I am sure someone has already commented on this, but it occurs to me that openness and analytics have a problematic relationship.

For instance, in the Intro to Psych course we are developing on several campuses, we’d like to use analytics to do things like nationally norm percent correct on questions. So you answer a question, and I see as a teacher what percentage of students nationally got that answer correct, and what other questions students that got that one wrong tended to also get wrong (or what support they found helpful).

That’s trivial to do in a centralized system. In a distributed world where everyone is working off of copies, and locally revised copies at that, it gets difficult — local alterations to question phrasing will corrupt your data….