Banca de QUALIFICAÇÃO: KELLY CHRISTINA DA SILVA MATOS

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
STUDENT : KELLY CHRISTINA DA SILVA MATOS
DATE: 25/04/2025
TIME: 08:00
LOCAL: Videoconferência via Gerência de Redes do CCET/UFRN
TITLE:

Analytical Triad in Violence Against Women: A Multimethodological Approach with VIVA/SINAN Data from Brazil to Rio Grande do Norte

KEY WORDS:

Violence Against Women; Data Analysis; Bayesian Networks; Logistic Regression; VIVA System (DATASUS).

PAGES: 69
BIG AREA: Ciências Sociais Aplicadas
AREA: Demografia
SUBÁREA: Política Pública e População
SUMMARY:

This doctoral thesis in Demography investigates violence against women in Brazil through three empirical studies, using data from the VIVA System (DATASUS) from 2009 to 2022. The articles analyze the phenomenon at different geographical scales—national, regional, and state—employing various statistical methods to identify patterns and factors associated with gender-based violence. The first article examines violence against women across Brazil’s five regions using logistic regression, identifying that sexual and psychological violence, as well as threats, are more frequent among women, particularly in intimate partner relationships. In contrast, violence involving firearms and sharp objects is more commonly associated with men. Despite socioeconomic and cultural differences between regions, a recurring pattern of violence against women is observed, reinforcing the need for integrated national public policies. The second article focuses on the Northeast region, conducting a descriptive and exploratory data analysis. The results reveal that nearly 80% of assaults occur in domestic settings and highlight racial disparities in victimization, as well as the high recurrence of violence—30% of victims have experienced aggression before. The third article takes a predictive approach, applying Bayesian networks to data from Rio Grande do Norte to forecast the occurrence of physical violence against women. The modeling shows that psychological violence is the main predictor of physical violence, with a conditional probability of 84%. Other relevant factors include intimate partner relationships and alcohol consumption. The findings of this thesis underscore the persistence of gender-based violence in Brazil and the importance of preventive strategies and specific public policies to mitigate its impact, considering the regional and structural particularities of this phenomenon.

COMMITTEE MEMBERS:
Externo à Instituição - Eli Narciso da Silva Torres - UNICAMP
Presidente - 1225734 - JORDANA CRISTINA DE JESUS
Interna - 1855608 - KARINA CARDOSO MEIRA