AOAV: all our reportsImpact on civiliansKnife crime

Can knife homicides be predicted?

Detective Superintendent John Massey is the Head of the Protecting Vulnerable People Department at Cambridgeshire Constabulary and author of a major research paper on knife homicide predictors in the UK: “Forecasting Knife Homicide Risk from Prior Knife Assaults in 4835 Local Areas of London, 2016–2018”. AOAV interviewed him about his insights into knife assaults and murders.

AOAV:           John, can you just outline what you did in terms of research?

John Massey:  Yes, certainly. I was studying for a master’s degree at Cambridge University and the essence of the research I did was to analyse around 10 years’ worth of homicide data within London. I looked at a whole series of different features around the geography, the demography, of where and how the offences have taken place. Most specifically, for the purpose of this conversation, the final results that I was able to glean, identify a correlation between non-fatal knife events and fatal knife events in the subsequent year. To really cut to the chase in terms of what I was able to identify was that there was a really strong correlation between small geographical areas called LSOAs – Lower layer Super Output Areas. These are in simple terms, very small geographical areas with around 800 households – And a strong correlation between non-fatal knife assaults in those areas in what we call year one, and then in the subsequent year, there being a fatal knife assault taking place. And to the extent that you’re twice as likely to have a knife enabled murder, in an area where there is a non-fatal assault in the first year and in an area where there are six or more non-fatal knife assaults, the next year you are 1400 times more likely to have a fatal knife assault. So that was the essence of the research. I studied 10 years’ worth of homicide data, knife enabled homicide data in particular and one particular year from 2016 to 2017, I studied 3500 non-fatal life assaults.

AOAV:           So, in a sense, putting it in a layman’s terms, this is almost a canary in the mine warning for violence reduction, you can identify where there may be a lethal event coming in the near future.

John Massey:  I think that’s a good way of putting it. I think it is as far as we’re ever able to give ourselves indicators of a higher probability and I wouldn’t go any stronger than that. It’s not a predictor, of course it’s not. But what it does do is give us the ability to target more efficiently and put our resources in certain areas. As you say there are areas, very small geographical areas which is an important point in itself because they’re more controllable, they’re more releasable because of their geographical size. There is a greater chance that these particular areas will yield sadly, a fatal knife event. It’s not a clear predictor, 100% of course not, but it does give you a scientific and intelligence-led basis to target your meagre resources perhaps in smaller places in the knowledge that they’re the right places to be patrolling and the right places to be policing.

AOAV:           Could you quickly just explain the sort of methodological approaches, how did you ensure that you were trying not to include abhorrent events? Or things that weren’t necessarily, you know, obviously, freak freak moments or a sudden uptake of extreme violence where maybe, three people were killed in a very short setting by the same knife? How did you make sure that your methodology was rigorous?

John Massey:  In terms of rigour, it really is looking at each individual offence in isolation, but at the same time, and I’ll be really clear about it, it is aggregating that data together. You have to understand that we are looking at knife-level homicides, we are looking at homicides, per se. I obviously looked at some of the features of other types of homicide, the other types of knife-enabled crime, I did aggregate those together but what I haven’t included are any other CT related events.  Some have different topological features, they may be drug-related or gang-related, they may be domestic abuse-related, but they are all knife related. In other areas, it does become a little harder to make these correlations because mercifully, there aren’t the numbers of knife enabled assaults or knife-enabled homicides, but there are within London within that period of time. So, for example, in 2017/2018, there was almost a 100% increase in the number of homicides in that year overall but one of the previous 10 years. So, you do have some sufficient data and some sort of statistical power to look at the data with some integrity and know that you’re not capturing as you say, freaky bends. What you’re capturing is tragically business as usual of crime trends over a course of 10 years.

AOAV:           And just to be clear, what was your geographical spread? Where did you look?

John Massey:  Across the whole of London, across the whole of the metropolitan area for a 10-year period.  Over that time, I think there were 1156 homicides and there are 4835 LSOAs of these small areas. Over that 10-year period, for instance, as a statistic, I think almost just shy of 30% of all homicides were clustered within just 3% of those LSOAs. So, while mercifully, even the 1156 homicides, perhaps compared to certain other cities around the world, isn’t an extraordinarily high number, it certainly is to the person and people and families affected by that. Absolutely. What you can actually say is that well over 70, almost 80% of LSOAs in a 10-year period didn’t experience a single homicide. Now you can look it up from either perspective, any single homicide is absolutely one too many but as a trend across a large metropolitan city such as London and the size of the community population that’s a comparatively lower number. Statistically, it gets a little more challenging. But you can say that the offence has been at knife-enabled crimes and certainly homicides themselves are clustered in a small number of geographically small areas.

AOAV:           Roughly, for how many people stabbed and survived are people stabbed and die when it comes to the sort of violence we’re talking about?

John Massey:  The specific data that we got, within 2016/2017 there were 53 knife-enabled homicides and 3500 or so stabbings, per se. So, in terms of this simple question, how often does a stabbing cause death, it is probably 1 in 66.

AOAV:           So, knife crime or knife violence is significantly less lethal than gun violence, but it’s still overtiming with the sheer ability to get your hands on knives, it is much more ubiquitous.

John Massey:  It is tragically, and we saw as it was really emphasised by statistics in 2017/2018, with an almost 100% increase in knife-enabled homicides, you can see that there was almost an epidemic of knife-enabled crime taking place, not necessarily within typological subsets of being robber related or gang-related per se. But knife level crime full stop. As you say that the ubiquity of it was really apparent and it wasn’t thankfully, paralleled by a particularly high increase in gun enabled crime, it was quite exclusively around knife-enabled typology.

AOAV:           Which arguably is a reflection of how hard it is to get your hands on a gun, or a handgun, at least in the UK.

John Massey:  I would say so, I’d be quite clear about that. I think most observers of it would echo your thoughts in that it does demonstrate the strength, robustness, and effectiveness of our gun laws in the UK. Others will point to the fact that when you’re focusing simply on knife-enabled crime, that’s not the headline, the headline is obviously the fact that there was a really steep increase and a real problem around knives. I don’t think anyone would deny that, but certainly, as a secondary story, there’s something to be taken from the fact that it hasn’t tipped over into an escalation of gun violence.

AOAV:           The media always and obviously focuses on deaths. But do you think that there needs to be more focus arguably on non-lethal knife crime? Because as you seem to be suggesting it is this precursor indicator warning.

John Massey:  I’m not sure I would necessarily say that there needs to be more. I’m confident in terms of what I’ve seen within Cambridgeshire Constabulary and having worked in the Met, I think there are phenomenal efforts made by the police, by our partners, partners in the community, by the criminal justice system, the Crown Prosecution Service. Huge, huge amounts of money and an enormous amount of time and energy is put in by local authorities in preventative substrings, in preventative measures, and in prison programmes, a phenomenal amount is done within the prison system. I think what we’re seeing and what we saw, was a particular phenomenon at a particular time; it remains far far too high. Knife crime is an issue that government acknowledges, that all police services acknowledge and is linked in more often than not to capitalise drug usage and acquisitive crime. Can more be done? Absolutely more can be done. Everything I see practically on a weekly basis detracts from the phenomenal efforts that are being made, I can give everyone a real assurance that there is no end, and we continue to have the energy put behind trying to reduce and arrest those figures.

AOAV:           What do you think from a research perspective might need more focus on? What deserves more research?

John Massey:  I think around what works in terms of deterrent programmes if I’m honest. I think, from what I’ve seen lots of research that mirrors mine in terms of identifying, if not necessarily hotspots per se, but areas and trends and patterns geographically where knife crime is taking place. I think that gives us the evidence of where to intervene and where to put preventative measures. I think there needs more research as to what is more effective in the longer term. High visibility patrolling, for example, has a strong body of evidence to suggest that crime will be dispersed, crime will be eliminated in geographical areas, and it won’t just move around the corner. It’s an effective, robust tool. However, more so what works in particular young people’s lives in dissuading them and keeping them from crime and deterring them from falling to county lines. So, I think more research needs to be done now in terms of what is really effective and what is really going to give them long term solutions to deter young people from falling into acquisitive crime because I think that is still a real challenge. We see far too many young people being absorbed into a life of crime, being coerced into a life of crime by recruiters effectively who are increasingly professional and scrupulous in how they operate. So, yes, I think research around how we can more effectively keep young people safe from falling into a life of crime is obviously imperative.

AOAV:           And finally, do you think your research has actually had tangible real-life policing impact?

John Massey:  I certainly hope it has. It’s something that I’ve delivered to several different conferences and presented fairly widely. Hopefully, the article itself with Professor Sherman gained some traction. It’s certainly something which we trialled in the Met for a period of time, it’s something that we’re looking at now, in Cambridgeshire, and hope other forces do as well. I’ll be really open about this; it is something which in various different other forms is already being done in different ways. People have for long periods of time in policing, police have been targeting geographical areas, targeting hotspots. I think what my research showed is the value of targeting areas that have a particularly high prevalence of non-fatal knife crimes in these small geographical areas, these LSOAs. I think, within Cambridgeshire and within the Met, we are increasingly looking at crime, looking at violent crime at an LSOA level. It’s a much more granular way of looking at it instead of looking at boroughs and wards, which traditionally we looked at, which gives you a pretty opaque picture in terms of what’s actually happening. As everyone knows boroughs are huge areas, there could be quarter-million people of us living there. You have such a spread of demography and geography that you have such a difficult job in isolating what the real causal factors are.  Within an LSOA these are areas which perhaps are the size of two football pitches, a little more than that. You can police an area like that with some confidence to understand what’s taking place and then eliminate the factors within it that are causing a crime. So, I think hopefully, that that has had an impact. I think increasingly we are seeing lots more forces looking at crime in that granular detail.

AOAV:           Well, thank you very much for giving your time today.