Abstract
When comparing crime rates across several locations or analysing general trends, it is usually assumed that police statistics are an imperfect, but relatively constant indicator of the phenomenon. This study delves into the possibility that this may not be the case. Multilevel models are used to analyse the recording rate of six types of offences for the 33 police districts of a large Canadian city between 2005 and 2009. Results demonstrate that geographical and time factors unrelated to the offence committed may influence a police officer's decision whether or not to record certain types of offences. In some cases, official statistics do not provide a meaningful indicator of crime for comparative purposes.
Get full access to this article
View all access options for this article.
