Data and Approach

• Data: Combined Europe Media Monitor (EMM) (134 million articles, 466 sources across 123 countries) with EM-DAT disaster database.
• Dyadic Dataset: Build a cross border daily panel of country mentions as a measure of cross border media coverage between country pairs.
• Event Study: Econometrically estimate, for each unique natural disaster, an event study to identify traces of differential reporting from media across the world on a country affected by a natural disaster within a few days of disaster start.
• Focus analysis: Study what drives the variation in the estimated reporting increase across country pairs.
• Key Variables: Explore patterns in detected increases in reporting focusing on disaster type, fatalities, duration, and measures of country-pair connections (social ties, cultural overlap, etc.).

Headline findings

• Skewed reporting by natural disaster type: Reporting increases are most notable for particularly “newsworthy” disasters (earthquakes, accidents, wildfires) vs. climate-change-induced disasters (storms, floods, extreme temperatures).
• Death gradient: Reporting increases strongly informed by the number of fatalities attached to a disaster.
• Country connectedness matters interactively: Measures of country connectedness primarily shape reporting increases in interaction with the number of fatalities.

Simulation Approach: Select a disaster country and disaster type, then a focal country to see how “equivalent attention” patterns unfold. This asks what number of hypothetical fatalities in a focal country would trigger similar reporting increase compared to a baseline (e.g., 100 fatalities) in the disaster country.

Once all three selections are made, the map and chart below will highlight what number of fatalities in each region would generate a similar amount of attention in the focal country's media compared to the baseline disaster of 100 fatalities.