Police-recorded crime vs the Crime Survey: Which number should you use?

There is no single perfect crime count. Each source answers a different question.

In short

Compare street-level records, police-recorded crime and the Crime Survey for England and Wales, including what each source measures and omits.

Police-recorded crime

Use it for police-system demand

Best for offences known to and recorded by police, local operational patterns and offence groups not well covered by household surveys. It can change when reporting or recording practices change.

Use the UK crime map to inspect the underlying fields, and keep the visible source month beside any screenshot or comparison.

The Crime Survey for England and Wales

Use it for experiences in households

Best for consistent long-term trends in selected offences experienced by people living in households, including incidents not reported to police. It excludes businesses, tourists and people in communal establishments.

Street-level crime data

Use it for local distribution

Best for exploring approximate local patterns and published outcomes. It is a separate extract from the datasets used for official ONS totals, so figures can differ.

The responsible answer

Name all four comparison fields

State the source, geography, reference period and denominator. If two measures disagree, explain why rather than selecting the larger or more dramatic figure.

Choose the source from the wording of the question

Let the question choose the dataset

Use street-level records for approximate local distribution, the recorded-crime series for offences known to police across consistent areas, and the Crime Survey for experiences among its sampled population. For a question about convictions or sentencing, none of those is sufficient: use the relevant court or Ministry of Justice table. Writing the question first prevents a convenient dataset from quietly changing what is being measured.

Quick checklist

Check these points before you publish a comparison

  • Use survey evidence for experiences covered by the sampled population.
  • Use recorded-crime evidence for offences known to police.
  • Use court or Ministry of Justice tables for convictions and sentences.

Sources and further reading

Editorial note

This guide explains statistics, not individual safety or legal liability. We update source links when official methodology changes.

Continue with a related crime-data guide