Recently published results from the PISA 2006 Report show the existence of significant differences in performance scores among students from different Spanish regions participating in this evaluation. The aim of this paper is to use the information provided by this dataset in order to identify the causes of those divergences after controlling for educational inputs and environmental variables. For this purpose, we explicitly consider that education is a multi-input multi-output production process subject to inefficient behaviours, which can be identified at student level using a parametric stochastic distance function approach. Our findings suggest that La Rioja and Castile-Leon are the most efficient regions in Spain while Andalusia, Catalonia and the group composed of the regions that do not participate in the PISA with an extended sample are the worst. In addition, we conclude that most divergences in efficiency are attributable to students who are immigrants or repeating some course. In contrast, other factors such as class size or the type of school ownership do not seem to account for differences in students’ performance.