QUANTITATIVE RESEARCH METHODS WORKSHOP: Walter Mebane (U Michigan), “Election Forensics: Frauds in the 2015 Elections in Turkey?”
“Election Forensics: Frauds in the 2015 Elections in Turkey?”
Walter Mebane, Professor in the Departments of Political Science and Statistics, University of Michigan, Ann Arbor
Description: I review various methods for using polling station eligible voter and vote counts to diagnose frauds and other anomalies in elections, using the legislative elections in Turkey in 2015 as a primary example. The methods include statistical techniques implemented in the Election Forensics Toolkit, including especially a positive empirical model of election frauds (a kind of finite mixture model). I discuss the extent to which the methods may be able to distinguish consequences of strategic voter behavior from consequences of frauds.
Walter R. Mebane, Jr., is Professor of Political Science and Professor of Statistics at the University of Michigan, Ann Arbor. He works on political methodology and American politics, especially elections. His current primary project is Election Forensics, which aims to develop statistical and computational tools for detecting anomalies and diagnosing fraud in election results. He is writing a book on this topic. His work in this area includes papers about the 2000 presidential election, a report written for the Democratic National Committee analyzing the 2004 presidential election in Ohio, and analysis of election fraud in Russia and of likely fraud in the 2009 election in Iran. Recently with support from USAID he created the Election Forensics Toolkit, which is a prototype that implements various of the diagnostic tools, and applied the methods to vote count data and geocoded data from several countries.