We call for greater transparency and speed in investigations of scientific misconduct. For the benefit of the scientific community, swift assessment of the validity of published results should be distinguished from legally convoluted verdicts that concern wilful deceit.

Journal publishers and academic institutions have differed markedly in their response to recent cases of alleged misconduct. For example, the committee investigating publications by the Dutch social psychologist Diederik Stapel (Nature 479, 15; 2011) published an interim report within two months and is releasing its findings in under a year (www.commissielevelt.nl). And a panel investigating the papers of another social psychologist, Dirk Smeesters (Nature 487, 18; 2012), reported its results within a week of his resignation. Such prompt and detailed reporting contrasts with that of other protracted enquiries that have culminated in retractions of papers.

Web-based initiatives such as Retraction Watch (www.retractionwatch.com) can help by offering timely, although informal, alerts about suspect papers. Even if fabrication or falsification cannot be proved beyond reasonable doubt, statistical indications of problems with the data — such as gross errors or remarkably consistent findings — are scientifically relevant.