My first research job was spent at the Johannesburg Magistrates’ Court in an overflowing cupboard of a room housing inquest files. Every day I sorted the male from the female non-natural deaths and then analysed the women’s cases.
A geography of sorts developed in my head: one of a Johannesburg where the borders were marked by the balconies, underpasses and streets where women’s bodies had been found and another informed by the detailed topography inscribed by violent death – tram track injuries consistent with a sjambokking*, stab wounds rendered fatal by their depth and placement, and the peculiar interior pathways travelled by bullets as they ricocheted and bounced within bodies.
Stranger still was the discovery that this violence was not the handiwork of strangers, but of husbands, boyfriends and lovers. Home, it turned out, was where the hurt was.
This particular study, conducted in 1995, was the first in South Africa to explore female homicide. It calculated that every six days a woman in the country’s Gauteng province would be killed by her intimate male partner.
A woman murdered every six hours
A more detailed national female homicide study was later undertaken by South Africa’s Medical Research Council, the Centre for the Study of Violence and Reconciliation and the University of Cape Town which found that approximately half of all women murdered in 1999 died at the hands of their partners.
And, instead of one woman dying every six days, it revealed that a woman was being murdered every six hours. This translated into four women killed every day or, in more scientific terms, a prevalence rate of 8.8 per 100,000 of the female population and one that was also six times higher than the global average for such killings.
This large proportion of women killed by their intimate partners is consistent with trends internationally and as a phenomenon is termed “intimate femicide”. It also needs to be understood as domestic violence at its most lethal and extreme. Given how widespread domestic violence is in South Africa, it is unsurprising that the prevalence of intimate femicide turned out to be so high.
The 1999 female homicide study findings became dated and so a second study was conducted in 2009. This was also necessitated by the fact that the police do not routinely disaggregate homicide figures by sex. Even when they do provide such figures, they are so decontextualized that we know nothing about the relationship between perpetrator and victim, nor the circumstances under which the killing occurred – information crucial to developing evidence-based policies and programmes around violence, as well as assessing the impact of such programmes.
Why correct sampling is crucial
The sheer volume of homicides in South Africa ensures that it is a near-impossible task to study all murders committed in any one year in the country. For this reason researchers sample a sub-set of cases from the total population of homicide cases instead. How this sample is ultimately selected is crucial. If it is unrepresentative of the total population, it produces skewed, unreliable findings.
Because all victims of non-natural or suspicious deaths must undergo a post-mortem examination, mortuaries were chosen as the first source of information about female deaths for both the 1999 and subsequent 2009 study. Again, because all mortuaries could not be visited, they were stratified according to the number of autopsies performed annually.
Once clustered as small, medium or large in size, a sampling fraction was then used to determine the number of mortuaries visited in each category. Individual female homicide cases were then identified from the autopsy registers housed at the selected mortuaries and data extracted from the accompanying autopsy reports. This was further supplemented by interviews with detectives investigating the cases chosen.
The 2009 study showed the number of female homicides to have declined from 24.7 per 100,000 in 1999, to 12.9 per 100,000 in 2009. However, the decrease in female homicides was far more noticeable for non-intimate female homicides than intimate femicides.
Non-intimate female homicides declined from 8.6 per 100,000 in 1999 to 4.2 per 100,000 in 2009, while intimate femicides declined from 8.8 per 100,000 to 5.6 per 100,000.
The leading cause of female homicide in South Africa
Although the decline in non-intimate homicide is statistically significant, that for intimate femicide is not. In other words, over the decade, there was a real decrease in the number of women killed by strangers, family members, friends and acquaintances but no significant decrease in the number of women killed by their husbands, lovers and boy-friends. As a result, by 2009, intimate femicide had become the leading cause of female homicide in South Africa.
Drawing on media reports for the period May 2012 and June 2013, Steve Hofmeyr and Sunette Bridges, have disputed this research by producing 55 cases of white women they claim were killed by unknown black men and one report of a white woman killed by her intimate partner. The latter was Reeva Steenkamp, who was shot and killed by her boyfriend, the South African paralympian athlete Oscar Pistorius.
Relying on the underlying claim that the statistical ratio of non-intimate femicide to intimate femicide is 55:1, Hofmeyr and Bridges have alleged that black men pose a very considerable risk to white women – an argument that must be examined critically.
Interrogating the numbers
First, it is important to note that their figures are drawn from unverified reports published over a 14-month period, not a 12 month period.
Their figures also need to be located within the number of female homicides overall – an admittedly inexact exercise given that the police have yet to release crime statistics for 2012/13. In the absence of these figures, the most recent numbers will be used instead. These state that 2 286 murders of women over the age of 18 were recorded by the police between 1 April 2011 to 31 March 2012. The 56 cases referred to by Hofmeyr and Bridges would therefore account for roughly 2% of female homicide cases nationally.
Because this percentage is much lower than the proportion of white women in the general population, it suggests that either white women face a very low risk of homicide, or that Hofmeyr and Bridges have excluded a significant number of cases from their argument. Considering their methodology, the latter explanation is more likely.
Their method of gathering information also needs to be interrogated. Where did they get the names? Were they supplied to them by their various followers on Twitter and Facebook or did they examine all newspapers in the country, online and print, both daily and weekly, national and local? If not, which newspapers did they select and what informed their choice? In other words, what was included and what was excluded? Further, did they review their chosen newspapers in a routine and consistent way – that is every day or every third day or once a week? Again we need to know this to understand what may have been included and what may have been left out.
Claims make light of intimate femicide
But even bigger questions arise around their choice of source: newspaper reports. Journalists largely obtain their crime-related information from the South African Police Service. This means that whatever information journalists have been provided has first been filtered by the police.
Information about homicides is then filtered again in the newsroom where journalists and editors decide whether or not a particular crime is newsworthy. Those that do fit the criteria of newsworthiness will receive coverage while those that do not will fall by the wayside. Newspaper reports, therefore, do not accurately reflect either the prevalence or incidence of homicide.
Hofmeyr and Bridges’ claims therefore cast no doubt on the research showing the majority of women to be murdered by their intimate partners and, by extension, that the majority of white women are murdered by their white male partners.
However, Hofmeyr and Bridges’ 55:1 ratio is very telling for another reason and that is for underscoring quite how negligible, unremarkable and mundane both the police and print media consider intimate femicide to be. This is the real outrage of this debate: that Hofmeyr and Bridges wish to make light of intimate femicide in favour of racial scare-mongering and that the media treats these deaths as so unworthy of coverage.
* A sjambok is a heavy leather whip, traditionally made in South Africa out of hippopotamus or rhinoceros hide. These days they are more commonly made of plastic.
Lisa Vetten is an independent researcher specialising in gender and violence and was involved in both the 1995 and 1999 femicide studies referred to in the article. On a number of occasions she has also appeared in court to provide expert testimony on intimate femicide.