Texto para Discussão, Instituto de Pesquisa Econômica Aplicada (IPEA) 1345
The present research aims to contribute to the area of poverty studies by appraising the potentiality of traditional research methods and the Structural Equation Modeling (SEM) as proper instruments for analyzing poverty situations according to a relational perspective. The theoretical concepts of poverty have gradually developed towards the idea that it is a complex and multidimensional problem, which involves interrelated social and economic factors. Empirical research methods about the subject, however, do not appear to reflect such complexity. Under this point of view, one can say that the main methodologies usually supporting poverty studies - such as synthetic indices, social-indicator systems and regression analysis - are not satisfactory ways of treating the phenomenon relationally. Considering the state of the art of quantitative methodologies for social research, SEM is regarded as an innovative method that can provide helpful ways of approaching social phenomena relationally. SEM is seen as the most advanced technique in terms of its ability to articulate social variables, as it allows for a simultaneous estimation of the relationships among all factors associated with a phenomenon. This paper examines SEM's analytical features as a relational method by using a conceptual approach, in which the traditional methods are also discussed, in order to theoretically appraise their limitations and potentialities as relational methodologies. The findings show that, indeed, SEM is a more powerful instrument for carrying out a relational analysis of poverty than the traditional quantitative methodologies, since unlike those, SEM is able to empirically reflect the articulate complexity of recent conceptual formulations.