Crime statistics – what’s next?

ONS has been working on a “step change” in its presentation of the crime statistics for England and Wales.  As Iain Bell explains, the new approach emphasises clear explanation of the latest trends using the best possible source of information on each type of crime.         

In November, I blogged about the trends in overall crime and the need for ONS to make a step change in our presentation of crime statistics. Later this week, we will present the latest crime statistics. In advance of this, I am setting out the approach we will be taking and the role of different sources in presenting a clearer picture of what’s happening to crime. This is based on using the best sources of information currently available to us to show what is happening to different crime types in the short and long-term and our best overall assessment of recent changes in the levels of crime.

In presenting crime statistics, we go beyond what is reported to the police. Most offences – about 60% – are not reported to the authorities. We also cover a wide range of circumstances and from crime against citizens and households, to those against businesses.  Incidents range from quite rare but high-harm crimes such as homicide and rape to more common lower harm crimes (such as minor criminal damage or petty theft).  This means no one source will ever give a complete picture of all circumstances.

So, our first question must be “Is it meaningful and helpful to measure change in total levels of crime?”

To answer this, it is worth thinking of a different set of statistics. In the world of health, nobody would ask “What is the total change in levels of disease?”. We recognise that there are a range of diseases with different impacts on people’s lives. We ask questions like “are we living longer?”, “are we healthier than before?” and “which types of disease are becoming more and less common?”

Levels of impact

Crime in many ways is similar. There are many types of crime with different levels of impact. Our role as ONS is to help answer questions like:

  • What types of crimes are increasing and what types are decreasing?
  • What is our overall likelihood of being a victim of crime?
  • What is the best overall assessment of changes in crime in England and Wales?

It is this thinking that will inform how we present our statistics later this week.

We will be placing the emphasis on clear presentation of the latest trends using the best possible source of information on each type of crime. We will also provide an overall assessment of crime trends. We use two main sources of crime statistics, each with different strengths and weaknesses to present this.

The Crime Survey for England and Wales is very good at measuring long-term trends and higher volume crimes. It is less good at measuring lower volume crimes – particularly in the short-term.

While the Crime Survey is our preferred source for many types of crime, in some circumstances and for specific offence types, we choose to supplement the survey figures with data from our second main source, police recorded crime. So, what do the recorded crime data tell us that the survey can’t? The police data cover some types of crime, such as homicide, which cannot be covered by the Crime Survey. Recorded crime can also be good for identifying changes in low-volume higher-harm crime types such as gun and knife crime which can’t be reliably measured by the survey because they occur relatively infrequently. The police data can also be a better source for identifying emerging trends in crime as the lag time in the data is much shorter.

But let me be clear. Changes in the total volume of police recorded crime are not a measure of changes in overall crime. For many types of crime police recording is not a reliable measure of change in volume of crime.

Changes in the way the police record crime, related to a tightening of recording standards, have impacted heavily on trends in lower level violent crime. Some higher-harm crimes including sexual offences have also been affected in this way. Police figures on sexual offences and domestic abuse offences are also thought to have been influenced by increases in victims’ willingness to come forward and report the offence. Policing priorities can also be important; for example, recent reductions in the number of recorded drug offences is more likely to reflect changes in priorities than falling levels of use.

This is not to say that police recorded crime is not a useful statistic. They give some important insights into crime trends in some areas. They also provide a good picture of crime-related workload being handled by the police. They also help indicate changes in the volume and nature of cases the criminal justice system will be handling further downstream.

As well as using these two main sources of crime data we’ll also continue to supplement these wherever we can with other sources. We already use data from hospitals to help in understanding violent crime trends, and industry data on bank account and card fraud. We’ll continue to look for further opportunities to build the best picture of changes in across the range of crime types.

Iain Bell is Deputy National Statistician for Population and Public Policy