Education, as a process of human capital accumulation, is essential for a society’s economic
growth. Given its relevance and bearing in mind that public spending is not exclusively devoted
to financing the education system, it is essential to analyse and identify the different obstacles
that hinder school efficiency. Throughout this paper we have carried out an efficiency analysis of
Canarian schools for the period 2018-2019 using data provided by the Canarian Agency for
University Quality and Educational Assessment. Specifically, a three-stage data envelopment
analysis (DEA) based on the slack measure (SBM) and combined with stochastic frontier analysis
(SFA) has been applied to find the benchmarks of inefficient schools, understand their
characteristics and explore the different opportunities for improvement. Using this non-
parametric methodology, we can overcome one of the drawbacks of the DEA approach by
enabling us to consider the influence of contextual variables and random errors in the assessment
of efficiency. In particular, using mathematics and reading scores as outputs, teacher-student ratio
and an index of the quality of schools’ educational resources and some contextual variables such
as socio-economic status, we have tried to find out how efficiency varies when contextual
variables are introduced, and which contextual variables have the greatest influence on schools’
inefficiency.