Banca de DEFESA: GUILHERME LIRA RAPOSO

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
STUDENT : GUILHERME LIRA RAPOSO
DATE: 08/04/2026
TIME: 14:30
LOCAL: Google Meet
TITLE:

Regression Models for Global Religious Disbelief Rates

KEY WORDS:

Religious Disbelief Rate; Regression Models; GJS Regression; PL Regression;
Intelligence Quotient.

PAGES: 60
BIG AREA: Ciências Exatas e da Terra
AREA: Probabilidade e Estatística
SUMMARY:

The general factor of intelligence, known as the factor g, was proposed by Spearman
(1904) as a common underlying cognitive ability present in all intellectual tasks. This concept
underlies measures such as the intelligence quotient (IQ), used to represent the average
cognitive ability of populations. Among the studies in this field, the research of Cribari-Neto e
Souza (2013) stands out, which employs the beta regression model to analyze the rate of
religious disbelief across several countries, using average IQ as the explanatory variable, with
data extracted from Lynn, Harvey e Nyborg (2009). In this work, we investigate alternative
regression models belonging to the Generalized Johnson SB class proposed by Lemonte e
Bazán (2016) and the Power Logit model by Queiroz e Ferrari (2023) to conduct this analysis. In
this sense, the present study seeks, among all the models presented and adapted to handle
proportions, the one that best fits the modeling of atheism rates as a function of the average
IQ of countries, aiming to identify the model with the best goodness of fit to represent this

relationship, thereby contributing to the use of unconventional models for the analysis of
social data expressed as rates and proportions.

COMMITTEE MEMBERS:
Presidente - 2310429 - ARTUR JOSE LEMONTE
Externa ao Programa - 2083593 - LUZ MILENA ZEA FERNANDEZ - nullExterna à Instituição - MARIANA CORREIA DE ARAUJO - UFPB