Desempenho na educação básica a partir das Capacidades Estatais dos municípios goianos

Abstract

How can we universalize the assessment of State Capabilities of municipal bureaucracies that are endowed with relative autonomy to measure the performance of municipal basic education public policies? Based on this challenge, this paper investigates the theme of State Capabilities in a macro context of assessing municipal structures in Goiás. To this end, the literature on State Capabilities was systematized with the aim of defining the history and generations of formulation of the theory. A methodology was also developed to calculate a State Capabilities Index (ICE) for the 246 municipalities in Goiás based on information from public databases – especially the Munic surveys of the Brazilian Institute of Geography and Statistics (IBGE) and surveys by the Court of Auditors of the Municipalities of the State of Goiás (TCM-GO), taking the year 2018 as a reference. The data were treated by Winsorization to exclude outliers and submitted to an expansion procedure. The results of the ICE and its indicators were correlated with the public IDEB scores of the municipalities of Goiás using Pearson, Kendall and Spearman correlation tests in order to establish correlations between the ICE and the results of the municipal public education policies measured by the IDEB. No relevant statistical correlation was found between the municipal State Capacities (ICE) and the IDEB results based on the established methodology. Finally, reasons that may justify this finding are discussed, such as the inability of the IDEB to translate the results of public education policies and the need to improve quantitative research that aims to measure municipal State Capacities. The research is exploratory in nature, presenting an initial model of analysis on educational issues based on a theme that is still little explored: municipal state capacities. The work contributes to the reflection on the methodology of quantitative research, especially regarding data treatment and evaluation of results through boxplot graphs, distribution/normality curve (Gauss), dispersion and cluster maps, all prepared in the R software accompanied by the respective scripts and databases for replication.