Analysis as part of a problem-solving approach to homicide.
Analysis forms part of the SARA model and can be used as part of a problem-solving approach to homicide.
Having defined your local homicide problem and identified its common characteristics through scanning, you can progress to analysis. Analysis helps you to understand the causes of your problem by using available data to identify the circumstances that are generating or maintaining it.
There are many causes of homicide. It’s often not feasible for the police to influence:
- long-term macro-level drivers of homicide, such as demographics or inequality
- meso-level factors, such as an age group's alcohol consumption
But intervening in many of the short-term micro-level factors often sits in the police remit. In setting your analysis goals, it’s therefore useful to aim for micro-level factors, which tend to manifest at the individual, interpersonal and community levels.
A common term in the problem-solving literature is identifying pinch points and your analysis is aiming to identify these. A pinch point is a causal factor in homicide that can reasonably be addressed through policing practice. It’s a common event or circumstance that’s present in or precedes most of these incidents. This might be an event that increases an individual’s vulnerability or a predictable/scheduled removal of a safeguard or a failure of that system.
Problem analysis triangle
The problem analysis triangle can be used to think about the identification of pinch points.
Inner layer
Three things must be present for a homicide to occur:
- a person who is motivated to commit a homicide (the equipped offender)
- an accessible victim (the target)
- these two coming together in a place and at a time where capable guardians are absent (the place)
In the case of industrial homicide, the perpetrator of the homicide need not be present. But the mechanism by which the homicide occurs – such as a fire – must be present to fatally affect the victim.
Middle layer
Capable guardianship can be any person or thing that can prevent the homicide from occurring. Examples of this include police, security guards and even bullet-proof glass, as guardianship does not have to be physical. The sound of a child in a home may prevent a domestic abuse incident from becoming fatal, or witnesses may be able to de-escalate a violent situation simply by their presence.
The middle layer of the problem analysis triangle reminds us to look at actors beyond the immediate homicide event that might be able to deter or pre-empt the homicide. These actors might be:
- guarding a victim (guardian). Examples of guards include police domestic abuse liaison officers or child welfare officers
- handling perpetrators (handler). Handlers might include probation officers or family members of a perpetrator
- managing locations where homicide is a possibility (manager). Managers might include nightclub security staff or housing officers
When victims have already come to the attention of the police and allied services, as is often the case with homicide, this layer of the triangle is an essential part of the causal pathway to homicide and a crucial component for response.
At this level, data sharing is important. Homicide reviews have repeatedly identified that a lack of joined-up data sharing between services who ‘guard’ victims and services who ‘handle’ perpetrators have failed to share information that could have been central to identifying and managing risk of homicide earlier.
Outer layer
The outer layer of the triangle focuses on ‘super controllers’ who can influence the activity of the actors in the middle layer. This might be a local authority that increases the availability of domestic abuse shelter facilities, or legislators who support or prohibit data sharing.
A careful analysis of a specific crime problem using the problem analysis triangle can be an enlightening experience in the process of developing a response. The triangle emphasises that there are many levers that might profitably be pulled and actors engaged in the interests of preventing homicide.
The important point to recognise is that not all causes of your local homicide can (or need to) be addressed at once. Good problem solving might examine an issue from several perspectives and result in the implementation of a range of interventions that target different outputs (or pinch points) of the analysis.
Crime scripts and trajectories
Homicides are events about which we often have unusually detailed information. Case files and homicide reviews will often provide timelines of events as well as contextual and background information.
Working on the assumption that there is at least some consistency across specific types of homicide, you may be able to construct an event script for the homicide type you have chosen.
Naturally there will be some variation when dealing with small numbers of incidents, even when homicide types are closely matched. A problem solver seeking a pinch point must be able to find the common signal in the noise of the data. Examples include the following.
- In post-separation domestic homicide, was there a trigger event that preceded perpetrators returning to the family home?
- If there has been a spate of knife crime homicides involving machetes, did the perpetrators access their knives through a local underground market?
The challenge here is to avoid the black swan fallacy of extrapolating from one rare event to another. Thinking critically about the relative importance of signals you have identified and seeking independent advice is important if your response is to be effective.
Multi-agency responsibility for a multi-agency problem
The evidence on the ‘cycle of violence’ (Butler and others, 2020) and the impact of ‘adverse childhood experience’ (Fox and others, 2015) on later violent behaviour is compelling.
It seems beyond doubt that much of today’s violence has its origins in the childhood experiences of ten and fifteen years ago. This link is not just about street violence, it seems to endure across types – witnessing domestic abuse as a child is strongly associated with both domestic abuse perpetration and victimisation as an adult.
If the levers available to police are not likely to be effective in isolation, a problem-solving approach to homicide must look beyond policing. Even when police think beyond their usual constraints to develop creative solutions, police may not be the best organisation to deliver the intervention. Having trusted community and statutory partners to be the public face of interventions can greatly affect how they’re received by the public.
It’s also important to remember that partnership working has its limits. Services have different cultures and priorities that are sometimes incompatible. If partnership activities are too broadly defined, wide initial enthusiasm will result in weakening commitment from many partners. Organisational commitment is also vulnerable to changes in personnel.
In establishing partnerships for a specific intervention:
- try to keep the partner membership as lean as possible
- involve partners thoroughly in the design of the intervention
- try to ensure full written organisational commitment
- consider a sunset clause (expiry date) for the partnership so that it does not outlast its value
Data sharing for homicide prevention
- Clear guidance on data protection and general data protection regulation (GDPR) rules are required for all agencies involved in homicide prevention. Concerns over culpability in the instance of data loss outweigh the potential value of data sharing for prevention.
- Prevention efforts would benefit from a coordinated, multi-agency, national data-sharing infrastructure.
- Crime analysts would benefit from guidance on use of data to determine risk and identify patterns.
- Analysts involved in homicide problem solving should have a working knowledge of common logical fallacies in interpreting data and causal inference, such as the law of small numbers, the base rate fallacy and the ‘table 2’ fallacy.
Next steps
Once your analysis is complete, you can move on to the next stage of the SARA model – response.