Abstract:
Reading and writing are two of the most important skills acquired by young learners. However, approximately 4–10% of the German population suffers from dyslexia. If not treated adequately, reading and spelling disorders negatively affect children’s academic, personal, and social development. Thus, the interest in digital tools to efficiently support literacy acquisition outside of class or therapy has grown.
This thesis addresses three main aspects to expand the current state of research on digital tools to support reading and spelling development. First, I present our novel digital game-based spelling training called “Prosodiya” to improve literacy skills in German primary school children. Prosodiya differs from similar approaches in that it systematically teaches orthographic knowledge in combination with the awareness of syllable stress. I investigate the feasibility, effectiveness, and validity of Prosodiya in a randomized controlled field trial with 116 German second to fourth graders with mainly poor spelling skills. The training was carried out at home over a short period of 9–10 weeks. Results showed significant improvements in syllable stress awareness and spelling abilities for trained and untrained word material. Prosodiya was also reportedly easy to use, motivating, and provided good game experience, proving its feasibility for use at home. Further, the validity of our novel pedagogical approach was confirmed in correlation analyses investigating the relationships among syllable stress awareness, reading, spelling, and training performances. Prosodiya may therefore expand the traditional pool of training methods.
Second, I address one challenge of interaction design in educational applications, namely choosing the appropriate interaction style for use with children – which is not trivial. For this, I compare drag-and-drop, point-and-touch, and simple touch for writing words in a mobile spelling game. I evaluate the perceived workload, user experience, and writing time of 25 German third and fourth graders (8–11 years). We found that touch was the fastest, best rated, and most preferred interaction style. We also found small advantages for drag-and-drop over point-and-touch, which runs counter to some research and recent design guidelines.
Third, I explore the potential and limitations of automatic generation of language learning material in two domains. I investigate the utility of text-to-speech tools to automatically generate minimal pairs (e.g., beaver vs. peaver) for use in language learning systems to foster learners’ phonological awareness. I present our novel approach to improve the pronunciation of artificially generated German pseudowords. The results of an online study showed that distinguishing the lexical word from its pseudoword counterpart was equally successful when the minimal pair was generated by our method or produced by a human. In addition, I further present COAST, a web-based tool for easy and automatic visual enhancement of syllable structure, word stress, and spacing in reading material using resources of natural language processing. COAST’s feasibility and usability were validated in user tests.
All aspects contribute to the current state of research on digital learning tools and pose promising results to support children’s reading and spelling development.