How can we use data science to better understand public attitudes towards the criminal justice system?
Through our partnership with Imperial College London, Thomson Reuters Labs – London was approached by the UK Ministry of Justice (MoJ) to explore different data-driven methods for understanding public attitudes towards the criminal justice system.
The MoJ are constantly seeking to become a more evidence and data-driven department and, at a recent gathering of their policy team, wanted an insight into the opportunity of harnessing the power of data for their justice policy-making process.
Opinions are data
Social media platforms have exploded in number, size and impact in recent years, replacing the soapbox in the way that aircraft replaced the horse carriage – bigger, better, and way more powerful.
Never before has it been so easy to tell the world what you had for lunch, who you’ll be voting for, or whether you think that justice has prevailed in the latest criminal court case.
Apart from acting as a trampoline to your thoughts, social media is also where more than half of news consumers worldwide get their news today. The nature of these platforms enables users to share, discuss and comment on the news – voicing their opinion on real matters in a public forum on a large scale. The sheer volume of activity means that social media are now invaluable for comprehending the public view. We now have the ability to understand what the nation thinks, by treating each opinion as a data point.
Access to information shapes attitudes to justice
A recent crime survey by the MoJ found that “Access to information, whether reflected in levels of education or in newspaper readership, appeared to shape attitudes to justice.” There was a distinct difference in a person’s confidence in the fairness of the criminal justice system, depending on the newspapers they read. This suggests that the information people access can change their view of what is considered central to the functioning of modern society.
Social media forms offer raw opinion which cannot be captured in conventional news media. And if it is also the main medium we tap to get our news, surely this represents a new perspective on “access to information”, influencing our attitudes to justice. However, whilst social media benefits from the lack of an editorial filter, it also suffers from it; there is a lot of noise to wade through. It gives rise to a number of other confounding factors as well – many users on social media often don’t notice the brand or source of the news they read. This finding perhaps explains the proliferation of fake news articles, making it harder to draw actionable conclusions from the response online.
Analysing public reactions to convictions and sentencing
Taking guidance from the MoJ, we explored the data behind the information and opinion in news and social media surrounding two criminal cases in the UK.
The first was a child abuse case in Rotherham, South Yorkshire, which garnered a lot of news coverage over 2016 as well as a significant reaction from the public on Twitter to the conviction of the perpetrators. By using natural language processing, we were able to show how the story was presented in the news – the volume of articles, the geography of coverage, the topics mentioned, and the sentiment associated with various legal actors, such as the local police force and crown court. Following news of the conviction, we also analysed the activity on Twitter, which brought up a notably different view; tweets covered a range of topics not found in the news, providing important insight into popular opinion.
The second criminal case we analysed, a double murder in Spalding, Lincolnshire, made headlines when the perpetrators were sentenced. It did not elicit a lot of news coverage then, but it did attract strong opinions on Twitter from those who thought the sentence of 20 years to be too short for the crime committed. Twelve days later the case referred to the Attorney General for being “unduly lenient”. During this period, we were unable to find any major news publications covering the case. Analysing social media here provided invaluable knowledge that would not have otherwise been available.
Data-driven, evidence-based policy
Our analysis was presented at an awayday for the MoJ policy team at Imperial College London’s Data Observatory, a 313 degree immersive environment where the attendees were able to get up-close to the data and scrutinise what was most interesting to them. We aimed to give an insight into the opportunity for government to leverage data science, and provoke thoughts about the role of news and social media to inform policy design. Although our analysis focused on two specific cases, the same approach could be taken towards any number of cases in any legal jurisdiction around the world that are identified to be significant; whether on specific legal topics that are subject to particular scrutiny or to make comparisons between a range of similar cases.
While we may not entrust the future of criminal justice to the social media hivemind, using these data-driven techniques can help ensure that the policies developed serve the people they are designed for.
Thomson Reuters has been partnering and adding value to the MoJ in a number of different ways. Most recently, an on-site team was deployed to integrate C-Track case management software in the Royal Courts of Justice, vastly increasing the volume and quality of digital information being captured.
Download our white paper on The Future of Courts.