Skip to main content

A framework for analysing learners' engagement with online courses to detect different communities of learner behaviour

Venue: Birkbeck Main Building

This presentation introduces a mathematical framework for analysing temporal data relating to learners’ engagement with online courses for the purpose of identifying different communities of  learner behaviour.   This differs to methods in common use in that distinct learner communities are not pre-determined or identified subjectively but are detected algorithmically during the data analysis.   An initial analysis was performed on data relating to 81 adult learners studying six online post-graduate modules in management studies.  This analysis identified six distinct communities, where four communities were identified as learners with distinctly different learning behaviours such as `distributed' and `massed' learning, and low task completion. The remaining two communities contained single outlier learners that exhibited highly sporadic approaches to their learning. The position of high/low performing students within these communities was then investigated.  All low-performing students were located in a single community whereas the high performing students were distributed across the four larger communities.  Our analysis is also able to show that the behaviours of learners on an extended course (>6 months) evolve over time; indicating that learners fall into their natural learning behaviours as later times. 


Dr David Lefevre is Director of the EdTech Lab at Imperial College Business School

Contact name: