Résumé
"The study of crime and place recognizes the important interplay between the physical landscape and criminal activity. In doing so, research in this area has shown substantial concentrations of crime amongst micro geographic units, such as street blocks. Despite these revelations, little research has examined whether such criminal concentrations persist over time. The developmental trajectory of criminal activity on street blocks was originally studied in Seattle, Washington. This dissertation replicates that seminal study by examining crime volumes on the streets of Vancouver, British Columbia, over a 16 year period using a group-based trajectory model (GBTM). Going further, this research also applies a non-parametric technique, termed k-means to address various limitations inherent to the GBTM method. The major findings reveal the majority of street blocks in Vancouver evidence stable crime levels, with a minority of street blocks throughout the city showing decreasing crime trajectories over the 16 year period. Both statistical techniques found comparable patterns of crime throughout Vancouver. A geographic analysis of the identified crime trajectories revealed linear concentrations of high, medium and low decreasing trajectories throughout the city, with the high decreasing street blocks showing particularly visible concentrations in the northeast part of Vancouver. Overall, the results confirm the original conclusions from the Seattle study in that many street blocks evidence significant developmental trajectories of crime and that the application of trajectory analysis to crime at micro places is a strategically useful way to examine the longevity of crime clusters. The results did not support the existence or stability of bad areas, but did find ‘bad streets’. It is recommended that police and public safety practitioners pay close attention to the varying levels of criminal activity on street blocks when developing place-based crime prevention initiatives."--Abstract.