Abstract:

Learning to program is considered to be a difficult activity [1]. As in other subjects, flipped classroom approaches are therefore increasing in Computer Science [2]. For a successful flipped classroom, there is a need for asynchronous learning material, for example videos, which can also be used in distance learning. Educational videos for learning programming can be produced in various styles. These differences can for instance relate to the production effort or how the code is displayed. During Live-Coding the development of a program code is shown as a process. Depending on the preparation mistakes can occur, which can be shown to the learners and solved afterwards or cut out in the post-production if no further attempts to write the program without mistakes should be undertaken. Live-Coding videos that should appear to be perfect from many lecturers’ point of view need extensive preparation and at least some production effort.

In this study low effort Live-Coding videos with nearly no post-production, Live-Coding videos with some (post-)production effort and an explanation video with finalized code are compared to a commented code development in Jupyter Notebooks and pure commented code. Jupyter Notebooks offer the possibility to interact with the content, e. g. to copy, change and run the code and are suitable for adjustments by the instructors without high effort. Students from a introductory Computer Science university course got a random learning material about a graph problem and should then perform a transfer coding task on a similar problem.

Besides the learning transfer success, also the students' preferences and attitudes towards the different learning material types were analyzed. A total of 85 students with varying levels of prior knowledge completed the study. The results suggest that the type of learning material has no significant effect on the learning transfer success, although not all participants were able to solve the rather difficult transfer task. The success correlated with the previous knowledge of the participants across all groups, but no significant differences could be found between the groups with different learning materials. Despite the concerns of some lecturers, especially low production effort seems not to be a major problem. Some students even mentioned that occurring errors can be helpful in order to understand typical mistakes. Additionally, students have preferences regarding the style as videos and more extensive insights in the development are preferred to pure commented code.

References:
[1] Tony Jenkins. 2002. On the Difficulty of Learning to Program. In 3rd Annual Conference of the LTSN Centre for Information and Computer Sciences.
[2] Brett A. Becker and Keith Quille. 2019. 50 Years of CS1 at SIGCSE: A Review of the Evolution of Introductory Programming Education Research. In: Proceedings of the 50th ACM Technical Symposium on Computer Science Education (SIGCSE '19). Association for Computing Machinery, New York, NY, USA, p. 338–344. https://doi.org/10.1145/3287324.3287432

Keywords:

Videos, programming, live coding, worked examples, Jupyter Notebook, production effort.