In October 1999, at a press conference in Washington and in the November issue of National Geographic, the magazine unveiled a small feathered fossil from Liaoning Province, China, hailed as a transitional creature linking ground-running dinosaurs to birds. It was informally christened Archaeoraptor liaoningensis. The fossil was not a single animal at all. It was a chimera — a composite assembled, largely with glue, from pieces of at least two different species and several specimens, joined on a slab to look like one body. Within months it stood exposed as one of the most embarrassing fakes in modern paleontology.
The exposure was swift and came from inside the science. The Chinese paleontologist Xu Xing, a co-describer of the specimen, grew suspicious and on 20 December 1999 emailed his collaborators that the fossil was a forgery; in early 2000 he located in a Chinese institution the counterslab of the tail, proving it belonged to a separate animal. CT scans by Timothy Rowe at the University of Texas had already shown the slab to be made of mismatched parts. The tail was eventually identified as that of the dromaeosaur later named Microraptor; the front half was a primitive bird subsequently named Yanornis. The “link” was an accident of adhesive.
The deeper failure was editorial, not geological. The manuscript describing the fossil had been rejected by both Nature and Science, the discipline’s leading journals, yet National Geographic — a popular magazine, not a peer-reviewed one — published the find anyway, written up by its own art editor, Christopher Sloan. The specimen had been smuggled out of China in violation of that country’s antiquities law, sold at the Tucson gem show in February 1999 for a reported 80,000 dollars to Stephen Czerkas of a private dinosaur museum in Utah, and rushed to print on a wave of excitement about the genuine, and genuinely revolutionary, feathered dinosaurs then emerging from Liaoning.
That last point is the irony the case is remembered for. The dinosaur–bird link Archaeoraptor was meant to dramatize is real; Liaoning’s authentic fossils have since confirmed it overwhelmingly. The forgery did not invent a false idea so much as fabricate a counterfeit of a true one, and in doing so it handed critics of the science a propaganda gift. National Geographic commissioned an outside investigation and published a candid post-mortem, “Archaeoraptor Fossil Trail,” by the journalist Lewis M. Simons, in October 2000.
In Seoul, South Korea, between 2004 and 2005, the veterinary scientist Hwang Woo-suk announced in the journal Science that he had done what no laboratory had managed before: derived human embryonic stem cells from cloned embryos, tailored to individual patients. The first paper, published on 12 March 2004, claimed a single stem-cell line created by somatic cell nuclear transfer; the second, published on 17 June 2005, claimed eleven patient-specific lines and promised rejection-free cures for incurable disease. None of the stem-cell lines was real. By January 2006 an investigation by his own university had concluded that the data and images in both papers were fabricated, and Science retracted them.
The exposure came not from the peer-review system that had certified the work but from outside it. A whistleblower among Hwang’s collaborators alerted the South Korean television program PD Su-cheop (PD Notebook), broadcast by MBC, which in November 2005 questioned both the ethics of the egg supply and the integrity of the results. As public scrutiny mounted, anonymous scientists posting online identified duplicated photographs of supposedly distinct cell lines. Seoul National University convened an investigation panel on 17 December 2005; within days it found that nine of the eleven 2005 lines were fabricated, and on 10 January 2006 it reported that all eleven were invented and that the 2004 line was fraudulent as well. Science retracted both papers on 11–12 January 2006.
Hwang had been more than a scientist; he was a state-sponsored symbol. South Korea had named him its first “Supreme Scientist,” issued a commemorative postage stamp showing a person rising from a wheelchair, and underwritten his work as a point of national prestige. That investment turned the fraud into a public trauma. When MBC first aired its doubts, advertisers withdrew and the broadcaster faced a wave of nationalist backlash for attacking a hero — a reaction that briefly protected the fraud from the very questions that would dismantle it.
The damage was both ethical and evidentiary. The eggs underpinning the research had been obtained improperly: junior researchers in Hwang’s own lab had donated, and women had been paid, in breach of bioethics norms — an estimated 2,000-plus oocytes were consumed across the program. Hwang was fired by Seoul National University in March 2006, indicted on 12 May 2006, and convicted on 26 October 2009 of embezzlement of research funds and violations of South Korea’s bioethics law, though acquitted of the core fraud charge; he received a suspended sentence. One claim survived scrutiny: Snuppy, the Afghan hound his team produced in 2005, was independently confirmed as a genuinely cloned dog.
At Bell Labs in Murray Hill, New Jersey, between roughly 1998 and 2001, the young German physicist Jan Hendrik Schön published a torrent of results that appeared to remake condensed-matter physics. He reported transistors built from single organic molecules, superconductivity in organic crystals, and a string of other firsts, at a tempo that astonished his field — in 2001 he was an author on a newly published paper, on average, roughly every eight days, in the most selective journals in science. Almost none of it was true. In 2002 an investigation concluded that Schön had fabricated or falsified data across at least sixteen separate cases, and his cascade of breakthroughs collapsed into one of the largest fraud episodes in the history of physics.
The exposure turned on the data themselves rather than on any failed replication. In the spring of 2002, physicists outside Bell Labs — among them Lydia Sohn, then at Princeton, and Paul McEuen at Cornell — noticed that graphs in different Schön papers, supposedly describing different experiments at different temperatures and in different materials, contained identical traces of random electronic noise, down to the same minute, irreproducible wiggles. Real measurements do not share their noise; the curves had been copied. Once the duplications were flagged, Bell Labs’ parent company, Lucent Technologies, convened an external committee chaired by the Stanford physicist Malcolm Beasley.
The Beasley committee reported on 25 September 2002. Examining 24 allegations across 25 papers, it found compelling evidence that Schön had committed scientific misconduct in at least 16 of them, while explicitly clearing his co-authors of misconduct. Schön had, in numerous cases, been unable to produce raw data, primary samples, or laboratory notebooks; the committee found his explanations unconvincing and the pattern of manipulation unmistakable. Bell Labs dismissed him immediately. In the months that followed, the journals unwound the record: Science retracted eight papers and later two more, Nature retracted seven, and Physical Review and Applied Physics Letters retracted others.
The reckoning extended to his credentials. In 2004 the University of Konstanz revoked the doctorate it had awarded him, citing dishonorable conduct, and Germany’s research funding agency barred him from its grants and committees. Schön contested the revocation through years of German litigation; the courts ultimately upheld it, with the Federal Administrative Court sustaining the decision in 2013. What remained, once the fabricated papers were withdrawn, was close to nothing: a celebrated body of work that had reported phenomena no one — including Schön — could reproduce.
In the Department of Social Psychology at Tilburg University in the Netherlands, the celebrated professor Diederik Stapel was suspended on 7 September 2011 after three junior researchers reported that his data looked too good to be real. They were correct. Stapel, born in 1966 and one of the most decorated social psychologists in Europe, had for years not run the experiments he published. He invented the numbers, typed clean spreadsheets at his kitchen table, and handed the fabricated files to the doctoral students whose careers he was supposedly building. There were no subjects, no surveys, no field sites — only a respected man with a laptop manufacturing the exact results his hypotheses predicted.
The verdict was delivered quickly and overwhelmingly. Tilburg announced its findings at a press conference on 31 October 2011, and a joint investigation by three committees — convened at Tilburg, Groningen and the University of Amsterdam under the chairmanship of psycholinguist Willem (Pim) Levelt — published its final report, Flawed science: The fraudulent research practices of social psychologist Diederik Stapel, on 28 November 2012. The committees identified at least 55 publications containing fabricated or manipulated data and found fraud entangled in at least ten doctoral dissertations he had supervised. By December 2015 the count of formally retracted papers had reached 58, making Stapel one of the largest single cases of data fabrication in the history of science.
Unusually for so consequential a fraud, much of it was confessed rather than merely caught. Stapel admitted the fabrication, surrendered his 1997 doctorate from the University of Amsterdam in November 2011, and in 2012 published a memoir, Ontsporing (“Derailment”), later translated as Faking Science, narrating his slide from cutting corners to inventing entire studies. In June 2013 he reached a settlement with Dutch prosecutors that spared him a criminal trial in exchange for 120 hours of community service and the forfeiture of benefits equivalent to roughly a year and a half of salary.
What makes the case a landmark is not only its scale but its target. Stapel did not fool a credulous public with a monster or a relic; he fooled the peer-review machinery of his own science. The same committees that condemned him turned their criticism outward, faulting a “verification culture” in social psychology that prized novel, media-friendly results over the unglamorous work of checking whether they were true. The fraud was personal, but the openings it exploited were structural.