Audit Commission

Local agencies hold a variety of data that provides the basis for compiling a better picture of the problems facing neighbourhoods. The second section of each example profile uses GDI to analyse the incidence of crime and anti-social behaviour by household group.

The table below contains a list of the incident data provided to the Audit Commission by local agencies in the five CDRP areas. Sharing of personal information between local agencies under the CDRP for the purposes of reducing and preventing crime and anti-social behaviour is enabled and governed by the provisions of section 115 of the Crime and Disorder Act 1998. Confidentiality of personal information can be maintained in a neighbourhood profile by using a GDI based analysis of incident data, similar to that used in the example profiles.

The record types listed in the table below are a broad mix of primary, secondary and tertiary incident data. For incident data to be useable for either geographic or geodemographic analysis, each incident record must include a valid geocode. For the data provided to the Commission, the last column of the table indicates the extent to which local agency data contained valid geocodes. An explanation of geocodes follows the table.

Source: Audit Commission fieldwork

While most, but not all local agencies shared incident data with the Audit Commission, not all individual records included a valid geocode. The following are examples of valid geocodes.

• The full (eight digit) postcode.
• Eastings and Northings.
• The Ordnance Survey grid reference.

The most precise geocodes are Eastings and Northings, because this locates a record to its exact geographical location. To avoid complexity in first steps towards geocoding data, local agencies should ensure that a full postcode is included with each record. Records that include a valid geocode can be analysed using both geographic information systems (GIS) and GDI. To gain the full benefit of a GDI based analysis of crime and anti-social behaviour reduction data, local agencies should ensure that victim, offender and witness records all incorporate a valid geocode.

In developing neighbourhood profiles, a useful preparatory step for local agencies and CDRPs is to review the following:

• the range of incident data that each agency holds;
• the method by which data are stored and quality is assured;
• the extent to which all records include a valid geocode;
• the flexibility by which data is extractable for exchange and analyses between CDRP members; and
• the structure and purpose of analysis to support decision-making.

The local agency with the greatest technical skill and capacity should lead a joint exercise where individual agencies lack the expertise or resources to undertake this assessment.

Teams compiling neighbourhood profiles for the first time may encounter the following problems that may constrain the extent of an initial profile.

• Potentially useful data excludes a valid geocode. Local agencies should ensure that data recording includes an appropriate geocode as a standard practice.
• Potentially useful data are held in bespoke information systems with limited export or reporting functions. Local agencies should ensure that plans to update or replace systems holding incident data are scrutinised by the CDRP.

Profiles require periodic refreshment, and at each refreshment, local agencies should aim to benefit from continually improving data quality.

Local agency incident data that included a valid geocode was analysed for the Audit Commission by UCL to determine the distribution of risk for each of the 11 main household groups. The results of these analyses are presented in a standard form in each example profile.

Source: UCL

The key information in the table is the Index Value for each of the 11 main household groups. The above table describes how the indices are calculated. A group index value of 100 indicates that the group incidence is the same as the average for all groups. A group index value of 150 indicates that the group incidence is 50 per cent above the all group average. Likewise, an index value of 200 indicates a group incidence that is twice the all group average. Conversely, a group index value of 50 indicates that the group incidence is 50 per cent below the all group average. To understand how these indices work in practice, read Section 2 in any of the example profiles.

The indices show the extent to which each main household group has been a victim of different types of crime and anti-social behaviour. For example some household groups, with small shares of total ward population, had disproportionately high levels of victimisation.

For the example profiles, the data applies only to the two electoral wards for each CDRP area. The risk shown for each household group in the example profiles is based upon the incident data provided, and is included in this guide for indicative purposes only. The principles guiding this analysis can be widely applied, but the findings shown in the example profiles cannot be applied more broadly.

¹ Chainey S, Ratcliffe J, 'GIS and Crime Mapping', Wiley, 2005.