Retractions of popular papers ‘often go unnoticed’ 3 years ago Analysis argues that science has a ‘critical problem’ because its self-correction process is slow and ineffective