How effective are policy interventions to fight crime and how valid is the policy theory that underlies them? This is the twofold research question addressed in this article, which presents an evidence-based evaluation of Dutch social safety policy. By bridging the gap between actual effects and assumed effects, this study seeks to make fuller use of the practical relevance of evidence-based evaluations. The results reveal promising interventions and mechanisms for policy practice. In addition, the chosen approach advances current practice using evidence-based outcomes to distinguish plausible from implausible policy assumptions. An urgent need is signaled for governments to provide substantive reasoning for their policy choices. Many of the assumptions that correspond with evidence from impact evaluations relate to general prevention, often found within the strategy of situational prevention. Contrary to any general preventive or short-term effects, the assumptions regarding long-term specific prevention through law enforcement cannot be taken as read.