Game design elements in serious games: Analysis of a chatbot for medical education

Abstract

Serious games are defined as games pursuing a primary objective beyond mere entertainment and are often used in educational contexts to foster intrinsic motivation and improve learning outcomes. A frequently referenced motivational theory underlying serious games is the self-determination theory, which proposes that intrinsic motivation is fostered by satisfying the three basic psychological needs for autonomy, competence, and social relatedness. Prior research has already attempted to assign specific game design elements to these needs, as they are responsible for affecting users' intrinsic motivation in serious games. In medical education, serious games provide a safe, risk-free environment for training a wide range of skills and competencies, including medical history taking, which is an integral part of the curriculum. This thesis aimed to examine the theoretical foundation of using a chatbot for medical history taking embedded in a serious game to extend the list of theory-based game design elements with the element chatbot. <br/> A first study developed a serious game with an inherent retrieval-based chatbot relying on free open-text entries and conducted a first usability and user experience evaluation. The second study compared the newly established chatbot with a previously existing keyword-based retrieval-based chatbot. Autonomy during the use was measured twofold: The objective autonomy in terms of students' exploratory behaviour was assessed by means of the serious games' process data while the subjective autonomy was assessed via questionnaires. Results indicated that the chatbot relying on free open-text entries allowed for more exploratory behaviour while no significant differences were found for the subjective autonomy. A third study was conducted focusing on the effectiveness of material for self-directed learning on history taking in chatbots. Therefore, the two chatbots of the second study were used and it was assessed whether students profited from an interposed guideline between two sessions. Results indicated that students using the chatbot with free open-text entries achieved overall but not significantly higher scores for their history taking in both sessions. Only students using the keyword-based retrieval-based chatbot improved significantly between the sessions. A fourth study was conducted, in which students took histories with a generative chatbot. Their entries were analysed with focus on indicators of empathy and students' perception on subjective autonomy was assessed. The results indicated that the generative chatbot enabled students to show written empathic reactions and to report high levels of autonomy afterwards. However, due to the small sample size these results have to be regarded as preliminary. <br/> In conclusion, this dissertation provided initial findings on the theoretical foundation of chatbots as game design elements. Building on the definitions of game design elements, it can be assumed that a chatbot can be regarded as such due to its interactive character. While chatbots must be selected according to their context of use, greater degrees of freedom may enhance students' subjective autonomy. It is further discussed whether other needs, such as the need for competence, may be addressed and ideas for consequent research are proposed.