Exploring Italian students’ performances in the SNV test: a quantile regression perspective

Antonella Costanzo, Domenico Vistocco
(2018) Classification, (Big) Data Analysis and Statistical Learning, (eds. Mola F, Conversano C, Vichi M). Studies in Classification, Data Analysis, and Knowledge Organization, pp 117-126. Springer

Amazon DOI  pdf

Over the past decades, in educational studies, there is a growing interest in exploring heterogeneous effects of educational predictors affecting students’ performances. For instance, the impact of gender gap, regional disparities and socio-economic background could be different for different levels of students’ abilities, e.g. between low-performing and high-performing students. In this framework, quantile regression is a useful complement to standard analysis, as it offers a different perspective to investigate educational data particularly interesting for researchers and policymakers. Through an analysis of data collected in the Italian annual survey on educational achievement carried out by INVALSI, this chapter illustrates the added value of quantile regression to identify peculiar patterns of the relationship between predictors affecting performances at different level of students’ attainment.