For about one month, I have been organising a self-phased course seeking to help students to use computers in qualitative research. This includes programs such as Atlas.TI, DiscoverText as well as more data science related tools. The course is designed for PhD students at the Faculty of Social Science, with a total enrolment of 46 students. I conducted a survey on the course to understand how students experience it and continuously improve the course – 16 students responded the survey.
The course is planned to be self-phased, allowing people to do exercises when they have time or a clear need to develop methodological tools emerges. However, student responses highlighted that they expected the course to provide more structuring and scheduling for the course work. Students also commented that they “just need to find the time” to start the course properly. Thus, it seems there is a clear demand to help students’ self-regulation more than the self-phased studies do. In the end, this is not that surprising, studies on MOOCs have suggested that deadlines have much value to ensure students can work on the course. At the same time, other students thanked for the flexible course arrangements, so the course design approach is not trivial.
Change on course practice: During the first week of each month, I will organise a short (30-45 minutes) introduction to the course and propose a four-week schedule to completing the course tasks for each module. We have already piloted social hours to support peer-interaction and I will keep organising them on the third week of each month. However, the introduction sessions will be kept voluntary.
Course example data
Some students highlighted that they would rather work on their own data on open and closed coding modules. As I have wanted to introduce stages requiring collaboration (such as re-coding the material or comparing the codes) I have chosen it is more convenient if we all work on the same material. Furthermore, this way we do not have issues related to people sharing their own research material, where privacy regulations may limit how well the data can be shared even within the university.
Choosing the data for course was a non-trivial task. As this is not a research topic for anyone, I wanted to keep the data condense, both in terms of number of units of analysis as well as length of each unit of analysis. I really hope that the workload on the data side is rather limited as the course is focused on tools only.
The data must also be on a topic that is familiar enough for students to engage with. The datasets were chosen from BBC to make them at least somewhat relevant to most students (and ensure they were written in English). For closed coding, a set of 55 news headlines is taken from their websites. Similarly, the three articles on the crisis Megan & Harry and the Royal Family are authentic articles published on BBC site, albeit being brief.
Naturally, one would never conduct actual social science with such a small dataset. Instead, this is a toy dataset to help us work with computer programs, understand their features, and envision how they could benefit one’s own research project. The principles are not that different with more realistic data.
Change to course practices: Explain what is written above on the first introduction lecture and write it more clearly as part of the assignment materials.
Course materials raised both praises “The videos and explanations work well.” and “The materials are so bad.”, but sadly students did not further explain what works well or what makes materials bad.
Change to the course practice: During the first introduction lecture, highlight the opportunity to provide direct feedback on the materials via the website even more strongly and improve the material based on concerns emerging.