RMP.2:2 - Problem

Method prestige often chooses the design before the question’s criticism demand is understood. Randomization, statistical significance, interviews, simulation, triangulation, systematic review, or AI-assisted synthesis becomes a ritual answer rather than one candidate relation. The resulting design may be detailed but unable to expose the error that matters.

Premature selection hides missing owner results: a construct is not related to an observable, a source is inadmissible, an instrument cannot discriminate the change, a site cannot be entered safely, participants cannot be recruited legitimately, or the authority needed for the proposed Work does not exist. The team continues into protocol work and discovers too late that a different design—or no study—was required.

Indiscriminate holding wastes the other way. If a remaining need only assigns an already-required record locator or selects a schedule within a qualified fixed window, holding the family decision adds no epistemic protection. The practitioner needs a completed branch whose boundary can be tested, not a confidence adjective.