Replication Data for: Election Fraud: A Latent Class Framework for Digit-Based Tests

Digit-based election forensics typically relies on null hypothesis significance testing, with undesirable effects on substantive conclusions. This paper proposes an alterna- tive free of this problem. It rests on decomposing the observed numeral distribution into the ‘no fraud’ and ‘fraud’ latent classes, by finding the smallest fraction of nu- merals that either needs to be removed or reallocated to achieve a perfect fit of the ‘no fraud’ model. The size of this fraction can be interpreted as a measure of fraud- ulence. Both alternatives are special cases of measures of model fit–the π∗ mixture index of fit and the ∆ dissimilarity index, respectively. Furthermore, independently of the latent class framework, the distributional assumptions of digit-based election forensics can be relaxed in some contexts. Independently or jointly, the latent class framework and the relaxed distributional assumptions allow to dissect the observed distributions using models more flexible than those of existing digit-based election forensics. Reanalysis of Beber and Scacco’s (2012) data shows that the approach can lead to new substantive conclusions.