Opening, Polling, Counting: Deterring Election-Day Fraud

International election observation has been scrutinized over the past two decades by scholars trying to explain its aim, scope, and consequences. As it is now a standard practice in democratizing countries, scholarly literature has almost reached a consensus on the ability of observers to deter election-day fraud, one of the most positive unintended consequences of election observation. Using data from the 2004 Ukraine presidential election, this article extends our knowledge through a natural experiment on polling station-level election results. I will show that observation during counting does not have a stronger effect than observation during opening and/or polling. Results support previous studies showing the usefulness of election observation in reducing election-day fraud but add an important finding that deserves further scholarly attention.