- Does past light crime (i.e. one broken window) lead to more light crimes (i.e. more broken windows)?
- Does past crime (i.e. one broken window) lead to more severe crimes?
They show that spillovers from light crime to more severe types of crime are quite limited empirically, so if the goal is to reduce severe socially costly crimes, then resources are best spent directly targeting those crimes (rather than trying to crack down on less severe crimes). In short, they suggest that police should focus on preventing murders and rapes rather than broken windows.