Monday, 27 October 2008

Lies, damned lies and (crime) statistics

By Dr Bill Dixon

‘Statistics are like politicians … they tell lies.’ These are the words of ‘Michael’ reacting online to a story in the Daily Mail about the revelation that police forces up and down the country have been undercounting the most serious types of violent crime. With his very 21st century distrust of politicians it is ironic that ‘Michael’s’ words echo those of the 19th century Prime Minister, Benjamin Disraeli, who once said that there were ‘lies, damned lies and statistics’.

But is such scepticism justified? Are the crime statistics published by the Home Office every quarter, and then for a full year each July, really just ‘damn lies’ – or worse? It’s almost second nature for criminologists to be critical of crime statistics, whatever their source and ‘Michael’ would have been nearer the mark if he’d said that crime statistics are like cigarettes: they should always carry a health warning. Crimes can’t be counted like sheep because what counts as crime is both constantly changing – think of homosexuality and rape in marriage. And, at least for criminologists, they are endlessly controversial.

According to the Daily Mail, the admission that the police have been ‘downplaying’ (note the none-too-subtle hint of a conspiracy in that word) serious violence has dealt a ‘devastating blow’ to public confidence in official crime figures. Later in the story Shadow Home Secretary, Dominic Grieve, weighs in with the view that the figures ‘fatally undermine government spin that violent crime [is] getting better’. ‘If you can’t count a problem’, he adds, ‘You can’t combat it’.

Crime statistics, as criminologists like to put it, are socially constructed. In other words, they are the outcome of very complicated social processes and should never be (mis)treated as ‘hard facts’. It should come as no surprise then as many as 17 of England and Wales’ 43 police forces have been categorising offences involving intent to cause grievous bodily harm where no such harm is actually caused as ‘other violence against the person’ rather than as the ‘most serious violence against the person’. Leaving aside the possibility that the Daily Mail and the Shadow Home Secretary might be using this revelation as a handy stick with which to beat the government, what their shock and outrage betrays is not so much a world-weary cynicism about crime statistics but a touching faith in the ability of numbers to reflect social reality. But for ‘government spin’ and the pressure put on police forces to meet crime reduction targets, they seem to believe that crime could be counted accurately, and then combated effectively.

If crime statistics are neither ‘damned lies’ nor ‘hard facts’, what do they tell us? Well, despite the failings documented with such loving care by the Daily Mail, and seized on with such relish by Mr Grieve, they can tell us something. And what they tell us is a rather more optimistic story than headlines about a 22% increase in violent crime might suggest. In fact, the most recent annual figures (for 2007/8) published in July indicate that violent crime, indeed all crime, has fallen by more than 40% since 1995. These statistics, derived from the British Crime Survey (BCS) rather than police figures, rarely make the headlines but almost certainly represent reality more accurately than the fevered imaginings of Opposition politicians and middle market tabloids obsessed with the idea that Britain is an increasingly violent and lawless place.

Like all statistics – and cigarettes – these figures should be treated with caution. But – unlike cigarettes – it would be very foolish to make no use of them at all.


Read Matthew Hickley’s story about the police under-recording violent crime in the Daily Mail online at

For Home Office crime statistics for the quarter to June 2008, go to
And for the full year April 2007 to March 2008, go to

No comments: