Hi there. I wonder if you are familiar with design of experiments (DoE) or factorial designs from the statistics branch.

With a DoE/ factorial design setup you can compare classrooms based on different factors that are set up for each classroom. Outcome results such as class average grades can be used for comparison.