13/08/2020

IMPA helps TSE guarantee sanitary safety in elections.

Foto: Marcelo Camargo/ Agência Brasil

Ensuring sanitary safety during the voting days of the 2020 municipal elections, scheduled for November 15th and 29th, is also a mathematical task. IMPA researchers Roberto Imbuzeiro and Paulo Orenstein are working with the statistics department of the Superior Electoral Court (TSE) on a project that seeks to define voting times, attempting to minimize the spread of the new coronavirus among poll workers and voters. The project also includes researchers from the Institute of Education and Research (Insper), the University of São Paulo (USP), and the Oswaldo Cruz Foundation (Fiocruz).

During the first stage, which began on Monday (3), researchers will evaluate and improve the scenario simulations that were carried out by the TSE statistics group. Among the measures studied are the possibility of extending voting hours by at least one hour and the recommendation that the first three hours of voting, from 8 am to 11 am, be reserved for people over 60 years of age, considered a risk group for Covid-19.

They work with a database from the TSE (Superior Electoral Court) that gathers information from all electronic voting machines in Brazil from all elections. The voter is not identified, but it is possible to know how long each person takes on average to vote, the time each person takes to register before voting, and the average time each machine remains idle.

“The voter is sovereign, so it’s not possible to determine that each group votes at a specific time. The most that can be done is to make recommendations.” Furthermore, each municipality and electoral zone has its own demographic particularities. The IMPA researcher cites as an example some polling stations in Bahia where 90% of voters are over 60 years old. “In these cases, it’s pointless to try to encourage the elderly population to vote at separate times.”

After the first stage, the TSE (Superior Electoral Court) should provide an opportunity for the research group to begin developing a new method that will analyze voter behavior trends during elections more deeply through modeling, statistics, and probability.