Do crime statistics mask a growing problem with anti-social behaviour?
"The reason why people disagree with these figures is that they see in their personal lives that life is getting less safe and more disorderly.
"Many of the things they see don't get recorded. Travelling home on the train full of menacing lads doesn't get recorded in the crime figures, but it might not have happened 10 years ago."
Peter Hitchens, Today Programme, 24 May 2012
One of the reasons given by Daily Mail columnist Peter Hitchens for his distrust of crime statistics are the gaps that are left in the figures by incidents that don't make it onto police records.
Mr Hitchens denied that he was 'placing anecdote before evidence', so what evidence do we have about those incidents that go unrecorded on the crime figures?
Has Peter Hitchens got a point?
Yes, to a degree. He can certainly point to evidence that shows the public often seem to value anecdotal experiences of crime more highly than national crime statistics.
The UK Statistics Authority's review of public trust in crime figures found that "anti social behaviour is likely to affect public perceptions about the extent and nature of crime, and people's views on their personal risk."
It concludes: "The public tend to take a more pessimistic view about national trends than local."
This is borne out by the findings of the British Crime Survey (BCS), which has asked respondents whether they thought crime was rising or falling since 1996:
This has consistently shown that a majority of the public think that crime is rising nationwide, while the BCS itself has largely shown falling levels of crime over the same period.
Why do people think crime is rising when the statistics show otherwise?
Mr Hitchens' explanation for this is that people see more low-level crime and anti-social behaviour - his 'menacing lads' on the train - which leads them to feel less safe, but which doesn't feature in the national crime figures.
Again, there is some truth to this.
As the Statistics Authority's review notes:
"Most summary offences do not have to be notified to the Home Office and do not therefore enter into the recorded crime statistics. These tend to be dominated by motoring offences (in volume terms) but include a number of other categories such as drunkenness or social security offences."
The impact that the absence of this data has on public perceptions of crime isn't necessarily simple however.
In fact, the Association of Chief Police Officers has even suggested that it may inflate the risk of crime. It said in its 2006 Submission to the Statistics Commission:
"Due to … many of these offences being more minor crimes coupled with how they come to police notice, ACPO believe they do not contribute to deliberations on trends in crime as much as may be assumed and that they are distorting the overall performance picture and with it public perceptions of the true risk of crime."
However while these more minor offences don't feature in the national crime statistics, they do feature in the Home Office's crime mapping service, so it isn't absolutely true to say that they aren't recorded.
Can we measure anti-social behaviour?
Not very easily. Unlike crime, anti-social behaviour is only loosely defined.
As an HM Inspectorate of Constabulary (HMIC) report notes, anti-social behaviour is something of a "'plastic' concept that can be stretched and moulded according to the demands of a situation."
This means that there is no single understanding of what constitutes 'anti-social behaviour' that is shared by all, which in turn means that any attempt to measure the number of incidents would be hamstrung by many different interpretations of the term.
A separate HMIC report in 2010 did suggest that there could be as many as 14 million anti-social behaviour incidents every year, based on the assumption that nearly three quarters of them went unreported.
As Full Fact found at the time, this wasn't a figure that we could trust, as it used only a selection of types of incident collected in a way that was not necessarily comparable.
Is anti-social behaviour becoming more common?
Probably. HMIC has concluded that anti-social behaviour incidents "have grown and evolved in intensity and harm" over the past 20 years.
Furthermore it claims that the size of the problem might have been masked by "an increasing acceptance or "defining down" of ASB that we should not have grown used to."
HMIC itself concedes that "people do not make 'hard and fast' distinctions between crime and anti-social behaviour", and it is reasonable for Peter Hitchens to suggest that a perceived rise in anti-social behaviour could undermine the public's confidence in the lower levels of crime reported by the national crime figures.
However this in itself does not make the national crime figures unreliable. In fact, the UK Statistics Authority has noted that the UK data is better at capturing anti-social behaviour than many other countries, as it notes that "there is also less emphasis in many other countries on the link between crime and anti social behaviour, with less effort to collect information about the latter."
However the nature of anti-social behaviour makes it very difficult to record national figures reliably, and while the datasets on crime available to us are far from perfect, they do allow for the kind of analysis of trends in notifiable crime that currently isn't possible for anti-social behaviour.