This study examines the effect Project Green Light Detroit (PGLD), an integrated CCTV program, on crime at commercial and non-commercial city parcels in Detroit, MI.
Methods
A quasi-experimental design was used by implementing a difference-in-differences model with adjustments made for variation in treatment timing and treatment heterogeneity.
Results
Findings from the study indicate that PGLD increased reporting of property crimes at some participating locations but did not significantly impact violent or disorder crimes. Most of the impact of PGLD was attributable to locations that joined the program early in its implementation.
Conclusions
Studies examining treatment effects that are implemented over time should adjust for variation in treatment timing and treatment heterogeneity. Several new statistical methods exist that can implement these in a variety of software packages.