London Murders: a Predictable Pattern?
London murders: a predictable pattern?
Four stabbings to death in a single day. Ninety murders in 7 months. Shocking figures—or are they? Knife crime makes the headlines almost daily but are Londoners really at increased risk of being murdered? David Spiegelhalter and Arthur Barnett investigate— and find a predictable pattern of murder.
Violence in London attracts headlines. After four people were murdered in separate incidents in London on July 10th, 2008, BBC correspondent Andy Tighe said “To have four fatal stabbings in one day could be a statistical freak”1. But could it? And on July 28th thelondonpaper had the front page headline: “London’s murder count reaches 90”. Of course every single murder is a tragedy for those whom it affects, and it is little comfort to them what patterns we may detect in the figures. But for the benefit of the police service and broader society we may wish to consider whether these “shocking” numbers in the media are evidence that things are really getting worse. Each year the London Metropolitan Police record around 170 homicides, and this has been stable for the last 5 years2. Box 1 explains what is covered by “homicide”—here we shall use “murder” as an equivalent term. We shall focus on all murders, whatever the method of killing: Box 2 provides details about the cause of death and briefly considers stabbings alone. Each of these murders is an individual crime that cannot be predicted. It may appear strange, but this very randomness means that the overall pattern of murders is, in some ways, quite predictable. Using some basic probability theory, and assuming that the level of violence remains the same, we can answer the questions coming out of the two stories given above: first, four murders in a day is unusual but not extraordinary. We would expect it to occur around once every 3 years. And, second, we would have expected, on statistical grounds, 93 murders by July 28th—so the count of 90 is not at all surprising. We can also
Four stabbings to death in a single day. Ninety murders in 7 months. Shocking figures—or are they? Knife crime makes the headlines almost daily but are Londoners really at increased risk of being murdered? David Spiegelhalter and Arthur Barnett investigate— and find a predictable pattern of murder.
Violence in London attracts headlines. After four people were murdered in separate incidents in London on July 10th, 2008, BBC correspondent Andy Tighe said “To have four fatal stabbings in one day could be a statistical freak”1. But could it? And on July 28th thelondonpaper had the front page headline: “London’s murder count reaches 90”. Of course every single murder is a tragedy for those whom it affects, and it is little comfort to them what patterns we may detect in the figures. But for the benefit of the police service and broader society we may wish to consider whether these “shocking” numbers in the media are evidence that things are really getting worse. Each year the London Metropolitan Police record around 170 homicides, and this has been stable for the last 5 years2. Box 1 explains what is covered by “homicide”—here we shall use “murder” as an equivalent term. We shall focus on all murders, whatever the method of killing: Box 2 provides details about the cause of death and briefly considers stabbings alone. Each of these murders is an individual crime that cannot be predicted. It may appear strange, but this very randomness means that the overall pattern of murders is, in some ways, quite predictable. Using some basic probability theory, and assuming that the level of violence remains the same, we can answer the questions coming out of the two stories given above: first, four murders in a day is unusual but not extraordinary. We would expect it to occur around once every 3 years. And, second, we would have expected, on statistical grounds, 93 murders by July 28th—so the count of 90 is not at all surprising. We can also
References: 1. BBC News Online (2008) Brown pledge to tackle stabbings. (Available from http://news. bbc.co.uk/1/hi/uk/7502569.stm.) 2. Coleman, K., Kaiza, P. Hoare, J. and Jansson, K. (2008) Homicides, firearm offences and intimate violence 2006/07. Home Office Statistical Bulletin, Jan. 31st, 2008. (Available from http:// www.homeoffice.gov.uk/rds/pdfs08/ hosb0308.pdf.) 3. Flanagan, R. (2008) Independent Review of Policing by Sir Ronnie Flanagan—Final Report. London: Home Office. (Available from http:// police.homeoffice.gov.uk/ publications/police-reform/Review_ of_policing_final_report/.) 4. Preece, D. A., Ross, G. J. S. and Kirby, P. J. (1988) Bortkewitsch’s horse-kicks and the generalised linear model. The Statistician, 37, 313–318. 5. Spiegelhalter, D. J. and Best, N. G. (2004) Shipman’s statistical legacy. Significance, 1, 10–12. David Spiegelhalter is Winton Professor for the Public Understanding of Risk at the Statistical Laboratory in the University of Cambridge, and Senior Scientist in the MRC Biostatistics Unit Cambridge. email david@ statslab.cam.ac.uk Arthur Barnett is a statistician in the Department for Business Enterprise and Regulatory Reform (BERR) Strategic Policy Analysis directorate (SPA) working with the Risk and Regulation Advisory Council (RRAC). email arthur.barnett@berr.gsi.gov.uk 8 march2009