Structural equation modeling and student evaluation of teaching: Inputs of the PLS Approach

Abstract

Theorists such as Jöreskog (1973), Keesling (1972) and WILEY D.E. (1973) are respectively improved the structural equation models with latent variables in order to examine multiple causal relationships. The two most widely answered procedures in the literature on structural equations are the LISREL method and the PLS method. The PLS method, as Wold (1982) is recommended is based on the use of Partial Least Squares regression techniques for estimating models of structural equations. Indeed, structural equation models are the most suitable for modeling the Student Evaluation of Teaching (SET). The approach (SET) consists of evaluating the quality of the teaching system, based in part on the judgments of students concerning aspects related to the course, and considered to be of primary interest for the educational management of the university. We will concern ourselves in this article to demonstrate how the PLS method would be a relevant method for evaluating the quality of teaching based on student judgments. The PLS approach is suitable for exploratory approaches that do not require a well-founded theory and allow the estimation of partial models, and this is much more consistent with the objective of teaching evaluations by students who are interested in describing the opinions of students. 

Keywords: Structural equation modelling, Partial least squares, Student Evaluation of Teaching, Lisrel

JEL Classification : H11, H19

Type du papier : Theoretical Research