San Francisco

Scientific journals have been striving of late for greater openness about the competing financial interests of their contributors. But a spat over a paper in Science suggests that clashing views over what makes a conflict of interest means full transparency is hard to achieve.

Earlier this year, Australian activists noticed that a 2002 paper on the spread of herbicide resistance from transgenic canola to nearby fields (M. A. Rieger, M. Lamond, C. Preston, S. B. Powles and R. T. Roush Science 296, 2386–2388; 2002) did not mention that two biotechnology firms — Monsanto and Aventis Crop Sciences (now owned by Bayer) — paid nearly 20% of the costs of the trials.

Alerted to the fact by a reporter for an Australian television programme in early May, Science contacted the authors for an explanation. Science requires contributors to declare financial ties that might be construed as influencing the outcome of their research.

The authors responded that they did not view the company funding as a conflict of interest. Industry co-sponsors don't participate in the design or conduct of the study, nor are they permitted to vet the findings or stop publication, claims co-author Christopher Preston, a molecular ecologist at the University of Adelaide. “I refuse to participate unless I can call the shots,” Preston told Nature.

Although Science concluded that the funding did not amount to a conflict of interest, it has now revised its disclosure policy as a direct result of the incident, according to a statement provided to Nature on 23 June. Now, all funding sources must be revealed in the paper's reference section, Science says.

Many journals ask about conflicts of interest, but some authors don't realize that they have them, says Nicholas Cozzarelli, editor-in-chief of the Proceedings of the National Academy of Sciences. For instance, one contributor didn't feel that his position as chief scientist of a firm that was supporting his academic work needed to be mentioned as he didn't think the paper would affect the company's stock price, Cozzarelli says.

Nature has requested and published details of competing interests for every paper accepted since October 2001. Despite the rising number of researchers with ties to industry, of the 1,300 or so papers published under the policy, only 50 have declared competing interests.