Why 'No Evidence of Effect' Is Not the Same as 'Evidence of No Effect'

A Subtle but Consequential Logical Slip

It is extremely common, even in otherwise careful clinical discussion, to hear "there's no evidence this works" used interchangeably with "this has been proven not to work" — these are, in fact, very different claims, and conflating them is a well-recognised logical error.

Absence of Evidence Is Not Evidence of Absence

A now-famous clarification in the medical literature makes precisely this point: a lack of statistically significant findings, particularly from small or underpowered studies, does not mean a treatment has been shown to have no effect — it may simply mean the available research has not yet been sufficiently large or well-designed to reliably detect an effect that genuinely exists (Altman & Bland, 1995).

Why This Distinction Matters for Newer or Less-Studied Treatments

This is particularly relevant for treatments that are newer, less commercially profitable to study, or simply harder to research — a limited research base reflects limited research investment and methodological difficulty just as often as it reflects a genuinely ineffective treatment, and the two possibilities require very different interpretation.