Methods: Data from the MacArthur mental health court study was analyzed. The data included approximately 4,000 residential information from 741 offenders with mental illness from four counties in CA, MN, and IN. Participants were interviewed twice (baseline and six months) and objective arrest data were obtained 12-18 months pre- and post-court involvement. Vegetation data were derived from classification from remote sensing satellite images. The overall percentage of tree canopy of each block group were derived by averaging the fractions of all pixels located within a block group and then linked through GeoID. To control for the excessive zero in objective arrest data and different level of data structure (i.e., individual factors as level 1; vegetation as level 2; site as level 3), multilevel negative binomial regression was conducted to examine the effect of vegetation on a number of arrests. Three different models were tested (i.e., Model 1 with vegetation only, Model 2 with individual and vegetation, Model 3 with individual, vegetation, and site).
Results: Offenders without SMI had .33 times arrests, while offenders with SMI had .36 times arrest during a 12-month period. Bivariate analyses indicated that study participants without SMI resided in more vegetated areas (6.31%) compared to those with SMI (5.71%), but they were not statistically different. Multilevel negative binomial regression indicated that there is no significant effect of vegetation on arrests among offenders without SMI across study models. However, for those with SMI, a significant variation of vegetation on arrests was found among all study models (p<.001).
Conclusion and Implications: Findings suggest that vegetation abundance is significantly associated with the number of arrests, indicating that vegetation has a positive effect on reducing the criminal behaviors only among offenders with SMI. Given the fact that practice and policy almost exclusively focus on individual factors of crime, implications include a re-evaluation of environmental factors (e.g., vegetation) associated with crime to incorporate comprehensive methods of crime prevention.