Update on population research

Report: 11
Date: 12 June 2001
By: Commissioner

Summary

This report is to inform the Committee of the difficulties and current developments into research of population data for the purposes of monitoring BVPIs and stop and search data. The committee is asked to consider how the development of reliable population data can be taken forward.

A. Recommendations

1. The Committee notes the difficulties with current resident population estimates in performance monitoring.
2. The Committee considers the recent developments in research into effective population data.
3. The Committee campaigns for national research to continue into gaining reliable national population data.
4. The Committee considers whether further research into population data is required.

B. Supporting information

Background

1. Population data is used in many Best Value Performance Indicators (BVPIs) as a denominator (ie: per 1000 population) for crime levels, stop and search data, complaints, incidents and expenditure. In order to compare forces, BVPIs use a base of 1000 population, for example total crimes recorded per 1000 population. The BVPIs using population as a denominator are shown at Appendix 1.

2. The current population data used is resident population taken from the 1991 census. The census provides information for BVPIs on population for police authority areas and for basic command units (BCUs), for total population and population by ethnic group.

3. As the 1991 census becomes out of date relatively quickly, the total population data by borough and force area is updated yearly from the Office of National Statistics (ONS) mid-year population estimates. These estimates are for total population only. Ethnic groups are estimated by applying the same proportions in each group from the 1991 census results to the totals from the mid-year estimates.

Problems with current population data

4. It is felt by the police service, the Home Office and independent researchers that the current population estimates have a number of problems. These are described below:

• the census will be between 2 and 12 years out of date. In particular, proportions of ethnic minority groups in a force area are likely to have changed over a 12-year period and hence applying the same proportions over that period is likely to provide inaccurate population data by ethnic group. Although the 2001 census has just been carried out, it will be two years before publication and will become out of date in a few years;
• there is some evidence that the census is under-representative of ethnic minority residents;
• population measures for the BVPIs based on census data provide information on resident populations and hence exclude day time population in terms of commuters or visitors (an important factor for London) or the homeless;
• population may not always be the best comparator, for example autocrime may be better compared by numbers of cars.

5. A recent scoping study by the Newcastle University Centre for Urban and Regional Development Studies (CURDS), noted that forces with high daytime populations compared to their resident populations "might argue that comparisons of their crime rates against other forces and partnerships, using resident population as a denominator in such crime rates, are unfair".

6. Recent Home Office research on profiling populations available for stops and searches found that: "The research shows that measures of available populations are very different from resident populations (as measured by the 1991 census). Most significantly, in areas with high stop and search activity, young men and people from minority ethnic backgrounds tend to be over-represented in the available populations".

7. The Home Office has also stated that research should 'aim to compare the profile of searches with the population actually available for searches. This provides a direct measure of ethnic biases in the street level decision-making by officers'.

Current developments in population research

8. There are two distinct national areas of population research being carried out. Firstly, national research into reliable comparative data for crime based BVPIs. Robust national data would allow comparisons between forces to be made by the MPS/ MPA and also by the Audit Commission, Home Office and Department of the Environment, Transport and the Regions (DETR). Secondly, the development of robust population data for stop and search analysis, a high profile topic distinct from the other BVPIs. In addition, there has been research carried out regarding London in isolation, which may be useful internally for means of analysing stop and search data and for comparing performance across boroughs.

National research for BVPI crime population data

9. With regard to the research on crime BVPIs, the MPS supported by the MPA convinced the Home Office to commission research from CURDS into investigating the feasibility of estimating adequate effective population measures across England and Wales at the local level. The CURDS produced a paper in May 2001 recommending that the 'population at risk' should provide the basis for crime dominators and that research be carried out to calculate this, including different denominators for different types of crime and taking into account daytime populations.

10. However, at a joint working group held to discuss these issues, it was concluded that there was no satisfactory way forward at present, since the population at risk approach would still not capture non-residents, for example tourists and because it would not be possible to break that data down to BCU level. Accordingly the police service would end up with different BVPI denominators for force/BCU indicators. Further work would also be expensive and it was felt by some that there are other more pressing priorities for funding on best value issues.

11. The issue of appropriate populations for the BVPIs is of particular concern to the MPS and City Police. Other forces have not expressed disquiet (although forces who are affected by high influxes of tourists may have a comparable case to make). ACPO and the APA are not pushing for this issue to be taken further.

12. The working group agreed that:

• the Home Office would discuss with the DETR whether there was a wider consensus on the need for improved population bases across the range of BVPIs;
• APA would discuss with the Local Government Association (LGA) whether there was a local government view on the limitation of resident population;
• MPS and MPA would consider any alternative approaches that might be pursued.

The possibilities will be considered again in early June. The recommendations of the group will go forward to the Home Office working group on Police Performance and Best Value who will make a final decision.

National research on stop and search population profiles

13. Regarding stop and search population profiles, the Home Office has plans for a further research project "Measuring and Achieving Fairness in Police Use of Searches". This follows on from the work done in "Profiling Populations Available for Stops and Searches" – MVA and J. Miller (2000), and the Home Office commitment to the Lawrence steering group to take this work further. The project is aiming to address a number of key questions about how fairness can be measured and achieved in relation to searches. One aspect of the research will be to develop a methodology for measuring street populations.

14. This research is planned to start later this year, and could run for around twelve months before reporting in about two years time. This project is therefore unlikely to provide a short-term solution to MPA and MPS concerns around the use of resident population as a basis for performance monitoring with regard to stop and search. However, given that this work is to be progressed and funded nationally, there seems little advantage to be gained in investing MPA/MPS resource in pursuing this locally.

London based research on population data

15. The London Research Centre (now part of the Greater London Authority (GLA) and renamed Data Management and Analysis Group) has carried out the most comprehensive research into population profiles for London. Using Census data and ONS mid-year estimates, as well as migration, housing and other trend data, they produce annual estimates and projections of London resident populations by borough and ward. At borough level this data is broken down by ethnic group, age and gender.

16. The GLA data has been used by the MPS to provide a better population profile for stop and search data. However, the data is still limited to resident population and does not show transient or daytime populations. Work will be on going with the MPS and MPA to develop the use of this data and how GLA data can be developed to include transient populations. MPA officers sit on the functional bodies demographics research group and will monitor developments and consider how data could be adapted for the needs of the MPS.

17. As part of the Policing London project, Dr Marion Fitzgerald has produced "borough profiles" which give a picture of the policing environment (including ethnicity) for each borough.

18. A further option could be for the MPS to commission or carry out research to gain reliable population data for London. However, given that existing data has not provided reliable estimates this would involve new research and is hence likely to be expensive and would take some time.

Summary

• Unless the Home Office can be persuaded to fund further research, the data used for BVPIs (other than stop and search data) is unlikely to change. Any comparisons made between forces or BCUs will hence be unreliable.
• National research is being carried out to profile populations for monitoring stop and search. This could provide more reliable data in two years time.
• The MPS and MPA are considering how GLA population data could be better used for monitoring of MPS data, particularly stop and search.

C. Financial implications

None.

D. Background papers

• A feasibility study into developing a measure of effective population for use within the performance management framework for police and local authorities. Professor Mike Coombes, Centre for Urban and Regional Development Studies
• Statistics on Race and the Criminal Justice System: 2000, Home Office

E. Contact details

The author of this report is Johanna Gillians.

For information contact:

MPA general: 020 7202 0202
Media enquiries: 020 7202 0217/18

Appendix 1: BVPIs using Population Data

BVPI125 - Total crimes recorded per 1000 population and percentage detected

BVPI129 - Number of offenders dealt with for supply offences in of Class A drugs per 10,000 population

BVPI130 - Number of public disorder incidents per 1000 population

BVPI132 - Number of road traffic collisions involving death or serious injury per 1000 population

BVPI138 - Number of Police and Criminal Evidence Act (PACE) stop/searches of white persons per 1000 population and percentage leading to arrest

BVPI139 - Number of Police and Criminal Evidence Act (PACE) stop/searches of minority ethnic persons per 1000 population and percentage leading to arrest

BVPI140 - Number of substantiated complaints under Police Authority Act 1996 by detainees per 10,000 people detained

BVPI127a - Violent crimes per 1000 population and percentage detected

BVPI127b - Robberies per 1000 population and percentage detected

BVPI128 - Vehicle crimes per 1000 population and percentage detected

ACPIA2a - The number of racial incidents recorded by the authority per 100,000 population

Send an e-mail linking to this page

Feedback