Abstract
Scientific misconduct obstructs the advance of knowledge in science. Its impact in some disciplines is still poorly known, as is the frequency in which it is detected. Here, I examine how frequently editors of ecology and evolution journals detect scientist misconduct. On average, editors managed 0.114 allegations of misconduct per year. Editors considered 6 of 14 allegations (42.9%) to be true, but only in 2 cases were the authors declared guilty, the remaining being dropped for lack of proof. The annual rate of allegations that were probably warranted was 0.053, although the rate of demonstrated misconduct was 0.018, while the rate of false or erroneous allegations was 0.024. Considering that several cases of misconduct are probably not reported, these findings suggest that editors detect less than one-third of all fraudulent papers