Authoring gamified intelligent tutoring systems.

Abstract

Intelligent Tutoring Systems (ITSs) have been drawing the attention of academics and practitioners since early 70’s. There have been a number of recent studies in support of the effectiveness of ITSs. However, it is very common that students become disengaged or bored during the learning process by using ITSs. To explicitly consider students’ motivational aspects, researchers are increasingly interested in using gamification along with ITS.However, despite providing individualized tutoring to students and some kind of support for teachers, teachers have been not considered as first-class citizens in the development of these kinds of systems. In order to contribute to the active and customized use of gamified ITS by teachers, three technical problems should be considered. First, designing ITS is very complex (i.e., take into account different theories, components, and stahekolders) and including gamification may significantly increase such complexity and variability. Second, gamified ITS features can be used depending on several elements (e.g., educational level, knowledge domain, gamification and ITS theories, etc). Thus, it is imperative to take advantage of theories and practices from both topics to reduce the design space of these systems. Third, in order to effectively aid teachers to actively use such systems, it is needed to provide a simple and usable solution for them. To deal with these problems, the main objective of this thesis is to design an authoring computational solution to provide for teachers a way to customize gamified ITS features managing the high variability of these systems and considering gamification and ITS theories/practices. To achieve this objective, we identify the variability space and represent it using an ontology-based feature modeling approach (OntoSPL). We develop an integrated ontological model (Gamified tutoring ontology) that connects evidence-supported game design elements in the e-learning domain as well as gamification theories and frameworks to existing ITS concepts. Finally, we develop an authoring solution (named AGITS) that takes into account these ontologies to aid teachers in the customization of gamified ITS features. We evaluate our contributions by conducting four empirical studies: (1) we perform a controlled experiment to compare OntoSPL against a well-known ontology-based feature modeling approach. The results suggest that our approach is more flexible and requires less time to change; (2) we evaluate the ontological integrated model by using an ontology evaluation method (FOCA) with experts from academic and industrial settings. The results suggest that our ontologies are properly targeting the knowledge representation roles; (3) we evaluate non-interactive versions of the designed authoring solution with 59 participants. The results indicate a positive attitude towards the use of the designed authoring solutions, in which participants agreed that they are ease to use, usable, simple, aesthetically appealing, have a well-perceived system support and high credibility; and (4) we also evaluate interactive versions (scratch and template) of our authoring solution with 41 teachers. The results suggest that teachers can use and reuse, with a high acceptance level, an authoring solution that includes all the complexity to design gamified ITS.