Written evidence submitted by the Centre for Public Data (OUS0019)

1. About us

The Centre for Public Data (CFPD) is a non-partisan organisation with a mission to strengthen the UK’s public data [1]. We aim to reduce gaps in data that harm civil society and business. We support legislators and policymakers to improve data coverage and quality, via practical interventions in legislation, codes of practice and governance. We are happy to discuss any of these issues further: contact@centreforpublicdata.org

2. Our evidence

Our evidence focuses on the Committee’s questions “How does the public access information on sentencing?” and “What could be done to improve public understanding of sentencing?”[2].
As background to the Committee’s call for evidence, the Sentencing Academy’s 2022 report, Public Knowledge of Sentencing Practice and Trends, found that the public’s knowledge of certain aspects of sentencing is poor [3]. It is also often argued that poor public understanding of sentencing can affect public confidence in the wider criminal justice system [4].

We suggest that the official sources of data for sentencing in England and Wales could be strengthened to improve the public understanding of sentencing. Our response makes the following recommendations:

The Ministry of Justice (MoJ) should take steps to make its sentencing statistics more clearly signposted, documented and promoted, so that the public is aware of the data that is available.
In addition, the MoJ should extend its public statistical releases to allow breakdowns of average sentencing length at individual court level, by offence type and demographic factors. This would be of significant public interest, ensure that the Government meets its commitments to act on the recommendations of the Lammy Review, and meet the Sentencing Council’s statutory duty to publish sentencing data at local area level.
In addition, MoJ sentencing data available via the Data First programme should be extended to include anonymised judge identifiers. This would be of significant research interest.

Background: available data

Publicly available data and statistics on sentencing

The Ministry of Justice (MoJ) publishes detailed sentencing data for England & Wales in its annual statistical reports. These are based on the Court Proceedings Database (CPD). The online Crown Court data tool [5] and Magistrates’ court data tool [6] are published in Excel pivot table form, and can be used to break down sentencing outcomes and length by each of the the following characteristics:

Type of offence (detailed offence code)
Guilty / not guilty plea (Crown Court data only).
Type of defendant (individual or company)
Location (43 Police Force Areas)
Demographics of individual defendents (sex, age and ethnicity)

Since 2008, the MoJ has also published sentencing statistics and data tools specifically for knife and offensive weapon offences in England and Wales [7], which additionally allow the public to break down by age range and sentencing outcomes by the number of previous convictions the defendant received for knife-related offences.

The MoJ also publishes quarterly data on sentencing in its regular statistical release, the ‘Criminal Justice Statistics Quarterly’ (CJSQ)[8], though this is so high-level as to be largely meaningless (e.g. the average custodial sentence in months, across all types of offence).

Other statistical publications appear to draw on the MoJ’s annual reports, e.g. the GOV.UK ‘Ethnicity Facts and Figures’ service provides some high-level statistics on average sentence length by offence type and ethnicity.[9]

The Sentencing Council, an independent non-departmental public body created in 2010 to promote greater transparency and consistency in sentencing, publishes a range of high-level statistical bulletins on offences, including number of convictions and average sentence length [10]. These do not appear to provide more granular information than is available in the MoJ’s public statistics. Again, the Court Proceedings Database (CPD), maintained by the Ministry of Justice, is the data source for its bulletins.

Information available to researchers

Since 2021, the MoJ’s secure ‘Data First’ programme has made case-level data on sentencing available to accredited researchers for pre-approved use [11]. This dataset links together data from multiple sources, including court, prison and probation data, and includes variables about the defendant, case, and outcomes, including the type and length of sentence. Researchers are currently using this data to study, for example, ethnic disparities in sentencing in the magistrates' and Crown Court [12].

Previously available data and statistics on sentencing

From 2011-15, the Sentencing Council commissioned the Crown Court Sentencing Survey (CCSS) to inform sentencing guidelines by the Council. Sentencers in Crown Courts were asked to complete a survey for each sentenced case, including questions on aggravating and mitigating factors, and completion rates were approximately 60%[13]. The scope of the survey was welcomed by researchers [14] as it provided “unprecedented detail” to inform “more accurate and informative empirical research on sentencing”[15]. Aggregate data from 2011-15 is still available online [16].

Recommendation 1: MoJ sentencing data should be better signposted

The MoJ’s data tools (paragraph 5 above) are a rich source of information about sentencing. These powerful data tools make it possible for the public to ask detailed questions such as “What was the average custodial sentence length received for a particular offence?”, or “What percentage of defendants pleading guilty and not guilty received a custodial sentence?”, and view the trends broken down by age, sex and ethnicity, for different police force areas across England & Wales, between 2017 and 2021.

By international standards, the data made available by the MoJ is relatively detailed (although as noted below it lacks one important element). An international comparison by Gormley et al. found that: “the MoJ data are more comprehensive than those found in most other jurisdictions”[17].

The question then arises why the information is not being more widely used to inform public understanding. We suggest that this is partly because the data tools could be made easier to use. As Gormley et al. comment, the tools “are rather inaccessible to wider readers looking to gain a good understanding of current sentencing trends”.

An easy and low-cost improvement would be for this data to be more clearly signposted and made easier to use, for the public and journalists. We recommend the following:

The MoJ’s statistical releases should include some published series based on the sentencing tools (preferably those series of high interest to the public, such as average sentence length for serious sexual offences), and these should be discussed (with appropriate context, following the guidance in the UK’s Code of Practice for Statistics) in the commentary published alongside the annual releases. These will be more likely to be reported upon by journalists, and will be more accessible to members of the public who lack Excel knowledge.
The MoJ should invest time in outreach - i.e. writing blog posts about using the data tools, and trends in the sentencing data.
The data tools should be published with a more search-engine-friendly name, preferably on their own web page, so they are easier for members of the public to find.
The Excel pivot tables are only currently available to Windows users, and should be made available for Mac and other operating systems.

Recommendation 2: MoJ should allow sentencing statistics to be broken down at court level

While these MoJ statistics allow the public to answer some questions about sentencing, it is not possible to use them to understand a question of significant public interest: whether disparities in sentencing exist between particular courts. This was identified as a need (and accepted by the Government) several years ago following the Lammy Review.

2a. The Government should implement the Lammy Review’s recommendations on sentencing data, as it committed to do

In 2017, the Lammy Review, an independent review into the treatment of and outcomes for Black, Asian and Minority Ethnic individuals in the criminal justice system, recommended that the public should be able to access sentencing information broken down by demographic characteristics including gender and ethnicity for individual offences at individual courts (our emphasis)[18].

The Lammy Review explained that making this information accessible would help identify whether there are areas of the country where BAME defendants faced disparities in sentencing outcomes compared with White defendants for similar offences at the same court. By making this information publicly available, sentencing decisions would be placed under greater public scrutiny and sentencing inconsistencies would be identifiable at an appropriate level of analysis.

The Government accepted this recommendation [19] and stated in an official response: “[P]roduced and updated annually in May [2020]… were the sentencing and offence tools which break ethnicity data down by demographic characteristics in response to the Lammy recommendation”[20]. The Government also told the House of Commons in 2020 that actions relating to the recommendation have been ‘completed’[21].

However, it does not appear the Lammy Review’s recommendation has been met. The sentencing tools allow breakdowns by Police Force Area, but not at individual court level.

We have corresponded with the Ministry of Justice in an attempt to clarify whether court-level data have been published, but did not receive an answer.

Without court-level data, it is not possible to scrutinise the sentencing practices of particular courts, and identify local disparities, as the Lammy Review recommended.

2b. Publishing court-level data would also fulfil the statutory duties of the Sentencing Council

The Sentencing Council of England and Wales has a statutory duty under section 129(1) of the Coroners and Justice Act 2009 [22] to publish statistics on sentencing patterns from the Magistrates’ and Crown Courts in local justice areas across the country. Yet concerns have been raised that the Council is failing to meet this statutory obligation. Gormley et al. have argued that although the Council has been successful in its function of developing sentencing guidelines, it has made little progress in collecting and publishing sentencing data at the local level [23].

Likewise, section 129(2) of the Coroners and Justice Act 2009 assigns the Council powers to “promote awareness of matters” relating to sentences imposed by courts in England and Wales [24]. Leading academics have expressed the opinion that the Council will need to take on a more proactive role in developing and distributing sentencing statistics to satisfy this function [25].

Accordingly, we recommend that the MoJ should begin publishing sentencing information at the individual court level (for both Magistrates’ courts and the Crown Court), including breakdowns by age, sex and ethnicity. This would improve public understanding of sentencing, fulfil the statutory duties of the Sentencing Council, and meet the Government’s commitment to implement Recommendation 12 of the Lammy Review. In practice, individual courts could straightforwardly be included as an extra filter option within the MoJ’s existing sentencing data tools.

Recommendation 3: MoJ sentencing data should include information on judges

A significant amount of research exists to suggest that the demographics and characteristics of individual judges have an impact on sentencing outcomes, and this may be the factor that drives discrepancies at court level [26] [27] [28] [29]. In the Crown Court in England and Wales, research from Pina-Sánchez et al. found that sentencing inconsistencies were overall more affected by the characteristics of the judge presiding over the case than the location of the case [30].

Pina-Sánchez et al. identified the need to better understand the impact that the characteristics of judges have on sentences. They concluded that progress in sentencing research has been handicapped by data gaps and that the lack of identification of judges in statistics has “limited the study of sentencing disparities in England and Wales”[31].

Similar calls for better data collection on judges have been made outside of England and Wales. In the United States, where some judges issue sentences twice as severe as other judges in the same court [32], senior judges have called for sentencing data to incorporate judges’ characteristics and their sentencing decisions [33].

Currently, neither the MoJ’s public statistics nor (it appears) the data available to accredited researchers via the Data First programme provide information about the characteristics of the judges or magistrates making sentencing decisions.

We recommend that the MoJ should start to make anonymised judge identifiers and information on characteristics available to accredited researchers in its Data First programme. The information provided on judges’ characteristics could include: sex, age, time on the bench, whether they rotate across courts, and type of judge (e.g High Court judge, circuit judge, magistrate). We make this recommendation because of the high level of research interest in this issue, and its evident importance to improving the public understanding of sentencing disparities.

August 2022