You've made a great point.
I'd like to see them report how many deleted posts were in treatment versus control. If it was symmetric, I'm more willing to pay attention to their reported treatment effects. Your theory (mostly) relies on a proposed asymmetry.