Just read Niall Sclater’s literature review for JISC: Code of practice for learning analytics: A literature review of the ethical and legal issues. The report asks a lot of important questions. And it mentions, albeit in passing in the “Rationale” section, what for me is the key issue: “Greater trust and a better relationship with the people you collect information about.”
But that word “relationship” masks the hard human issue about education and learning that no amount of data will resolve. Ultimately “authentic” relationship is between embodied (or formerly embodied, in the case of dead writers) individual humans. I don’t want to get all Bakhtinian about it, but while analytics may help ask questions, they will not provide the solutions.
A problem we uncover when surveying students (either “failing” students, or students on “failing” courses) is that they do not feel that their teachers know them or care about them. I do not think these students will be impressed if the response is to fix the broken relationship with an algorithm.
At best analytics can help start a conversation. People have to be willing to take the conversation on. That is: people, not institutions. Analytics may help pinpoint areas where hard work needs to be done, but are not likely to make that work any easier.
Management wants “the” answer to achieving particular mandated targets in order to protect or augment resources (or consolidate and preserve “power”). People (learners and teachers) are not trusted to do this “consistently” (as if they ever could!). So it is hoped that algorithms will: a) discover the right answer and then b) enforce a solution in line with targets and outcomes (which shift with the political wind).
We must not let loss of trust be simply collateral damage.
Other important contributions to the discussion