Letters:Statistics: unemployment, crime, graphs and independent schools

Sir: The Home Office has an exemplary record of maintaining the integrity of the crime statistics that it publishes. I have long experience of working with criminal statistics, mainly within the Home Office, and have no knowledge of any attempts t o massage the figures. On the contrary, the Home Office Research and Statistics Department has done a very great deal to improve popular understanding of crime statistics and their shortcomings.

Crime statistics recorded by the police can be especially misleading as victims' preparedness to report crimes varies over time and place - as do police recording practices. Thus the statistics recorded by the police (and published by the Home Office) may not reflect the underlying trend. For this reason, the Research and Statistics department mounts regular sweeps of a very large sample survey of the general population, the British Crime Survey. This provides a second measure of crime, which - though subject itself to sampling error and other forms of measurement error - is free of the distortions to which police figures are prone. The two measures together enable much more informed judgement of crime trends than the police statistics alone.

In your timely examination (19 December) of government statistics, I would not want the achievements of the Home Office in this field to go unnoticed.

Yours sincerely, MIKE HOUGH Professor of Social Policy South Bank University London, SE1

20 December