Opiskelijoiden aihekohtaiset vaikeudet oppia tietorakenteita ja algoritmeja

Data structures and algorithms (DSA) are fundamental concepts in computer science, as they provide basis for e cient computation. An introductory DSA course is mandatory ffi for students in an undergraduate computing curriculum similarly to an introductory programming (IP) course. Investigating students’ difficulties in learning DSA would advance teaching of this common topic. Knowledge of misconceptions, students’ understandings which deviate from the scientific consensus, could be utilized to increase the effectiveness of DSA instruction. This dissertation investigated students’ misconceptions and other difficulties specific in learning DSA. A systematic literature review produced a catalog of 70 students’ misconceptions related to DSA subtopics and recursion. Previously, a questionnaire instrument was designed to diagnose whether a certain student’s difficulties relate to missing knowledge of IP or DSA. In this dissertation, the instrument was evaluated by think-aloud interviews with DSA students, confirming that the instrument can differentiate difficulties between IP and DSA, but not diagnose difficulties on IP/DSA subtopics reliably. Students’ misconceptions of Dijkstra’s single-source shortest paths algorithm were studied extensively with mixed methods. A key instrument were visual algorithm simulation (VAS) exercises, where student simulates the execution of Dijkstra’s algorithm by interacting with a graph visualization. Several methods were used to analyze students’ errors during VAS activity: think-aloud interviews, automatic detection of exercise states with certain properties, and modeling systematically repeating errors with algorithms. Based on the observations, students confused elementary graph algorithm concepts, or their simulations resembled other graph algorithms instead of Dijkstra’s algorithm. This thesis provides instructors with a catalog of DSA misconceptions from the literature and new misconceptions of Dijkstra’s algorithm, both utilizable in DSA course instruction. The catalog points to future directions of DSA misconception research. Furthermore, the VAS-relevant mixed methods methodology supports investigation of students’ difficulties in visual algorithm simulation exercises.