Building composite indicators of learning conditions through principal component analysis: an exploratory study

Abstract

Comparisons of education systems rely mostly on student achievement, whereas few Moroccan studies provide composite indicators of the conditions in which learning takes place at the school level. This paper builds four composite indicators using principal component analysis (PCA): school resources, teaching practices and conditions, socio-economic and digital environment, and student engagement. The study covers 150 Moroccan schools and 30 indicators (8, 7, 7 and 8 per dimension). The first component explains 65.29%, 59.19%, 60.93% and 60.68% of the variance, and the second eigenvalue remains below 1 in every dimension. Scores rescaled from 0 to 100 show large dispersion across schools, with standard deviations between 23.75 and 26.45. Each dimension can therefore be summarised by a single index, but no conclusion about school performance itself can be drawn without an independent outcome variable. The paper discusses the limits of the method and proposes a validation agenda.

JEL Classification : I21, I24, C38

Paper Type: Research Article: Exploratory Empirical Study