What-if analysis combining Fuzzy Cognitive Map and Structural Equation Modeling

Abstract

Nowadays, public funded universities have to increase their traction in acquiring students and in gaining and retaining their commitment, satisfaction and loyalty. In the area of information management, the Structural Equation Modeling (SEM) technique has been widely applied to perform inter-construct causal analysis (i.e. Commitment, Quality of Services, Loyalty, etc.) on specific observed indicators in order to model and analyze student loyalty in educational institutions. Nevertheless, SEM does not support what-if analysis contrived to consider decision-making scenarios, such as: what happens to student loyalty if managers adopt a quality-based strategy, and so on. This work proposes to combine SEM results with a Fuzzy Cognitive Map (FCM) to support what-if analysis and to determine the best strategy to adopt for the case study of the Relationship Quality-Based Student Loyalty Model (RQSL). Specifically, the causal models retrieved from SEM will be exploited as the input map of concepts enabling FCM. Subsequently, FCM is performed considering different input configurations corresponding to specific managing strategies (e.g. investing in quality of service, or quality of teaching, etc.) activating/deactivating various input concepts in order to simulate implications on the output constructs (e.g., Loyalty, Commitment, etc.). The results of the experiments provide reasonably good estimates of the impact on student loyalty deriving from investing in each specific factor and provides a helpful analysis that will support decision making. Furthermore, this study highlights the opportunities deriving from the cross-fertilization between the management and computer science domains.