Although studies have typically found that exposing people to factchecks can help combat misinformation, the effectiveness of labelling—the strategy used by the most widely used social networks—is less clear. Using online survey experiments conducted in Brazil, India, and the UK (N=1000 in each country), we find that: (i) the effectiveness of labels in reducing the perceived accuracy and likelihood of sharing false coronavirus information varies by country, with stronger effects in India and negligible effects in the UK; (ii) individuals with high levels of conspiracist ideation and trust in social media are more likely to perceive false information as accurate and express intentions to share it; however, (iii) labels are neither significantly more nor less effective for these groups. These findings suggest that while labelling false posts can “move the needle” in addressing misinformation, the effects are modest compared to fact checks and may be context-dependent. This study highlights the importance of tailoring misinformation interventions to specific cultural and social contexts, providing insights for policymakers and platforms seeking to mitigate the spread of false information.