Near repeat victimisation – the observation that
future criminal incidents are located close in
space and time to previous incidents – has been
demonstrated for burglary, car crime, gun
shootings and insurgent activity. However these
analyses have been largely aspatial and atemporal
with scant attention paid to the spatial and
temporal distribution of these close pairs (but
see Johnson and Bowers (2004) for an exception).
This study attempts to discover whether patterns
exist when near repeats are mapped. For
instance, are near repeats uniformly distributed
across space or are they linked in some
configured to form chains? In addition, we are
interested in whether mapping near repeats
reveals insightful pattern not available from studying all incidents.
Graph theory is used to model events (connected
on the basis of their temporal and spatial
proximity to each other). A range of descriptive
statistics are computed at the node, chain and
network levels which are then compared to a
derived expected distribution. Certain node
features were observed at very different levels
than would be expected on the basis of
chance. The practical implications of these
patterns are articulated, as is a method of modelling node attributes.
Audio (MP3 File, Podcast Ready)
Copyright. All Rights Reserved.