MULTI-LEVEL ADAPTATION OF AN EDUCATIONAL GAME TO INDIVIDUAL STUDENT’S GAMEPLAY, KNOWLEDGE AND EMOTIONS
Application of games for educational purposes is the subject of growing attention due to their potential to improve learning process and outcomes through increased enjoyment, involvement, motivation, and emotions. Emotions are fundamental to learning because they influence perception, attention, decision making, motivation to learn, as well as acquisition and retrieval of knowledge. Games for the knowledge assessment has emerged as a promising research area since games are all about constant assessment by providing challenges and giving feedback. In many cases, knowledge assessment can lead to negative emotions, e.g., anxiety, fear, if the student is not confident about his/her knowledge level or if the selected challenge level is not matched appropriately to the student’s skills.
Consideration of emotions during the knowledge assessment allows identifying situations possibly leading to negative learning outcomes.
Students usually differ regarding their interests, preferences, personality, and also learning needs. All these aspects can be considered to provide adapted instruction. However, differences of individual students are seldom considered in educational games and emotions as an adaptation source are employed even rarely. In general, adaptation can be ensured at two levels: macro-level, prior to learning, more static, and micro-level, during learning, dynamic. Such adaptation granularity is implemented more often in intelligent tutoring systems, which are created with the aim to provide personalized instruction for an individual student. Similar adaptation approach can be implemented also in educational games.
Currently, development of a game with integrated emotion recognition based on facial expressions for the Artificial Intelligence course is carried out. The main aim is to assess students' knowledge through adapted game elements leading to increased motivation to learn and achieve higher results not only in the game but also in final exams, in which failure rate has increased notably in the last few years.
To ensure macro-adaptation, before game starts, the student’s prior knowledge, preferred learning style and gaming skills, e.g., beginner, expert, is considered by combining these parameters with the student’s personality represented as Big Five personality traits. Personality traits give information not only about the student’s personality but also about default mood or tendency to some specific emotions, preferred learning style, teaching approach, and motivation towards achievements, as well as personality traits allow identifying players' types, for example, mastery or performance-oriented.
As gameplay starts, micro-adaptation in the real time based on dynamic parameters is initiated. Emotions are also considered as one of the parameters, since they are occurring and changing during the gameplay depending on the perception of game elements, challenge level, the student’s knowledge, and playing skills, achievements and provided feedback. Emotions serve as an information source for identifying being in the flow state, which is considered to be optimal for the learning. Based on the flow model, an occurrence of some emotions, e.g., boredom, anxiety, represents mismatch between challenge (task difficulty level) and skills.