Identify prior-to-subsequent event effects in large conflict event corpora

Determine, in large corpora of international conflict event data, how specific prior events affect the probabilities of subsequent events, in order to identify the sequential dependencies that drive escalation and de-escalation processes.

Background

The paper argues that existing events datasets in international relations have struggled to reveal the sequential "begats"—how earlier events shape later ones—despite large volumes of coded events. Researchers often resort to binning events by counts and time windows, which obscures causal and predictive linkages between specific event types over time.

The authors highlight that without understanding how prior events influence subsequent events, it remains difficult to adjudicate between theoretical frameworks of escalation (e.g., spiral vs. deterrence) and to diagnose crisis processes. Pinpointing these dependencies is therefore crucial for moving from descriptive event streams to explanatory models of crisis dynamics.