How to use Crime Stats when creating a Crime Risk Assessment


Crime Risk Assessments are an important step in assembling the evidence for a CPTED recommendation or report. Crime Risk Assessments are based on crime statistics for the location - usually reported crime incident data. A standardised risk  format is used to show the likelihood of crimes. It can be a challenge to convert crime statistics to risk likelihood. There is a simple way...

What follows aligns with the Australian and New Zealand Standards - the ISO 31000 standards family, in particular ISO/IEC 31010:2009 – Risk Management – Risk Assessment Techniques.

Here is a typical 1-5 scale of  risk likelihood as used in a Crime Risk Assessment.

  1. Very rare or not possible
  2. Unlikely
  3. Possible
  4. Likely
  5. Almost certain, or certain, to happen

The challenge is to convert different kinds of crime incident data into the risk likelihood levels.  The problem is that crime data is generally in incidents/year, but the risk is different depending both on the type of crime and the scale of the context.

The problem

Imagine...

Crime incident data of 30 cars stolen from a car park in the course of a year. For a car park with 10,000 cars per day (3, 650,000 car parkings per year) then viewed from the position of someone thinking of parking their car in that car park, the risk of theft is almost non-existent - i.e. level 1 - very rare

Now consider if only 30 cars were parked in the car park over the last year and all 30 had been stolen. From the point of view of the car owner thinking of parking their car in that carpark, the risk is very high, i.e. Level 5 - Almost certain, or certain, to happen

Similarly, some single crime incidents affect only one person and others affect many (e.g. anti-social behaviour of a group of individuals in a busy mall). In the latter case, the incident is only counted once but affects many so the risk per individual of being a victim of that crime is higher  than the crime incident count.

The Method

The method has four steps:

  1. First choose one type of crime and identify the period of time the data is for
  2. Identify the number of persons specifically  at risk of direct consequences from that crime over that time in that location
  3. Compare (divide) the number of crime incidents with the number of persons from 2.
  4. Put yourself in the shoes of one of those persons and ask yourself about the likelihood of the risk.

Example 1

In a particular short length of street in a city over the last year there have been 12 robberies.  You go out to that street and identify that around 20 people per minute pass through it for around 12 hours each day five days each week. for 52 weeks per year.  That is around 3,744,000 persons in that year of whom 30 got robbed. From the viewpoint of a person walking down that street the likelihood of getting robbed is again Level 1- almost non-existent.

Example 2

In a busy village main street which has much traffic there are five shops and 2 public telephones. Every 3 months there are about 5 incidents of malicious damage to the shops and the public telephones. In this case, the persons specifically at risk from the consequences are the shop owners and and those responsible for the public phones. The number of shoppers  and the amount of traffic may be high and the level of the incidents may be low, but from the point of view of the shop and phone managers, the likelihood of malicious damage occurring in any three month period are high, i.e Level 5 -  Almost certain to happen 

Example 3

In a a busy shopping centre, anti-social, loud  and threatening behaviour from groups of unruly sports fans upset shoppers. This most commonly happens at weekends, but can also happen at other times. The risk picture of this kind of incident is that for one incident, many individuals can be adversely affected (the same is true of terrorist related incidents).  Although there may be only a few incidents per week and thousands of shoppers each week, the relative numbers of shoppers exposed to the anti-social and threatening behaviour is much higher than the number of incidents and a significant proportion of  the total number of weekly shoppers. In this case, the relative likelihood, depending on the exact circumstances, will be in one of the middle levels of  risk likelihood, e.g. Level 3 - Possible

Take Aways

In identifying risk likelihood levels from crime incident data focus on the relative likelihood for that crime over that time and in exactly that context for those who are directly adversely affected by the consequences.


© 2019 Design Out Crime and CPTED Centre. All Rights Reserved.

Please publish modules in offcanvas position.