German Newspaper Forces Correction at Science Magazine, While Americans Ignore Study That Undermines Market Theory for Pandemic Origin

Latest revelations add evidence that American science writers are allergic to journalism. Science Magazine accused of “careless and unprofessional handling of statistical methodology.”

9 minute read

While American science writers continue to ignore new research that calls a study published by Science Magazine “flawed” and “invalid,” a German newspaper forced the journal to issue a correction to their study, authored by Arizona researcher Michael Worobey, that argued the COVID pandemic began in a wet market.

Science Magazine corrected the Worobey study last October, but the Frankfurter Allegemaine reported a couple weeks back that their reporters forced the correction after sending several queries to the journal.

The Frankfurter Allegemeine is Germany’s third most read newspaper and seems to have a similar bent to the Wall Street Journal, with conservative opinions but news that leans more liberal. In their article, Frankfurter Allegemaine’s journalists noted that a new study concluded that the statistics in Worobey’s Science Magazine research were “blatantly nonsensical” and found two experts to back up these claims.

One said Worobey’s Science Magazine study was inadequate, while another academic accused Science Magazine of “careless and unprofessional handling of statistical methodology.”

Because the Frankfurter Allegemeine is such a prominent newspaper, I worked to get a translation into English, which is provided below. But first a little background.

Politics at Science Magazine

This saga began in February 2022, when Chinese researchers posted a preprint online that concluded the Huanan wet market in Wuhan, China, was either the pandemic’s origin or a spreading event, meaning someone who was sick entered the market and infected others.

The Chinese paper was led by George Gao, the former head of China’s disease control agency. Attempting to counter the Chinese findings, a group of American and other Western researchers rushed out two preprints with contrary findings the next day.

Led by Michael Worobey with the University of Arizona, the first preprint dismissed the idea that someone brought the virus into the market and concluded that the pandemic began in the Huanan market. These findings were buttressed by the other preprint also hastened online, and led by Jonathan Pekar, from the University of California at San Diego.

As I reported at the time for UnHerd, New York Times reporters Carl Zimmer and Benjamin Mueller were obviously working closely with Worobey and Pekar, because they soon posted a long science article that prominently featured their preprints, while burying the Chinese scientists’ paper several paragraphs into the story.

Here’s the headline and first paragraph of that New York Times article.

But burying the Chinese study to favor Worobey and Pekar was not Zimmer and Mueller’s only sin.

They also failed to report the financial conflicts of interest of the Western scientists and did not emphasize that the preprints were not peer reviewed. I also noted that Carl Zimmer had engaged in questionable reporting in the past, scurrying a story into the Times based on a preprint, without reporting that a study published by Chinese researchers undermined the preprint and his own reporting.

Some months after the Times’ questionable journalism, Science Magazine published the two Western studies that downplayed the possibility of a lab accident. Since that time, Science Magazine has worked to protect those studies in order to downplay a possible lab accident in Wuhan.

As I previously reported, two researchers found a statistical error in the Science Magazine study published by Pekar. But Science Magazine refused to publish the research that found the error. “It's just another nail in the coffin of [laughs] free thinking and science,” the researcher told me.

Meanwhile, Science Magazine’s Jon Cohen wrote a hit piece on George Gao and the Chinese researchers who argued the market was likely a spreading event, not the pandemic’s origin. The Chinese researchers’ study was eventually published in Nature, a competing journal to Science Magazine.

The English translation of the Frankfurter Allegemaine article is below with a few caveats: No translation can be perfect, especially a newspaper story, in part because foreign newspapers write differently. American reporters write short, punchy sentences, a style that became prominent in the writings and novels by former reporter Ernest Hemingway.

But French and Spanish journalists compose sentences that would seem to drift on if they were in English, consuming dozens of words and several lines of writing, making it hard for American readers to follow a translation. The Frankfurter Allegemaine reporting more closely resembles the American style with shorter sentences, but the paragraphs are huge, long and blocky.

To make reading easier, I have broken up some of the paragraphs into smaller chunks.

---

DISCUSSION ABOUT “SCIENCE” STUDY

Criticism of Statistical Analysis on the Origin of Corona

BY HINNERK FELDWISCH-DRENTRUP - Updated on 01/28/2024-10:06

Two years ago, the question of where the corona virus originated seemed largely resolved. Statisticians are now criticizing earlier evaluations.

It was the clearest answer to date to an elusive question that still preoccupies people around the world today: How did the corona pandemic come about? Was it natural zoonosis, i.e. transmission from wild animals to humans, as coronavirus expert Christian Drosten from the Charité assumes is plausible for fur farms? Or possibly a laboratory accident?

“The Huanan market was the epicenter of the emergence of SARS-CoV-2,” wrote a team led by biologist Michael Worobey from the University of Arizona in a preprint on February 26, 2022. New research points to the market “as the origin of the pandemic,” reported the New York Times on the same day: The preprint and a second paper in which Worobey was involved represented “a significant salvo” in the debate about the start of the pandemic.

There is now an "extremely clear picture" that the pandemic originated in the market, the newspaper quotes Worobey as saying. Epidemiologist Thea Fischer from the University of Copenhagen also said that the question of a zoonotic origin had been “clarified with a very high level of evidence.”

Virologist Jesse Bloom from the Fred Hutchinson Cancer Research Center was more cautious: The conclusions could be true, but the quality of the data used was not sufficient. An analysis was based on around 160 cases from December 2019, which were extracted from a map of a WHO report (see graphic). It is clear that this is only a portion of all cases, as the virus had already spread before and certainly not all of the cases were known. Furthermore, If the first cases were connected to the Wuhan Institute of Virology, the WHO would hardly have found this out given the Chinese regime's censorship.

The two preprints reported further evidence: More corona-positive samples were found in the area of the wholesale market where wildlife traders were located. And according to the second preprint, there were most likely two transmissions to humans.

Around six months later, the two studies were published in Science Magazine. The one on the market as the epicenter has now been downloaded more than 800,000 times; according to an evaluation linked by Science Magazine, there are hundreds of media reports on this.

But now a paper by statisticians from Germany and Hong Kong in the Journal of the Royal Statistical Society Series A attacks the study as invalid. According to Dietrich Stoyan, statistician emeritus from the TU Bergakademie in Freiberg, and his colleague Sung Nok Chiu from the Hong Kong Baptist University, the Science Magazine study does not prove that the market is the starting point of the pandemic.

In any case, statistical analyses can hardly determine causes, only correlations, says Stoyan.

Test is “blatantly nonsensical”

To evaluate the early cases, Worobey's team used spatial statistics methods, an area in which Stoyan is an expert. They wanted to determine whether the Covid cases were closer to the market than would be expected if the market had nothing to do with the outbreak. To do this, they calculated that the patients' homes shown on the WHO map were, on average, a good four kilometers away from the market. Those with a known connection to the market, such as those who worked there, were a little further away than others.

To analyze whether these distances indicate a connection, Worobey's team compared them with the distance of randomly generated point clouds to the market.  These were generated according to the population density of Wuhan, with densely populated areas occurring more frequently than those where fewer people live. To account for the fact that older people are more likely to fall seriously ill and be recognized as Covid cases and taken to hospital, the researchers also took into account the age distribution in different areas of the city.

In fact, the average distance between the cases from the WHO report and the market was significantly smaller than that of randomly generated locations. In the New York Times, Worobey was certain: The fact that the cases were centered around the market was most likely not a coincidence.

But the chosen statistical test was “blatantly nonsensical,” says Stoyan.

Always calculating the distance to the market with the randomly generated point clouds automatically creates large distances, which is expected and not surprising. A realistic model for disease clusters is also missing.

Stoyan was “deeply disappointed” that the prestigious magazine Science published the article in this way. Statistics experts would have recognized the analysis as nonsense. Now he hopes for a scientific discussion on this.

Neither Worobey's article nor his own could prove nor refute the zoonosis hypothesis.

Independent experts confirm the criticism

Worobey defends himself against criticism: the article by Stoyan and Chiu is “shockingly flawed,” he tells the Frankfurter Allgemeine. They seemed unaware of the analyses included in his study. Worobey doesn't want to provide details about this, even when asked; he refers to a response to the study that will follow.

At the request of this newspaper, two independent experts in spatial statistics analyzed Worobey’s work. René Westerholt from TU Dortmund says he came to a similar assessment as Stoyan and Chiu. The methodology of the Science Magazine study is inadequate: On the one hand, the analysis of the cases cannot rule out that locations near the wet market are a possible origin. On the other hand, it does allow for some of the conclusions that are drawn.

Although some of statements in the Science Magazine study were partly restrained, this was undermined by very strong interpretations that were used throughout the entire paper—even in the title. Although it provides strong evidence of early infection events near the market, the conclusions drawn “could not be made in full and to the extent communicated.”

Evgeny Spodarev of Ulm University also says Stoyan and Chiu's conclusions are correct: the spatial statistical analysis in the Science study is "flawed in multiple ways." He speaks of a “careless and unprofessional handling of statistical methodology.” Worobey's team assumes that the source of the Covid pandemic is local and immobile.

The market is not necessarily responsible for many cases: there is also a large train station and a shopping mall nearby. Only a few hundred meters away is the Wuhan disease control authority, where, according to a report by CNN, experiments were carried out on an unknown coronavirus—according to intelligence sources, the US Department of Energy, referred to these experiments and tends to believe that Covid-19 spread through a laboratory accident, even if the evidence is poor. Other American intelligence information referred to the Wuhan Institute of Virology, which conducted research on coronaviruses with partners from the USA. But evidence of a laboratory accident remains lacking, and the majority of American intelligence agencies lean toward a natural origin.

It is undisputed that a superspreading event occurred in the market. George Gao, then head of China's disease control agency, wrote this in a preprint that was published shortly before the Worobey preprints. The virus was therefore “most likely” carried to the market by people, since it had spread weeks before December 2019.

Gao's team reported that no virus material was found in samples from animals at the market. However, later analyses of the gene sequences he published found traces of wild animals such as raccoon dogs, which can transmit SARS-CoV-2, along with those of the virus. But that doesn't have to mean that they brought it to the market: they could also have been infected there, or both could have been mixed in the samples.

According to a spokeswoman, Science Magazine sees “no need for changes” to the Worobey study at this time and has carried out a “thorough investigation.” She did not answer when this happened or whether statisticians were involved in the original review.

But there have already been small corrections: First of all, the first sentence of the article said that the Chinese government had notified the WHO on December 31, 2019, about cases of severe lung disease in Wuhan—although the UN organization was not informed by official bodies, but found media reports about this.

When asked by Frankfurter Allegemeine for evidence of this claim, Science corrected the article. There was also a correction to the Pekar Science article published at the same time with Worobey as co-author: an error in the software code caused effects to be overestimated.

---