La potencialidad de la Regresión Logística Multinivel. Una propuesta de aplicación en el análisis del estado de salud percibido

Resumen

The development of different software is facilitating the extension of multilevel models in different disciplines of social sciences in recent years. These quantitative techniques help us to observe characteristics of individuals and the contexts in which they interact in the same analytical model. It let to regard the effect of several explanatory variables on a phenomenon considering both the differences between individuals and the available information that they are grouped above in different hierarchical. Although initially it’s thrived in models of linear dependent variables, it could also be used in models with categorical outcomes. The Multilevel logistic regression (MLR) can facilitate estimation of complex models with a dichotomous categorical dependent variable type. The current article is a brief explanation of this technique. We seek to indicate that it can be used in different areas of sociological analysis, especially in inferential statistics. Despite its apparent complexity, thanks to the recommendations of some authors, we can utilize in a feasible and to facilitate the provision of information on the association of variables and the context of a phenomenon. In the second part of this text we offer an adjusting exercise of multilevel logistic models. The intention was to observe the incidence of individual variables such as working conditions, occupational conditions and the domestic-family demands that working people have, together with variables of the labour market and social protection of the 27 EU members on perceived health status. The main results indicated already that the worst working conditions and temporary employment were associated with poor health with significant gender inequalities between countries. These conclusions were noted by some occupational health studies. This study allowed going beyond the individual analysis and observing that a percentage of the differences between employees can be explained by different variables of welfare between countries.