We implement a spatial probit model to differentiate states with a lottery from those without a lottery. Our analysis extends the basic spatial probit model by allowing spatial dependence to vary across geographic regions. We also separate the spatial effects of neighbors versus non-neighbors. The methodology provides consistent and efficient coefficient estimation in light of the simultaneity in spatial dependence. We find evidence of spatial dependence and spatial heterogeneity in lottery usage, and we find that spatial patterns differ significantly by geographic region. The importance of spatial dependence in state lottery usage suggests the need to consider spatial effects in empirical models examining the use of any policy tool by subnational governmental units.