Is violence in decline?

We know a lot less about trends in violence than we think and our explanations for their rises and falls are not up to much either, Richard Garside argues.

By: 
Richard Garside
Date: 
Thursday, 24 April, 2014

The ongoing apparent fall in violence continues to perplex commentators and criminologists alike.

Just today the latest quarterly update of the Crime Survey for England and Wales found a 22 per cent fall in violence incidents to December 2013, compared with the same period last year.

Yesterday the Violence and Society Research Group at Cardiff University estimated that violence in England and Wales declined by 12 per cent between 2012 and 2013, based on an analysis of hospital accident and emergency admissions data.

These are dramatic and large declines. For many they are also distinctly counterintuitive. Many expected various types of crime, including violence, to rise during the recession. The result has been a proliferation of theories and speculation masquerading as evidence and analysis.

Joe Shute's piece in today's Daily Telegraph has done us a service in bringing together the main themes of the recent discussion.

First we get a multiplicity of statistical measures all pointing to the apparent 'fact' that violence is in decline. The Crime Survey for England and Wales (CSEW) and the Violence and Society Research Group (VSRG) study both feature. So does the United King Peace Index, which concluded that Broadland in Norfolk is the safest place in Britain. The author also throws in some long-term historical comparisons: we're a lot safer now than in 18th Century Norwich or 14th Century Oxford apparently.

These different data all offer interesting and useful perspectives. They do not, individually or taken collectively, get anywhere close to making a definitive case that violence is in decline in England and Wales, or the United Kingdom more broadly.

The CSEW, for instance, is shot through with omissions and methodological weaknesses that belie its status as a 'gold standard' in relation to crime trends.

The experience of domestic violence, for instance, is not captured well by the CSEW, while sexual violence is not measured at all. Violence experienced by women, children and young people, vulnerable adults, those in institutional settings, the homeless, migrants, the poor and marginalised are variously not measured at all, or done so only very partially.

Add to this the methodological problems with data collection and the CSEW begins to look pretty ropey. Those who have experienced most violence are far less likely to be interviewed for the survey, if only because they are going to be less inclined to invite a complete stranger, however friendly, into their home for a one to two hour interview involving the sharing of a lot of potentially sensitive and traumatic experiences.

The total figure for violence in the CSEW - 1,537,000 in the latest report - is reassuringly precise. It is worth remembering that it is an estimate based on the experiences of around 36,000 people. This is a relatively large sample size. It remains, though, an estimate.

It would be better if the CSEW authors restricted themselves to detailing the self-reported victimisation experiences of the 36,000 or so people interviewed, leaving the statistical extrapolations to others.

In general, those who hold forth confidently about the falls in violence are white, middle-aged, middle class males who, for a number of reasons - economic, social, demographic to name three - are far less likely to be victims of violence than they were as teenagers. Unsurprisingly they are inclined to believe official data that reinforces their own experiences.

The CSEW omits far more violence victimisation experiences than it covers. Violence victimisation is far more complex than the abstractions of this highly partial, methodologically flawed, survey.

This lack of reliable data does not stop the proliferation of theories and speculation about the apparent fall in violence, not least of all because those offering up their theories assume the data are reliable. Shute's article offers a fair smattering of them:

  • A decline in binge drinking, improved multi-agency prevention work and CCTV cameras, according to Jonathan Shepherd from the Violence and Society Research Group.
  • We're evolving to become less aggressive, according to the psychologist Stephen Pinker.
  • We imprisoned our way to lower violence, according to think tank director David Green.
  • There are a 'growing number of influential studies' suggesting the decline in lead pollution has made us less violent.

I'm inclined to take Jonathan Shepherd's thoughts more seriously than Stephen Pinker's or David Green's. The lead pollution theory, on the other hand, makes the classic error of attributing a major social change to a beguilingly simple biological mechanism. If you start from the position that a simple explanation for a socially complex change will not stack up, you are unlikely to be proved wrong.

Shute also quotes the former chief constable Crispian Strachan, who with impressive candour remarks 'we really don't know why crime is going down'.

While I am sceptical of talk about overall rises and falls in crime, Strachan does point to a fundamental problem with the current public discussion on crime trends. Our data on different crime types is not what it's cracked up to be and most of the current explanations for observed crime trends are variously unevidenced and unconvincing.