Using the Affective Computing Technique to Evaluate the Performance of Digital Game-Based Learning

• Main Authors: Yi-Lin Tzeng, 曾奕霖
• Other Authors: Chih-Hung Wu
• Format: Others
• Language: en_US
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/98328193823899059532

碩士 === 國立臺中教育大學 === 數位內容科技學系碩士班 === 101 === Digital Game-Based Learning (DGBL) is thought to be an effective tool for learning, but the empirical evidence to support this assumption is still limited and contradictory. For verifying the possibility of playing digital game to learn the Newton's laws of motion, this study used a quasi-experimental design to examine the effectiveness of Digital Game-Based Learning (DGBL) and traditional static e-learning on students’ learning attention, affective experiences, cognitive load, academic achievement and problem solving skills. In phase 1, when student learning, their physiology signals were measured by affective computing technique for analyzing their learning states. After learning, a posttest of learners was conducted to find the differences in academic achievement between DGBL and static e-learning. In phase 2, this study found that learner had difference problem solving strategies between different working memory capacity learning styles (active vs. reflective, sensing vs. intuitive, sequential vs. global), learning environments (DGBL vs. traditional static e-learning), major (non-science vs. science) and problem-solving performance (low performance vs. high performance). The results showed five major findings. The first finding, the DGBL group has more cognitive load via proof of affective computing technique. The second finding, the DGBL group has better academic achievement but have no significant differences between DGBL and static group. The third finding, the high working memory capacity learning style group, the DGBL group and high problem-solving group easily to find out the key factors. The fourth finding, Successful problem solvers inspected the factors in a different pattern from unsuccessful problem solvers. The fifth finding, learning style (active vs. reflective) showed significant various ways of eye movement when learners solving problem. This study suggested future educators can provide the enough feature of DGBL environment. In addition, we found that educator can measure leaners’ learning style by affective computing technique. This is helpful for developing an adaptive learning system. In the future work, we propose the future studies can adopt more physiology signals to measure learning state of learners and provide better DGBL environment.