Did police activity reduce crime?
by
Marilyn Chilvers
NSW Bureau of Crime Statistics and Research

In this presentation, some data issues and methodological challenges faced by a statistician working in the government sector are highlighted.

In April 2000, the NSW Bureau of Crime Statistics and Research publicly reported the first substantial drop in crime in NSW in more than a decade. Reactions to the news varied in the significant public and political scrutiny that followed. Were NSW Police justified in claiming credit for the crime reduction? Or could this be a sign that citizens had simply lost faith in police and stopped reporting crimes? Or, worse, were police not recording the crimes reported to them so that crime trends would look better? We decided to investigate.

The results of the two major projects that comprised our research are presented in this paper. Using simple statistical procedures with data that were readily available, we undertook an analysis of the relationship between crime trends and policing (and potential under-reporting or under-recording). Firstly, using alternative data sources, we show that the police recording of offences are verifiable. Secondly, using interrupted time series analysis, we examine the causal relationship between arrests and crime, controlling for other factors shown in the literature to affect crime trends. We find that there is some justification for the claim that police activity contributed to the crime reduction in NSW.

Date received: April 3, 2002