This research studies protests after fraudulent elections in a collective action framework, examining the impact of the potential cost, benefit and likelihood of success of protest on the occurrence and intensity of protests. Quantitative analysis of fraudulent elections in about 100 countries from 1990 to 2004 shows that the odds of protest after fraudulent elections are greater when the level of state repression is moderate with a possible backlash effect of high repression, when the opposition is united, and when international monitors denounce election results. The analysis only partially supports the benefit of protest argument. Also, the research uses case studies from Eurasia (Armenia, Belarus, Georgia, and Russia) and mini-case studies from Africa and Latin America to study in more detail the effects of the factors identified in the quantitative analysis and to identify overlooked but important explanatory factors using a set of extensive interviews conducted in the United States and during fieldwork in Armenia, Georgia, and Russia with politicians, domestic and international election monitors, and country experts.