J Epidemiol Community Health 66:e2 doi:10.1136/jech.2010.112300
  • Research report

Men's exposure to human rights violations and relations with perpetration of intimate partner violence in South Africa

  1. David R Williams5
  1. 1Yale School of Public Health; New Haven, Connecticut, USA
  2. 2George Washington University School of Public Health and Health Services, Washington, DC, USA
  3. 3Harvard Humanitarian Initiative; Cambridge, Massachusetts, USA
  4. 4Department of Psychiatry University of Cape Town, Cape Town, South Africa
  5. 5Harvard School of Public Health, Boston, Massachusetts, USA
  1. Correspondence to Professor Jhumka Gupta, Global Health Epidemiology, Yale School of Public Health, 60 College Street, Suite 400, New Haven, CT 06510, USA; jhumka.gupta{at}
  1. Contributors JG conceptualised the current study, led the analysis and wrote the paper. ER assisted in writing the paper and the interpretation of data. JK assisted in writing the paper. DJS and DRW are both principal investigators of the study from which these data were obtained.

  • Accepted 8 November 2010
  • Published Online First 8 December 2010


Background Despite widespread apartheid-related human rights violations (HRV) and intimate partner violence (IPV) in South Africa, research investigating the influence of HRV on IPV perpetration is scarce.

Methods This study analysed data from the South Africa Stress and Health Study, a cross-sectional survey conducted from 2003 to 2004 with 4351 South Africans examining public health concerns associated with apartheid. Analyses were restricted to men who had ever been married or had ever cohabited with a female partner. Logistic regression was used to examine associations between experiences of HRV and lifetime physical IPV perpetration.

Results A total of 772 South Africa men met the study criteria (389 liberation supporters and 383 government supporters). Adjusted logistic regression analyses indicated that among liberation supporters, a significant association existed between experiencing major HRV (AOR 2.40, 95% CI 1.20 to 4.81), custody-related HRV (AOR 6.61, 95% CI 2.00 to 21.83), victimisation of close friends/family members (AOR 3.38, 95% CI 1.26 to 9.07) and physical IPV perpetration. Among government supporters, a significant association was observed between experiencing HRV (AOR 2.99, 95% CI 1.34 to 6.65) and victimisation of close friends/immediate family (AOR 5.42, 95% CI 1.44 to 19.02) and IPV perpetration.

Conclusion This work indicates the importance of men's experiences with HRV with regard to IPV perpetration risk. Future work is needed to understand the mechanisms underlying the observed relationships, particularly regarding mental health and gender norms as suggested by current literature, in order to inform interventions in South Africa and other regions affected by politically motivated conflict.

Political violence (eg, state-perpetrated violence, repression, genocide, torture, forced disappearance of family members and massacre) is a major human rights violation (HRV) that continues to overwhelm many societies across the globe.1 2 Among the numerous public health concerns that have been documented among conflict-affected populations (eg, poor mental health, negative birth and reproductive health outcomes, mortality),3–7 gender-based violence against women is receiving growing research attention.8–10 Much existing work, however, has largely focused on sexual violence perpetrated against women by armed groups. Less work has specifically examined violence perpetrated against women by male intimate partners (ie, intimate partner violence; IPV).

Experts have hypothesised that IPV may occur more commonly among those affected by political conflict and violence.2 Furthermore, high frequencies of IPV victimisation among women have been reported in settings rife with political conflict.11–13 However, only a handful of studies have specifically examined the association between political conflict exposure and IPV perpetration.14 15 These studies documented an association between political conflict exposure and higher odds of IPV perpetration among immigrant and Palestinian samples.14 15 Given these recent findings, existing work documenting a high prevalence of IPV among combat veterans in the USA,16–18 along with the well-established impact of IPV on the health of women and girls worldwide (eg, HIV/AIDS, negative maternal and child health, poor mental health),19–22 more work is needed to examine the influence of political conflict exposure on the perpetration of IPV against women. Such work can help inform better our understanding of the broad range of health issues that impact families and communities affected by political conflict, as well as how to better address such public health concerns.

The influence of political conflict on IPV is of particular concern in South Africa, given the high prevalence of IPV,23–25 health outcomes associated with IPV (eg, HIV/AIDS),19 26 and a history of politically motivated conflict and HRV surrounding the legacy of apartheid.27 Apartheid formally began in 1948 and continued to 1994. Under the apartheid system, the ‘non-white’ majority were subjected to ‘white’ minority laws that imposed forced segregation in terms of residence, medical care, education and public services, and triggered protests, violent uprisings and state-sponsored violence and repression. Thus, given that both IPV and a history of HRV are both highly prevalent in South Africa, and ethnographic work has called for greater consideration of the context of Apartheid-related violence victimization on men's use of violence against women,28 studies investigating the relation between HRV and IPV are needed but remain scarce in this region and elsewhere. Such work is needed to advance the current understanding of male-perpetrated IPV in this context, which can ultimately inform intervention efforts that can both reduce IPV and its deleterious health sequelae. The purpose of this study is thus to utilise a nationally drawn sample of South African men to examine the association between exposure to HRV during South Africa's apartheid era and IPV perpetration.


Study participants

This study utilised data from the South Africa Stress and Health Study (SASH)27 a national investigation of South Africans examining the prevalence and correlates of mental health concerns. This study obtained a three-stage, area probability sample of 4351 non-institutionalised adults (>18 years old). Area-based sampling frames were developed from enumeration area maps. The 85 783 geographical enumeration area units were divided into 53 strata based on province, urbanicity and racial group; 960 enumeration areas were selected. The 2001 South African Census was used to develop listings of homes/dwelling units within each enumeration area, in which a random sample of five homes/dwelling units was collected. Fieldwork supervisors and interviewers contacted each household/dwelling unit and obtained informed consent from a single adult respondent randomly selected using the Kish procedure.29

Of the 5088 adults selected, 4351 agreed to participate (85.5% response), of whom 1733 (39.8%) were men. Inclusion criteria for the current study were as follows: men who reported: (1) ever being in a married/cohabiting relationship; (2) a response on the physical IPV perpetration question; (3) a response as to whether they were a supporter of the liberation struggle or the government during the apartheid era; and (4) a response to all HRV items. In total, 889 men reported ever being married/cohabitating. Fifty-five men did not provide information on IPV perpetration. An additional 20 men did not report whether they supported the liberation struggle or the government during apartheid. Finally, 42 men did not provide complete information on HRV, thus restricting the final sample to 772 men.

All interviewers were trained in field research methods and administration of the paper and pencil version of the world mental health composite international diagnostic interview.30 Interviewers administered in-person surveys to consenting respondents during prescheduled appointment times from 2003 to 2004. All interviews were conducted in private areas with respondents. The questionnaire was offered in seven languages (English, Afrikaans, Zulu, Xhosa, northern Sotho, southern Sotho and Tswana). To ensure consistency of constructs across languages, questionnaires were translated into English and back-translated into each South African language; two pre-tests with 50 participants were also conducted.

All research protocols were approved by the human subjects committees of the University of Michigan and Harvard Medical School. A single project assurance of compliance from the Medical University of South Africa was approved by the National Institute of Mental Health.


All variables were assessed by self-report. Demographic variables (age, family income, education, marital status, employment) and physical IPV perpetration were assessed by items from the world mental health survey, a measurement instrument used in over 25 countries and all WHO regions.30 In particular, one item was used to assess physical IPV perpetration. Referring to their most recent or current female spouse or cohabiting partner, men were asked to indicate if they did the following: pushed, grabbed, shoved, threw something, slapped, or hit. All responses regarding physical IPV perpetration were dichotomised as never/ever.

Human rights violations

The HRV questionnaire developed for SASH aimed to elicit experiences with HRV by both supporters and critics of the apartheid regime.27 Extensive pilot work was conducted to capture exposure to HRV during the apartheid era; in-depth interviews were conducted with 90 South Africans from diverse backgrounds. Content analysis was then used to analyse emergent themes from these interviews; a formal interview with structured questions was then created. In addition, development of this instrument was informed by reading transcripts from the Truth and Reconciliation Council hearings. The structured interview was re-piloted to clarify any difficulties.

The pilot work revealed that during the apartheid era, different types of HRV were experienced by supporters of the liberation struggle and supporters of the government. Therefore, based on this pilot work, the SASH study created a different set of HRV questions for liberation supporters and a different set of HRV questions for government supporters.

During the survey interview, respondents were then asked a single item to self-report whether, during the apartheid era, they were supporters of the liberation struggle or the government. Tailored HRV questions were then asked of men who indicated being supporters of the liberation struggle and the government. In particular, government supporters were asked if: (1) they were ever victimised by HRV because of their political activities (eg, criticised by others, physically beaten or injured, faced with their home being burned, faced with other property being burned, victimised by political activists in any other way); and (2) their family or close friends experienced HRV because of the respondents' political activities (eg, detained, sexually assaulted, imprisoned, physically beaten or injured, killed). All responses were dichotomised as ever versus never.

Men who reported being supporters of the liberation struggle were asked if, because of their political activities: (1) they ever experienced major HRV (eg, were physically beaten, witnessed someone being necklaced, deliberately attacked by dogs, placed under house arrest, visited at home by the special branch police); and (2) they ever experienced HRV while in political custody (eg, body forced into painful positions, electric shock, extreme heat, withholding of food, unwanted sexual touching, isolation); (3) their family or close friends experienced HRV (same question that was asked of government supporters). Major HRV were coded as a three-level categorical variable (0, 1 or 2 and 3+). HRV while in political custody and HRV experiences of family friends were dichotomised as ever/never.

Statistical analysis

Because separate questions were asked of men who indicated being supporters of the liberation struggle and supporters of the government, all analyses were conducted separately for each subsample. The administration of separate questions to each subsample precluded a pooled analysis of the entire study sample. While this decreased our overall statistical power, we were able to investigate our association of interest, the association between experiences with HRV and IPV perpetration.

A prevalence estimate for physical IPV perpetration in the most recent marriage/cohabiting relationship was calculated separately for both supporters of the liberation struggle and the government, followed by bivariate demographic differences in physical IPV perpetration using χ2 analysis (significance p<0.05). We then conducted crude logistic regression analysis to examine bivariate associations between exposure to HRV and physical IPV perpetration. In particular, among liberation supporters, we investigated the association between exposure to the following HRV and physical IPV perpetration: major, custody-related and experiences of close friends/immediate family. Among government supporters, the association between IPV perpetration and being victimised by HRV and having friends/immediate family experience HRV was examined.

Separate multivariable logistic regression models were then constructed to examine these associations among liberation struggle supporters and government supporters, after controlling for demographics. Following rationale described by Miettinen and Cook,31 demographic variables that have been proposed or identified as correlates of IPV perpetration were included in the multivariable model as covariates and simultaneously entered. To preserve statistical power, all demographic covariates were dichotomised in adjusted models. OR were calculated and significance was assessed by calculating 95% CI (p<0.05). Due to the complex survey design, the weighting and clustering of the data were taken into account in data analyses by the Taylor series linearisation method using STATA version 9.32


Sample demographics

Overall, 50.4% of participating men indicated they were supporters of the liberation struggle, while 49.6% reported being supporters of the government during South Africa's apartheid era. A detailed breakdown of sample demographics is displayed in tables 1 and 2. As liberation supporters and government supporters were separately analysed, demographics are reported for each of these groups, respectively.

Table 1

Demographics and associations with physical IPV perpetration among South African men who were supporters of the liberation struggle (n=389)

Table 2

Demographics and associations with physical IPV perpetration of South African men who were government supporters (n=383)

Prevalence of physical IPV perpetration and associations with demographics

Overall, slightly over one in three (31.4%) male liberation supporters and nearly one in four (23.1%) male government supporters reported perpetrating physical IPV against their current or most recent female partner. Among liberation supporters, χ2 analyses indicated that marital status was associated with physical IPV perpetration, with men who were separated, divorced, or widowed reporting a greater percentage of IPV perpetration (66.2%) in comparison with currently married men (25.6%) and men who were in a cohabiting relationship (35.2%). No other significant relationships between physical IPV perpetration and other demographic variables considered in the current study were observed among either liberation or government supporters.

Relations of experiences with HRV and physical IPV perpetration

Liberation supporters

In bivariate analyses, experiencing a high number of major HRV (three or more) was significantly associated with greater reporting of physical IPV perpetration against a current or most recent female partner relative to experiencing no major HRV (crude OR 2.37, 95% CI 1.18 to 4.77), with 49.2% of men in this group reporting the use of such violence. No significant differences in IPV perpetration were detected between men who reported experiencing one or two major HRV compared with those who experienced no major HRV. In unadjusted logistic regression, we also observed that, relative to men who did not experience HRV while in political custody, men who did report such victimisation while in political custody were nearly twice as likely to perpetrate physical IPV (75.0% vs 29.7%, crude OR 7.09, 95% CI 1.95 to 25.79). Finally, compared with men who did not report that their close friends/family members were victimised by HRV, men who indicated that their close friends/family members were victimised by HRV because of their (respondents') own political activities were more likely (50.0% vs 31.6%) to report IPV perpetration. Crude logistic regression indicated this difference in IPV perpetration to be significant (crude OR 2.79, 95% CI 1.21 to 6.45). After controlling for age, race, education, employment, income and marital status, exposure to a high number of major HRV (AOR 2.40 95% CI 1.20 to 4.81), custody-related HRV (AOR 6.61 95% CI 2.00 to 21.83) and victimisation of close friends/immediate family members (AOR 3.38, 95% CI 1.26 to 9.07) continued to be significantly associated with IPV perpetration in adjusted logistic regression (table 3).

Table 3

Crude and adjusted logistic regression for associations between exposure to HRV and physical IPV perpetration among liberation supporters

Government supporters

Crude logistic analysis revealed that men who reported being victimised by HRV because they were government supporters were over three times as likely to report IPV perpetration than their counterparts who did not report such victimisation (44.2% vs 19.8%, crude OR 3.22, 95% CI 1.59 to 6.54). Significantly higher IPV perpetration was also observed among men who indicated that their close friends/immediate family experienced HRV because of their own political activities (56.3% vs 21.7%, crude OR 4.65, 95% CI 1.44 to 14.99). The relationships between those victimised by HRV as a result of being a government supporter (AOR 2.99, 95% CI 1.34 to 6.65), having close friends/immediate family experience HRV because of their own physical activities (AOR 5.42, 95% CI 1.55 to 19.02) and increased IPV perpetration persisted in multivariable logistic regression analyses (table 4).

Table 4

Crude and adjusted logistic regression for associations between exposure to HRV and physical IPV perpetration among government supporters


In this South Africa-based study with current or ever married/cohabiting men, we demonstrate for the first time, a significant relationship between men's experiences with HRV during apartheid, and the perpetration of physical violence against a current or most recent female partner. This association was evident regardless of whether men reported being supporters of the liberation struggle or the government during South Africa's apartheid era. Our findings corroborate the few existing data on political conflict exposure and men's increased IPV perpetration.14 15 Furthermore, our work is also consistent with existing literature that has documented a high prevalence of women's victimisation from gender-based violence in conflict-ridden settings.2 12 13 The current study thus strengthens existing evidence regarding the importance of considering the context of political conflict as an important factor in influencing men's use of violence against female intimate partners.

Building upon earlier work on political conflict and men's use of violence against female partners, the current study investigated victimisation from multiple forms of HRV violations and noted their effects on men's IPV perpetration. While more research is needed to clarify the mechanisms that underlie the observed associations, mental health sequelae associated with experiencing such atrocities are likely to play an important mediating role. Much research in South Africa and other global contexts has documented associations between HRV victimisation and post-traumatic stress disorder and other negative mental health outcomes (eg, depression).3 33 34 Poor mental health has received considerable attention as an important predictor of men's perpetration of IPV.35 36 Future work is needed to investigate how HRV may indirectly influence South African men's IPV perpetration by poor mental health sequelae.

Another factor that may play an important mechanistic role is perceived racial discrimination. Like the USA, South Africa is also a racialised society with a legacy of racial discrimination during and after apartheid27 37; thus investigating the role of discrimination in the relationship of interest is also warranted in South Africa. Given that recent evidence in the USA has indicated a link between perceived racial discrimination and IPV perpetration among urban African-American men,38 research is probably needed to examine how racial discrimination may also play an important role in HRV and related IPV perpetration. Such research may provide information that may be applicable to other settings that have been affected by politically motivated conflict or HRV, as such violence largely occurs along ethnic or religious lines.9 39 Moreover, experiencing racial discrimination has also been shown to be linked to a host of mental health concerns in South Africa and other nations where ethnic and racially motivated violence and discrimination exist.40 The cumulative effect of experiencing a HRV and racial discrimination may thus collectively influence men's violent behaviours against their female partners. Future research must be devoted to clarifying these potential pathways.

Another critical factor to consider is gender norms. Research in South Africa and elsewhere has established that patriarchal gender norms are linked to increased male-perpetrated IPV against women.24 41–43 Men who experience HRV may feel that their masculine identity is threatened, and thus try to exert their power over their female partner by using violence against them.44 Therefore, while patriarchal gender norms are likely to be a stronger predictor of men's IPV perpetration, existing social norms and structures disfavouring women may be exacerbated when men face such HRV and oppression, as has been discussed in situations of prolonged armed conflict.13

Study findings must be considered in the context of several notable limitations. Given the cross-sectional nature of the study, causality can not be established. In addition, survey items are self-reported, and may be influenced by social desirability or recall bias. However, social desirability is likely to lead to underreporting of violence experiences compared with overreporting. Another important limitation is our inability to control for exposure to multiple forms of HRV due to limited power, nor were we able to examine the potential effects of cumulative exposure to HRV and associations with men's IPV perpetration. In addition, we were unable to control for other violent exposure, such as general community violence (eg, robbery, muggings) as well as adverse childhood experiences. Furthermore, while the survey took place well after the apartheid era in South Africa, we are unable to determine whether the IPV perpetration in the current or most recent marriage/cohabiting relationship occurred after exposure to HRV in all cases. Finally, the current assessment of IPV perpetration is limited to physical IPV. Given the high prevalence of sexual IPV in South Africa,45 future research should also investigate associations between HRV exposure and sexual IPV perpetration.

These findings notwithstanding, our study is among the first to show the relation between HRV victimisation experiences and IPV perpetration—and the first study to examine this issue in the South African context. Future research is needed to investigate the mechanisms involved in this association (eg, mental health, racial discrimination, exacerbation of existing social norms disfavouring women), which may contribute to this important public health concern. Such research will be critical to informing public health programmes directed at reducing men's use of IPV. Very few programmes currently exist in areas affected by political conflict that focus on working with men to change their behaviour with regard to IPV perpetration. Our findings indicate that explicitly addressing men's experiences with HRV may be an important component of such programming in areas affected by widespread violence and conflict. Such programming may include structural interventions (eg, addressing socioeconomic losses), as well as group and community mobilisation focusing on promoting psychosocial health and changing gender norms. Given the immense global health impacts of both politically motivated conflict and IPV, future study is necessary to understand better this relationship and develop appropriate strategies to both reduce such violence and minimise the associated health consequences.

What is already known on this subject

  • High levels of male-perpetrated intimate partner violence against women have been well documented in South Africa.

  • Politically motivated violence during South Africa's apartheid era was widespread.

  • Earlier work in other settings (US-based immigrants, Palestine) has documented an association between politically motivated violence and IPV.

What this study adds

  • Exposure to HRV during apartheid was associated with an increased likelihood of men's perpetration of IPV.

  • Our findings underscore the importance of considering the context of political conflict when developing programmes to prevent and reduce men's use of IPV against women in settings affected by a history of political conflict.


The South African Stress and Health Study was carried out in conjunction with the WHO world mental health survey initiative. The authors would like to thank the staff of the World Mental Health Survey for assistance with instrumentation, fieldwork and data analysis.


  • Funding The activities of the World Mental Health Survey Initiative were supported by the US National Institute of Mental Health (grant R01MH070884), the John D and Catherine T MacArthur Foundation, the Pfizer Foundation, the US Public Health Service (grants R13-MH066849, R01-MH069864 and R01 DA016558), the Fogarty International Center (grant FIRCA R01-TW006481), the Pan-American Health Organization, Eli Lilly and Co, Ortho-McNeil Pharmaceutical, Inc, GlaxoSmithKline and Bristol-Myers Squibb. The South Africa Stress and Health Study was funded by grant R01-MH059575 from the US National Institute of Mental Health and the US National Institute on Drug Abuse, with supplemental funding from the South African Department of Health and the University of Michigan. JG's work was partly supported by award T32MH020031 from the National Institute of Mental Health. The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the US National Institutes of Health. A complete list of publications related to the World Mental Health Survey Initiative can be found at

  • Competing interests None declared.

  • Ethics approval This study was conducted with the approval of the University of Michigan, Harvard Medical School and the Medical University of South Africa.

  • Provenance and peer review Not commissioned; externally peer reviewed.


Free sample

This recent issue is free to all users to allow everyone the opportunity to see the full scope and typical content of JECH.
View free sample issue >>

Don't forget to sign up for content alerts so you keep up to date with all the articles as they are published.