Unmasking publication bias in the COVID-19 pandemic

During the COVID-19 pandemic, face masks were widely promoted and mandated as a key non-pharmaceutical intervention, despite mixed and often weak evidence regarding their effectiveness in reducing viral transmission. This study systematically examines publication bias in the scientific literature on masks published between 2020 and 2024. Using a comprehensive literature search, 1,284 peer-reviewed articles were identified, manually classified into 19 thematic categories, and coded according to their explicit or implicit stance on mask effectiveness. Overall, 89.5% of publications expressed a positive stance toward masks, whereas only 1.6% reported negative conclusions and 7.6% were neutral. Notably, even articles unrelated to mask efficacy frequently included affirmative statements endorsing mask use. In contrast, randomized controlled trials, which constitute the highest level of clinical evidence, were the only category in which negative conclusions outnumbered positive ones. The results reveal a pronounced asymmetry between the strength of the available randomized evidence and the overwhelmingly positive portrayal of masks in the broader literature. I discuss several mechanisms that may have contributed to this bias, including publication selection bias, overreliance on observational studies, conformity statements and virtue signalling, social and institutional pressures, and self-censorship driven by perceived prosocial motives. Taken together, these findings suggest that mask research during the COVID-19 pandemic represents a clear and substantial case of publication bias, with important implications for evidence-based public health policy and the integrity of scientific discourse.