Well over a million papers were published last year in journals whose contents are recorded by the Institute of Scientific Information. Against that level of activity, tens of incidences of fraud and other severe research misconduct around the world is a statistically small number. But the implications of such cases for individual institutions, and even for a national scientific community, can be extraordinarily costly and damaging — in time, in possible litigation, in the smearing of colleagues' reputations, and in public confidence in science. That is why, for example, recent cases of fraud at the laboratories of the prestigious French institution INSERM and of the equally prestigious Max Planck Society have proved so significant and painful.

Acting promptly, decisively and transparently is what has not happened at INSERM but what, commendably, did happen in the Max Planck Society's investigations into a recent prominent case of fraud — to the credit of Hubert Markl, head of the Max Planck Society, but also to the credit of the Deutsche Forschungsgemeinschaft (DFG), which has provided one of the best overviews of misconduct and how to deal with it (see www.dfg.de). It is encouraging that other European agencies, including INSERM, are now collaborating with their German counterparts and with others who have already set standards — in Scandinavia and Britain, for example.

Following principles adopted in the United States, the DFG's document includes the threat that, in future, an institution lacking adequate procedures for handling misconduct might not receive funds. But pre-empting misconduct is just as important, and the DFG is therefore to be commended for hinting that the same penalty could apply to those institutions that lack good guidelines for the proper conduct of research itself. One prosaic but essential possibility is an insistence that institutions keep good laboratory notebooks and also archive them for a guaranteed ten years — such notebooks are often crucial in understanding what has gone wrong.

In any country, there will inevitably be both explicit and implicit attempts to undermine such goals. Explicit, because many laboratories are run by people who are under great competitive pressure and claim that they have neither the time nor the inclination to introduce new approaches to the management of their colleagues. Implicit because in biomedical research, where extreme misconduct tends to occur most frequently, funding is increasing — to an extraordinary extent in the case of the US National Institutes of Health (see page 734). Such a beneficent environment is in itself not conducive to tightening up the quality of one's act, unless the funding agencies insist on it as a condition for receiving support.

Much of this comes down to an issue of leadership and of institutional accountability for the use of public funds. Regrettably or not, many academic researchers find themselves required to achieve institutional targets — long gone, for them, are the relaxed days of disinterested enquiry for its own sake. But too often, a sense of direction and accountability has not trickled down the lab hierarchy. Younger scientists are still left in isolation and without much supervision or feedback, and the principles of good person-management — appraisal and career development — are frequently in short supply.

In short, like it or not, the pressure of misconduct cases coming from one direction, and a greater sense of institutional purpose coming from the other, are forcing researchers to act in a way that requires accountability to be more formalized, in both research and employment practice, within laboratories. The question is whether institutions are willing to grapple with the commitment this requires in both time and staff, and whether funding agencies will be prepared to contemplate the additional costs. Presumably, in the absence of enforcement from above, they will not.