DS Crime > Information on data > EN > Revamp

Important note to users

When analysing these statistics, please keep in mind that the data refer only to crimes recorded by authorities and consequently reported to the police by victims and witnesses, among other things. Inferring crime occurrence from official crime figures can therefore be misleading. Furthermore, definitions and counting of official crime vary between countries, and comparisons between countries can therefore be misleading.

Available time series

Crime statistics have been published by Eurostat since 2008. In addition, some historical data are available for the period 1993-2007 (number of police-recorded crimes by type, police-recorded homicides in largest cities, police officers, prison population) and for 1950-2000 (total number of police-recorded crime).

Data collection and checks

Authorities in each country produce the basic data. These include data on police, prosecutors, courts, and prisons. Each country establishes its own criminal laws, crime definitions, law enforcement, legal proceedings, and legal response. All these elements impact the national crime statistics.

By agreement, EU and EFTA countries as well as candidate countries and potential candidates send their official crime figures to Eurostat. In practice, the figures from the different authorities are collected into one file for each country. This is done once a year, together with the Survey on crime trends of the United Nations Office on Drugs and Crime.

No records of individual crimes or criminals are sent to Eurostat.

The figures requested are the annual totals, in absolute numbers, for each country. Following a checking procedure, some countries are asked about figures that have unexpected values. When the figures are confirmed or corrected, they are uploaded into the Eurostat database.

In addition, Eurostat asks all the national authorities for additional information on how the figures are calculated and whether these are compliance with definitions and counting rules. This information is included in the metadata file  Crime and criminal justice, but is not used to adjust the data.

Eurostat doesn't carry out any editing, estimation, imputation, weighting, seasonal adjustment, or other statistical modifications.

Comparisons

Comparing countries and years

Directly comparing crime figures between countries may be irrelevant or invalid, resulting in misleading inferences or incorrect conclusions. This is because criminal justice systems, crime definitions, and crime statistics can vary substantially between countries. A change in a crime figure from one year to the next does not necessarily imply a change of crime levels. The underlying cause could be a change in the response of the law enforcement agencies towards the crime.

As a general rule, comparisons should focus upon trends rather than on levels. This recommendation is based on the assumption that the characteristics of the recording system of a country remain fairly constant over time. However, there could be exceptions as methods might change which will cause breaks in time series. These are indicated by the flag ‘b’ for values in the datasets.

Comparing crime figures between countries or years may be affected by differences or changes in:

  • legal systems and criminal justice systems
  • legislation, criminal law, and legal definitions
  • the efficiency of police, prosecution, courts, and prisons
  • recording practices (input, process, output)
  • recording, reporting, and production systems
  • national crime definitions and international statistical definitions
  • statistical units and statistical populations
  • statistical definitions, reference times, counting methods, and calculation methods
  • crime reporting rates

Users are strongly advised to consult the metadata for each country to obtain information about definitions, counting rules for offences and persons, and other aspects.

Comparing small figures

‘Small figures’ means:

  • low frequency (absolute number of crimes)
  • low rate (crimes per 100 000 people)
  • small proportion (e.g., % of women)
  • small change (from one year to the next)

Total figures from complete administrative records are considered precise and don't have sampling errors, as is the case for survey-based statistics. However, there may be other random errors in the data that can make small figures unreliable.

Some relative changes from one year to another may be substantial simply because the absolute figures are very small.