‘It is all explainable,’ says scholar accused of data fabrication

Danielle Dixson, an associate professor of marine science at the University of Delaware, received blow after blow last year.

On 8 July, Proceedings of the Royal Society B issued a correction to a 2016 article she co-authored. It was on how anemonefish respond to bleached and unbleached host anemones – an important question regarding coral reefs.

On 29 July, according to documents Dr Dixson provided to Inside Higher Ed, a faculty committee report found her guilty of research misconduct, in that paper and elsewhere. The same day, provost Laura Carlson said the university intended to fire her.

“The evidence establishes both incompetence and gross irresponsibility. Either basis warrants termination,” Professor Carlson wrote in a letter to Dr Dixson, who has tenure. “As you are aware, the university’s Research Misconduct Investigation Committee concluded, after a thorough investigation, that you committed research misconduct in the form of falsification and fabrication.”

Science’s news department, which had already run a lengthy story on questions about the work of Dr Dixson and a colleague in 2021, got its hands on a “heavily redacted” draft of that report. That’s according to the article it then ran, headlined “Star marine ecologist committed misconduct, university says”.

In August, Science’s editorial side retracted a 2014 academic article. Dr Dixson was the lead author.

“In August, the University of Delaware informed us that the data in Figures 1A, 2, 3 and 4 were questioned and that they no longer have confidence in the validity of the data,” said Science’s retraction notice. “In agreement with the recommendation of the University of Delaware, Science is retracting the paper.”

In December, Behavioral Ecology placed an “expression of concern” about “the credibility of the data” on a paper Dr Dixson co-authored. But the new year brought her two wins.

A Faculty Senate committee report she provided to Inside Higher Ed unanimously concluded that she should not be fired, and the Proceedings of the Royal Society B published an editor’s note on the paper it had corrected, saying its own investigation “concluded that the evidence in support of claims that these data have been fabricated/manipulated, and hence are unreliable, are too weak to warrant retraction of the paper”.

However, University of Delaware president Dennis Assanis plans to fire Dr Dixson regardless this September, when her paid administrative leave ends, according to another document she provided.

“As I expressed in my response to your initial recommendation, I see very serious issues with Dr Dixson’s research practices and find her after-the-fact (and shifting) explanations to be implausible,” Professor Assanis wrote in March to the chair of the faculty committee that opposed firing Dr Dixson.

“Indeed, Dr Dixson’s actions are such that I believe the only appropriate outcome in this matter is to terminate Dr Dixson’s appointment,” he wrote. “I understand this is a different outcome than the one reached by the Hearing Panel and the FRR [Faculty Rights and Responsibilities] Committee, but I do not see how Dr Dixson can teach our students to be ethical researchers or how the results of future research projects conducted by Dr Dixson could be trusted.”

He wrote that he could not “allow a faculty member who has engaged in research misconduct, including data fabrication, to remain at the university”.

Mark Serva, the associate professor who chaired the hearing panel, declined to comment. Faculty Senate president Nancy Getchell did not respond to requests for comment. The university did not provide documents and said it did not comment on personnel matters.

Dr Dixson has taken issue with the Research Misconduct Investigation Committee’s work. “All of it is explainable,” she said. “But I was never given a chance to explain it.

“The evidence I needed to clear myself, I didn’t have access to through no fault of my own,” she said, adding that the university broke one of her hard drives.

“The level of disdain that they had for me before they even met me was quite high,” she said of the committee members.

She also complained of a “pretty calculated attack” by Timothy Clark, an associate professor at Australia’s Deakin University who, alongside others, raised alarms about her research.

Dr Clark said he had been intrigued by some of her research on ocean acidification impairing coral reef fish behaviour, but he eventually became concerned about the findings. In 2020, he and others published a paper in Nature casting doubt on the results.

He is not backing down in his criticism and his support for the Research Misconduct Investigation Committee’s findings.

“The idea that she could have exonerated herself by going through spreadsheets with them is nothing short of ridiculous,” Dr Clark said. “The data patterns that are in the spreadsheets – there’s no explanation for them besides copy and pasting data.

“Her spreadsheets are just rife with copy and pasting, so there is really no other explanation apart from data fabrication.

“There’s been a tremendous amount of work that’s gone into compiling all of the evidence and handing it to the university and the journals on a silver platter,” he said. “And, for the most part, most parties have done nothing with that pile of evidence. The University of Delaware’s investigation was a shining light because they actually did a thorough and relatively transparent investigation.

“If Delaware overturns this, then that’s just another nail in the coffin of the future of robust science,” Dr Clark added.

The Research Misconduct Investigation Committee wrote a roughly 50-page report, including passages such as this: “The committee repeatedly questioned the respondent (Dixson) about the lack of research record-keeping materials. How could respondent not have kept records of experiments performed by herself and her students? This is a clear requirement of standard research practice across scientific fields, regardless of whether the research is funded by federal agencies. At one point in the interview, she was asked specifically whether she kept lab notebooks. She responded affirmatively. However, later in that same interview, respondent offered various reasons for the absence of lab notebooks including that she did not keep lab notebooks, which appears to be a clear contradiction. She also stated that her graduate students took the lab notebooks with them when they left, or it was not standard in her field to keep lab notebooks, or that she was not trained to keep lab records. But laboratory related research and field planning must have been documented. Mr. [Paul] Leingang, her former graduate student, supplied photographic evidence of one of respondent’s fluming notebooks from November 2019 where fluming data were recorded using cues generated by Dr Jennifer Biddle. This suggests that the respondent chose not to provide any lab notebooks that did exist, perhaps because of discrepancies that might be found therein.” 

The committee, whose chair did not respond to requests for comment, concluded that it “was repeatedly struck by a serial pattern of sloppiness, poor record-keeping, copying and pasting within spreadsheets, errors within many papers under investigation and deviation from established animal ethics protocols. This pattern was discernible across the studies we evaluated and throughout our investigation.”

But in a roughly five-page report, the Faculty Senate’s faculty hearing panel found that the provost failed to meet the burden of proof to terminate Dr Dixson’s tenure, or even to establish that she had committed research misconduct.

“The respondent herself admitted that she made errors in data recording, data handling and data copying,” that committee said. “The initiator (the provost) failed in her obligation, however, to present clear and convincing evidence of research misconduct and did not establish that the respondent’s deviations from accepted, scientific community research practice were done in a premeditated or negligent manner, or without regard to the consequences of her actions. The Hearing Panel notes that the respondent did not benefit from the errors, given that the results did not affect the published results. In fact, the opposite has clearly occurred, given the negative publicity and harmful impact on her reputation.”

Dr Clark said of this exonerating report that “no whistleblowers have been contacted for comment, so it sounds like it’s been a very one-sided affair”.

“My problem is with bad science,” he said. “And nowhere near enough scientists speak up when they see something wrong going on.”

James Cook University, the Australian university where Dr Dixson earned her PhD, said in an email that “an investigator was externally appointed to examine allegations referred to JCU by the Australian Research Council in relation to alleged research misconduct. The allegations were made by anonymous parties in regards to research conducted at JCU.

“The investigator’s report found no evidence of research misconduct and recommended the matter be dismissed.”

This is an edited version of a story that originally appeared on Inside Higher Ed.