Corrections are Effective for Science Misinformation

In their impactful pre-registered meta-analysis, Chan and Albarracín aimed to determine the degree to which the public updates science-relevant misinformation following a correction (e.g., p.1514, paragraph 3). Based on an impressive 74 studies and 205 effect sizes, the authors concluded that “attempts to debunk science-relevant misinformation were, on average, not successful (d=0.19, P=0.131, 95% CI −0.06 to 0.43)”; p.1514), with the effect of corrections “smaller than those identified in all other areas” (p.15171; e.g., politics and health). In this commentary we show that the reported null effect was due to the inappropriate pooling of two distinct effect types into a single estimate. This clarification is necessary because meta-analyses are often perceived as the ‘gold standard’ of evidence, and numerous papers have cited Chan and Albarracín1 as evidence that corrections have little-to-no effect on science-relevant misinformation (35% by our estimates; see Supplementary Table 1).