Efficiency of secondary schools in Portugal: a novel DEA hybrid approach

Abstract

This paper is aimed at casting new light on the assessment of secondary schools’ performance in Portugal by considering data from the Programme for International Student Assessment 2015. We use the Weighted Russel Directional Distance approach and establish an equivalence model between this Data Envelopment Analysis model and the super-ideal point model, which is then combined with an interactive algorithm. Besides, a procedure to perform robustness analysis has been proposed, allowing decision-makers to understand how sensitive the efficiency of each school to data variation is. Overall, our findings suggest that the average efficient public school has mean scores across all competences above the scores attained by the average public national school, but below the OECD average. In addition, we have concluded that schools operating with the lowest efficiency scores usually need to make a greater effort to improve results in reading when compared to maths and science competences. Finally, schools more often selected as a reference in terms of best practices and which are classified as robust in terms of efficiency are not necessarily performing with high test results, highlighting the need of seeking new benchmarks more consistent with the preferences of the decision-maker.