Abstract
The challenge of automatically screening for potential school shooters involves several difficulties. In this paper, we present a simple and interpretable methodology for screening for potential school shooters through (1) the psychological textual signature of the shooter and (2) Jaynes approach for measuring the weight of evidence. We have tested our proposed approach on a dataset of texts written by shooters and non-shooters alike (N = 5047). Our major finding is that the methodology can successfully support the screening for potential shooters in interpretable way. The major implication for stakeholders is that there is great potential in developing screening systems for improving the safely of schools. However, developing such a system is a project that must be actualized within an integrated system of "command and control".