The concept of a return on research investment has acquired a sharper edge since the global financial slump began. But an assessment of those returns should include more than knowledge for its own sake and economic growth — as highlighted by a timely series of articles in the latest issue of the journal Minerva.

Under the intellectual and editorial leadership of the policy scientists Barry Bozeman of the University of Georgia in Athens and Daniel Sarewitz of Arizona State University in Tempe, the journal presents case studies that analyse a broader way to measure returns on investment: public values. These public values include not only the commonly discussed knowledge and economic criteria, but also information useful to decision-makers, participation in agenda-setting by stakeholders and communication to the public in general. In fields as diverse as hurricane research and nanomedicine, the contributors demonstrate how initial ambitions to target these values get lost as projects unfold. Too often, the pull of the heavyweight 'science for its own sake' and economic agendas crushes wider intentions.

The notion that society considers any advance in knowledge to be inherently good cannot be sustained.

Policy-makers, funders and scientists should take note. For example, a paper by Ryan Meyer, also a policy scientist at Arizona State University, focuses on the failure of the US government's Global Change Research Program to deliver broad public value (Minerva 49, 47–70; 2011). Basing his studies on public statements and private interviews with researchers and political decision-makers, Meyer says that US climate programmes have in the past two decades benefited from public investment of more than US$30 billion, but have largely failed to produce information and participation in the forms that policy-makers and the public wanted. The notion that society considers any advance in knowledge to be inherently good — even if the science fails to meet the objectives and priorities it was meant to address — cannot be sustained, says Meyer.

Coincidentally, at a workshop last week in Hamburg, Germany, a gathering of climate scientists, policy experts and philosophers of 'post-normal science' articulated a similar perspective. Science becomes 'post-normal' when facts are uncertain, stakes high, values in dispute and decisions urgent; in such cases, societal needs must be taken into account to avoid costly mistakes. Climate research certainly fulfils this definition. But, according to the workshop participants, most climate researchers continue to act as if purely scientific values are, and will always be, adequate to set the agenda.

Instead, researchers and, especially, their funders must embrace the idea that public and stakeholder participation can help to define research priorities. And they must do more to track and communicate all outcomes. Policy-makers need to ensure that those with direct needs for climate-related information — businesses, regional planners, government departments — have a greater say in the kind of services and knowledge that they expect publicly funded researchers to produce, and in assuring the quality and relevance of what is delivered.

The agenda here is to broaden rather than restrict climate research: to ensure that, alongside studies driven by the priorities of creative scientific imaginations, research also enhances public confidence in society's use of science. A fuller consideration of these issues will also maximize science's public value, in the form of successful collaboration across disciplinary boundaries, delivery of information useful to stakeholders, and transparency and well targeted communication.

Grant-giving agencies need to take a fresh look at how to address these issues. Lay representation in decision-making, as has already been tried on a limited scale at the US National Institutes of Health, is one way to increase public involvement. Previous exercises in engagement on nanotechnology have led to positive, albeit restricted, outcomes (see Nature 448, 1–2; 2007).

But the problem goes deeper than public engagement. The Global Change Research Program has involved stakeholders at various stages, but, says Meyer, “it is not clear that the US currently has the institutional capacity to achieve the public values promised about climate science”. If that is indeed the issue then, given the sheer scale and breadth of US capacity in these topics, one has to wonder whether anyone will be able to achieve what these authors are seeking.

As is often the case, policy researchers are good at identifying problems but slower to reveal how to improve matters. To his credit, Meyer offers some practical ways for the programme to improve the leverage of, and incentives for, a broader participation in delivering public value.

More importantly, these studies highlight a significant deficit in current typical appraisals of science and technology outcomes. They should serve as cautionary tales about the danger of scientists' interests, deliberately or otherwise, becoming too dominant in determining outcomes. And they introduce ways to assess failures in social returns on investment that, one can only hope, will help to improve science's public value.