In 2005, a physician-scientist name John Ionnidis published a study titled “Why Most Published Research Findings are False.” He used a statistical modeling and simulation approach to attempt to quantify the likelihood that a hypothetical study would result in a true finding (rather than a false positive). The article makes a series of claims about the strength of various kinds of study formats based on the model, which had not been empirically tested. Ionnidis concluded his article with an outrageous claim: "for most study designs and settings, it is more likely for a research claim to be false than true."

The article ignited a firestorm of comments and criticism. The most salient critique, articulated by Steven Goodman and Sander Greenland, detailed severe methodological problems with Ionnidis' modeling approach. Ionnidis had lumped every possible kind of bias together into a single parameter, even though they have different effects on a potential study's likely outcome. They also noted that Ionnidis had chosen values for the bias parameter for different kinds of studies without any underlying justification or empirical validation, and that regardless of the value, it was almost impossible for any study to come out "true" using his model.

John Arnold, a billionaire and former Enron executive, came across the paper several years later. If he had read the critical replies, he ignored them. In his view, science was rife with low-quality work that couldn't be trusted. After founding what is now Arnold Ventures in 2008, he and his wife Laura decided they would use their fortune to identify and support work they felt they would be able to trust: work they considered open, replicable, and truth-seeking.

What followed has been framed as a yearslong battle to improve the quality, reproducibility, and robustness of scientific research, particularly in highly complex fields like social psychology, nutrition science, and preclinical research. It is also the story of how the Arnolds, through their influence over what kind of work gets funded, picked winners and losers in research. In at least once case when the data didn't confirm John Arnold's beliefs, they simply picked a new winner.

John Arnold began his career at Enron trading oil and gas options. He was regarded by others at the company as an incredibly skilled trader, and he rose to head of the fixed price trading desk by 2000, five years after he graduated from Vanderbilt University. As Enron was busy collapsing, John Arnold was busy collecting an $8 million bonus from the company days before it declared bankruptcy in December 2001. He was one of the only high-level executives to emerge from the massive scandal unscathed.

His trading practices nonetheless came under legal scrutiny. In May 2001, seven months before Enron filed for bankruptcy, Arnold was deposed¹ in a lawsuit filed by Enron against Anatara Resources. The lawyer for the defense asked a series of questions about how Enron had "unwound" a broken contract with Antara, including how Arnold arrived at his pricing of gas futures that he provided to Enron's trade execution team. As part of the unwinding of the contract with Antara, Arnold had provided Enron with pricing data used in the termination process. Some of Arnold's predictive pricings covered years for which no public data was available on pricing.

His next deposition in May 2005 by the same lawyer was far more aggressive². The lawyer peppered Arnold with questions about his trading practices. Did former employees at Enron try to litigate his $8 million bonus, alleging price manipulation? Had he advised other traders to use the "Enron bat" to move the market? Had he reported prices or volumes to any index that were false? Did he help execute a fake trade one day after the Antara contract was terminated? Did he realize that the Federal Energy Regulatory Commission had identified 73 trades made by Arnold that it considered fraudulent wash trades?

To every single question, Arnold replied "I assert my rights under the Fifth Amendment."

Arnold would spend the next several years making even more money running his new hedge fund Centaurus. His firm saw massive returns (over 300%) in its early years, and by 2010 he stepped down from the company to focus on philanthropy. After having refused to give forthright and honest answers about his own potentially unethical and illegal actions at Enron, John Arnold set out to make science more trustworthy.

The Arnolds made contact with Brian Nosek, a social psychologist working at the University of Virginia, in July 2012. Nosek had been building what he called the Replication Project, a consortium of scientists assembled to perform what are known as replication studies. In these studies, scientists attempt to repeat experiments and analyses that have already been published, ideally resulting in the same (or very similar) conclusions. What began as an email soon became a $35 million grant to create the Center for Open Science (COS).

COS targeted 100 different research studies for their initial replication attempts. The results were sobering: according to their statistical analyses, only 36-47% of the original studies replicated successfully. While the numbers looked damning, the authors were circumspect in their interpretation:

Because reproducibility is a hallmark of credible scientific evidence, it is tempting to think that maximum reproducibility of original results is important from the onset of a line of inquiry through its maturation. This is a mistake. If initial ideas were always correct, then there would hardly be a reason to conduct research in the first place. A healthy discipline will have many false starts as it confronts the limits of present understanding.

The paper inevitably set off another furious debate. The paper itself spurred a testy back-and-forth between COS and another group of researchers, who argued that the COS researchers made several statistical and methodological errors in their analyses. The debate grew acrimonious: in a separate incident, Daniel Gilbert, one of the scientists who critiqued the COS paper, lamented the "shameless little bullies" who had failed to replicate the work of Simone Schnall, an experimental social psychologist.

While the researchers argued with each other, a media frenzy erupted. The failures to replicate these findings contributed to the declaration of a replication crisis. It was not just that some studies failed to replicate: a media narrative emerged declaring that entire fields of science could no longer be trusted.

The crisis was apparently not restricted to social psychology. In 2012, researchers from Amgen and the M. D. Anderson Cancer Center reported that they could only successfully replicate 11% of what they called "landmark" studies in preclinical medical research. The list of studies they attempted to replicate was never made public, and the researchers cautioned that a portion of the original findings "might not hold up" because they were selected for their novelty. The report was published as a "comment" without undergoing any peer review. In exchange for borrowed reagents, the replicators also signed non-disclosure agreements that barred them from identifying the specific papers publicly. The claims nevertheless extended the appearance of crisis into the medical sciences.

The impact of the internecine debates over replicability has largely been constructive, at least from the scientists' perspectives. The practice of preregistration, where a study's methodology undergoes peer review before experiments begin, has become more common. The literature around so-called "questionable research practices," which include practices such as selective reporting and p-hacking, is growing. Most importantly is the fact that these changes have been spurred by researchers themselves. While some media outlets published by-lines insisting science is no longer self-correcting, scientists themselves were engaging in that very process.

The Laura and John Arnold Foundation's investment in the Center for Open Science thrust metascience (the science of doing science) into the limelight and supported the development of new approaches to ensure replicability. John Arnold had set out to spur on the science reform movement, and it seemed like the investment, and his contrarianism, was paying off. His attempt to revamp nutrition science would end much differently.

In 2011, John Arnold wrote an email to Gary Taubes after listening to a libertarian podcast that featured him as a guest. Taubes, a journalist and author with no formal scientific training in nutrition, was a proponent of the "carbohydrate-insulin hypothesis," which stated that obesity could be explained not by consumption of too many calories or dietary fats but by consumption of excess carbohydrates. While conventional wisdom and published work supported the former ideas, there was little evidence supporting the C/I hypothesis.

In Arnold's email to Taubes, he wrote "From the little I know about the science of nutrition, your study makes a lot of sense." Here Arnold did far more than express admiration for a researcher with attractive ideas. He revealed a quality shared universally among contrarians: a lack of expertise combined with a sense that they have stumbled upon something esoteric and exciting. Arnold became a billionaire by placing winning bets at his hedge fund Centaurus. He was convinced he had found another winner.

Arnold began funding Taubes' nonprofit NuSI in 2012, which began recruiting a scientific team to conduct a study of obesity in a small cohort of men. The team, the Energy Balance Consortium (EBC), initially met with Taubes and Peter Attia, a physician-scientist and co-founder of NuSI, to discuss the study plan, and they signed a contract that guaranteed scientific independence for EBC when it came time to draw conclusions and publish results. EBC recruited a cohort of 17 men for their pilot study, which were kept inside a ward and fed specific diets to determine whether, or how, lower amounts of carbs would influence their energy expenditure rate.

Two years later, EBC presented their preliminary results to NuSI: a very modest difference in energy expenditure. Taubes immediately critiqued the EBC's methodology, insisting that the scientists should have made several changes to the "standard American diet" used in the first four weeks of the study. He listed other complaints, including that there had been no control group. NuSI had met multiple times with the EBC scientists to finalize the study design before the experiment began, and now Taubes insisted the entire study had been done wrong. He called the pilot a "failure."

The road to publication was turbulent. The EBC continued to clash with NuSI, even sending an email demanding that NuSI respect their independence to carry out their work without interference. Attia resigned, as did Kevin Hall, a National Institutes of Health scientist who until that point led the EBC. At a poster presentation session in 2016, Hall publicly stated that this work, along with a previous paper, essentially falsified the C/I hypothesis. The final results were published the same year. Hall would eventually lead a randomized trial in 2021 that struck another blow against the idea.

The Arnolds' foundation could have admitted defeat. They could have shifted their investments to align with the prevailing evidence. Instead, the foundation doubled down: they stopped sending money to NuSI and began soliciting proposals from other scientists to study the C/I hypothesis. Like a gambler rolling the dice over and over until a win, the Arnolds' used their ample supply of capital to keep the C/I hypothesis alive long after the evidence had killed it.

America is experiencing an extraordinary degree of wealth inequality. At the same time, government programs to fund public science are being corruptly starved of the resources they need to operate. Laboratories that relied on federal grant dollars to operate have shuttered or stopped recruiting new staff.

Against this backdrop, billionaires like John Arnold market themselves as the good guys. They're using their wealth to fix our problems, improve our science, and boost public trust. They insist that their talent for wealth accumulation translates to discernment in the public sphere.

John Arnolds' refusal to follow the evidence where it actually leads is enough for us to doubt his trustworthiness even if one discounts his questionable conduct at Enron's trading desk. It would be simple enough to conclude that the problems of John Arnold are simply his alone. The problems with social Darwinism go much deeper than one man. The accumulation of massive amounts of capital and power in the hands of private individuals is a glaring structural problem that poisons public life with misguided solutions, biased narratives, and most importantly a complete absence of public accountability. When billionaires insist on giving imprimatur to junk science, we lack meaningful recourse to push back.

The most effective way to stop the subversion of the public sphere by the rich is to strip them of their immense wealth.

Update 12:29 pm Aug. 4 2026: The following post by Jeffrey Lockhart gives some more background on the Arnolds and the science reform movement. The article was a jumping-off point for my research into John Arnold.

Notes

¹Oral Deposition of John Douglas Arnold, May 31, 2001. Sunbelt Reporting and Litigation Services. Accessed from Scribd Aug. 3, 2026.

²Oral and Videotaped Deposition of John Arnold, May 6, 2005. Team Litigation Company. Accessed from Scribd Aug. 3, 2026.