ENTERTAINMENT COMPUTING, cilt.42, 2022 (SCI-Expanded)
This article describes the design, implementation, and evaluation of an educational mobile game. Pickstar aims to make vocabulary learning training more accessible to families with ASD, in a way that increases engagement by offering a multisensory experience and by adapting to the level of each player. The game includes two learning approaches: the classical levels advancement approach where the levels gradually increase in difficulty and Dynamic Difficulty Adjustment (DDA) approach that adapts to the skill level of the player as determined via feature extraction. Through the AdaptiveMiniMax algorithm, Pickstar's DDA feature ensures that everyone has the same winning rate of approximately 70%, encouraging players of all levels to play again. The engagement rate of children with ASD in the learning process was 88.4% higher for DDA compared with classical levels. Usability evaluation of Pickstar, based on ISO 25062:2006, on caregiver interviews concluded that Pickstar not only decreases caregivers' teaching and assessment loads, but also increases the time children spend on learning. Caregiver interviews, usability testing, and engagement assessments show that Pickstar can be used with minimal guidance to improve vocabulary skills of individuals with ASD and can mitigate caregiver problems in accessing good-quality, engaging, and affordable education for their children.