In efficiency studies, inputs and outputs are often noisy measures, especially when education data is used. This article complements the conditional efficiency model by correcting for bias within conditional draws, using the m out of n bootstrap procedure. With a unique panel dataset, we estimate managerial efficiency, which accounts for nondiscretionary variables, and explain efficiency
... [Show full abstract] differentials of adult education programs in Flanders. Our results suggest that the characteristics of learners in a program matter for managerial efficiency, and that teacher characteristics are strongly correlated with efficiency differentials, as more homogeneity in the teacher workforce appears to result in higher program efficiency.