State-level educational performance in Brazil: A MCDM approach taking a governance perspective

This study presents an educational performance assessment of Brazilian state-level units using the technique for order preference by similarity to the ideal solution (TOPSIS). Although this technique is a well-established multi-criteria decision making (MCDM) model that can be applied across diverse areas, it has yet to be applied to an analysis of primary and secondary education. We assess how state-level units in Brazil perform in terms of primary and secondary educational inputs and outputs during the period from 2013 to 2017. We handle epistemic uncertainty with regard to weight definition using maximal information entropy. Additionally, a Tobit regression approach on performance scores is developed. Results show that region of the country plays an important role in determining educational performance. Moreover, results indicate that gross domestic product (GDP) is positively related to education scores while infant mortality is negatively associated with educational performance.