The EDEN Annual Conference in Zagreb beginning of June gave LACE project the opportunity to promote the idea of building a community around issues of Learning Analytics and Educational Data Mining. We presented papers, explored synergies with other projects, and engaged in workshops discussing how learning analytics should be brought to fruition within schools, higher education and the workplace. For me, the highlight of these meetings with colleagues showing a keen interest in LA was the dialogue that took place in the session where I presented the LACE paper. The framework of critical dimensions of learning analytics developed by Greller & Drachler and used by LACE to structure the discourse resonated well with the following presentation by Paul Prinsloo (@14prinsp) and Sharon Slade (@SharonSlade) titled «Educational triage in higher online education: Walking a moral tightrope». I’ll return to the concept of a triage in a second, first stressing the need to look at Internal Limitations and External Constraints when promoting the benefits of learning analytics.
The story of the closure of inBloom, a Gates and Carnegie Foundations backed US operation to help students by providing more personalised learning, makes it very clear that ignoring the non-technical and ethical aspects of sharing educational data does have risks. (See The Economist for a news report on inBloom and this link for how a lobby group presents the challenges.)
Prinsloo and Slade’s question was how do we make moral decisions in education when resources are (increasingly) limited? Before accepting that learning analytics was the way to go, they invited us to make a detour to the First World War trenches and the concept of ‘triage’, inspired by a quote from Manning (2012):
“Are students walking around with invisible triage tags attached, that only lecturers can see? Is this fair? Or is it just pragmatic? Like battlefield medical attention, lecturers’ attention is finite. And as class sizes and workloads increase, it is becoming scarcer.”
Triage is balancing between the futility or impact of the intervention juxtaposed with the number of patients requiring care, the scope of care required, and the resources available for care/interventions. Four moral principles guide medical triage:
- Respect patient autonomy
- Justice – care not determined by privilege, status, race, gender
Prinsloo and Slade propose to adapt these principles for ‘educational triage’, implying that in the case of learning analytics there will be noise and casualties along with benefits also for the ones most likely to be wounded in the first line.
They suggest that
- Student and institutional autonomy are situated and bounded. Students success and retention are not the sole responsibility of either students or the institution. Both parties should be acknowledged; the student autonomy should be respected, but so should the long-term sustainability of the institution.
- Beneficence – in the best interest of the student. No access without success.
- Non-maleficence and transparency. Transparency should characterise not only the analysis, but also the diagnosis, prognosis and outcome.
- Distributive justice. Ethic origin, sex, social status and ability to pay should not play a role. However, «it is immoral not to take into account the historical impact of some of these factors in consisering the classification of students in educational triage,» say Prinsloo & Slade.
We cannot afford not to use data, and knowing that our data and algorithms cannot be complete or non biased as we know that we by applying learning analytics on a bigger scale will be walking the moral tightrope. Prinsloo and Slade brought to this discussion a concept of educational triage that has a potential to personalise the discussion without loosing sight of necessity and societal and educational benefits of developing this new class of learning technologies.
Manning, C. (2012). Educational triage. Web log post, Wednesday, March 14, 2012. Retrieved from http://colinmcit.blogspot.co.uk/2012/03/educational-triage.html