9.2 This chapter examines the state of crime in the South West by considering both crimes recorded by the police and the results of the BCS. These two datasets each have important advantages, and when they are examined in conjunction they are especially strong, but it is important to understand the differences between them. This is outlined in section 9.2.1.
Using Data: Recorded Crime and the British Crime Survey
9.2.1.1 Recorded crime figures are compiled from police returns to the Home Office and are the most readily available measures of the incidence of crime. However, they do not necessarily indicate the true extent of crime: offences that are recorded by the police are merely a subset of crimes that are reported to them and those are a subset of the crimes that actually occur (since many crimes are never reported to the police). Home Office research suggests that only around 41% of crimes are reported to the police.
9.2.1.2 The decision to report an incident to the police depends on a number of factors, typically influenced by the victim's personal circumstances, attitude to the police and the type of crime. The BCS, by contrast, counts crimes experienced by interviewees in the 12 months prior to interview, regardless of whether they have been reported to or recorded by the police. It is a victimisation survey in which adults aged 16 years or over and living in private households in England and Wales are asked questions about offences against their household (such as theft or damage of household property) and about personal offences (such as assault). The survey also explores people's perceptions of crime and anti-social behaviour, contact with the police and drug misuse. However, certain crimes such as murder, where the victim cannot be interviewed, and 'victimless crimes' (e.g. possession of drugs), are excluded from the BCS. Equally since the BCS is a measure of crime against private households, commercial crime is excluded. The survey was also extended in January 2009 to include 4,000 interviews with children aged 10-15, although it is currently too early for any results from these interviews to have been made available.
9.2.1.3 The BCS estimates are subject to sampling error, which means that differences between estimates from successive years of the survey or between population sub-groups may occur by chance. Tests of statistical significance are used to identify which differences are unlikely to have occurred by chance. In this chapter where significance levels are mentioned this is at the 5% level (the level at which there is a one in twenty chance of an observed difference being solely due to chance).
9.2.1.4 Volume of crime is related to, amongst other things, the size of an area's population, as the more people that live in an area, the more potential victims and/or offenders there could be. Police recorded crime statistics record where an offence takes place and to standardise the data for differences in population size, crime data is expressed as a rate per 1,000 resident population. However, rates that are based on resident populations can overestimate levels of crime in areas where a large number of people visit from outside (e.g. seaside towns such as Bournemouth during the summer months).
