By: Johann Larusson, Former Research Scientist, Pearson
We should, as we do already, put our very best effort towards guaranteeing that student privacy is at the heart of everything we do in education, especially when using technology to facilitate teaching and learning. Student privacy is a concern and part of many conversations about personalized learning and adaptivity.
How do adaptive programs use student data? By digging into data use, we can see where student data privacy needs to be protected.
An adaptive learning program is based on statistical methods informed by variables (data points) that describe student activity and outcome. In order to develop personalized learning pathways, adaptivity does not need to know any specific personal identifying information for a given student. The program needs to be able to collect and summarize historical data points with newer evidence, and know that they all belong to the the same student. But it does not need to know, nor does it inform the program, that it is a specific student or teacher. The data is telling a story for some “one thing”. Think of adaptivity as being a statistical method working towards identifying trends in behavior for a given a population of users or estimating probabilities of success for certain events or goals for a “type of user”.
Of course, at some point, in the design of an educational system, a connection needs to be made between what is being recommended and an actual student. How would we otherwise communicate the valuable recommendations to the right person? This is the sensitive part and this is also where human-made rules come into play.
To do adaptivity right, we need well-designed systems, with well thought out instructional designs, that adhere to the current rigorous security standards. But, we also need people — the human element — to understand what can be done with the data and what should not be done with the data. Having access to data brings responsibility.
To the adaptivity itself, it is all about numeric values and probabilities. It is not about identities. Embedding it in the right design has enormous potential.
I talk more about personalized learning through adaptivity in this video: