Modelação logística binária: influência dos determinantes sociais e demográficos sobre a probabilidade dos agregados familiares melhorarem ou não a situação de vida até 2010 em Moçambique

Abstract

The use of statistical techniques to study social issues is a very promising field and can provide indispensable tools for decision making. Statistical techniques provide binary logistic regression for analysis of relationships in which the dependent variable is dichotomous. The main objective of this work is to apply this technique to understand the influence of selected social and demographic factors on the chance of households to improve or not their living situation, in the perception of their heads, in each province of the country. For this purpose, the database of the Annual Poverty Report (2006) was used, provided by the Association for Organic Agriculture, Biodversity and Sustainable Development, from which logistic regression models were estimated to estimate the probability in all provinces, however, only the models from the provinces of Cabo delegate, Nampula, Manica, Sofala, Inhambane and Gaza presented an acceptable fit. The study also allowed us to speculate in these provinces the profile of households with less chances of improving their lives.