3 - External Validation of Protest Event Analysis

Summary

Protest Event Analysis is attractive to social movement scholars because it is an unobtrusive technique that can cope with a large amount of unstructured data. However, it has also drawn some major criticisms such as its susceptibility to reporting bias. This chapter sets out to engage with these criticisms by discussing and defending our choice to rely on international news agencies publishing in English and by identifying the strengths and weaknesses of our dataset compared to two external datasets. On the one hand, we find that our dataset is more precise than the Integrated Crisis Early Warning System (ICEWS) data, which suffers from duplicated events and misclassifications of the protest form. As opposed to the ICEWS data, our dataset actually reveals the major waves of demonstrations across Europe. On the other hand, we evaluate our semi-automatically compiled data with help of data derived manually from national newspaper articles and news reports for Spain, Germany, the Netherlands, the UK, Hungary, and Poland. This comparison shows that our data have a slight tendency in the direction of covering larger events and events taking place in the capital more than other events.