Software Engineering for Dynamic Game Adaptation in Educational Games

Educational games and game-based assessments have evolved over the past several years and are continuing to evolve. They promote student engagement in the learning process by creating an interactive environment where they can learn in a fun and challenging way. This gives them the potential to yield diagnostic information to educators and feedback to students. During the game play process, game-based assessment (GBA) can be used to assess the learning imparted by the game to the students. A common strategy for GBA has been to utilize surveys and built-in quizzes to measure student learning during the game play. However, this impacts students’ attention negatively as they need to change their attention from game play to the assessment and back. Stealth assessment provides a natural alternative for assessment of learning without breaking the delicate flow of engagement. It aims to blur the lines between assessment and learning by weaving them together within the game. Stealth assessment uses game play interaction data to build inferences about student performance and learning. As an advantage, it provides ways to assess hard-to-measure constructs such as learning proficiency, critical thinking, persistence, and other twenty-first-century skills. Designing and developing an educational game takes time, and repeating the process for every new content or concept can be inefficient. The authors provide a framework called content-agnostic game engineering (CAGE) that can be used to create multiple learning contents within a single game by reusing already developed educational game mechanics. CAGE helps reduce time for creating an educational game by building content-agnostic mechanics that could be used across multiple content topics. It does so by separating the game into three components of mechanics, content, and student modeling that operate independently. Additionally, stealth assessment can be integrated into CAGE as a part of the student model and can also be content agnostic as a way to demonstrate the advantages of adopting a CAGE-based development framework. While CAGE can work with multiple content domains, it cannot work with every domain. The limit is decided by how the game mechanics are implemented. In this chapter, we discuss the software practices to implement CAGE architecture and ways to embed stealth assessment in a content-agnostic way.