Daily chronicle – 5th day of the Leadership School
By Julie Bu Daher & Laura Infante Blanco
First session of day 5 of the leadership school
Panel on legal and ethical issues: do we need a global framework?
Anne Boyer welcomed everyone on this fifth day of D-transform reminding us that since the first day of the LS, legal and ethical issues came to the speakers minds so this panel should be at the core of a HE strategy.
C. Stückelberger (Globethics.net) facilitated this extremely relevant panel with a focus in the possibility of a global framework, guiding principles, legally binding conventions or a code of ethics/practise.
Mandla Makhanya (Unisa) had a clear opinion about the first subject: it is not possible to have a global framework for data collection, because of the diversity of HE institutions around the world and the local cultural, social and political environments in which they are embedded. Another important point he remarked was the awareness of the impact of the learning analytics outside universities. And he got some thoughts for the students, reminding that they are not only in the centre of the discussion but their point of view is critical on the process.
J. Breivik (BI Norwegian Business School) was aware about the consequences of uncontrolled use of educational data. In her opinion, learning is a process of growth, change, development and stakeholders have to keep in mind the aims and stakes of data usage. Systems may know more about us than what we know about ourselves. She stated that in Norway, It’s difficult to set a local legal and ethical framework, and an international one is a big challenge. She was also aware of usage of educational information outside the university, and she illustrated this with an example of a student with dyslexia whose data could be maliciously used by health insurance companies.
The floor was then given to Célia Zolynski (UVSQ) who spoke about the main aims of regulation such as making stakeholders aware of the problem of the protection of personal data, ensuring that protection and providing algorithms transparency, and all of those have to be applied case by case, respecting the proportionality principle. To achieve this goals, technicals and organizational measures have to be taken, for the whole duration of the project. She enunciated the main issues when using Big Data, like the asymmetrical relation between the institution, teacher and learning, the specific right of individuals for privacy, protection, being informed, rectification, profiling, being forgotten or not being discriminated. She proposed that institutions should adopt a solution involving the learner in the data processing, sharing data, within an “ethics by design” plan of cooperation.
Then it was the turn of Thomas Toulotte (UVSQ) who went in depth with the legal issues already stated. There are many situations in which LA techniques will require the use of personal data. When facing the problem of personal data, regulation provides some mechanisms to undergo this issue, like anonymisation or pseudonymisation, some rules for sharing between countries with the same level of protection, and a definition of consent for the collecting of data. There are also specific rights to take in account for underrage, like a stronger right to forget, and a distinct setting in the case of research activities. LA activities are usually related to profiling students and the use of prediction algorithms. In the first case, even if specific policies for this haven’t been developed yet, they are subject to the rules of consent. In the second one, algorithms are subject to policies concerning transparency and neutrality.
N. Sclater (JISC) was the last to contribute before the coffee break. He started illustrating the main issues with an example of initiatives to create an open source format to collect learners data which died during experimentation because of parents concerns. He stated that university aims are different from those of commercial organizations and they should be treated differently. Other points to take into account are the wrong conclusions of LA because of wrong data, the issues of predicting and presenting predictions to students which do not need it, model validation importance, harmful emerging effects for the motivation of students because of their “being monitored” feeling, and the reinforcement of discrimination attitudes which can occur when the recommendations are not presented in a positive way. He concluded that even if we use modern techniques of LA we have to always keep in mind that students are not numbers, and they are much more complex than the models we can infer from them.
C. Stueckelberg closed the panel thanking the contributors and inviting participants not only to use modern LA techniques but to combine it with other modern but not necessarily technological learning like serious games or deep-listening techniques.
Keywords: “ethics by design”, personal data protection, right to rectification, profiling, anonymisation, consent
Second session of day 5 of the leadership school
Panel on teachers, students and staff involvement: some success stories
The facilitator Christine Wihak (Thompson Rivers university) started the panel by giving a small introduction about LA and its implementation, and she asked what could encourage teachers who are teaching in small classrooms with a small number of students to implement LA.
The first talk in the panel was presented by Jordi Conesa Caralt (UOC). Jordi gave an introduction about his university and presented LA and their challenges in implementing them as it is a strict campus of 60,000 students, and they have difficulties in bringing their projects in real.
He discussed the main agents affected by LA: teachers, students and parents and the important correlations between them.
He described some analytics systems for teachers, one example is the “emotional thermometer” that determines the emotional climate of the classroom from students’ messages. He also described some analytics systems for students, one example is ICT flag that includes gamification where the system provides award when a student wins like removing one question from the exam and other awards to motivate students to use the system and improve their level. Finally he described analytics systems for university staff, one example is UOC index that provides API for data warehouse to facilitate the obtaining of aggregated data. He concluded that the right use of data can support learning, teaching and managing and that the difficulty is not only in creating analytics systems but also in implementing them in real environments.
The second talk was presented by Sarah Guri-Rosenblit (Open university of Israel). She gave an introduction about her university and described LA and the existing gap between descriptive and prescriptive phases. She detailed three projects at her university addressing students in the Arab sector, enhancing study skills, writing workshops to novice students and addressing potential early drop-outs. These projects are pilot interventions that would be implemented on a large scale.
The third talk was presented by Wim Van Petegem (EDEN). Wim described LA and involvement of teachers, students and staff. He described that students always tell teachers “challenge us” and consider that technology is not fully exploited in order to achieve optimal learning. He talked about technology trends from MOOC’s, tablet use, game-based learning, learning analytics and formative analysis. He confirmed that there should be new ways of teaching and learning and that teachers should be involved in LA and share knowledge. Then he described LACE (Learning Analytics Community Exchange) that is supporting a growing European community of interest in using data to improve education and training in schools, universities and the workplace.
Finally, he invited teachers to challenge students with all these successful research results and innovations in LA that should be motivating and challenging to them.
After this last session, Anne thanked all members for participating in the second D-Transform leadership school and sharing their knowledge and experience in the domain.
Keywords: formative analytics, game-based learning, LA implementation, student engagement.
Tips of the day:
- Stakeholders should be aware of the stakes of personal data protection
- We need a global ethical and legal framework but the implementation would be very difficult as each HE institution has a local context
- It is important to develop new ways of teaching and learning through LA.
- We should get teachers involved and help them share knowledge
- We should challenge students and show them the positive impact of LA implementations in their educational systems
- Data is everywhere around us, but the right use of this data is very important to support learning
- The correlation of LA agents: teachers, students and parents is vital for ensuring LA success
- The main challenge is not only to create analytic systems but also to implement them in real environments