Simple logic warnings improve climate argument evaluation across beliefs and partisanship

Collective climate action requires shared beliefs aligned with scientific consensus, yet fallacious information leads people to misjudge arguments. The mechanisms behind these failures remain unclear. Here we isolate the independent effects of partisan identity, prior beliefs and cognitive limits on argument-quality detection. Across four preregistered studies (N = 4,000), we introduce a protocol that orthogonally manipulates argument quality and direction across realistic contexts. In neutral contexts, prior beliefs dominated judgements, with partisanship and argument quality having smaller effects. Contrary to identity-protective accounts, embedding arguments in hostile partisan contexts did not increase partisan bias or impair argument-quality detection. Instead, partisan attacks reduced reliance on prior beliefs, prompting participants to move away from their prior positions. Finally, minimal logic warnings indicating possible fallacies enhanced discrimination between weak and strong arguments. These findings show climate argument evaluation is context sensitive and can be improved with simple logic warnings.