A Case-Control Method for Studying Protest Participation and Other Rare Events: An Application to Ukraine’s EuroMaidan

• In Event: Politics of Opposition

Abstract

Abstract Studies of individual protest participation confront a variety of inferential challenges. Representative surveys capture few protest participants, are biased by respondent recall, and provide only post-hoc measures of other covariates. Surveys of protesters offer a larger sample size, minimize problems of recall, and effectively verify participation. However, they have limited utility for understanding the causes of protest participation, because focusing on protesters introduces selection on the dependent variable. In this paper I show how a variant of the standard case-control design, used in individual-level rare events studies in epidemiology but ignored to date in political science, enables researchers to estimate the probability of protest as a function of individual-level characteristics. In this approach, researchers combine two distinct samples---one where the outcome is measured along with relevant covariates and the other where relevant covariates are measured but the outcome is not. After describing the statistical setup for this design, I demonstrate its value by applying it to data from Ukraine’s EuroMaidan protests. The results show that the group which occupied the square during the protests final phase was strikingly different than the earlier participatory core of nonviolent activists. They also reveal that western Ukrainian opponents of an EU agreement were more highly mobilized than western Ukrainian supporters in the weeks leading up to Yanukovych’s ouster.

Author

• Bryn Rosenfeld, University of Southern California