How will the A/B testing be overseen? Considering the limits to trying different combinations due to a lack of students you probably want to go at it most effectively. I am curious how that will be done. Including de:biasing the overseers.

Might be overthinking on my side. It sounds really exiting already.