AbstractWhen paternalism is deemed morally justified, weak paternalism—which restricts itself to assisting the target of paternalism realize his own preferences—is the preferred (less problematic) alternative. In determining the appropriateness of weak paternalism, the level of certitude of the paternalist regarding the correctness of her assessment of the true preferences of the one‐paternalized is obviously a crucial factor. Yet in the ethics of paternalism this parameter has escaped systematic treatment. This paper aims to initiate discussion on this indispensable consideration for weak paternalism. Analysing a real‐life dilemma of paternalism in healthcare, the paper focuses on the theoretical question of how the paternalist can optimize her certitude by combining personal knowledge of the individual patient with population data on treatment refusal/consent of patients facing similar decisions. The paper presents an outline of a decision‐making scheme that can be valuable in medical ethics and beyond.