Part of why we struggle to understand school shootings is because there isn’t enough data available about these extremely rare events. A recent study published in Nature Human Behavior describes a carefully curated dataset for school shootings between 1990-2013, created from existing data and original data sources.

In their analysis, the authors of this paper found that the rate of school shootings increased from 2007-2013. They also found data that suggested increased shooting rates were correlated with increases in unemployment rates. This finding indicates that high levels of economic distress may lead to increases in school-related gun violence.

Previous research on school shootings has resulted in contradictory claims because there hasn’t been a single, coherent dataset. Instead, multiple datasets with different inclusion criteria have made the resulting findings difficult or impossible to compare since they analyze fundamentally different information. To solve this problem, the authors made a new collection of school shooting data, which resulted in the inclusion of almost 400 events. Their criteria for inclusion in this dataset are: