In May 2021 we reported on a controversy:
Does ocean acidification alter fish behavior? Fraud allegations create a sea of doubt . . .
[Biologist Philip] Munday has co-authored more than 250 papers and drawn scores of aspiring scientists to Townsville, a mecca of marine biology on Australia’s northeastern coast. He is best known for pioneering work on the effects of the oceans’ changing chemistry on fish, part of it carried out with Danielle Dixson, a U.S. biologist who obtained her Ph.D. under Munday’s supervision in 2012 and has since become a successful lab head at the University of Delaware . . .
In 2009, Munday and Dixson began to publish evidence that ocean acidification—a knock-on effect of the rising carbon dioxide (CO2) level in Earth’s atmosphere—has a range of striking effects on fish behavior . . .
But their work has come under attack. . . .
Fish fight!
Seriously, this story disturbed me, not just because of the possible evidence of scientific fraud, but because some of the buddies of the people who wrote the controversial papers did the offense-is-the-best-form-of-defense thing and attacked the critics. As I wrote at the time:
The news article continues:
The seven [critics] were an “odd little bro-pocket” whose “whole point is to harm other scientists,” marine ecologist John Bruno of the University of North Carolina, Chapel Hill—who hasn’t collaborated with Dixson and Munday—tweeted in October 2020. “The cruelty is the driving force of the work.”
I have no idea what a “bro-pocket” is, and Google was no help here. The seven authors of the critical article appear to be four men and three women. I guess that makes it a “bro pocket”? If the authors had been four women and three men, maybe they would’ve been called a “coven of witches” or come up with some other insult.
In any case, this seems like a classic Javert bind. Sure, the critics get bothered by research flaws: if they weren’t bothered, they wouldn’t have put in the effort to track down all the problems!
Also this bit, which we’ve heard in various forms so many times before:
“This is a simple human error, not fraud,” he says. Many other data points are similar because the methodology could yield only a limited combination of numbers, he says. Munday says he has sent Nature Climate Change an author correction but says the mistake does not affect the paper’s conclusions.
Bad data but they do not affect the paper’s conclusions, huh? It kinda makes you wonder why they bother collecting data at all, given that the conclusions never seem to change.
News!
From Retraction Watch, we learn that one of these papers has been retracted and that the University of Delaware released a report finding the researcher to be guilty of research misconduct. The researcher’s lawyer described her as a “‘brilliant, hardworking female scientist’ who was ‘targeted’ by a group of scientists who ‘chose to “convict” Dr. Dixson in the court of public opinion’ by sharing their accusations with a Science reporter last year.”
What about the brilliant female scientists who were bothered by the research misconduct? Don’t they count? And what about the brilliant female scientists whose excellent submissions were rejected from the journal because the place of their article was already taken up by the fraudulent paper? If you’re doing legit science it can be hard to compete with cheaters: with fake data you can make all sorts of amazing claims. Perhaps next the lawyer will be describing Lance Armstrong as a brilliant, hardworking male cyclists who was target by a group of envious competitors.
And then there’s the “court of public opinion” thing . . . Give me a break. If the lawyer’s so upset about the court of public opinion, why is she throwing out these inflammatory accusations? Beyond all this, if you publish a paper in a scientific journal, you’re already out there in the “court of public opinion.” A published paper is published, i.e. public. Not to mention that this researcher had no problem going on NPR. If you don’t want your scientific claims to be scrutinized, then the solution is simple: don’t publish them, don’t post them, don’t claim you have data when you don’t, etc.
Anyway, good that the University of Delaware is taking research misconduct seriously. We don’t always see that.
Fun stuff
Here’s one detail to give a small sense of what was going on:
The draft report also found misconduct in a 2016 paper on whether anemone fishes can sniff out the condition of potential host anemones, published in Proceedings of the Royal Society B by Dixson and marine ecologist Anna Scott of Southern Cross University in Australia. Again, the timeline was implausible, the committee concluded. Collecting the data would have taken 22 working days of 12 hours, it wrote, “working continuously without any breaks or doing any preparation work, recalibration, cleaning, bucket switchouts,” and so on. Yet the paper said the study was done in 13 days, between 12 November and 24 November 2014.
Kind of reminds me of the Pizzagate scandal, all those contradictions, all those supposed data that never could’ve existed.
Not that it really matters – but I could imagine two mechanisms here. One might be straight ahead fraud motivated purely by goals of personal/professional advancement. Another might be because the scientists are concerned about climate change, resulting in a toxic combination of intention to deceive and motivated reasoning.
If it’s the latter mechanism, it’s notable that outcomes such as this can only move towards in the direction of reinforcing the existing partisan obstacles to mitigate climate change. Not that it will really move the needle much (the existing partisan gridlock is very difficult to meaningfully increase or decrease), but it can’t help.
These papers received high profile coverage because of the invocation of a tipping point. There is plenty of good work in climate science of course, but none of it has to do with tipping points.
Predicting a tipping point is actually three claims, (1) that the system currently exhibits elements of stability, (2) that the system is exiting the regime where stability governs behavior, and (3) that our understanding of the system control mechanisms is sufficient to predict a transition to instability. In general with climate, the level of understanding about the system that would be necessary to make this type of prediction is lacking.
The editors at the highest profile journals like Nature and Science seem utterly unaware that these papers are making extraordinary claims with ordinary results. In most cases the tipping point papers contain some basic research that does move our knowledge base forward, but with a few added paragraphs about a tipping point that is not supported by the work. Take away the bit about the tipping point, and the paper no longer satisfies the requirements for inclusion in a high-profile journal. It is a sad state of affairs.
Meanwhile…Merry Christmas to the blog denizens!
This fish controversy is just the tip of a massive iceberg crashing into “the science.”
“Lack of communalism during the pandemic fueled scandals and conspiracy theories, which were then treated as fact in the name of science by much of the popular press and on social media. The retraction of a highly visible hydroxychloroquine paper from the The Lancet was a startling example: A lack of sharing and openness allowed a top medical journal to publish an article in which 671 hospitals allegedly contributed data that did not exist, and no one noticed this outright fabrication before publication. The New England Journal of Medicine, another top medical journal, managed to publish a similar paper; many scientists continue to heavily cite it long after its retraction.”
“Heated but healthy scientific debates are welcome. Serious critics are our greatest benefactors. John Tukey once said that the collective noun for a group of statisticians is a quarrel. This applies to other scientists, too. But “we are at war” led to a step beyond: This is a dirty war, one without dignity. Opponents were threatened, abused, and bullied by cancel culture campaigns in social media, hit stories in mainstream media, and bestsellers written by zealots. Statements were distorted, turned into straw men, and ridiculed. Wikipedia pages were vandalized. Reputations were systematically devastated and destroyed. Many brilliant scientists were abused and received threats during the pandemic, intended to make them and their families miserable.”
John Ioannidis, Tablet, August
The whole essay is worth reading because Ioannidis is strictly non-partisan and a scientist with a great reputation, until it was trashed during the pandemic. We are indeed in an unprecedented regime of censorship and cancellation. As Musk says all the conspiracy theories about Twitter and Federal law enforcement and the CIA were true.
John Ioannadis’s reputation rightfully took a thrashing during the pandemic because he made a major mathematical error that he refused to back down from. People may be trying to make him personally miserable, and that’s unfair and uncalled for, but a hit to his reputation and vociferous criticism of his bad work is right and necessary.
There are no scientists especially those who have published as much who haven’t made errors. The reason he was singled out for a caning was because he didn’t toe the political line on covid. It’s happened to a lot of scientists. Read the Twitter files to appreciate what has happened. The latest one is on covid censorship. The Government is censoring truths they don’t like and spreading misinformation.
It’s not a question of making errors. It’s a question of accountability, doubling down on errors, and explicitly politicizing the science. And indeed, Ioannidis takes a hit on all accounts.
When a world renown epidemiologist makes a fundamental error such as extrapolating from an unrepresentative sampling to go on a TV campaign to downplay a virus amid an ongoing pandemic to wrongly equate that virus to the seasonal flu, despite abundant evidence at that time he was obviously wrong. Yeah, that’s a problem.
Again, it’s not just a matter of making an error. It’s a matter of engaging in poor science. If he had just acknowledged that error of scientific process early on, and moved on from that, I doubt that very many would have continued to be so critical or personal. Instead, he has continued to build more science upon a faulty foundational orientation, and to blame others rather than accept accountability.
It’s a sad state of affairs.
You misrepresent what happened and show your biases are strong. As I recall he was extrapolating from the Diamond Princess data which was the best available at the time. His body of work during the pandemic is on the whole good.
There was a need for balancing the very misleading data being propagated in the media implying a CFR of 15% or even more.
The question is why focus on this when the CDC is vastly more culpable in terms of propagating obvious pseudo-science on masks, vaccinating children, school closures and many other issues? The reason must be political bias. It’s vastly more important that there be accountability for the CDC as they are very powerful and can harm people on a vast scale.
> As I recall he was extrapolating from the Diamond Princess data which was the best available at the time.
He extrapolated from unrepresentative sampling there, and went on to do so later in the Santa Clara study.
It’s fine to do these investigations. Absolutely. You conduct science with the best data you have available.
What’s not fine is to use unrepresentative sampling for extrapolation, and then on that basis to go on a national TV campaign to say that the virulence of COVID is on par with the seasonal flue, as he did when he used the Santa Clara study as a basis for promoting a globalize IFR.
It’s fine to make mistakes. What’s not fine is to contradict the fundamentals of basic epidemiology.
Even Wikipaedia is more objective than you Josh.
“In an editorial on STAT published March 17, 2020, Ioannidis wondered whether the global response to the COVID-19 pandemic may be a “once-in-a-century evidence fiasco” and asked for obtaining more reliable data to deal with the pandemic.[7] He estimated that the coronavirus could cause 10,000 U.S. deaths if it infected 1% of the U.S. population, and argued that more data was needed to determine how widely the virus would spread.[129][5][7] The virus in fact eventually became widely disseminated, and would cause more than one million deaths in the U.S.[130][129][5] Ioannidis expressed doubt that vaccines or treatments would be developed and tested in time to affect how the pandemic would unfold.[131] Marc Lipsitch, Director of the Center for Communicable Disease Dynamics at the Harvard T.H. Chan School of Public Health, objected to Ioannidis’s characterization of the global response in a reply that was published on STAT the next day after Ioannidis’s.[132] Ioannidis later stated that early in 2020 he wrote about needing more data, without that meaning he was mocking those who worried about COVID-19, and that he was elated with the quick development of vaccines and treatments, and the scientific progress made since 2020.”
Just doing the math. 10,000 fatalities if 1% infected would mean 1 million fatalities if almost everyone was exposed. Looks rather prescient to me.
And of course, no sample is ever perfectly “representative” in every way.
He said during an interview that the virulence of COVID was more or less on par with the seasonal flu.
If you want to put up some money we’ll make a bet and then I’ll do the work to track the quote down.
He reached reached conclusion based a specific reference to the Santa Clara study, where he globalized an from an IFR he based on the Santa Clara data, which was not adjusted for stratification by basic predictors of health outcomes like SES and/or race/ethnicity.
What was funny (but incriminating) is that they then went on to do a study with MLB where they got results that were inconsistent with the Santa Clara study (suggesting a higher IFR) and as a way to explain those unexpected findings, they pointed to the lack of representativeness of their sampling for the MLB study.
That’s terrible science.
Andrew –
Apologies. I’ll stop now.
Nothing you said Josh detracts from the fact that Ioannidis was very right in what Wikipaedia quoted. You have no context for how “serious” any non-representativeness actually was. It’s just your non-expert guess.
I do suspect that the Santa Clara study came out on the low end, given that ultimate PFR’s in the US were often around 0.3. That’s just slightly higher than the upper confidence interval from that study. But it was so early in the pandemic that there were more uncertainties. There are still a lot of very big uncertainties such as how many people went to the hospital with a heart attack, caught covid there and tested positive and became a covid death.
Science is a process that is very imperfect and to fixate as you do on “errors” that had no effect on the progress of science or policy is childish.
On the whole Ioannidis’ large opus during spring 2020 turned out to be pretty accurate but not perfect, unlike that of the CDC, who continue to this day to support pseudo-science on surgical masks (still required in health care facilities and US government buildings), vaccinating children (a very serious one), school closures, and the effectiveness of the new bivalent boosters.
Bottom line is that science took a massive hit during the pandemic and it may take a long time to recover if it ever will.