At last month's meeting of the American Association for the Advancement of science, there was an inspiring talk about how the open sharing of scientific data could provide new avenues for research. But the same session also provided a cautionary tale of all the factors that can get in the way of effective sharing of data. That talk came courtesy of Vernon Asper, a researcher at the University of Southern Mississippi. Asper normally studies natural hydrocarbon seeps on the floor of the Gulf of Mexico, and found himself dragged into the media spotlight amidst a swirl of competing interests as he tried to study the oil spill.
Asper said that there was a clear "truth" about the spill that everybody was interested in: how much oil was spilling into the Gulf, and at what rate. But, in the absence of any way of directly measuring it, everybody was forced into relying on indirect ways of estimating the flow. If the gulf oil spill were a situation where nobody had money riding on the final outcome, these estimates might be combined to provide a rough final number along with a sense of the uncertainty associated with that number. Unfortunately, this wasn't a case where nobody cared.
According to Asper, there were three groups that had a vested interest in the final value of the amount of oil spilling out into the gulf. The companies (like BP and Halliburton) that had been drilling the well wanted the number to be small. The media, which can attract eyeballs through drama, wanted the number to be large. Scientists, in general, just wanted the actual number.
Here, the scientists shared an ally in the government, which also wanted the most accurate number possible. Unfortunately, the government and scientists had different ideas about what to do with the numbers they were generating. Researchers, as is their tendency, wanted to share it through collaborations and publications. The government, however, was preparing for the inevitable court case against the companies involved, and wanted to keep the numbers it generated private until they could be used in the legal arena.
The end result of all these competing interests was a complete absence of coordination. Research ships (including one used by Asper) gravitated to the site, but nobody knew who was doing what. He described how they'd identify a research vessel using radar, then call them on the radio to informally ask what sorts of samples the ship was there to obtain. This ad-hoc system helped, but there was still a lot of cases where the same work was done by multiple vessels. At the same time, if one ship's experiments failed due to technical reasons, there was no mechanism for them to let anyone else know, which might have allowed another ship to make up for the loss.
From Asper's perspective, things really went bad when the data gathering ended. "The press was waiting to greet the boat when it came in at 5am on a Sunday morning," he said. None of the researchers he was with had any expectations that this would happen, and no media training to prepare them for it. "We talked as openly as we would to another scientist," Asper said, saying he was "amazed someone finally cared" about the sort of research he did. His university encouraged him to blog about the experience, and members of the media also found his blog.
The reports that resulted, however, often eliminated a lot of the caveats and limits that Asper placed on his own results. NOAA felt compelled to respond, and termed some reports "misleading, premature, and, in some cases, inaccurate," a response that increased the mistrust between researchers and the agency that funded them. Eventually, this also gave NOAA a black eye. As the caveats were reduced by further studies, it turned out the original reports were in the right ballpark, leaving NOAA looking like it had attempted to downplay the truth.
(Meanwhile, the media attention has made it harder for Asper to publish his work, since journals have rules against publishing papers that describe previously publicized work.)
Asper also described how even something as simple as equipment choice can produce seemingly contradictory results. He showed how a piece of hardware that tended to land rapidly on the bottom of the Gulf would produce a picture showing no contamination of the sediments there, simply because it would blast away the surface layers when it hit. Equipment that was designed to land gently, in contrast, was used to show that a film of hydrocarbons had been deposited down-current of the blown well.
Circumstances like this can easily confuse press reports that are filed on a tight deadline and prepared by people without a background in this sort of research. Placed in an environment where different parties have vested interests in pushing orthogonal "truths," and they can be used to push an agenda.
Asper wrapped up with some lessons specific to these circumstances: data gathering needs to be planned more carefully, the data need to be shared, and the analysis of it needs to be decoupled from issues of legal liability. But the larger lessons were implicit in his description of the media frenzy that followed the oil spill. There's no way to predict when some quiet corner of the scientific endeavor is going to be turned into front-page news, and we could do a far better job of working with scientists to make sure they're ready to handle things when it happens.
Listing image by Photograph by usm.edu
reader comments
28