Nonparametric methods for stratified C-sample designs: a case study

Abstract

Several parametric and nonparametric methods have been proposed to deal with stratified C-sample problems where the main interest lies in evaluating the presence of a certain treatment effect, but the strata effects cannot be overlooked. Stratified scenarios can be found in several different fields. In this paper we focus on a particular case study from the field of education, addressing a typical stochastic ordering problem in the presence of stratification. We are interested in assessing how the performance of students from different degree programs at the University of Padova change, in terms of university credits and grades, when compared with their entry test results. To address this problem, we propose an extension of the Non-Parametric Combination (NPC) methodology, a permutation-based technique (see Pesarin and Salmaso, 2010), as a valuable tool to improve the data analytics for monitoring University students’ careers at the School of Engineering of the University of Padova. This new procedure indeed allows us to assess the efficacy of the University of Padova’s entry tests in evaluating and selecting future students.