Investigating Knowledge Structures in Computer Science Education

To be knowledgeable in a subject matter is a prerequisite of being competent. Experts typically possess densely connected structural knowledge of the concepts in their field of expertise. Therefore, investigating the knowledge structures of learners remains a central aspect of educational research in computer science as well as in other subjects, despite the ongoing trend towards outcome and competence oriented assessment. This thesis presents concept landscapes a novel way of investigating the state and development of knowledge structures using concept maps. Instead of focusing on the assessment and evaluation of single maps, the data of many persons is aggregated and data mining approaches are used in analysis. New insights into the “shared” knowledge of groups of learners are possible in this way. The educational theories underlying the approach, the definition of concept landscapes, and the accompanying analysis methods are presented. Since both data mining techniques and concept map collection are well suited for computer-based approaches, three software projects have been realized in the course of this thesis. They allow computer-based drawing and assessing concept maps, the subsequent analysis of concept maps and concept landscapes, and the extraction of salient concepts and propositions from texts. The methods and tools have been applied in three research projects that investigate structural knowledge in computer science education. Among others, the structural knowledge of CS students entering university has been analyzed. It has been found, that attending the compulsory school subject “Informatics” in Bavaria has visible effects on the knowledge structures of beginning students. The results of these studies can be taken in order to better understand the possibilities, limits, and effects of teaching methods, curricula, or learning materials in computer science education and other subjects.