Retraction Watch

Tracking retractions as a window into the scientific process

Data dispute forces journal to pull paper about rare bird on valuable land

with 2 comments

coverThe authors of a paper about the density of an endangered bird on valuable Texas lands are disputing the journal’s decision to pull the paper after other scientists objected to some of its contents.

The paper, published online in August by the Journal of Field Ornithology, compared different models for estimating the number of golden-cheeked warblers that nest in Texas. According to the Austin American-Statesman, some of that land includes highly sought-after woodlands. Developers had used estimates from recent years, which suggested that the bird population was in relatively good health, to petition to open that land for real estate development. But the new paper concluded that those estimates were too high.

After some scientists who produced the more rosy estimates claimed that the some of their data had been included in the newest paper without permission, the editor in chief decided to retract it.

That was in November. There’s no retraction note for “Density influences accuracy of model-based estimates for a forest songbird.” We contacted the editor in chief, Gary Ritchison, to ask why. He told us,

I have nothing to say.

The first author on the paper — Lisa O’Donnell, a biologist for the city of Austin — told us that the authors did not agree with the retraction:

We’ve evaluated 3 different models. because of concerns with one the data sets we used, you throw out the whole paper?

The issue extends beyond the use of some of the data. The American-Statesman explained the political implications of those models:

…the 10-page study concluded that several recently developed models that predict the population of the golden-cheeked warbler — including one created by Texas A&M University researchers — exaggerated the bird’s abundance. The numbers are significant because the warbler’s low population contributed to the federal government’s decision, in 1990, to list it as an endangered species, erecting steep legal barriers to development of its habitat.

The paper was scheduled to be published in the December edition of the journal. But that was before the Texas A&M researchers, some of them no longer at the university, caught wind of it. Working behind the scenes, the Aggie biologists lobbied to have the paper spiked….In correspondence and interviews, the A&M researchers cited technical scientific errors and breaches of academic ethics as motivation for their campaign to scuttle the Austin study. “It’s plagiarism,” said Neal Wilkins, one of the original A&M authors.

The American-Statesman noted how expensive it would have been for the state to accept that the warbler’s numbers were in more jeopardy than previous estimates:

The Aggie model being tested by the city was featured in a 2012 paper funded by the Texas Department of Transportation. In it, the A&M biologists concluded the warbler’s population to be nearly 20 times higher than earlier counts. The paper was divisive from the start, with some scientists, including U.S. Fish and Wildlife Service officials, dismissing it as flawed.

Still, TxDOT has used the A&M model in its road planning through warbler territory. Pro-development opponents of the bird’s endangered status, including Susan Combs — who as state comptroller actively resisted endangered species listings — also seized on the findings. Citing the A&M paper as proof federal protections were no longer needed, six months ago they petitioned Washington to de-list the warbler, which would open up tens of thousands of acres to more intensive development.

Concerns about the data from the 2015 paper came from researchers who collected it on a preserve that was under the purview of the city of Austin. In an email to Ritchison, obtained by the American-Statesman, Bret Collier from Louisiana State University (but formerly at Texas A&M) explains that O’Donnell’s paper contains

what is typically considered unpublished data that stemmed from a direct email request, per a contractual obligation with said entity, from the lead author of the JFO manuscript to one of my collaborators.

The debate continued, according the American-Statesman:

A three-way correspondence among Ritchison, O’Donnell and several current and former Aggie researchers continued through the fall. The A&M biologists claimed that the city’s use of the data was, variously, inappropriate because it was previously unpublished; illegal because the Balcones contract described any gathered information as “co-owned” and so A&M should have been consulted; and unethical because O’Donnell’s paper didn’t properly credit the A&M authors.

O’Donnell told us that she believes the data were published, as they were part of a published model. And as co-owners of the data, the city had the right to publish them, she added.

Our paper was authorized by the city of Austin. They’re the ones that [co-own] the data.

But, another co-author on the original paper on the model told the Statesman that O’Donnell hadn’t tested the model correctly:

“It was for making estimates across the state, on a broad scale,” said Michael Morrison, a study co-author and professor in A&M’s Department of Wildlife and Fisheries Sciences. “It’s as if you were going to do polling for any issue or for any presidential candidate, and you sample across the state. You wouldn’t take that model and predict what it will do for this community in Austin.”

O’Donnell told us that she thinks the conversation about the model’s use should be carried out in the journal:

What makes the most sense to me is let the science prevail, and put this back in the science realm. Right now it’s in the media and the political realm. We would like for the journal to publish our paper. And have the [other scientists] provide a rebuttal. That makes the most sense to me.

But in November, the journal decided to retract the paper. In an email from Ritchison to O’Donnell, obtained by the American-Statesman, Ritchison told her:

…co-owned data cannot be used by one party without the consent of the other party. Thus, you should have requested permission to use the co-owned data in your JFO paper and, if necessary to gain that permission, allowed those who collected the data to either review the manuscript prior to submission and/or be included as co-authors of the manuscript. Unfortunately, therefore, I’ve decided that your paper must be retracted.

The last time O’Donnell from editor Ritchison was in mid-December:

The editors kind of shut down communication with us. We asked them if there’s an appeal process, they’re not responding at all. I don’t know how it works, not having had a paper retracted.

Some scientists are siding with O’Donnell in this debate, reports the American-Statesman:

…some observers have questioned the A&M group’s motives, wondering why scientists would work so hard to kill a study that seemed a legitimate contribution to a debate over the warbler’s future.

“The golden-cheeked warbler already has an uncertain future as more and more land is developed in the Hill Country,” said Joan Marshall, director of Texas Audubon. “That’s why the science is so important — to base listing decisions on the best science available. That’s what’s frustrating about Texas A&M’s suppression of the O’Donnell article.”

A bonus: Listen here as Retraction Watch co-founder Ivan Oransky discusses the retraction on the radio show Texas Standard.

Clarification 1/12/16 3:52 p.m. eastern: We’ve updated O’Donnell’s comments to clarify her position about the ownership of the data.

Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, and sign up on our homepage for an email every time there’s a new post. Click here to review our Comments Policy.

Comments
  • Tom Langen January 13, 2016 at 11:21 pm

    This does not seem right, at all. An invited commentary/rebuttal from the A&M team might be justified, maybe, but not a retraction. Given the policy implications of this publication, JFO appears to have handled this very poorly.

  • PJTV January 18, 2016 at 4:37 am

    Apparently authorship and data ownership issues have resulted in the retraction. However there are ‘policy implications’. Is it then not more important to retain the study and allow it to be scrutinized? Authorship and ownership issues are secondary and can be settled differently.

  • Post a comment

    Threaded commenting powered by interconnect/it code.