Crime stats: counting something that doesn’t want to be counted

Written by Oz Flanagan on . Posted in Features

Crime is notoriously hard to quantify. Whether through cover-up by the perpetrator or the reluctance of the victim to report the crime, the true picture of criminality is obscured by the very nature of the act. Gathering accurate data on crime is at an immediate disadvantage as soon as the act is committed. Yet law and order is one of the foundations of any society and having statistics on the level, nature and evolution of crime is crucial for the state to be able to function.

The UK has some of the best crime statistics in the world, but because of its significance in describing the performance of one of the government’s basic duties, scrutiny is inevitable. But criticism of official crime statistics has been building for some time now as the status quo has been called into question in the UK, and across the Irish Sea.

Crime statistics are split into two sources. The first is the crime survey that is carried out in England and Wales (and separately in Scotland) which records the public’s experience of crime. The second source is crime as recorded by police forces, and this is the source that has come in for intense scrutiny.

The trouble with police recorded crime statistics

Every police force in the UK keeps records of the incidents they are called to. When the police have investigated an incident, they then decide if a crime has taken place and apply a crime code to record the event. This is the raw data that is delivered to the ONS, but the recording and categorisation of crime is done by the individual police forces.

Last year, a parliamentary select committee heard evidence of the systemic under recording of crime because of ‘lax compliance with the agreed national standard of victim-focussed crime recording.’ That evidence prompted the UKSA to strip police recorded crime of its ‘national statistic’ status.

The select committee also recommended that ‘senior police leaders must emphasise data integrity and accuracy, not targets.’ This is now starting to be echoed by senior police officers, with the commissioner of City of London police Adrian Leppard, recently saying police figures provided only a 'small prism of the harm' and that police recorded crime is 'dead in the water'.

The reason why the police have faced this criticism lies in the process whereby an incident is revised once it is dealt with. When police officers on the ground deal with a call, the incident is logged as the type of crime reported to the police. However, there may be genuine reasons why this would have to be revised following police inquires.

For example, police could be called to a burglary in progress. But after investigation, they learn that the ‘burglar’, is simply a family friend gaining access to a property with permission. This incident would need to be revised so that the data did not reflect it as a crime.

Another example with burglary could entail an intruder attempting gain access to a third floor window via scaffolding. If the intruder failed to gain entry but damaged the window, this should be recorded as an attempted burglary. However, this could be downgraded to the lesser offence of criminal damage later in the process. Except this time, the change is not for the integrity and accuracy of the data, but to reduce the level of more serious crimes recorded by that police force.

{mbox:opinion/2015/flagged-domestic-violence.png|width=400|height=311|caption=Click to enlarge|title=Source: HM Inspectorate of Constabulary (2014), based on forces own definitions of domestic abuse and calls for assistance and the use of a domestic abuse marker on IT systems.}

The variation in methods of recording crime across the police forces is a major weakness in the data. For instance, domestic violence does not have its own specific code in recorded crime. To gauge its prevalence, a new process was initiated whereby incidents that police found to be domestic violence were to be ‘flagged’ as such. But this effort also showed the imbalance in recording practices across the police forces of England and Wales, demonstrated in the chart on the right.

These discrepancies between police forces are nothing new. In 2002, the national crime recording standard was introduced to try and harmonise data collection across police forces. This was after a report ‘found a recording rate that varied between 55% and 82% across the forces.’ The new rules sought to better record crimes, even if the victims account lacked supporting evidence. The effect was an uptick in crime and violence in particular, but these standards are the same ones that the select committee has now found ‘lax compliance’ with.

The ‘target culture’ of the New Labour era has left a legacy of creative accounting that is proving difficult to extinguish. The US crime drama The Wire dramatized this culture in US law enforcement a decade ago - one scene in particular brilliantly portrays the discussion of statistics among Baltimore's fictional senior police officers.

The home secretary, Theresa May said she would end the target culture as far back as 2010. Yet she still cites police recorded crime statistics as evidence of her department’s accomplishments, despite the well-known shortcomings of these statistics. Old habits die hard and falling police recorded crime, whether a national statistic or not, is still a handy crutch to use.

A problem that stretches north of the border and across the Irish Sea

The unreliability of the police recorded crime is now a problem that is manifesting itself further afield than just England and Wales. In Scotland, the Holyrood government is reported to be seeking to wash their hands of the statistics by transferring responsibility to Police Scotland. This comes after Scotland’s previous eight police forces (and their data systems) were consolidated into one single force.

Meanwhile, in the Republic of Ireland, the Central Statistics Office (CSO) suspended the production of crime statistics last year after an independent report found similar systemic problems in the Irish police force, An Garda Síochana.

Quarterly crime statistics have just been restarted in Ireland, but the CSO has conducted a review that found 18% of crimes reported to An Garda Síochana in 2011, never appeared on the computer system they use as the primary data source. The CSO’s methodology has not changed since the suspension, but the latest statistical release shows assault and burglary have increased by 8% in the period between March 2014 and March 2015. A difficult question is now posed for users of these statistics – have burglary and assault been increasing, or have An Garda Síochana tightened up how they record crime since the damning report last year?

This presents Irish statistics with the same problems faced by UK statistics following the introduction of the national crime recording standard in 2002. The changes were meant to improve the data, but played havoc with the time series for anyone looking at crime trends either side of the new rules.

Surveys - experience of crime is the only alternative

With reduced confidence in the statistics recorded by the police, it is left to the crime surveys to provide the best estimates. Unfortunately, the Republic of Ireland only conducts crime and victimisation surveys on an occasional basis, so they are forced to rely on police recorded crime for regular statistics.

The UK is fortunate enough to have an annual crime survey dating back to 1982 – now called the Crime Survey for England and Wales (CSEW). Its crucial difference to recorded crime statistics is that it gives us an indication of the public’s experience of crime.

It also retains its ‘national statistic’ status, making it (and its equivalent in Scotland and Northern Ireland) the most robust measure for overall crime. But it still has drawbacks as some crimes are harder to survey than others. Car theft is easy to track because people will easily recall such a big crime in their lives and it will also have to be reported to the police for insurance purposes. On the other hand, sexual assault is much harder to measure because of the reluctance of victims to reveal such incidents in a survey context.

The survey is conducted in face to face interviews along with a self-completion section, designed to encourage more truthful responses. When it comes to sexual assault, the response to these questions is so low that it is excluded from the surveys main estimates. An overview of sexual offending by the Ministry of Justice, Home Office and the ONS in 2013 found that, of the surveyed females who said they were victims of the  most serious sexual offences, only 15% said they reported it to the police.

Domestic violence is another crime that is difficult to measure. Criminologist Sylvia Walby, recently found that the ONS are placing a cap on the number of these incidents. This meant that ‘when the survey’s cap is removed, and the raw data examined, the number of violent crimes increases by 60%. The amount of violent crimes against women, and the amount of violent crimes by domestic perpetrators, both increase by 70%.’

A lack of detail

The main shortcomings of crime statistics come when you try and drill down to find out about specific acts of crime. In her work looking at domestic violence, Sylvia was unable to make use of police recorded violent crime because it is not broken down any further than whether an injury was inflicted or not.

She says she would ‘like to see all violent crime (and also sexual offences and threats) disaggregated by the sex of the victim and the relationship between victim and perpetrator.  The police collect this information as part of their investigations, so this isn’t an additional data collection exercise.’

How specific crimes are changing is important because criminals are innovating with technology. Bank robbery is now too risky compared with the illicit gains of fraud, but quantifying it is fraught with difficulty. To begin with the victim can be an individual, a small business or a global multi-national. So a fraudulent act will often never be reported to police and instead be written off as a bad debt. The latest statistics show a 17% rise in fraud year on year, but the ONS admit that it's difficult to judge to what extent this reflects 'an improvement in recording practices, an increase in public reports or a rise in actual criminality.’

Crime statistics only record fraud under one category, even though all reporting of fraud is now directed to a central agency called Action Fraud. This change took place in 2011 because individual police forces often did not have the resources available to investigate fraud allegations. The only statistics that are broken down are submitted by industry bodies such Cifas and Financial Fraud Action UK. But this is a snapshot of the type and volume of fraud that exists, and makes it difficult to make comparisons between credit card fraud versus internet scams.

Crime has consistently fallen for the past 20 years, but beyond this headline figure it’s very difficult to see how crime has changed. The internet, along with globalisation, has revolutionised the way fraud can be committed. But how prevalent it is and how much it generates for organised crime is almost impossible to accurately quantify, especially across borders.

This is a microcosm of the trouble with trying to count something that doesn't want to be counted. It’s worth repeating that the UK has some of the best crime statistics in the world, yet this task will continue to cause headaches for official statisticians for a long time to come.


The views expressed in the Opinion section of StatsLife are solely those of the original authors and other contributors. These views and opinions do not necessarily represent those of The Royal Statistical Society.

Official Statistics Public Administration Select Committee (PASC)

Join the RSS

Join the RSS

Become part of an organisation which works to advance statistics and support statisticians

Copyright 2019 Royal Statistical Society. All Rights Reserved.
12 Errol Street, London, EC1Y 8LX. UK registered charity in England and Wales. No.306096

Twitter Facebook YouTube RSS feed RSS feed RSS newsletter

We use cookies to understand how you use our site and to improve your experience. By continuing to use our site, you accept our use of cookies and Terms of Use.