Written Evidence from the UK Statistics Authority [CST0017]
Thank you for your invitation to submit evidence to the Committee's inquiry into crime statistics.
While in some cases the police service or Her Majesty’s Inspectorate of Constabulary (HMIC) may be best placed to respond to the Committee’s specific questions, I have set out below some general observations about particular themes raised in your letter. I also enclose a detailed note prepared by the Office for National Statistics (ONS), as the body responsible for the publication of the statistics, at Annex A.
The UK Statistics Authority has conducted a range of work on crime statistics. In 2010, the Authority published a monitoring review Overcoming Barriers to Trust in Crime Statistics, and the National Statistician subsequently undertook an independent review of crime statistics in 2011. This work culminated in the decision to transfer responsibility for the compilation and publication of crime statistics from the Home Office to the ONS in April 2012. The National Statistician’s review also identified the need for a transparent, independent advisory committee to consider any change in definitions, classifications or methodologies that may affect the statistics. The Crime Statistics Advisory Committee was established in late 2011 and has taken an active role, advising the Home Office over matters related to the collection of police recorded crime, the ONS over changes to the presentation of crime statistics, and HMIC over the inspection of police recording practices.
It is the Authority’s view that these have been important steps in improving the integrity and independence of the statistics; prerequisites for building public trust in them. We do, however, note that there is scope for further improvement.
While much of the evidence is anecdotal, the Authority considers that possible inconsistencies in crime recording by the police and the potential distorting effect of performance targets are a matter for concern. Trust in the integrity of these statistics requires users to have confidence that consistent decisions about crime recording are being made across forces, and over time.
The current lack of periodic independent audit of crime recording by the police is similarly a matter for concern. We, therefore, welcome the intention of the HMIC to carry out a national inspection of the integrity of crime recording starting in late 2013. We believe such inspections should become a permanent feature of the system.
Finally, the Committee asks whether recorded crime statistics should be classified as National Statistics. The Authority originally assessed these statistics when they were the responsibility of the Home Office, and they were designated as National Statistics in 2012. They are now undergoing reassessment. The Authority holds the view that – given the national importance of police recorded crime statistics – these statistics should be National Statistics. Compliance with the Code of Practice will ensure that the statistics meet the needs of users, and are produced, managed and disseminated to high standards.
ANNEX: NOTE PREPARED BY THE OFFICE FOR NATIONAL STATISTICS
The Home Office has been collecting information from the police on the number of police recorded crimes since the late 19th century. This series has been based on aggregate counts of the number of specific offences contained within the notifiable offence list (NOL). This list includes all indictable crimes (i.e. those tried at Crown Court) and all those that could be tried either at a crown court or a magistrates’ court. The police also record other crimes, or crime-related incidents, which are outside the scope of the main recorded crime collection.
To supplement the police recorded crime series, the Home Office initiated a separate survey – known then as the British Crime Survey (BCS) – which measured the extent of crime by asking people whether they had experienced crime. The survey aimed to assess how much crime went unreported to the police and to capture richer data (than available from the aggregate police recorded crime series) on the characteristics and experiences of victims of crime. The first BCS ran in 1982 and was repeated periodically through the 1980s and 1990s before the Home Office decided to establish the survey on a continuous basis in 2001. While Home Office analysts were responsible for the data analysis and publication of research and statistics coming out of the BCS, the data collection was contracted out to an independent survey research organisation.
Following a monitoring report published by the UK Statistics Authority in 2010, the Home Secretary asked the National Statistician to undertake an independent review of crime statistics. The terms of reference for this review included considering which independent body should assume future responsibility for crime statistics and to consider the data collection arrangements that should underpin any new arrangements.
The National Statistician recommended that responsibility for the publication of the statistics should be transferred to the independent Office for National Statistics (ONS), to demonstrate that the statistics were free from political interference. The National Statistician recommended that responsibility for the whole life cycle of the BCS, and funding, should also be seen to be independent of Ministers and that this should also transfer to ONS. However, it was recognised that, given the existing relationships between the Home Office and the police service, there were benefits to the Home Office in retaining responsibility for the collection and validation of police recorded crime data. This would also avoid an increase in bureaucracy in the police supply of data to both ONS (for crime statistics purposes) and to the Home Office (for other purposes).
The National Statistician’s recommendations were accepted by the Government and the new arrangements took effect on 1 April 2012. The ONS now receives police recorded crime data from the Home Office Statistics Unit on a quarterly basis, and analyses these data alongside Crime Survey data which is collected independently by the ONS.
There are currently no international standards for crime recording by the police. Recording practice in England and Wales is governed by the Home Office Counting Rules (HOCR) and the National Crime Recording Standard (NCRS). These rules set out whether, how, and when to record a crime. They are designed to provide a consistent approach across forces, and over time, in the recording and classifying of offences.
The HOCR requires that incidents reported to the police should be initially registered, irrespective of whether the report comes from victims, witnesses or third parties, and whether or not it relates to a crime. The NCRS was introduced in 2002 to provide a more victim-focused approach to crime recording and followed a report by Her Majesty’s Inspectorate of Constabulary (HMIC) which highlighted inconsistent recording practices across forces influenced by local police culture. The NCRS requires that an incident will be recorded as a crime if, on the balance of probability, the circumstances as reported amount to a crime defined by law and there is no credible evidence to the contrary. Once recorded, a crime should remain recorded unless there is additional verifiable information to show that a crime did not occur.
Alongside the adoption of the NCRS, forces were encouraged to undertake internal audits of crime records to ensure compliance with the new standards. This led to the creation of the role of a force crime registrar whose role it is to quality assure data and be the final arbiter with respect to disputes about the local interpretation of the HOCR. Force crime registrars are coordinated by a national crime registrar based at the Home Office, and the guidance with which they are provided includes a data quality audit manual.
On receipt of the data, the Home Office Statistics Unit checks the data for obvious errors or omissions and makes enquiries with forces in respect of any outliers or unusual trends. Forces may then revise and re-submit the data.
Regular independent audits of police data quality, carried out by the Audit Commission, took place between 2003/04 and 2006/07. There seems to be broad agreement that such inspections contributed to improvements in police crime recording. We understand that these audits ceased because the standard of crime recording significantly improved over the period. Since the cessation of such national audits, the HMIC has undertaken ad hoc inspections on specific aspects of crime recording by forces, but there have not been regular independent audits of crime recording in recent years.
Are crimes being recorded by the police when they should be? Are crimes being categorised correctly?
It is the view of ONS that these questions can be best answered through independent audits of crime recording.
However, an ONS analysis, published in January 2013[1], suggested there has been a reduction in the consistency of crime recording in recent years. The analysis examined trends at a national level in crimes being recorded by the police, against those estimated by the Crime Survey for England and Wales, exploring whether or not there was evidence of the two series diverging.
The analysis suggested that, prior to the introduction of the NCRS, a relatively large volume of reports made by victims to the police may not have appeared in the police recorded crime series, with as much as a 50 per cent shortfall each year. A step change was seen following the introduction of the NCRS in 2002, with the volume of comparable police recorded crime close to 90 per cent of the total estimated from the survey. This relationship remained fairly consistent up until 2007/08. A consistent downward trend in the ratio between the two series was seen up to 2011/12 with the latest figures (for 2012/13) showing no change. The police recorded crime series has thus been reducing at a faster rate between 2006/07 and 2011/12 than suggested by the survey which might suggest a gradual erosion of compliance with the NCRS (see Figure 1).
Figure 1: Ratio between CSEW reported incidents and crimes recorded by the police (in comparable sub-set)
ONS notes that the analysis does not definitively prove that there is mis-recording of crime. Nor is it able to examine the extent to which crimes, once recorded, are being categorised correctly. ONS therefore welcomes the forthcoming HMIC inspection, which will address these issues.
What are the factors which can influence police mis-recording of crime?
It is ONS’s view that there are a range of possible factors which may influence mis-recording of crime, including:
- the distorting effect of performance targets;
- a lack of awareness or training in recording standards, or confusion between recording and charging standards; and
- moves towards Neighbourhood Policing resulting in more crimes being dealt with informally and outside the crime recording system.
Are the right checks in place to ensure that the systems for recording crime function effectively and accurately? Has enough been done to ensure the integrity of crime data? What more should be done?
It is the view of ONS that force crime registrars play an important role in ensuring the integrity of data, appointed to act as a “final arbiter for the audit process, the interpretation of the counting rules, and assigning outcomes”.[2] In a review of crime and incident reports, published in 2012, HMIC identified “an objective force crime registrar with strong influence over local crime recording decisions to ensure performance pressures are not unduly exerted” as one of a number of factors which contributed to forces consistently making good crime and incident recording decisions.[3] At the same time, HMIC identified that in some forces the role of force crime registrar may not be sufficiently independent of those charged with delivering improvements in performance. Therefore, the role of the force crime registrar is an important one and there is a case for their role to be strengthened.
The process by which data supplied by police forces to the Home Office is checked and validated has been described above. The ONS undertakes further quality checks of data supplied to them by the Home Office for use in the National Statistics. These include checks to examine the internal consistency of the data and with previously supplied data. ONS also has representation on key Home Office working and strategic groups dealing with the police recorded crime data. This ensures the statisticians keep in touch with emerging issues and that the statistical implications of any changes to the series are properly considered.
However, without regular independent audit, it is difficult to assess whether or not crime recording systems are functioning effectively and accurately. The ONS would like to see the forthcoming inspection carried out by HMIC repeated annually so that users of police recorded crime statistics can be informed about their accuracy.
To what extent can policy-makers have confidence in the statistics which result from the recording of crime by police forces?
The answer to this question very much depends on what policy-makers (and other users) need the statistics for, and the extent to which recorded crime figures represent a good measure of the issue under consideration. ONS analysis mentioned above showed that the downward trend in levels of recorded crime over the last decade was validated by the independent Crime Survey, and therefore that users can have confidence in the broad trends.
The consistency of recording is likely to vary by type of crime. For example, it is thought that the vast majority of homicides come to the attention of the police and are recorded accurately. On the other hand, it is known that steps have been taken in recent years to encourage more victims of sexual offences to report crimes to the police and to improve the police handling of such matters. These are likely to have led to more of such offences being recorded.
The statistical commentary provided by ONS aims to provide users with an indication of the quality of the statistics and use is made of additional sources (for example, admissions to hospital in relation to violent assaults) to verify the broad trends shown in the Crime Survey and police recorded crime data.
November 2013