<< Fair and Balanced | Main | DNA Sampling of Arrestees >>

Crime Statistics and Unions

One of the big problems of doing research with people (unlike, say, electrons) is that you cannot neatly control all the variables that might affect your result.  In crime research, the outcome variable is usually the official crime statistics.  The input variables we want to study are usually such things as sentencing policy, number of police on the beat, economic factors, and things like that.  When something else alters the official crime number, unknown to the researcher, that can really mess things up.

From the June issue of Police Quarterly is a study on an "error term" that would never have occurred to me:

This article examines the effect that collective bargaining had on official crime statistics compiled between 1998 and 2009 in Montreal, Canada. Police officers collectively decided to reduce their use of record-discretion on two occasions, to increase administrative workload and pressure their employers. Considerable increases of recorded assaults and mischief were observed, despite no apparent variation of reported infractions measured by calls for service. Recorded and reported car thefts and burglaries displayed no significant variation. Results suggest that observed variations can be explained in terms of temporary differential treatment of specific incidents rather than increases of reported criminality or proactive police activity. This research challenges the reliability of official statistics as measures of crime and demonstrates that external circumstances can influence police recording practices.

Leave a comment

Monthly Archives