time series

Why you can't identify changes in crime by comparing this month to last month

This post first appeared on the Social Research Association blog. Understanding if some event is associated with a change in the frequency of crime is a common question, both in practice and in research. A crime analyst might need to understand the impact of a change in tactics, a local journalist might want to check the truth of the mayor’s claims that her policies are working, while an academic might seek to find patterns associated with trends in socio-economic changes.

Chart: Stop and search has fallen, arrests from searches less so

Police can search people they suspect of having prohibited items such as drugs or weapons. Since 2011, searches have decreased by 70%. Only 12% of searches lead to arrest, but this is increasing – total arrests from searches have decreased only 38% since 2011, suggesting police are getting better at targeting offenders. larger image | annotated R code to produce this chart Data source: Home Office, 2019

Chart: Stop and search has fallen, arrests from searches less so

Police can search people they suspect of having prohibited items such as drugs or weapons. Since 2011, searches have decreased by 70%. Only 12% of searches lead to arrest, but this is increasing – total arrests from searches have decreased only 38% since 2011, suggesting police are getting better at targeting offenders. larger image | annotated R code to produce this chart Data source: Home Office, 2019