01 Jun 2006: Mapping the Use of Guns in Violence Against Women: Findings From Three Studies, Lisa Vetten essay Mapping the use of guns in violence against women: Findings from three studies Lisa Vetten Lisa Vetten is the gender project co-ordinator at the Centre for the Study of Violence and Reconciliation (CSVR) in South Africa Published in African Security Review Vol 15 No 2 This article explores the role of firearms in acts of violence against women in South Africa, drawing on three datasets: one investigating the implementation of the Domestic Violence Act (DVA), the second exploring gang rapes, and the third documenting intimate femicide. In relation to domestic violence, it was found that while guns were referred to in one in four applications for protection, their removal was ordered in only two per cent of applications. Both a provincial femicide study and a national female homicide study found guns to be the leading cause of death for women killed by their intimate male partners and found that in the majority of cases, the gun was legally owned. In contrast, in the gang rape study it is more likely that the guns were illegally owned. The involvement of a firearm in gang rapes highlights the fact that guns not only fulfil the functional purpose of intimidation and injury, but also communicates power and masculine display. The three studies point to the need to train magistrates around the necessity to remove firearms in cases of domestic violence and that it is essential to challenge the symbolic associations between masculinity and power.  Violence against women is endemic to South Africa. Community-based prevalence studies find that domestic violence, in one form or another, affects as many as one in two women in some parts of the country.1 The consequences of domestic violence may be fatal, with women in abusive or controlling relationships being twice as likely to be infected with HIV than women in non-violent relationships.2 While a small proportion of women will kill their abusive partners, they are more likely to be the victims of homicide.3 Approximately half of all South African women murdered in 1999 were killed by their intimate partners, translating into a prevalence rate of 8.8 per 100 000 of the female population aged 14 years and older – the highest rate ever reported in research anywhere in the world.4  The true extent of sexual violence in South Africa is currently unknown. Statistics South Africa found that one in two rape survivors reported being raped to the police,5 while the Medical Research Council of South Africa (MRC) found that one in nine women reported being raped.6 Both studies clearly find rape to be under-reported, although their findings differ as to the extent of such under-reporting. On the basis of the above studies it can be extrapolated that the 55,114 rapes reported by the South African Police Service (SAPS) in their 2004/05 released data is an inaccurate reflection of the prevalence of rape in South Africa. It is more likely that between 110,000 and 495,000 actual rapes have taken place in that period.  This essay maps out the contribution of firearms to these statistics by reviewing five studies which, to varying degrees, have explored the role of firearms in violence against women. It opens by considering the role of firearms in a study of reported rapes involving two or more assailants. This is followed by a description of the methodologies and findings of two studies which, in different ways, explore aspects of domestic violence. The final two studies, one provincial and the other national, highlight the role of firearms in intimate femicide, or men’s killing of their intimate female partners. The essay concludes by discussing the implications of the various findings for policy and practice around firearms control. Gang rape and the use of firearms  Vetten and Haffejee conducted a small-scale retrospective census of all cases of rape and attempted rape reported by complainants 12 years old or older at six central Johannesburg police stations during 1999.7 One hundred and fifty-six cases, or just over one in four (or 27 per cent) of the 591 rapes reported at these six police stations during 1999 involved two or more perpetrators. The majority of gang rape cases involved two perpetrators (56 per cent), and in two cases there were as many as 12. Of the 162 cases, nine were recorded as attempted rape. Nine cases involved more than one victim.  In 40 per cent of gang rape cases the perpetrators were armed, usually with either guns (two per cent) or knives (17 per cent). While the weapon was primarily used to threaten the victim, in 14 per cent of cases it was used both to threaten and injure the victim. The study found that the greater the number of rapists involved, the greater the likelihood that at least one of the assailants was armed with some sort of weapon. Approximately one in four of the lone perpetrators were armed, compared to one in three of the rapist pairs and one in two of the rapist groups.8 The use of guns was also most likely in cases involving group rapists. Fourteen per cent of both the lone and pair rapists were armed with guns, compared to 25 per cent of the group rapists. It is unknown whether the guns involved in these cases were legally owned.  While the majority of victims (59 per cent) were dragged, pushed or shoved to the site of the rape, abduction at gunpoint was more typical of the cases involving group rapists. Of the 29 per cent of victims abducted by car, 40 per cent of such cases involved three or more perpetrators, as opposed to 20 per cent of cases involving two perpetrators and 19 per cent of cases involving one perpetrator.  In this study, the greatest proportion of rape cases involving guns was likely to involve three or more perpetrators. One would expect that force of numbers alone would be considered more than sufficient to subdue the victim, without the additional threat of a gun being necessary. However, the multiple perpetrators involved in group rapes may also be fulfilling the function of an audience. Thus the use of firearms in these situations may be as much to coerce and intimidate the victim, as it is to demonstrate toughness and power to male peers. Firearms and domestic violence  This section presents findings from two studies of domestic violence. The first is a retrospective, random review of applications for protection orders in terms of the Domestic Violence Act (No 116 of 1998) (DVA) registered in 2000 and 2001 at Alberton magistrate’s court, an urban court east of Johannesburg (1,537 applications) and Temba, a semi-urban court outside Tshwane Metropolitan Area (671 applications). A total sample of 2,208 applications were captured.9  Weapons were referred to in 813 (37 per cent) of the applications for protection orders. In at least one in four (25 per cent) cases a gun was used, in one in five applications (20 per cent) a knife was the weapon of choice, and in two per cent of cases a sjambok was referred to. Multiple weapons were used in one in eight cases, while a miscellany of weapons (including household items) accounted for the remaining 36 per cent of cases.  When guns featured, they were primarily used to threaten applicants (88 per cent), while knives were more likely to be used both to threaten and injure applicants (32 per cent of cases) than guns (10 per cent). This may be due to the fact that those who own knives may be more willing to injure than those who own guns. It is also possible that merely displaying a firearm is sufficient to intimidate victims and force compliance; there is no need to act on the threat – as may be the case with knives. Nonetheless, although weapons were used in more than one in three cases, orders for the police to remove those weapons were made in a scant two per cent of cases. To a large degree this low number of orders may be attributed to the fact that very few applicants (three per cent) actually asked that the weapon be removed.  Our findings corroborate those of a study of protection orders granted in Cape Town, Mitchells Plain and George in the Western Cape.10 The authors found that magistrates rarely ordered seizure of weapons in comparison to the number of times applicants referred in their applications to the use of weapons against them. It was also noted that applicants did not often request the seizure. The study also suggested that the lack of detail in the protection order forms around what constitutes a dangerous weapon may have made it difficult for police officers to confiscate the weapons in question.11  The second study with findings relevant to describing the use of firearms in domestic violence was conducted at the three prisons in Gauteng which house female prisoners: Johannesburg, Pretoria and Heidelberg.12 When the study began, 565 women (525 adults and 40 juveniles) were held at Johannesburg prison, 146 (137 adults and 9 juveniles) at Pretoria and 57 at Heidelberg. The total population across the three prisons was 768. Through random, proportional sampling of both adult and juvenile female prisoners a sample of 569 female prisoners in Gauteng was selected for the study, calculated at a 99 per cent confidence interval and three per cent margin of error.13 Questions around violence were adapted from the World Health Organisation’s (WHO) Questionnaire on Emotional, Physical and Sexual Violence. Adaptations were based around the need to tailor the questionnaire to the unique circumstances of imprisonment.  Sixty-three per cent of women in the sample had experienced physical abuse in their last relationship. Just over one in four of the women (26 per cent) who had experienced intimate partner abuse reported fractures or dislocations of their bones, and a further 15 per cent reported gunshot and stab wounds as a result of the violence. While the method of injury in the majority of cases was unrecorded, 17 per cent reported that a sharp object like a bottle or knife was used to injure them, while four per cent reported that a gun was used. Firearms and intimate femicide  Internationally, the murder of women by their intimate partners accounts for between 40 and 70 per cent of all female homicides.14 In South Africa approximately 50 per cent of female homicides in 1999 were perpetrated by an intimate partner.15 This section highlights findings in relation to firearms made by two studies exploring intimate femicide.  Vetten, Ngwane and Isserow (in progress) undertook a ten-year retrospective study of men’s killing of their intimate female partners in the period 1990-1999. A census was conducted of all cases reported at 106 police stations in Gauteng.16 These records were cross-referenced with information from the 17 regional courts, as well as the inquest court records kept at the regional courts. Research was also conducted at the two high courts, Johannesburg and Pretoria, as well as the National Archives.17 A search of all these sources identified a total of 941 cases having occurred during this period. This is likely to be an undercount, given that records had been destroyed at a number of courts and police stations, or were still with the investigating officer, pending finalisation.  In this study, blunt force (including hands and feet and blunt objects) accounted for 35 per cent of women’s deaths. Sharp force (in the form of knives and other sharp objects) brought about 29 per cent of women’s deaths. The same percentage of deaths was caused by guns. Notably, gun use increased overall for this ten-year period from 23 per cent of cases in 1990 to 41 per cent of cases in 1999. This is an increase in gun use of 78 per cent in real terms over this period.  In 11 per cent of cases, others such as adult family members, children, friends, work mates and passers-by were also attacked and/or killed. Guns were used in 47 per cent of attacks, followed by knives (23 per cent). In a further 14 per cent of cases (one in seven) men committed suicide afterwards, while a further three per cent attempted suicide. Guns were used in 71 per cent of suicides. As the number of firearm killings increased over the period studied, so did the number of homicide-suicides. Further research is required to test whether or not this increase may be attributed solely to the increasing availability and use of guns during this period.  The study also identified a number of factors associated with a gun-killing. These included being a white man (three times increased risk), being younger (risk decreased by 4 per cent each year), being employed and working in the security industry – including police and the South African National Defence Force (SANDF) (nine times increased risk) and working as a professional (four times greater risk) or working in a semi-skilled occupation (three times greater risk). Being married (90 per cent increased risk compared to other relationship types), as well as wanting to end the relationship (risk increased six times) also increased the likelihood of a gun homicide.  The second study discussed in this section was national and retrospective and collected data on female homicides in 1999 of women aged 14 and over.18 Data were collected from a proportionate, random sample of 25 mortuaries, yielding a sample of 951 female homicides. Initial information was gathered from mortuary records and then followed with either a review of the docket or an interview with the investigating officer.  Women who died at the hands of their intimate partners were more likely to be killed by a legal firearm than women killed by non-intimates. One in every five perpetrators of an intimate femicide owned a legal firearm compared to 3,5 per cent of non-intimate perpetrators. The study also provides insight into aspects of the relationship between the use of firearm in homicides and criminal justice system processes. Longer sentences of imprisonment resulted in both the intimate and non-intimate homicides when the perpetrator used a legal firearm. Implications for policy and practice  Firearms are clearly implicated in rape, domestic violence and intimate femicide. However, the risk they pose to women appears to be contingent upon a range of factors such as the perpetrator’s age, class and occupation, as well as the state and status of his relationship to the victim. Gun use even within the same type of crime would also appear to vary, as the study involving multiple rapists suggests. Thus, while controlling and regulating possession of firearms may well play a role in combating such violence, the effect of such policies on combating violence against women is also likely to be variable.  This review also points to a number of issues around the implementation of existing policy and law. First, that so few women request the removal of guns despite their frequent use in cases of domestic violence is cause for concern – particularly in light of the findings of the Gauteng femicide study and the national female homicide study. Can this be attributed to an insufficient number of women knowing of that clause in the DVA which allows for the removal of weapons? Or, within a context of fear of crime generally, do women believe the protective value of a gun in the home outweighs the personal danger to themselves? Or are they simply too fearful of their partner’s reaction to having his gun removed? Further, if women are not aware of this clause, are they any more likely then to be aware of Chapter 12 of the Firearms Control Act, which permits them to make submissions around their male partners’ competence to own firearms?  Whatever the answers to these questions, the findings underscore the need to train magistrates to order the removal of guns in cases of domestic violence as well as clarifying some wording in the DVA which may obscure the procedures for removing firearms. Additionally, it is necessary to finalise and publicise policy regulations regarding SAPS, SANDF officials’ and prison warders’ off-duty access to guns.  Finally, further study is required to better understand the meanings attached to firearms. As the gang rape study suggests, firearms serve more than an instrumental, functional purpose. It is entirely possible that they are also essential to performances designed to impress male peers with the protagonist’s masculine power and control. Further, as the Gauteng femicide study highlights, some groups of men are considerably more likely to value and own firearms than others. Thus it is also crucial to identify the kinds of masculinity most likely to be associated with gun use, as well as to embark on strategies designed to challenge these associations. Notes N Abrahams, R Jewkes and R Laubscher, I do not believe in democracy in the home: Men on relationships with and abuse of women, Medical Research Council Technical Report, Tygerberg, Medical Research Council, 1999. A 1997 study conducted in three of South Africa’s nine provinces found that 27 per cent of women in the Eastern Cape, 28 per cent of women in Mpumalanga and 19 per cent of women in the Northern Province had been physically abused in their lifetimes by a current or ex-partner. The same study investigated the prevalence of emotional and financial abuse experienced by women in the year prior to the study. This was found to have affected 51 per cent of women in the Eastern Cape, 50 per cent in Mpumalanga and 40 per cent in Northern Province (ibid). A study of 1,394 men working for three Cape Town municipalities found that approximately 44 per cent of these men were willing to admit to the researchers that they abused their female partners. K Dunkle, R Jewkes, H Brown, J McIntyre, G Gray and S Harlow, Gender-based violence and HIV infection among pregnant women in Soweto, Gender and Health Group, Men on relationships with and abuse of women, Medical Research Council Technical Report, Medical Research Council, Tygerberg. H Ludsin and L Vetten, Spiral of entrapment: Battered women in conflict with the law, Johannesburg, Jacana, 2005. S Mathews, N Abrahams, L J Martin, L Vetten, L van der Merwe and R Jewkes, Every six hours a woman is killed by her intimate partner: A national study of female homicide in South Africa, MRC Policy Brief No 5, June 2004. R Hirschowitz, S Worku and M Orkin, Quantitative research findings on rape in South Africa, Pretoria, Statistics South Africa, 2000. R Jewkes and N Abrahams, The epidemiology of rape and sexual coercion in South Africa: An overview, Social Science & Medicine 55(7), 2002, pp 1231-1244. L Vetten and S Haffejee, Urban predators, Centre for the Study of Violence and Reconciliation Gender programme, Policy Brief No 01, January 2005. This refers to those cases involving three or more rapists. V Schneider and L Vetten, Going somewhere slowly? Research report prepared by the Centre for the Study of Violence and Reconciliation Gender Programme, 2006. P Parenzee, L Artz and K Moult, Monitoring the implementation of the Domestic Violence Act: First research report 2000-2001, Institute of Criminology, 2001. Ibid. S Haffajee, L Vetten and M Greyling, Violence and abuse in the lives of women and girls incarcerated at three Gauteng women’s prisons, CSVR Gender Programme Research Brief No 3, February 2006. Ultimately, 348 women were interviewed with a further 28 refusing to be interviewed and the remainder of interviews lost to releases. To account for the difference in sample size and the number of interviews ultimately completed, data was weighted to account for instances of over- or under-representation, taking into account the relative sizes of the prisons and the number of juveniles, and was thus proportional to size. L L Dahlberg, E G Krug, A J Mercy, A Zwi and R Lozano, World report on violence and health. Geneva, World Health Organisation, 2002. Mathews et al, op cit. Eight police stations in Soweto, 15 police stations in the West Rand, 23 police stations in the East Rand, 20 police stations in Johannesburg, 28 in Pretoria and 12 police stations in the Vaal. The archives keep records of all high court matters completed prior to 1993. Ibid.
