Aplicação de dados em painel para tratar informações de mortalidade infantil

Abstract

This work aims to determine a linear or non-linear relationship between the Infant Mortality Rate (IMR) and a set of socioeconomic and coverage variables observed by the federative unit of Brazil from 2005 to 2010 using the panel data model (Panel Data ) linear and non-linear considering the fixed and random effects. The results show that the most suitable model is the fixed effect with logarithmic transformation in the dependent and independent variables. The dependent variable was the IMR and the independent ones: illiteracy rate, GDP per capita, proportion of people with low income, percentage of the population served by water supply network and the proportion of the population served by garbage collection. The nonlinear Binomial Negative models did not present satisfactory convergence in the estimation of the parameters for any model. The Poisson nonlinear model showed convergence only in the Pooled OLS and Randon (Random Effect Model) models. The conclusion obtained is that the independent variables that have a significant impact on IMR are illiteracy rate, per capita GDP and proportion of people with low income. R software was used to perform this data analysis with the plm and pglm packages.