This thesis examines children's thinking, learning and metacognition when designing their own computer games. The study aims to understand more about what kind of learning takes place, and how it emerges whilst children are authoring their own computer games. The aim is to get an insight into the cognitive processes students exercise that activates the 'thinking for learning', in particular in relation to the role of the teacher and digital game making activities as a learning space. Whereas mainly case studies and design-based research projects have been used as methodologies to study learning with digital game making, this study gives an ethnographic account by observing children's problem-solving activities from moment to moment. Field notes were collected by examining the language and the context children use for their 'self' explanations and group discussions, the gestures, the culture of their relationship with their teacher, peers and technology in their classroom settings. A metacognitive skills self-report instrument was created and used to investigate the metacognitive skills that children develop whilst working on their games. The data were collected for a period of eight months, through participant observations, in-depth interviews, informal conversations and video recordings of children's group discussions in a primary school in London. Learning logs and problem-solving sheets were introduced for the ten focus children to record their thinking when solving problems. During this research there were many opportunities to observe the changes in a child's reasoning over time, which provided an insight into children's mental activities. The study found that game design activities have many learning benefits for children. The main themes that are emerged from the study include metacognitive awareness; CT; learning in curriculum subjects; and developing transferrable 21st century skills. Furthermore, the role of conversation in triggering thinking processes and self-regulated learning are discussed using data from the study. Although the study provides insight into different aspects of learning during game design, it also highlights the difficulty in evaluating these different learning benefits. The results contribute to the growing body of knowledge about how to evaluate children's computational skills by providing a multiple evaluation model and a Metacognitive Skills Instrument (MSI) for measuring metacognitive skills that children develop whilst making their computer games. The challenges and limitations of these methods are discussed to form questions for the future studies.