Unifying casualty distributions within and across conflicts

Area of Science:

• Quantitative Sociology
• Conflict Studies
• Network Science

Background:

• Human conflict, including wars and individual violent events, exhibits complex size distributions.
• Previous research suggests power-law distributions for event sizes, but a comprehensive understanding across scales is lacking.

Purpose of the Study:

• To investigate the relationship between the size distributions of individual war events and entire wars.
• To determine if a unified mathematical framework can describe conflict sizes across different scales.

Main Methods:

• Analysis of fatality data to measure the 'size' of wars and events.
• Application of power-law statistical models to describe size distributions.
• Numerical simulations to explore the effect of aggregating event sizes into whole wars.

Main Results:

• Both whole war sizes and individual war event sizes are well-approximated by power laws.
• The power-law exponent for whole wars is smaller (flatter distribution) than for individual events.
• A simple aggregation process mathematically links the different power-law exponents.

Conclusions:

• The observed differences in power-law exponents are an artifact of aggregating smaller events into larger wars.
• This finding supports a unified, scale-invariant description of human conflict.
• Understanding conflict dynamics across scales is crucial for effective analysis and intervention.