Logo: to the web site of Uppsala University

n this article we present a new dataset: the Early Conflict Prevention in Ethnic Crises dataset (ECPEC). It contains data on operational conflict prevention in 67 ethnic crises in the period 1990—98 that vary in terms of both preventive action and crisis outcomes. The new dataset thus allows for the evaluation of the effects of different types of preventive measures and also gives an overview of who takes what measures and in what conflicts. The global overview shows some interesting patterns. Preventive activity in the escalatory phase of ethnic conflict is dominated by verbal attention and facilitation. Coercive measures are rarely employed prior to the outbreak of war. Preventive action is most common in Europe and the Middle East, while crises in Asia tend to receive comparatively little attention. Most of the preventive action is focused on a relatively small number of high profile cases like those in Northern Ireland, the Balkans, Afghanistan, and Iraq. Major Powers (with the exception of China), neighboring states, the UN, and regional organizations are the most active interveners. To illustrate the usefulness of a large-N dataset on preventive measures, we also present a first analysis of the effects of different types of measures. The findings suggest that diplomatic measures and relief efforts both have conflict dampening effects, while carrots (inducements) increase the likelihood of escalation to war. Other measures show no significant effects in this sample. The findings also show that third parties are more likely to intervene in conflicts that are more prone to escalate to war. This implies that unless we account for the propensity of third parties to intervene in the more difficult cases, we risk underestimating the effects of preventive measures.