Causality Exploration in Modeling Engineering Student Satisfaction

The goal of this paper is to develop a self-aware computational model aimed at analyzing student satisfaction in an engineering faculty. We examine whether student diversity, as well as student engagement, social events, and academic experience have a direct effect on student satisfaction and, consequently, retention and loyalty. A Confirmatory Factor Analysis to explore the associations between latent and observed variables led to the design of the Structural Equation Model. In our model, the following latent variables were positively associated with Student Satisfaction: Student Engagement, Academic Experience, Student Diversity, and Social Events. Also, the association between Student Satisfaction and Student Loyalty was positive. Other latent variables were tested such as Student Well-being and Student Support, but they did not provide a good result. Note that, unlike other papers in the area, we considered student diversity to make the model diversity-aware. The learned model provided the basis for the reasoning process behind engineering student satisfaction. In future work, this Structural Equation Model will become the basis for building the Bayesian network, the model that allows us to perform probabilistic inference and test various scenarios using the reasoning mechanism.