Testing heterogeneity in quantile regression: a multigroup approach

Abstract The paper aims to introduce a multigroup approach assess group effects in quantile regression. procedure estimates the same regression model at different quantiles, and for groups of observations. Such are defined by levels one or more stratification variables. proposed exploits computational test effects. In particular, bootstrap parametric permutation compared through artificial data taking into account sample sizes, comparing their performance detecting low, medium, high differences among coefficients pertaining groups. An empirical analysis on MOOC students’ is used show proposal action. effect two main drivers impacting performance, learning engagement, explored conditional self-paced courses with instructor-paced courses, offered EdX platform.