Learning analytics has the potential to change the way we respond to students facing challenges but who and how best to make the intervention? Julia Taylor reminds us of the need to look deeper into the data.
Interpreting data is a skill. It requires considerable knowledge and experience to ‘read’ the increasingly complex layers of data which could mean different things in different circumstances. The Jisc webinar: Data and Disadvantaged Students reminded us that the quantity of data is not the key issue. Intelligent consideration of the story that the data is designed to describe is essential. What pathway is indicated by data on successful students and what can be concluded and changed where students are less successful? Academic and support staff who want to make use of the data have a range of options in relation to interventions but what patterns of data indicate a student needs help and how to decide on the most effective response in each case?
Jisc’s workshop on Planning interventions with at-risk students at Aston in June was designed to unravel these issues with a series of practical scenarios that highlight different data sets where achievement, attendance or engagement decreased or varied significantly. Delegates were invited to interpret the scenarios and plan an appropriate intervention: examining what would trigger a response, what assumptions were involved, what type of intervention would be most effective and how this could be evaluated.
In the absence of historical data, the process is necessarily superficial but provides a starting point for discussing established patterns. Participants identified recognisable triggers: 55% attendance, below average VLE use for 3 weeks, requests for extensions and marks lower than 40%. Discussions raised issues such as the need for prevention rather than intervention. Improving attendance with more engaging teaching, more frequent formative assessment to avoid low marks.
The conclusion was that although the retention officer (or equivalent) would be a key player, module tutors, personal tutors and programme directors should be involved to determine if there was cause for concern that would mean reconsidering the curriculum, module or activity. This is as crucial where students have a disability or disadvantage as is an intervention with the individual as it supports more inclusive practice.
After all, is lower attendance always an indication of not coping academically – maybe not for those with caring responsibilities or long term medical issues to contend with? Is the student being disadvantaged by the assessment schedule and should adjustments be made?
It was interesting that interpretations varied depending on the role of those participating. There was an assumption from some staff that late or low participation required a firm response, opinion differed on the appropriateness of texting a reminder even down to the wording used. Staff in support roles feared that automated interventions could be demotivating and appear aggressive in comparison to face to face or an email invite to meet your tutor.
Even more interesting, many of the responses changed when the possibility of a disability or different reading of the students circumstances was put forward. Is there a need to examine this issue further? Could an interventions workshop based solely on students who face additional challenges help to inform these decisions? What was clear was the data may be misleading if not referenced to disability or disadvantage. However, recent GDPR legislation calls into question the use of ‘sensitive data’ for learning analytics without consent. Jisc has suggested some approaches that institutions could take:
- Not ask for consent for the use of non-sensitive data for analytics (our current understanding is that this can be considered as of legitimate interest or public interest)
- Ask for consent for use of sensitive data (which, under the GDPR, will be called “special category data”)
- Ask for consent to take interventions directly with students on the basis of the analytics
Although many institutions are not ready to make interventions based on the learning analytics data they currently collect, they are gathering data which could be useful in this context in the future. So a policy that seeks consent regarding the use of what will now be known as ‘special category’ data is an important step right now. Then we can begin to evaluate how student data can be used to shape an experience that works for disadvantaged and under-represented groups as well.
The next Jisc learning analytics network meeting at Aston on September 5th will include a cut down version of the interventions workshop.