Back in July, I wrote a blog entitled “A dangerous domestic violence myth is born,” which queried the claim made by journalists Alan Travis and Nick Cohen, and criminologist David Wilson, that the ongoing fall in the British homicide rate could be attributed to the simultaneous decline in the prevalence of domestic violence.
The piece was widely shared and republished elsewhere, and remains my most visited page on this blog. I also sent the link to the Travis, Cohen and Wilson, inviting them to respond. None of them did, but I had hoped that the false claim might have been nipped in the bud. I should have known better.
This week David Wilson returned to Comment is Free with his views on the horrific disappearance of April Jones. Before I go further, let me stress that I entirely agree with the meat of Wilson’s argument, which was that abduction-murder of children by strangers is exceptionally rare and should not distract us from the more immediate challenge of addressing violence and murder in the home, where the vast majority of child murders occur.
That said, there were claims made in the middle of his article which I found bafflingly detached from reality. Here’s the relevant section:
On average since the early 1970s, only six children per year have been abducted and murdered by strangers, and while that is still six children too many, this sad statistic is put into perspective when we remember that two children a week are murdered within the home.
But, here’s the good news – the numbers of murders are falling, and they are falling for one specific reason. Partly in response to pressure from campaigners, the police now treat domestic violence much more seriously than they once did. And if a man is hitting his female partner, it is probable that he will be physically, emotionally or sexually abusing his children too. So by insisting that the authorities take domestic violence seriously, we are protecting not only women, but children too.
Wilson uses the figure “only six children per year” abducted and murdered by strangers as a reassuringly low number. It struck me as quite astonishingly high. A check on Home Office crime figures reveals that over the past ten years, an average (mean) of eight children under the age of 16 have been murdered by strangers per year. The vast majority of those were not “abducted and murdered”, they were victims of stabbings or beatings in parks or playgrounds, drive-by-shootings, deadly robberies and so on. It turns out there are no official statistics on kidnap-murders of the type we are discussing here, so we are dependant upon press reports. A quick (and very unscientific) trawl of Google news reports suggests to me that a more accurate estimate might be six per decade, not six per year.
As for the next statistic, that two children a week are murdered within the home, this turns out to be no more credible. On this we do have home office statistics. Over the past decade, the number of children killed by a parent, carer or acquaintance averages 49, or fewer than one per week. It looks to me that, in trying to dispel myths, Wilson has vastly over-reported the risks of a child being murdered either by a sadistic kidnapper or by a family member.
I asked David Wilson for his source for these statistics over the magic of Twitter. He told me that his figures come from his own book, Innocence Betrayed, first published in 2002. I will take it on trust that the figures he quotes were once true, but I do find it disturbing that an article which purports to be about current trends depends upon data which turn out to be more than a decade old.
And then we come to my own personal bugbear, the claim that the numbers of murders are falling, and they are falling for one specific reason… the police now treat domestic violence much more seriously than they once did. And if a man is hitting his female partner, it is probable that he will be physically, emotionally or sexually abusing his children too.
There are so many flaws in this claim that it beggars belief. First, lets look at the implication that child homicides are committed by abusive men. The first factual problem here is that, unlike most violent crime, women do kill their children. Around a third of child homicides are committed by mothers (where it is an infant who dies, more than half the perpetrators are female). That alone blows a huge hole in Wilson’s hypothesis. Secondly, I’m unable to find any research that places child homicide typically in a wider pattern of domestic abuse. On the contrary, one of the distinguishing features of one subset of child killers, ‘family annihilators’ – men (usually) who kill their kids before committing suicide – is said to be that they rarely have a history of criminality and the families often appear to be stable and happy before the incident. Of course there are some cases of men who murder their children as part of a wider pattern of violence and abuse, but to attribute the child homicide rate to their behaviour alone is positively fanciful.
There is one final reason to be suspicious of a causative link between declining domestic abuse and the homicide rate of either children or intimate partners. The statistics simply don’t match the pattern.
I have graphed the total homicide rate over the past decade against the rates of both intimate partner deaths and child homicides.
Series 1 (blue) is the total murder rate since 2001. Series 2 (red) is the child homicide rate and Series 3 (green) shows intimate partner homicides (both male and female). It does indeed show a significant decline in the homicide rate (if you’re wondering about the spike in 2002/3, it is explained by the 172 victims of Harold Shipman all being recorded that year). However the overall trend for both intimate partner and child deaths remains stubbornly flat.
For illustration, I can map on the trend for domestic violence incidents, as recorded by the British Crime Survey. These are in thousands, but fit conveniently onto the same graph.
(Domestic violence incidents in 000s in purple.) What this shows is that the domestic violence rate and the homicide rate follow a very similar pattern, but as we all know, correlation does not imply causation. I won’t bore you with stats and graphs, but the rates for interpersonal assault and other violent crimes follow a similar pattern. It would be reasonable to conclude that we, as a society, and men in particular, are becoming less violent in all sorts of ways. This must not make us lose track of the fact that the great majority of victims of violent crime are nether women nor children, but adult males. The risk of violent death in the home for children and adults alike is not the outcome of this trend, on the contrary, it is a troubling exception.