The US National Science Foundation (NSF) is unique among the world's science-funding agencies in its insistence that every proposal, large or small, must include an activity to demonstrate the research's 'broader impacts' on science or society. This might involve the researchers giving talks at a local museum, developing new curricula or perhaps forming a start-up company.

The requirement's goal is commendable. It aims to enlist the scientific community to help show a return on society's investment in research and to bolster the public's trust in science — the latter being particularly important given the well-organized movements currently attacking concepts such as evolution and climate change.

Unfortunately, the very breadth of the requirement can leave researchers struggling (see page 416). Few of them have training in the activities involved — especially when it comes to education and outreach — and the NSF has not done enough to provide a support infrastructure to help.

Such an infrastructure does exist in embryonic form. For example, a few research institutions, including Stanford University in Palo Alto, California, and the University of Wisconsin–Madison, already have centres that aim to connect scientists with experts in teaching, education and public outreach, to equip them with the necessary skills and to disseminate best practices. And a few places, such as the University of New Mexico in Albuquerque, have developed workshops in which graduate students, postdocs and junior faculty members get professional training on how to interact with the public, media and government. Such efforts need to be expanded and institutionalized throughout the country.

Broader-impacts efforts also need to be better evaluated and rewarded. For example, the NSF should consider offering cash awards for the best broader-impact activities, the money from which could help to continue or expand the activities. This would motivate investigators to put greater effort into these endeavours, and would spread the word to other scientists about the sorts of activities that have proved successful.

Such initiatives would motivate what is really needed: a fundamental change in the culture of science to value not just achievement in the laboratory, but also work that makes science a part of people's lives.

The US Congress can help. The America COMPETES Reauthorization Act, which would extend an earlier boost given to the budgets of the NSF and two other science agencies, requires grant applicants to show that they have received support from their institutions in meeting the broader-impacts requirement. It also calls on the NSF to clarify the requirement's goals and to improve evaluation of the outreach activities. The act is being held up by political manoeuvring, despite strong bipartisan support. Congress should pass it without delay.

It is a truism to say that science and society are intertwined. But no relationship should be taken for granted. The NSF needs to help scientists show the world that their work is valuable.