A new tool for collecting and analyzing homicide data: Homicide Media Tracker Part 2
By Dr Nechama Brodie
Data about criminal incidents contained in media reports presents a huge opportunity for researchers to plug the gaps often left by police and judicial records. In Part 1 of this blog series, Dr Nechama Brodie summarized these potentials and introduced a concept for a new tool currently under development with support from the SAGE Concept Grant that will enable the collection, curation and analysis of this data.
In Part 2, she tells us more about the Homicide Media Tracker project: Why is it needed, what kind of data will the tool be able to collect, and what insights can this data afford researchers? Read on to find out what’s in store for the Homicide Media Tracker.
There are a number of existing initiatives that track regional crime through media and social media reports – most frequently femicides, and hate crimes – with the aim of raising awareness and increasing reporting of these crimes.
However, these trackers are usually opportunistic (i.e. they don’t necessarily search systematically through all media), they don’t gather data on the media reports themselves, nor do they necessarily aggregate multiple media reports, which makes it difficult to assess the robustness of their media sample. As a result, data from these trackers is often considered anecdotal, and cannot easily be integrated with other crime data such as police reports or epidemiological studies.
In a move to plug this data gap, the Homicide Media Tracker Project looks to create a digital tool that will enable more efficient manual collection and enhanced digital curation of homicide-in-media data from multiple media sources. This will also allow the data to be more easily saved, searched and shared.
This new tool will allow media researchers access to much larger and more robust data sets, instead of being limited to only a handful of articles or news titles, or having to rely on broad but shallow searches that use word frequency as a proxy for better data. This tool means that it will be possible to easily pull out hundreds or thousands of relevant news reports, which can be analysed for language and tone, narrative framing, news clustering, and countless other qualitative fields.
What kind of data will the Homicide Media Tracker be able to collect?
When a news story reports on or mentions a murder, the individual incident will be captured together with all other data about the crime mentioned in the same report or other news reports. This could include the name and age of the victim, any information about the perpetrator, and the place, cause and date of death. News stories may also contain information about the relationship between the victim and perpetrator (if they were intimate partners, family members, or strangers), and, eventually, whether or not a suspect was arrested, charged, or convicted.
The tool will also capture information about the news story itself: who wrote it, when and where it was published, and a link to its digital location (or a unique file number if it is hosted on another service), together with – ideally – a digital ‘hard copy’ of the report.
Visualization of data collection in the Homicide Media Tracker
The concept for the Homicide Media Tracker has been developed by researcher Dr Nechama Brodie, with its basis in a multi-year research project carried out for her doctoral thesis. The project used mixed-methods approaches to study femicide in South Africa by taking female homicide data from and about media reports mentioning femicides from 2012/13, and integrating this information with other statistical and epidemiological data on the phenomenon of femicide in South Africa.
No similar data yet exists for other homicides, or for other years. In South Africa men are killed at a rate of five to six times the femicide rate, making it an urgent problem.
What can we learn from this data?
It’s important to note that media coverage does not always reflect the reality of crime – a story about a child being abused by a stranger typically receives more news coverage than a child experiencing abuse at the hands of a close family member, even though the latter is statistically much more likely. Media coverage is also more concentrated in urban areas, where there are more media outlets. For these and other reasons, media reports of crime should never be assumed to be equivalent to the actual prevalence of crime.
So what can this data tell us? The Homicide Media Tracker will collect and categorize expanded information about individual homicide incidents while also annotating and collecting data about the media coverage linked to that event. That’s to say, the tool will provide data not only from the news story, but also about the news story. The tool will also do this at a scale that will make the media sample and data more representative and robust, and therefore more useful for multiple applications including academic research.
“The Homicide Media Tracker will collect and categorize expanded information about individual homicide incidents while also annotating and collecting data about the media coverage linked to that event.”
This information will allow us to study how the media reports on homicide, what proportion of homicides are covered in the press, and which homicides get more coverage than others.This information is important because many people get their knowledge about violent crime from media reports, and media coverage of crime can influence audience perceptions of who is most at risk as a victim, and who is most-feared as a perpetrator.
The data collected will also make it possible to study a diverse array of factors such as types of victims, outcomes of judicial proceedings and changes in media representations of criminal activity over time. Read more about this in Part 1 of this blog series.
What’s next for the Homicide Media Tracker?
Whilst the Homicide Media Tracker is being initially developed in South Africa, the goal is to make it adaptable for use in other countries, and even in other languages, without the need for expensive software or equipment. This would enable data collection to take place in environments outside of academic settings, and make the information accessible to many instead of restricted to a few.
In order for the Homicide Media Tracker to produce robust and representative data, it needs to be able to collect, annotate and store data at scale, and to store it in a centralized location. This requires new or adapted digital tools which can supplement or replace existing spreadsheet and annotation programs, and for the tools and data to be easily shared with other people. It is the intention that the Homicide Media Tracker is used for and by open data projects.
The Homicide Media Tracker project was awarded a SAGE Ocean Concept Grant of £2,000 earlier this year. This seed funding will be used for software licensing and storage costs related to getting the project from concept to pilot phase, and contribute towards basic branding, marketing and networking for the project.
A further benefit of the backing from SAGE is access to SAGE Ocean’s network of projects and grantees, which connects the Homicide Media Tracker project with a broader community that is developing or have already developed digital tools and platforms that could be useful to the Homicide Media Tracker.
about
Nechama Brodie is a journalist, author and academic based in Johannesburg, South Africa. She holds a PhD in journalism from the University of the Witwatersrand, where she is also a part-time lecturer. Her research work focuses on homicide in the Global South, integrating media sources with statistical and epidemiological data to gain new insights into interpersonal and criminal violence. Her latest book is Femicide in South Africa published by Kwela Books.