Here’s a very interesting post by law professor Frank Pasquale, asking if some kinds of analytics should be done at all, given the potential for harm.
We have the best of intentions for predictive learning analytics, but do these predictions improve learning outcomes, even in the best cases? Do we risk prejudicing teachers about the potential of students, or even activating stereotype threat in students themselves?
I hope we can show the value to learners and teachers of learning analytics, and minimize the possible harm. I’ve argued in the past that if we have data and analysis methods that could help learners, we may have an ethical obligation to pursue that potential. But I think we probably shouldn’t assume that any predictive learning analytics system would be “better than nothing.”