Tom, who completed his Honours project last year with us, received the award for best Honours project last week – Congratulations, Tom!
Tom’s thesis work was very interesting – he was looking at whether we could analyse student behaviour on our online resource and discussion forums, and find any correlations between their behaviour and their eventual success, or lack thereof, in their course. We looked at a range of social network analysis techniques incorporating network visualisation, as well as ranking algorithms (HITS, SPEAR, Page Rank) and resource access patterns with some good results.
As with all good experiments, some of our results demonstrated no clear correlation where we had expected to find one, and other results demonstrated clear correlations where we were not sure we would find anything at all. The most effective analysis that we completed was based on analysing resource access patterns in association with grade band – we were able to determine that students who achieved a higher grade accessed more resources and more frequently – the inverse was also true. We were also able to show that patterns become apparent early on in the course – potentially as useful early indicator of at risk behaviour.
Overall, our more complex measures did not prove to be significantly more useful or informative. But we certainly want to continue this work with larger scale student data and cross-course data (we only looked within individual courses) to see whether we can find any useful results.
We hope to publish this work shortly, as Tom is continuing on his work (in a related, but different project space) this year as a PhD student within the CSER group.