The concept of a conflict of interest is crucial to the proper handling of alleged scientific fraud. Investigators appointed to review misconduct allegations should not have close links to either the complainant or the respondent scientists, and the most enlightened institutional policies include clauses to guard against 'real and apparent' conflicts of interest — intelligent phrasing that makes it possible to query the wisdom of an individual's appointment to an inquiry panel without having to question their objectivity.

Most policies guard less well against another kind of conflict of interest: affiliations to an institution. Managers and administrators have an apparent interest in playing down the extent of any problems at their institutions. A finding of misconduct can result in bad publicity and financial penalties, not to mention the abrupt severance of professional relationships with colleagues who are often well liked and respected.

For taxpayer-funded research, the check on these institutional conflicts of interest should be government oversight. But in many countries, such oversight is patchy or nonexistent. Even in the United States, where a federal policy mandates oversight of institutions by government funding agencies, there are clear and surprising failings — as revealed by a lawsuit that brought to light details of a misconduct investigation overseen by the US Department of Energy (DOE; see page 20). No misconduct was discovered, but errors of judgement were.

Officials who adjudicate misconduct cases must take oversight more seriously.

The oversight of the investigation had several problems, including that the adjudicating officials did not read the inquiry's report or keep a final copy of it, and that they approved an outcome on the basis of meetings and telephone calls with a lab manager, rather than their own scrutiny of the evidence. The officials who oversaw the investigation were the people who funded the research in the first place — again, there is an apparent conflict of interest.

DOE officials say that their actions are consistent with federal policy, but they seem disconcertingly casual about it. Important decisions were taken informally by a small group of officials and an adviser who apparently shared a common interest: to see the matter quietly resolved. The procedure ought to be more formal, better documented and even adversarial, with the institutional managers required to satisfy officials whom they do not know and who have no stake in the case outcome. This is closer to the more careful oversight of alleged misconduct by both the Office of Research Integrity at the Department of Health and Human Services (DHHS) and the inspector-general of the National Science Foundation.

No wonder policy experts are highly critical of the DOE. One influential US watchdog group, the Project on Government Oversight (POGO) in Washington DC, contacted during Nature's reporting of the failures described above, is now preparing a public letter to Steven Chu, the US energy secretary, to argue that the DOE needs a professional office of research integrity. The DOE does have an Office of Inspector General, but it typically handles more conventional alleged financial waste and fraud — not scientific misconduct, which requires specialized expertise. The National Institutes of Health (NIH), funded by the DHHS, has a much higher incidence of misconduct allegations in research than does the DOE. POGO suggests — plausibly — that this is the result of a rigorous scientific oversight system that encourages whistleblowers to come forward, rather than of any greater integrity among DOE-funded scientists.

Having an office of research integrity at the DOE could remedy this situation, and might also be politically astute. In a climate in which conservatives are looking for reasons to cut science budgets, and in which energy and climate research is a hot political topic, the DOE needs to be proactive to prevent scandals about alleged scientific misconduct escalating. One needs only to think of the bruising congressional inquiries into NIH-funded science in the 1980s to show the dangers of taking research integrity for granted.

In the meantime, DOE officials who adjudicate misconduct cases must take oversight more seriously. Their goals should be to prevent retaliation against whistleblowing scientists who have raised concerns, to ensure that those who have been unfairly accused are fairly exonerated and to reach the truth about the integrity of the science they fund. A careful look at the evidence is the first step.