DIVERSION — A spreading procedure undermines the work
- Situation: A frequently reused procedure receives continuing support while its recipients report worse results.
- Question: Does support favor a misleading signal, do selection conditions favor the variant, or is another mechanism responsible?
- First useful result or blocker: A discriminating causal account and a feasible response comparison, not a label applied to its users.
- Mantra: Establish the consequence; trace support and continuation; distinguish diversion from selection; compare response targets; check what persists or changes afterward.
- Start with: DTM.1 → DTM.7 or DTM.5; use DTM.6/.3 for consequential interactions and spread, DTM.8/.9 for response, and DTM.4 for return.
- Stop or return: Retain competing explanations if the observation does not distinguish them. A protection proposal needs the actual means and authority to perform it.