Submitted Master Thesis of Martin Ulrich Breuer
Martin Ulrich Breuer presents his master thesis on the topic "Untersuchung von Problemlösestrategien und Kompilierverhalten von Studierenden in einer Serious-Game-Umgebung".
Since 2016 the serious game "Codescape" supports the learning of the programming language Java in the first semester of computer science studies. In the long term, adaptive components are to be developed in order to improve the overall quality of the game and especially the feedback in the game and to tailor it to different user groups.
The goal of this work was to investigate the students' behavior while playing Codescape. For this purpose, the research on problem solving strategies of novice programmers was continued by discussing the applicability of the definitions and procedures of different authors and by examining the respective occurrence of the problem solving strategies in the Codescape data set of the lecture Programming at the RWTH Aachen University from the winter semester 2017/18. A first exploratory analysis revealed differences in the context in which the data were collected as well as in the applicability of the definitions to the students' problem-solving processes. These differences induced the need to adapt the definitions. Problem solving strategies were defined which represent the behaviour of the players of the serious game. For this purpose a task was selected for which new problem solving strategies were defined step by step until a complete categorization of all students could be carried out.
For the automated determination of these problem-solving strategies, which partly overlap, metric values were derived automatically, based on a procedure from literature. In addition, the most applicable procedure of the literature, which assigns a pattern to each pair of events, was adapted to the context of the Codescape data set, implemented, extended and a possibility for future use was presented.
Additionally, tools for visualization and detailed investigation of problem solving processes in free programming tasks of the Serious Game were developed and an analysis process based on the collected raw data for a categorization of problem solving strategies was presented.