Determinants of blank and null votes in Brazilian presidential elections

• Grant number: 24/18409-7
• Support Opportunities: Scholarships in Brazil - Program to Stimulate Scientific Vocations
• Start date: February 10, 2025
• End date: March 29, 2025
• Field of knowledge: Physical Sciences and Mathematics - Probability and Statistics - Statistics
• Principal Investigator: Renata Rojas Guerra
• Grantee: Ryan Novaes Pereira
• Host Institution: Centro de Ciências Naturais e Exatas (CCNE). Universidade Federal de Santa Maria. Santa Maria , SP, Brazil

Abstract

The objective of this study is to identify the factors that influence the proportions of blank and null votes in Brazilian municipalities during presidential elections. The behavior of the variable of interest will be described, and unit regression models will be adjusted. The study on unit regression models will be conducted using the Generalized Additive Models for Location, Scale, and Shape (GAMLSS) framework for beta, Simplex, Kumaraswamy, unit Weibull, and reflected unit Burr XII regressions, considering submodels for both parameters indexed to these distributions. After selecting the best model for each class and applying diagnostic techniques, the most suitable random component and significant factors will be identified to explain the proportion of votes. Regarding the factors to be tested, the effects of the Municipal Human Development Index (HDI), disaggregated into the dimensions of income (HDI Income), longevity (HDI Longevity), and education (HDI Education), as well as the region to which the municipality belongs, will be evaluated. Demographic variables of the municipalities will also be tested. (AU)