Investigating femicide �in South Africa

A colloboration between Media Hack Collective �and Bhekisisa Centre for Health Journalism

Laura Grant, GIJC, September 2023

The story

#SayHerName: The faces of SA’s femicide epidemic

• Bhekisisa Centre for Health Journalism received
• Canon Collins Trust 2020 Sylvester Stein Fellowship
• Media Hack Collective brought on to do the data journalism component
• Bhekisisa did the narrative component

bhekisisa.org/article/2021-04-14-sayhername-the-faces-of-south-africas-femicide-epidemic/

Database of 482 murdered women

Interactive map and filterable cards with the story

No ‘ordinary’ murders

• 1 in 3 women were killed by their current or former intimate partner
• Almost half knew their killer
• 1 in 3 of the victims are unnamed in the reports
• In 1 in 4 cases there was little information about how the woman died

Backstory

From gender-based violence to femicide

• Initially, we wanted to look at gender-based violence, but we couldn’t find reliable data
• We decided to focus on femicide – female homicide – because murder is the most reliable crime statistic available
• But SA’s official crime statistics - made available quarterly and annually in Excel workbooks - don’t break down murder numbers by the gender of the victims

Finding data

Slide decks

• A slide in a presentation�deck contained an overview of�crimes against women and�children
• 13,815 women murdered in a �five-year period. (2015-2020)
• Average of 7 women a day

Is that high?

International comparisons

• In SA woman victims are ~13% of the total murder victims
• Global female homicide rate 2.3 per 100,000 women (2017)
• South Africa’s female homicide rate 9.5 per 100,000 women (2018)
• One of the 5 highest countries for which stats are available

-UN Office on Drugs and Crime Global Homicide Report 2019

Finding more data

Public access to information requests

• We asked for more police statistics on murdered women
• At which police stations were the murders reported?
• Are there ‘hotspots’?
• Request was unsuccessful

What happened to the murdered women?

Official police statistics

• Femicide is not mentioned in the official crime stats presentations
• Causative factors listed for murder that could perhaps relate more to women than men are:�- domestic related�- rape related�- witchcraft related

Who murdered them?

What the police statistics say

• Doesn’t cover all the murders
• No gender of the victims

Turning to the media

For info about the women behind the murder stats

• Searched mostly English-language online media reports
• Set time period of 1 January 2018 and 31 October 2020.
• Found reports on 482 murders.
• 233 of them women 18 years or older (which the crime stats covers).
• Reports for 4% of the 5,466 total murder

The Spreadsheet

How we collect our homicide data

• We collect all the information we can find about each murder
• 25 variables that describe the victim and the incident - also was someone arrested, were they convicted, what was the sentence?
• Plus a number of columns for sources and links to sources

Searching online media reports

Google searches

• For the story, we had to search retrospectively for media reports
• We limited the time period to 2 years and 10 months
• We started with search terms like woman, female, murdered and killed
• Looked for multiple reports: If we found a murder where a name was mentioned we searched for the victim’s name or another detail of the crime for other reports on that murder to verify the information
• Followed leads, such as social media pages set up to focus on victims of gender-based violence

Still searching

• 3 people did the searching over a period of about 3 months
• Did extra searches on murders in the dataset from time to time to look for more information
• Regularly cleaned the dataset to remove duplicates

Difficulties collecting the data

• Duplications - police didn’t name the victim or different spellings of same name
• Lack of detail about how or where the murder occurred
• Date of murder is not included in article
• No information on the age of the victim
• Cases reported on if they get to court

After the story was published

What happens to the data?

• We decided to keep updating the spreadsheet
• Initially we did Google searches once a month and added any new murders to the spreadsheet
• In 2022 we began using a Google news alert: murder OR killed AND woman AND site:za
• 7 alerts 1-10 September

Growing problem

• Murder numbers on their own, with no human faces, with no stories, allow us to distance ourselves from this problem
• Without reliable data how do you understand the extent of the problem?
• Where to mobilise resources?
• Who is most in need of protection?

Why are we doing this?

• Our dataset now contains 1,000 female murders
• It has limitations: we know it’s only a small percentage of the total murders
• It’s likely that certain types of murders will be reported in the media, eg, celebrities, young pretty women, gruesome or salacious
• The murders that are not reported have their own important stories – poor women, in rural areas whose stories are not heard.

Laura@theoutlier.co.za

Thank you

​