In 2021, 3.5 million prisoners were being held in pre-trial detention. Almost half of all reporting countries are operating prisons at over 100 per cent capacity. Meanwhile, concerns continue to mount over police accountability and behavior in many countries.
It is clear that, as noted by Ghada Waly, Executive Director of the UN Office on Drugs and Crime (UNODC), action “to strengthen the rule of law, make justice accessible to all, and build effective, accountable and inclusive institutions around the world” is needed.
But without strong data collection and management systems, achieving effective and humane criminal justice systems – one of the promises of Sustainable Development Goal 16 (SDG 16) – will remain out of reach.
To assist states to create such systems, UNODC has developed four sets of Guidelines for the Production of Statistical Data (on police, prosecution/courts, prisons, and governance of data).
UNODC sat down with Martijn Kind, lead statistician for the guidelines, on the importance of data in delivering and accessing justice.
The guidelines series provides guidance to Member States on the collection, production, dissemination and use of high-quality crime and criminal justice data for statistical purposes (ICCS).
Administrative data are collected in the day-to-day operations of all criminal justice institutions. These data offer a potential wealth of information that provide unique in-depth insights into the operations of the criminal justice system and can be used to improve resource use, enhance operational effectiveness, promote a culture of accountability and, ultimately, improve access to justice for all.
For example, accurately registering data on prison incidents can help authorities pinpoint the types of incidents that are occurring and aid the development of policies to prevent future incidents, ensuring the safety of both prisoners and staff members.
The guidelines offer a standardized framework that enables countries to collect, produce and disseminate high quality crime and criminal justice statistics that are relevant, accurate and more comparable. The guidelines go beyond just classifying offences – which is the purpose of the International Classification of Crime for Statistical Purposes (ICCS) – and encourage countries to produce statistics on a range of other topics, including human and financial resources, the criminal justice process and the professional conduct of staff.
The use of data is key in order to address issues in an informed manner. Evidence-informed policy can only be formulated if data are available. Without data and evidence, we cannot adequately identify problems, let alone begin to address them.
Decisionmakers are essentially flying blind if data are not available and could be forced to take decisions based on more limited, partial evidence or have to rely on their instinct.
Worse still, without data, there is no way of knowing whether the issue at hand is improving or gradually getting worse. For example, it is not possible to determine why there is a backlog of cases in the courts if no information on the number of pending cases is available.
Additionally, standardizing data is vital for the production of accurate and reliable statistics. If data are not standardized, they are also not directly comparable. This implies we would not be able to produce reliable statistics that can inform us on the phenomenon under investigation and we could end up comparing apples to oranges. Standardized data promote consistency and accuracy, reducing errors and inconsistencies that can arise by combining different sources of data.
In short, if we don’t have standardized data to produce statistics, we may misinterpret signals and draw faulty conclusions. This could ultimately lead to un- or misinformed policy decisions that may not address the issue at hand or, in the worst case, exacerbate the problem.
Standardizing data is also important to ensure data interoperability. This is particularly relevant in the criminal justice system where typically a wide range of actors are involved in the collection, production and dissemination of data. It is vital to ensure interoperability in such an environment to facilitate collaboration between stakeholders and the integration of data from various sources, leading to more comprehensive insights and more informed decision-making.
SDG 16 focuses on promoting peaceful societies, providing access to justice for all and building effective, accountable and inclusive institutions at all levels. The framework proposed in the guidelines offer key data for knowing whether these objectives are being achieved. In fact, one of the key objectives of the guidelines is to promote accountability and show the public how the criminal justice system is operating.
The guidelines also directly promote the collection of data that can be used for calculating several SDG indicators. For example, standardized data on intentional homicides that be disaggregated by a range of variables such as sex, age and the relationship between the victim and the offender are vital for the calculation of indicator 16.1.1 on the number of intentional homicide victims per 100,000. The guidelines for the prison system also encourage Member States to collect detailed data on the prison population that can be used to calculate indicator 16.3.2 on unsentenced detainees as a proportion to the overall prison population.
Cameroon is currently in the process of improving the governance of data related to intentional homicide. SDG indicator 16.1.1 poses numerous problems for the country due to the concepts used, the management of the different data sources and calculation method difficulties that stem from the lack of a common database for all homicide data.
The solution being pursued is to create a single national database on intentional homicide that integrates data from the two main sources: the national gendarmerie and the police. The new database should make it possible to expand data coverage, address double counting issues and, above all, clearly distinguish intentional from non-intentional homicides.
Once this project is successfully completed, authorities in Cameroon will have access to a single, harmonized dataset on intentional homicide that will give both law enforcement agencies and policymakers a better understanding of the reality of intentional homicide and the dynamics that are driving the trend.
Another example would be from the United Kingdom. Since 2017, the Home Office has required all police departments in the United Kingdom to record data on use of force by the police.
Officers must complete a “use of force report” each time they use force on an individual. A report should include the use of force tactics applied, reasons for using force, impact factors and the location and outcome of the event. The purpose of this is to improve information provided to the public on the different types of force used by the police and the context in which the use of force occurs.
These data also inform the work of the National Police Chiefs’ Council and College of Policing in enhancing tactics, training and equipment in order to improve the safety of officers and people with whom they come into contact.
Talking specifically about administrative data, there are a range of issues that can be at play in any national context. The main issues that are worth highlighting relate to a lack of data standardization and documentation, data access barriers and the use of legacy systems.
We talked about the importance of standardized data before. Which can be particularly challenging in the criminal justice system where each data provider may have their own system and processes in place for collecting data. Oftentimes, there is also a lack of documentation about how the data were produced and little to no metadata is recorded.
Data access barriers can add another layer of complexity when trying to manage administrative data sources. For example, performing record linkage and integrating data across multiple sources can be very challenging without having a solid data sharing agreement in place. Different agencies may also have varying policies and procedures regarding data access or there may be trust barriers that prevent institutions from sharing data. All of which can lead to the fragmentation of data across multiple agencies, making it that much more challenging to produce reliable crime and criminal justice statistics.
The guidelines offer a framework for improving the governance of data in the criminal justice system. The focus is on establishing cooperation between the different agencies, putting in place data sharing agreements, developing a national data strategy and improving the data architecture. All of which has to take place within existing legal and regulatory frameworks given the sensitivity of crime and criminal justice data. Building an interoperable system of criminal justice statistics promotes transparency and accountability within the criminal justice system, fostering public trust and offering a pathway to a fairer, more inclusive, efficient and effective system.
It is important to note that effectively collecting, analyzing and utilizing data is not only a technological challenge. Improving the governance of data is key and, crucially, investing in human capital. Without adequately trained staff with a high degree of data literacy, it will be difficult to produce and use the data and statistics generated.
Recent developments in artificial intelligence (AI) have produced a range of new applications in the criminal justice sector, including both simple digital automation and tools that rely on more advanced algorithms such as machine learning or natural language processing. These more advanced systems are used for a variety of tasks, such as biometric identification (e.g., facial recognition), predictive policing to identify hotspots and optimize the use of resources, risk assessments, and process optimization.
AI systems offer numerous benefits, such as the ability to rapidly process large amounts of information of different origins and formats, and to perform a wide variety of tasks, thus offering the potential to vastly improve both the efficiency and the quality of crime and criminal justice data analysis. AI systems can also aid justice sector staff by assisting in the execution of repetitive tasks and safeguarding their well-being by reducing their exposure to challenging material (e.g., child sexual abuse material).
However, the application of AI systems can involve a number of challenges and limitations. Chief among these is the risk of algorithmic bias which can replicate existing patterns of discrimination potentially reflected in historic data. Another consideration with the application of AI systems is the lack of transparency, since the processes leading to AI system outputs are typically difficult, if not impossible, to fully understand and explain (the “black box problem”).
For the guidelines, the key is that the quality of any analysis will always be dependent on the quality of the underlying data. If a machine learning algorithm is fed either poor quality or biased data, the resulting output will not be of much help and may even end up doing harm. The guidelines provide a tool to help countries collect, produce, disseminate and govern high-quality data on a range of themes. Only when this high-quality data is available are we at the starting point for effectively using AI or machine learning solutions.
The world is awash with more and more data and a key future development will be to make more sense of all of this data. This will require significant investments in technological solutions and – perhaps more importantly – investments in the governance of these data. This is exactly what the guidelines focus on: Improving both the quality and governance of the basic data that are needed to provide comprehensive insights and inform both practitioners and policymakers.
Read here the guidelines on police, prosecution/courts, prisons, and governance of data.
