk factors associated with domestic violence

In terms of associated risk factors, domestic violence was largely unrelated to most socioeconomic and demographic variables. An important exception was women's education: women with secondary school education or higher faced significantly lower risks of violence. Some of the most systematic associations to emerge from our study were between indicators of risk behaviour and domestic violence.

Onset of sexual relations

We found a significant link between the onset of sexual relations among women and the risk of recent domestic violence: women who become sexually active younger than 15 years faced almost twice the risks of recent violence as those who became sexually active at or after 18 years. The current data cannot answer whether women who become sexually active early are self–selected for subsequent abusive relationships or whether, as a result of early sexual activity, they may be less empowered to protect themselves against subsequent violence. Further research is clearly warranted on the specific mechanisms that place women who become sexually active early at higher risk of violence.

Alcohol consumption

The important role of alcohol consumption in domestic violence also emerged in our study. Women whose partners often drink before sex experience risks of violence almost five times higher than women with non–drinking partners. It is possible that alcohol consumption and domestic violence are two largely independent risk behaviours that characterize certain relationships. That 52% and 27% of women who reported recent domestic violence reported that their partner had consumed alcohol or had frequently consumed alcohol, respectively, supports the conclusion that alcohol may play a direct precipitating role in such violence. The link between alcohol consumption and domestic violence — and their joint and independent roles in HIV transmission — are a focus of current research within the Rakai Project.

Perception of HIV risk

The strong association between women's perceptions of their male partner's HIV risk and the women's risk of domestic violence is also of interest. Women who believe it very likely that their partner is at high risk of HIV face risks of violence almost three times higher than those who perceive their partners to be at very low risk. A plausible explanation is that women who perceive their male partner to be at significant risk of HIV infection may be reluctant to engage in sexual relations with this partner; this resistance may be met, in turn, with physical violence or coercion into sex by the male partner. Results from a previous study in Rakai, which showed a strong association between perceptions of the partner's perceived HIV risks and sexual coercion, provide indirect support for this hypothesis (42). Further support comes from the finding that refusal of sex by the wife was one of the most commonly cited reasons for physical violence (Fig. 2). In addition, women who perceived their partner as at high risk of HIV were more than four times as likely to report refusal of sex as a primary reason for domestic violence than those who believed their partner to be at low risk.^c The inter–relationships between perceived HIV status, sexual coercion and physical violence appear to be closely linked, and they require further elucidation through additional quantitative and qualitative studies.

Conclusion

The results of our study have potentially important implications for programmes aimed at preventing violence and HIV. Our findings show that interventions aimed at reducing alcohol consumption are likely to have important corollary benefits in terms of reducing levels of violence between intimate partners. They also provide indirect evidence that domestic violence may represent a significant factor in women's vulnerability to HIV acquisition in settings such as Uganda; this raises the possibility that current programmes to prevent HIV may be overlooking a key behavioural dimension of HIV transmission (12). Our results also suggest that little progress in reducing levels of domestic violence is likely to be achieved without significant changes in prevailing individual and community attitudes toward domestic violence. At the same time, our findings underscore the challenges associated with changing attitudes toward violence, given that highest levels of support for such violence were found among women and younger adults. At a more immediate level, findings from the present survey have led the Rakai Project to expand its agenda on HIV research to investigate the potential contributing roles of physical and sexual violence and to plan a programme to prevent violence within project communities.

Acknowledgements

Data collection for this study was supported by grants from the National Institute of Allergy and Infectious Diseases (R01 AI34826b and R01 AI34826S), the National Institute of Child Health and Human Development (5P30HD06826), the Rockefeller Foundation, and the World Bank Uganda Sexually Transmitted Infections Project. The authors gratefully acknowledge the helpful comments of Rachel Jewkes, Suzanne Maman and two anonymous reviewers and the research assistance of Jessica Gipson and Michelle Mancini.

Conflicts of interest: none declared.

Résumé

Violence domestique dans les régions rurales de l'Ouganda : données d'une étude menée dans la communauté

Bien que la violence domestique soit un problème de santé publique en augmentation dans les pays en développement, on ne dispose que de données limitées en provenance d'études représentatives menées dans la communauté. Lors d'une enquête sur 5109 femmes en âge de procréer réalisée dans le district de Rakai en Ouganda, il est apparu que 30 % des femmes avaient subi des menaces ou des violences physiques de la part de leur partenaire actuel, contre 20 % l'année précédant l'étude. Parmi les femmes ayant déclaré avoir subi des menaces ou des violences, trois sur cinq ont rapporté au moins trois actes de violence spécifiques au cours de l'année précédente et près de la moitié ont déclaré avoir souffert de traumatismes à la suite de ces actes. L'analyse des facteurs de risque souligne le rôle crucial de la consommation d'alcool chez le partenaire et sa perception du risque d'infection à VIH dans l'augmentation du risque de violence domestique à l'encontre de la femme. La plupart des personnes ayant répondu à l'enquête (70 % des hommes et 90 % des femmes) considéraient le fait de battre sa femme ou sa partenaire comme justifiable dans certaines circonstances, ce qui pose un problème majeur du point de vue de la prévention de la violence dans ce type de contexte.

Resumen

La violencia doméstica en la Uganda rural: evidencia aportada por un estudio comunitario

La violencia doméstica constituye un problema de salud pública cada vez más preocupante en los países en desarrollo, pero la evidencia aportada por estudios comunitarios representativos es limitada. Un sondeo realizado entre 5109 mujeres en edad fecunda en el distrito de Rakai en Uganda reveló que el 30% de las encuestadas habían sufrido amenazas físicas o malos tratos por parte de la pareja que tenían a la sazón, el 20% durante el año anterior a la encuesta. Tres de cada cinco mujeres que informaron de amenazas o malos tratos físicos recientes declararon haber sido víctimas de tres o más actos concretos de violencia durante el año previo, y poco menos de la mitad señalaron haber sufrido lesiones como resultado. El análisis de los factores de riesgo destaca la enorme importancia del consumo de alcohol del compañero y de su riesgo percibido en relación con el VIH como causas de aumento del riesgo de violencia doméstica contra la mujer. La mayoría de los encuestados — el 70% de los hombres y el 90% de las mujeres — consideraban justificable que en algunas circunstancias se golpease a la esposa o compañera, lo cual plantea un problema considerable para conseguir prevenir la violencia en esos entornos.

References

1. World Health Organization. WHO/WHD Violence against women: a priority health issue. Geneva: World Health Organization, 1997. WHO document WHO/FRH/WHD/97.8#.        

2. Heise LL, Pitanguay J, Germain A. Violence against women: the hidden health burden. World Bank discussion Paper No. 255. Washington (DC): World Bank, 1994.         

3. El–Zanaty F, Hussein EM, Shawky GA, Way AA, Kishor S. Egypt Demographic and Health Survey 1995. Calverton, Maryland: National Population Council and Macro International Inc., 1996.        

4. Haj–Yahia MM, Edleson JL. Predicting the use of conflict resolution tactics among engaged Arab–Palestinian men in Israel. Journal of Family Violence 1994;9:47–62.        

5. Profamilia. Demographic and Health Survey for Columbia. Calverton, MD: Macro International Inc; 1995.         

6. Kim K, Cho Y. Epidemiological survey of spousal abuse in Korea, Intimate Violence: interdisciplinary Perspectives. Washington: Hemisphere Publishing Corporation; 1992.        

7. Jejeebhoy SJ, Cook RJ. State accountability for wife–beating: the Indian challenge. Lancet 1997;349 Suppl I: sl10–s2.        

8. Ellsberg MC, Rena R, Herrera A, Liljestrand J, Winkvist A. Wife abuse among women of childbearing age in Nicaragua. American Journal of Public Health 1999;89:241–4.         

9. Schuler SR, Hashemi SM, Riley AP, Akhter S. Credit programs, patriarchy and men's violence against women in rural Bangladesh. Social Science and Medicine 1996;43:1729–42.        

10. Fikree FF, Bhatti LI. Domestic violence and health of Pakistani women. International Journal of Gynecology and Obstetrics 1999;65:195– 201.        

11. Martin SL, Tsui AO, Maitra K, Marinshaw R. Domestic violence in Northern India. American Journal of Epidemiology 1999;150:417–26.        

12. Jewkes R, Penn–Kekana L, Levin J, Ratsaka M, Schrieber M. Prevalence of emotional, physical, and sexual abuse of women in three South African provinces. South African Medical Journal 2001;91:421–8.        

13. Blanc AB, Wolff B, Gage AJ, Ezeh AC, Neema S, Ssekamatte–Ssebuliba J. Negotiating reproductive outcomes in Uganda. Calverton, Maryland: Macro International Inc. and Institute of Statistics and Applied Economics [Uganda], 1996.        

14. van der Straten A, King R, Grinstead O, Vittinghoff E, Serufilira A, Allen S. Sexual coercion, physical violence, and HIV infection among women in steady relationships in Kigali, Rwanda. AIDS and Behavior 1998;2:61–73.        

15. Watts C, Keough E, Ndlovu M, Kwaramba R. Withholding of sex and forced sex: dimensions of violence against Zimbabwean women. Reproductive Health Matters 1998;6:57–65.        

16. Coker AL, Richter DL. Violence against women in Sierra Leone: frequency and correlates of intimate partner violence and forced sexual intercourse. African Journal of Reproductive Health 1998;2:61–72.        

17. Gazmararian JA, Lazorick S, Spitz AM, Ballard TJ, Saltzman LE, Marks JS. Prevalence of violence against pregnant women. Journal of American Medical Association 1996;275:1915–20.        

18. Martin SL, Kilgallen B, Tsui AO, Maitra K, Singh KK, Kupper SL. Sexual behaviors and reproductive health outcomes: associations with wife abuse in India. Journal of the American Medical Association 1999;282:1967–72.        

19. Valdez–Santiago R, Sanin–Aguirre LH. Domestic violence during pregnancy and its relationship with birth weight. Salud Publica Mexicana 1996;38:352– 62.        

20. Curry MA, Perrin N, Wall E. Effects of abuse on maternal complications and birth weight in adult and adolescent women. Obstetrics and Gynecology 1998;92:530–4.        

21. Parker B, McFarlane J, Soeken K. Abuse during pregnancy: effects on maternal complications and birth weight in adult and teenage women. Obstetrics and Gynecology 1994;84:323–8.        

22. Petersen R, Gazmararian JA, Spitz AM, Rowley DL, Goodwin MM, Saltzman LE, et al. Violence and adverse pregnancy outcomes: a review of the literature and directions for future research. American Journal of Preventive Medicine 1997;13:366–73.        

23. Murphy CC, Schei B, Myhr TL, DuMont J. Abuse: a risk factor for low birth weight? A systematic review and meta–analysis. Canadian Medical Association Journal 2001;164:1567–72.        

24. Jejeebhoy SJ. Associations between wife–beating and fetal and infant death: impressions from a survey in rural India. Studies in Family Planning 1998;29:300–8.        

25. Schei B, Bakketeig LS. Gynaecological impact of sexual and physical abuse by spouse. A study of a random sample of Norwegian women. British Journal of Obstetrics and Gynaecology 1989;96:1379–83.        

26. Maman S, Campbell J, Sweat MD, Gielen AC. The intersections of HIV and violence: directions for future research and interventions. Social Science and Medicine 2000;50:459–78.        

27. Garcia–Moreno C, Watts C. Violence against women: its importance for HIV/ AIDS. AIDS 2000;14 Suppl 3:S253–65.        

28. Hoffman K, Demo DH, Edwards JN. Physical wife abuse in a non–Western society: an integrated theoretical approach. Journal of Marriage and the Family 1994;56:131–46.        

29. Mahajan A. Instigators of wife battering. In: Sood S, editor. Violence against Women. Jaipur: Arihant Publishers; 1990.        

30. Rao V. Wife–beating in rural South India: a qualitative and econometric analysis. Social Science and Medicine 1997;44:1169–80.        

31. Jewkes R, Levin J, Penn–Kekana L. Risk factors for domestic violence: findings from a South African cross–sectional study. Social Science and Medicine 202;55:603–170.        

32. Jejeebhoy SJ. Wife–beating in rural India: a husband's right? Evidence from survey data. Economic and Political Weekly 1998;33:855–62.        

33. Levinson D. Family violence in cross–cultural perspectives. Newbury Park: Sage Publications; 1989.        

34. Koenig MA, Ahmed S, Hossain MB, Mozumder ABMKA. Women's status and domestic violence in rural Bangladesh: individual and community–level effects Demography (forthcoming).        

35. Martin SL, Moracco KE, Garro J, Tsui AO, Kupper LL, Chase JL, et al. Domestic violence across generations: findings from northern India. International Journal of Epidemiology 2002;31:560–72.        

36. Olowo–Freers BPA, Barton TG. In pursuit of fulfillment: studies of cultural diversity and sexual behaviour in Uganda. Kisubi, Uganda: Marianum Press; 1992.        

37. Wawer MJ, Sewankambo NK, Serwadda D, Quinn TC, Paxton LA, Kiwanuka N, et al. Control of sexually transmitted diseases for AIDS prevention in Uganda: a randomised community trial. Lancet 1999;353:525–35.        

38. Lutalo T, Kidugavu M, Wawer MJ, Serwadda D, Zabin LS, Gray RH. Trends and determinants of contraceptive use in Rakai District, Uganda, 1995–98. Studies in Family Planning 2000;31:217–27.        

39. Straus MA. Measuring intrafamily conflict and violence. The conflict tactics (CT) scales. Journal of Marriage and the Family 1979;41:75–88.        

40. Wawer MJ, Gray RH, Sewankambo NK, Serwadda D, Paxton L, Berkley S, et al. A randomized, community trial of intensive sexually transmitted disease control for AIDS prevention, Rakai, Uganda. AIDS 1998;12:1211–25.        

41. Ellsberg M, Pena R, Herrera A, Liljestrand J, Winkvist A. Candies in hell: women's experiences of violence in Nicaragua. Social Science and Medicine 2000;51:1595–610.        

42. Koenig MA, Lutalo T, Zhao F, Nalugoda F, Kiwanuka N, Wabwire–Mangen F, et al. Coercive sex in rural Uganda: prevalence and associated risk factors. Social Science and Medicine (forthcoming).        

¹Department of Population and Family Health Sciences, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD, USA. Correspondence should be addressed to Dr. M.A. Koenig (email: mkoenig@jhsph.edu).

²Rakai Project, Uganda Virus Research Institute, Entebbe, Uganda.

³Institute of Public Health, Makerere University, Kampala, Uganda.

⁴Hielbrunn Center for Population and Family Health, Columbia University Joseph L. Mailman School of Public Health, New York, NY, USA.

Ref. No. 01–1579

^aThe wording of this question left open the possibility that multiple violent acts could have taken place within the same episode of violence.

^bDo attitudes condoning violence lead to an increased likelihood of domestic violence or does the occurrence of domestic violence lead to the increased legitimization of such violence among both men and women?

^cAmong women who perceived their partner as "very likely" to be at risk of HIV, 7.6% cited refusal of sex as a primary reason for violence during the previous year compared with 1.8% of women who perceived their partner to be "not at risk" of HIV.