Bumping in case anybody has any advice about whether, in your experience, in-class differentiation is comparable to single subject acceleration.