STACK Researchers' Network: Meeting on 3 February 2025

Participants

We will aim to have a 1-hr workshop at the STACK Conference in Durham.
George proposed a format based on the working group at ICME:
- Introduction to the e-assessment research agenda themes (5 mins)
- Discussion in small groups, with each group discussing a different theme (30 mins)
- Suggestions for the Researchers' Network - what should the network do to be helpful? (20 mins)
Montognon offered a couple of suggestions:
- that our group could identify key questions from the agenda to help narrow down the options.
- we could giving a short presentations to illustrate addressing some of the questions, e.g. on student errors.
The group were generally happy with this plan, and the suggestion of using our next meeting to try out the activity. This would also help us to identify key questions that we might want to highlight in the conference session.

During the meeting we discussed a couple of topics that are of interest to the group.

Lucy is interested in analysing the ordering of STACK tasks, using existing student response data
Andreas reported on some analysis like this at ETH on their Integral Trainer

Lucy highlighted the difficulty of inferring student errors from their final response - both identifying what error might have led to a particular incorrect answer, and being sure that a final response really comes from the student.
Andreas shared the experience of collecting students' working on a mock exam to analyse the sources of errors.
Montognon described an approach using LLMs to generate feedback on student errors.

Lucy suggested having more time for introductions at the meetings. Building on this, George noted that it would be good in future meetings to have time for people to report on updates, e.g. new research projects or outputs.
Fix a time in the week of 3 March
- Action: George to send out a poll on Zulip (and also email previous invitees to direct them to Zulip!)