Tracking the evidence for a ‘mythical number’
By Heather Strang, Peter Neyroud, and Lawrence Sherman
There is a widely-repeated claim that victims of domestic abuse suffer an average of 35 incidents before the first call to the police. The claim is frequently repeated by senior police officers, by Ministers, by government reports, by academics and by domestic abuse victim advocates. It has certainly influenced police behaviour in deciding how to deal with domestic abuse calls. Police are required to justify the proportionality of their arrest decisions, so if 35 were the right number in a presenting case, arrest will always be seen as proportionate: after all, culpability may not be clear this time but there have probably been at least 34 priors.
The question arises: where did this claim come from? Tracking down the evidence for this number proved to be an interesting chase back through publications over the past thirty years. Repeatedly, publications cite the number with no reference to specific research demonstrating the finding, only to other publications citing the number. Finally we traced an obscure Canadian government report about a 1979 study in a small city that reported on 53 women who had experienced domestic abuse to which police had responded. These 53 were interviewed two years after the incident and represented 24% of all victims the researchers had sought to interview, a worryingly small response rate with a high risk of sampling bias. They were asked how many times they had been assaulted prior to this particular call, from their earliest recollection until two years previously (a very difficult question to answer accurately): they were not asked how many times they had been assaulted before calling the police for the first time. The women had a wide range of responses to the question, from zero to 312, and the average (not the median or the mode) was calculated to be 35. Furthermore, these women reported being assaulted an average of three times in the two years after the incident – a number difficult to square with the reported average of 35 over an unspecified time period when police may or may not have been called for any or all or none of the prior incidents.
So 53 Canadian women, whose domestic abuse came to police attention in 1979, bear the entire weight of the claim so widely – and increasingly – relied on.
Domestic violence is a very serious problem for our society. Effective prevention and criminal justice interventions need the best data. We are certain that many women suffer years of misery before the police are called. Some, perhaps most, may never call. But we do not help by citing unreliable numbers about their suffering. To cite a number as exact as ‘35’ lends a veneer of precision which is false and misleading: a ‘mythical’ number that gives the impression that governments, academics, police and the public generally know a lot more than we actually do. We submit that mythical numbers need to be defeated by evidence-based discussion, informed by real data. Victims of domestic abuse deserve no less when we discuss how best to help them.
Peter Neyroud CBE QPM is the editor of Policing, A Journal of Policy and Practice, and Resident Scholar at the Jerry Lee Centre for Experimental Criminology, Institute of Criminology, University of Cambridge, Dr. Heather Strang, is the Deputy Director of the Jerry Lee Centre for Experimental Criminology, and Deputy Director of the Police Executive Programme, Professor Lawrence Sherman, is the Wolfson Professor of Criminology, Director of the Institute of Criminology, Director of the Jerry Lee Centre for Experimental Criminology, and Director of the Police Executive Programme, University of Cambridge.
The full article will be available this June in Policing, A Journal of Policy and Practice, volume 8.2. This peer-reviewed journal contains critical analysis and commentary on a wide range of topics including current law enforcement policies, police reform, political and legal developments, training and education, patrol and investigative operations, accountability, comparative police practices, and human and civil rights.