Pinpointing the amount of oil lingering in the Gulf of Mexico continues to be a source of frustration for journalists and scientists alike, with multiple, contradictory—if not necessarily “dueling”—research reports having been published on the subject over the last few weeks.

Last month, the federal government released an “oil budget,” which claimed that 74 percent of the crude had essentially been dealt with through skimming, burning, dispersion, evaporation, and other means. Last week, scientists from the University of Georgia contradicted the federal assessment, claiming that as much as 79 percent of the oil spilled remained in the Gulf. A few days after that announcement, scientists from the Woods Hole Oceanographic Institution reported in the journal Science that they had confirmed the existence of 22-mile-long underwater oil plume near the leaking wellhead in late June, a little more than two weeks before the well was finally capped. Because both the federal government and the University of Georgia’s estimates of the remaining oil were based on incomplete information and not peer-reviewed, many reporters (myself included) turned the Woods Hole paper published in Science to help settle their contradictory findings. That, says one of the authors of that paper, was a big mistake. In column for CNN published Wednesday, Woods Hole’s Christopher Reddy wrote:

Instead of being able to consider our results on the basis of the information alone (“just the facts, ma’am”), readers, viewers and listeners around the world were exposed to newspaper, TV and radio reports clouded with politically charged agendas that were premature at the least and outright wrong at the most.

I must have spoken with at least 25 journalists last week, and despite my every effort to explain our findings, the media were more interested in using the new information to portray a duel between competing scientists. The story turned into an us-versus-them scenario in which some scientists are right and others are wrong. Seeking to elucidate, I felt caught in a crossfire…

Even though my colleagues and I repeatedly avoided contrasting our results with previous NOAA estimates that some 75 percent of the spilled oil was already gone from the Gulf, much of last week’s coverage of our work made that a prominent part of the story.

By way of example, Reddy cited an article in The Washington Post, which reported that “Academic scientists are challenging the Obama administration’s assertion that most of BP’s oil in the Gulf of Mexico is either gone or rapidly disappearing — with one group Thursday announcing the discovery of a 22-mile ‘plume’ of oil that shows little sign of vanishing.” I quoted a similar sentence from The New York Times in a review of the coverage last week. Such reporting, Reddy argues, cast the Woods Hole paper as evidence that the NOAA estimates were wrong and that the University of Georgia was right:

Neither of these conclusions was ever meant to be drawn from our research on the oil plume. This reasoning implicit in the media coverage was not only premature, but it might turn out to be wrong.

Science does not work that way. It is incremental. It is not a house of cards where one dissenting view leads to a complete collapse. Rather, science is more like a jigsaw puzzle. Each piece is added. Occasionally a wrong piece may be placed, but eventually science will correct it.

As if to illustrate his point, a Science published a paper on Tuesday from scientists at the Lawrence Berkeley National Laboratory, which found that oil in the twenty-two-mile-long plume near the wellhead was—contrary to the Woods Hole data—biodegrading rapidly. The measurements upon which the Berkeley team based its paper were actually taken before the Wood Hole team measurements, but in a meeting in Seattle on Tuesday, the lead of the Berkeley paper, Terry Hazen, said that in the last three weeks his team hasn’t been able to detect the underwater plume at all.

What is a reporter to do, with so many competing analyses? The simplest answer might be to re-read Reddy’s media criticism in CNN and bear in mind that science is an incremental process, in which a single dataset rarely, if ever, settles a particular debate.

Curtis Brainard collaborated on this article. Eliott and Griffin are the directors of the Science Media Centres in Australia and New Zealand, respectively. Kelland is Reuters health and science correspondent for Europe, the Middle East and Africa.