Some data is better than others - a discussion of data quality in HMIC's Value for Money profiles
All the data used in HMIC's Value for Money profiles has been subject to additional checks either by the Home Office, the Chartered Institute of Public Finance (CIPFA) or HMIC. Checking has focused on identifying outliers and querying them with forces, which are then required to make a revised return to the Home Office or CIPFA.
The profiles rely on Home Office published crime data, survey data and staffing data and on the annual data return made by finance staff to CIPFA. Each page has a footnote showing the source of the data.
All the data is borrowed from others - a form of 'data recycling', so not a single new piece of data has gone into the profiles. It saves time and effort, but bringing information together from different sources sometimes means that in a few places the pieces don't fit perfectly and you need to be aware of these. There are two main areas:
Differences in definitions
CIPFA and Home Office data for total numbers in the workforce can differ. Differences in definitions are part of the cause. CIPFA takes average staff numbers over the year, while the Home Office is a snapshot at the end of each financial year. There are also differences in the types of staff included, with CIPFA including contractors and temporary staff as well as traffic wardens. We have tried to get the categories to match as closely as possible, and by excluding traffic wardens, CIPFA data more closely aligns with Home Office data. There are some forces where the figures don't match and these are shown. Where the difference is more than 2 percent then they are highlighted in red on pages 3 and 4 which show police officers and police staff costs.
Population figures for the Finance and Staffing sections are taken from CIPFA data while the crime statistics and sanction detections using Home Office populations (both mid-2007). This enables the data presented in these profiles to match other published sources.
The data is drawn from three main sources:
●Chartered Institute of Public Finance and Accountancy (CIPFA) data on income and expenditure. This data is supplied to CIPFA by police forces usually by force finance/accounts department and is published on their website - available by subscription.
●Home Office Annual Data Return (ADR) data (previously unpublished). This data is provided by forces to the Home Office and separates staff by their rank and allocated function.
● Home Office crime data up to and including 2008/09
Differences in the level of detail
The financial information provided by CIPFA complements the staffing data provided by the Home Office as it provides additional information on unit costs, overtime and non staff costs. However the expenditure summary on page 2 of the profiles compares the cost per head with other similar forces without any further adjustments to reflect different responsibilities, such as additional costs imposed by counter terrorism. (This issue mainly affects larger urban forces, but there are some county forces that may have additional duties as well. Some, for example, provide protection for members of the royal family.)
The financial information, in particular the summary on page 2, which includes all costs, must be looked at in the context of the information on page 17: Workforce by function, as this separates staff engaged in local policing functions which are comparable, from those staff engaged in national functions which are not comparable.
Treat large differences with caution
In producing the profiles we have noticed that for some forces, possibly because they are organised quite differently from others, there are significantly more staff in one function, which is often offset by significantly fewer staff in another function. We have noticed this in the figures for some forces, for example for the custody and criminal justice unit functions. Caution is needed where you spot these patterns. It may be that the two functions need to be considered together. An alternative explanation is that custody staff numbers are low because the function has been contracted out (in which case look at the supplies and services costs on page 7 to find out) or in other cases, custody staff are drawn from staff working in the community. Where this occurs it is hard to identify without further work.
Another area where you may see some major differences is where trends are shown. We draw your attention to two.
Financial trends for a four year period are shown on page 10 and sometimes you may notice that in one year the trend line falls to zero. This is because the data was missing from the CIPFA return. We have only gone back to forces when there is missing data for the latest period.
Staffing trends: The 'Changes in workforce' page also shows some large variations over the three year period. We have not shown the data if the changes have exceeded 50 percent reductions or more than a 100 percent increase. Some of the larger changes may be due to checking of the most recent data as we have not asked forces who have corrected recent data to check data from three years ago.
Some presentational points
Most of the information is standardised by comparing per thousand population. The average values displayed are the un-weighted averages of all the forces covered by the profile.
Differences can sometimes be looked at in more than one way. For example non-staff costs are shown per head of population in some charts and also as a percentage of staff costs in other charts. We have done this for two reasons. First, it is a better indicator, since most (but not all) non staff costs are influenced by staff: Premises costs are a good example. Dividing by staff costs also acts as a deflator and therefore provides a better comparison. For example, premises costs are higher in London, but then so too are staff costs.
Throughout the profiles the chart scales may differ, so visual differences may not be as significant as they appear. More specifically, as the Metropolitan Police data distorts chart scales, they have been excluded from the profiles for all other forces although it is included in the relevant averages, (the City of London is excluded for the same reason).
Looking to the future
The profiles are based on existing data sets, but during last year force finance staff developed new costing information (Police Objective Analysis) which became available in December 2009, too late for this year's profiles. In future we aim to make as much use of this data as possible, especially if it is made mandatory.
Lastly, if by using the profiles or by using the data you have any suggestions on how the profiles can be improved, we are eager to hear from you. We can be contacted at
Official Statistics
The Value for Money Profiles are an Official Statistics output and are prepared by IPF on behalf of HMIC under the principles set out in the Code of Practice for Official Statistics.
Most of the data have previously been published and the data sources are quoted on each page.
The 'Staffing by Function' data have been verified but have not been previously published, although they are closely associated with the National Statistics police service strength dataset collected annually by the Home Office as part of the Annual Data Return (ADR) process. In addition, the 999 calls data are collected as part of the ADR process, but have not undergone the same verification processes as the other data.
Statisticians in the Home Office have been consulted on the development of these Value for Money Profiles and they are published with the agreement of the Home Office's Head of Profession for Statistics.
[this is a copy of the text from downloaded on 14/04/10]