Maintaining integrity in science seems to be in fashion. US officials are preparing to review policies on scientific integrity drafted by some 19 government agencies. Canada is overhauling its research-integrity policy, and the UK Research Integrity Office is looking for new sources of support after losing government funding. The European Commission has also started to discuss how to include scientific integrity in its research framework agreements.

The US policies will apply not only to agency scientists, but also to political appointees, managers and public-affairs officials. This broad scope is deliberate and intended to avoid a repeat of the alleged abuses of science that occurred under the administration of former US President George W. Bush (see page 262).

Still, to frame such policies so broadly does bring potential trouble — how should officials apply them to scientists? Vague requirements, such as to approach research objectively and to welcome constructive criticism, are good professional codes of conduct — but they must not be interpreted as a stick with which politically conscious managers can beat scientists to suppress inconvenient scientific findings.

It would be too easy for officials to violate the intended spirit of a loosely worded scientific-integrity policy by claiming that scientists had violated the letter of it. The United States classifies scientific misconduct as falsification, fabrication or plagiarism, and the broad concept of integrity must be coordinated carefully with that narrow definition. And, of course, good policies are not enough if they are not implemented wisely. To that end, US federal agencies must continue to improve their practices through training for managers of scientific staff, and ensure that they turn to the proper scientific expertise when misconduct is alleged.