There are many ways to manipulate the results of a vote, and some of them can be detected with statistical tools, as a team of scientists has shown in a work published in the journal PNAS.
Researchers have developed a mathematical model that detects two types of voter fraud. The first case, which they have called ‘gradual fraud’, occurs when ballots of one candidate are added and those of others are withdrawn. The second, called ‘extreme fraud’, takes place when 100% of the electoral participation collects the majority of the votes in favor of a specific party or candidate. After applying the model to electoral results data from different countries, the scientists observed thatsome statistical characteristics were repeated in all fair electoral processes. “When these statistical properties are not observed, one can intuit the presence of something incorrect within the electoral process,” Stefan Thurner, one of the authors, from the Vienna Medical University, explained to SINC. Using this information, the researchers have developed a test to detect electoral fraud.
On the other hand, the authors have warned that the fact of observing these anomalies does not necessarily imply the presence of fraud, since sometimes the test detects strange parameters in fair electoral processes. However, the test may indicate cases where “should be carefully analyzed if there is anything strange in the process“Thurner explained. Those responsible for the study believe that the model can help make elections more fair in democratic countries and believe that” organizations such as the Organization for Security and Cooperation in Europe or the United Nations should create a website where are the electoral data of each nation “.
