The Phrase "No Evidence" Is A Red Flag For Bad Science Communication

Argues that 'no evidence' is used to mean two opposite things — 'plausible and probably true but not yet formally studied' (no evidence of human-to-human COVID transmission) and 'we have hard evidence this is false' (no evidence 45,000 died of vaccines) — and that conflating them corrodes trust in science journalism (the reader who watched 'no evidence' claims flip to true repeatedly won't be reassured about vaccines). Deeper: there is no consistent meaning of 'no evidence,' shown by four cases that each demand different treatment of journal-articles-vs-informal-evidence (the parachute paper; alien-abduction eyewitnesses; homeopathy's 90 studies; Henry VIII's spleen). Real truth-seeking is Bayesian. Prescription: say 'no evidence either way,' or 'Scientists: X doesn't work' (centering fallible scientists rather than faux-objective certainty), or best, actually investigate and engage the argument people believe.

Quality 74 · Strong. Strong band, upper. A sharp, memorable, frequently-cited point about science communication, with a clever four-example demonstration that the folk concept of 'no evidence' is incoherent and a constructive fix. Upper-Strong: an important single-idea essay rather than a sweeping treatment.

Claude’s paradigm shift 45 · Moderate. Moderate. Applies Bayesian epistemics (and his own 'two types of no evidence') to science-journalism language; the equivocation diagnosis is a fresh, exportable framing.

Real-world impact 3 · Moderate. A sharp, frequently-cited point about science communication — 'no evidence' incoherently means both 'unstudied but probably true' and 'proven false,' corroding trust — with a clever four-example demonstration and a constructive fix. Conceptual influence within science-communication discourse, no material change — modest RWI.