Examining manual and semi-automated methods of analysing MOOC data for computing education

Michael Morgan, Aletta Nylén, Matthew Butler, Anna Eckerdal, Neena Thota, Päivi Kinnunen

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review


We examine a semi-automated approach to the analysis of data from MOOC discussion forums. Previous research had analysed a sample of discussion forum data and developed a manual analysis framework, however this process can be very time consuming, especially given the class size of some online courses. Therefore it is important to investigate appropriate and automated analysis techniques to improve timeliness of analysis and to reveal the topics that emerge from a semi-automated process. An analysis of a data set from a coding MOOC in 2015 using the automated Structural Topic Modeling (STM) technique in R is described and contrasted against a manual analysis conducted on a segment of data from the same course in 2014. The types of analyses available and the relevance to computing education research is highlighted, with a focus on providing a discussion of the contrasting capabilities of each approach. The aim is to enable computing education researchers to assess the relevance of these techniques for further work.

Original languageEnglish
Title of host publicationProceedings - 17th Koli Calling Conference on Computing Education Research, Koli Calling 2017
Number of pages5
VolumePart F133134
ISBN (Electronic)9781450353014
Publication statusPublished - 16 Nov 2017
MoE publication typeA4 Article in a conference publication
EventKoli Calling - International Conference on Computing Education Research - Koli, Lieksa, Finland
Duration: 16 Nov 201719 Nov 2017
Conference number: 17


ConferenceKoli Calling - International Conference on Computing Education Research
Internet address


  • Data analysis.
  • MOOC
  • Online discussion
  • Programming

Fingerprint Dive into the research topics of 'Examining manual and semi-automated methods of analysing MOOC data for computing education'. Together they form a unique fingerprint.

Cite this