Klimek et al (2012) Statistical detection of systematic election irregularities.pdf

Statistical detection of systematic
election irregularities
Peter Klimeka, Yuri Yegorovb, Rudolf Hanela, and Stefan Thurnera,c,d,1
aSection for Science plex Systems, Medical University of Vienna, A-1090 Vienna, Austria; bInstitut für Betriebswirtschaftslehre, University of Vienna,
1210 Vienna, Austria; cSanta Fe Institute, Santa Fe, NM 87501; and dInternational Institute for Applied Systems Analysis, A-2361 Laxenburg, Austria
Edited by Stephen E. Fienberg, Carnegie Mellon University, Pittsburgh, PA, and approved August 16, 2012 (received for review June 27, 2012)
Democratic societies are built around the principle of free and fair as a fraud detection method is subject to controversy (10, 11).
elections, and that each citizen’s vote should count equally. Na- The problem is that one needs to firmly establish a baseline of
tional elections can be regarded as large-scale social experiments, the expected distribution of digit occurrences for fair elections.
where people are grouped into usually large numbers of electoral Only then it can be asserted if actual numbers are over- or un-
districts and vote according to their preferences. The large number derrepresented and thus, suspicious. What is missing in this con-
of samples implies statistical consequences for the polling results, text is a theory that links specific fraud mechanisms to statistical
which can be used to