Predicting non-state terrorism worldwide

Area of Science:

• Political Science
• Criminology
• Data Science

Background:

• Non-state terrorism causes thousands of global deaths annually.
• Accurate local, short-term terrorism predictions are crucial for policymakers.
• Existing models often rely solely on historical event data.

Purpose of the Study:

• To develop and validate predictive models for local non-state terrorism.
• To assess the efficacy of theoretically informed predictors versus historical data alone.
• To identify key drivers of terrorism at local and regional levels.

Main Methods:

• Utilized publicly available data for model development.
• Incorporated structural and procedural variables into predictive models.
• Compared performance of theoretically informed models against historical event-based models.

Main Results:

• Models with structural and procedural predictors accurately forecast local non-state terrorism a week in advance.
• Theoretically informed models consistently outperformed models based only on past terrorist events.
• Identified and interpreted significant local drivers of global and regional terrorism.

Conclusions:

• Theoretically informed models offer a powerful tool for predicting political violence.
• Publicly available data can be leveraged for effective, policy-relevant terrorism forecasting.
• This approach enhances understanding and mitigation of non-state terrorism.