Mixed linear models as statistical tool for the analysis of the municipalities performance index of Goiás

Abstract

The objective of this paper is a methodological discussion regarding the application of Mixed Linear Models for the analysis of the Municipalities Performance Index (IDM) of Goiás for the years 2012, 2014, 2016 and 2018. Thereby, it was performed a literature review about of Mixed Linear Models, as well it was described the context that permeated the application of the model that concerns the socioeconomic situation of the cities of Goiás in the selected years. In addition to the literature review, data were collected from Instituto Mauro Borges Statistical Database of Goiás. The results indicated that the mixed linear models adequately modeled the effects of covariates on the IDM. It is concluded that in database which the same experimental units are collected more than once, there is a need for statistical modeling that can capture variance between and within groups. Thus, in the analysis of the IDM, the variables budget balance, GDP per capita, telephone and internet infrastructure, distribution of doctors per thousand inhabitants and formal employment had positive impacts on the Municipalities Performance Index of Goiás. On the other hand, enrollment in school units negatively impacted the IDM.

Keywords: Mixed Linear Models. Municipalities Performance Index of Goiás. Regional development. Longitudinal Data.