Irreproducibility and scientific truth

scientific truth
© Olivier Le Queinec

Ute Deichmann from Jacques Loeb Centre for the History and Philosophy of the Life Sciences, Ben-Gurion University of the Negev, Beer Sheva, Israel, explores the compelling issues of irreproducibility and scientific truth

In 2004, Hwang Woo-suk published the highly spectacular claim to have cloned patient-specific embryonic stem cells in the journal ‘Science’. However, his reputation was soon destroyed, when it turned out that many of his data were made up. Fraud is one of the factors that lead to the irreproducibility of experiments, a phenomenon that is today widely discussed as a problem in the medical sciences, social sciences and psychology, but increasingly in the hard sciences too.

According to the National Association of Scholars in the U.S., the pervasiveness of this challenge is clearly visible in the surprisingly large number of articles in reputable peer-reviewed journals that have turned out to be invalid or highly questionable. According to medical scientist and metascientist John Ioannidis (2005) of the Stanford School of Medicine, large parts of scientific research findings in medicine and probably other sciences as well are simply “false”. Focusing largely on the use and abuse of statistics in data-driven research, he mentions among other things, the willingness to publish studies reporting small effects, intellectual prejudices and competition among researchers, especially in fashionable areas of research.

Moreover, enormous publication pressure, competition and industrial profit interests in the publication systems of today are often used as explanations for the frequent occurrence of irreproducibility. However, irreproducibility is not only a problem of globalised and profit-oriented research today, but was already widespread in the first half of the 20th century.

An example is a spectacular claim by Linus Pauling, a highly renowned American chemist and Nobel laureate, of having produced artificial antibodies in 1942 based on his 1940 ‘instruction’ theory of the generation of antibodies, according to which antigens ‘instruct’ antibodies to take their specific shape (Deichmann 2021a). Pauling’s experiments were shown not to be reproducible by prominent colleagues shortly after and his theory became obsolete later on. However, Pauling, convinced of the correctness of his theory and experiments, never disavowed them, and they have continued to be cited approvingly by others until today. Irreproducible scientific beliefs in fashionable fields and promoted by prominent scientists sometimes have a long life and seriously impact scientific practice.

This raises several questions: how can the fact that irreproducibility is such a widespread phenomenon in various sciences be reconciled with the idea of science as the arbiter of truth about nature? Does the controversy about the reproducibility of scientific results undermine the authority of science? Do we have to call into question well-established scientific facts, such as the efficacy of vaccination?

To answer these questions, we have first to understand that the failure to replicate experiments can be caused by very different factors, such as scientific fraud or insufficient scientific methods, in particular, the flawed use of statistics, which have to be assessed and handled in different ways. Most cases of irreproducibility seem to occur in areas of scientific application where lucrative results are expected, leading to sloppy research and fast publication, such as in cancer studies, psychology, epigenetics and drug discovery.

Second, it is important to remember that there is more to good science than reproducible experiments. Novel ideas and theories often arise based on already confirmed knowledge. An example is Francis Crick’s and Vernon Ingram’s collaboration on the molecular basis of a genetic disease in the 1950s that was the first bridge at the molecular level between the formerly separate fields of genetics and chemistry. This integration of different approaches together with logical considerations played a major role in the generation of Crick’s ‘sequence hypothesis’, according to which the specificity of a piece of nucleic acid is expressed solely by the sequence of its bases, and this sequence is the code for the amino acid sequence of a particular protein.

This hypothesis became fundamental to molecular biology and later, big data biology, and was based on very little new experimental evidence, in the words of Crick, “the direct evidence [for the sequence hypothesis] is negligible.” He held instead that “the psychological drive behind this hypothesis is at the moment independent of such evidence.” His “drive” was based on knowledge also beyond the direct field of research, rational analysis and vision (Deichmann 2021b). Of course, it did not replace a subsequent thorough experimental verification of the sequence hypothesis, but this example illustrates clearly that science is about more than reproducible experimental data.

Third, history shows that despite incorrect theories, experimental flaws, and short-lived fashions, much knowledge has remained robust for a long time, such as Newton’s laws, the mechanisms of nuclear fission and fusion, cell theory, and the structure and function of DNA. Examples of reliable research today include mRNA that was discovered in 1962, first tested as a vaccine in 1993 (for influenza in mice) and today used as a vaccine in virus-related epidemics. Progress in understanding the Hepatitis C viral life cycle in combination with newly developed RNA technology paved the way for new anti-Hepatitis C virus therapy and even cure. The new powerful gene-editing tool ‘CrispR-cas 9, (proposed in 2012) was based on the discovery of a new type of repeated DNA sequences in bacteria some 25 five years earlier.

Research leads to reliable scientific knowledge

These examples show that with all the spectacular failures and successes that are now making journals’ headlines, it should not be forgotten that the basis for success is the intricate and often not spectacular basic research. This research is arduous and fallible, there are sometimes dead ends, and, since science is conducted by humans, there is also misconduct. Despite all this, this research has led to reliable scientific knowledge.

Scientific truth is a contested concept in the history and philosophy of science that is often dismissed outright. In contrast, philosopher of science Yemima Ben-Menachem recently came to the interesting conclusion that “the occasional failures of science provide a much better argument for realism (the belief that scientific models and theories reflect aspects of reality) than its success” (Ben-Menachem 2021). It is indeed interesting that it is failures, such as irreproducibility and flawed theories, their disclosure and correction that best confirm that the search for truth in the meaning of reliable knowledge has remained the fundamental goal of science.


  • Y. Ben-Menahem. 2021. Realism and the Failure of Science: Reversing the No Miracle Argument. Academia Letters, Article 152.
  • U. Deichmann. 2021a. Template theories, the rule of parsimony, and disregard for irreproducibility – the example of Linus Pauling’s research on antibody formation. Historical Studies in the Natural Sciences 51/4, 427–467.
  • U. Deichmann. 2021b. Data, theory, and scientific belief in early molecular biology: Pauling’s and Crick’s conflicting notions about the genetic determination of protein synthesis and the solution to the ‘secret of life’, HYLE – International Journal for Philosophy of Chemistry 27, 25-46.
  • J. P. A. Ioannidis. 2005. Why most published research findings are false, PLoS Medicine 2, 696-701.


Please note: This is a commercial profile

© 2019. This work is licensed under CC-BY-NC-ND.

Contributor Profile

Professor, Director
Jacques Loeb Centre for the History and Philosophy of the Life Sciences at Ben-Gurion University
Website: Visit Website


Please enter your comment!
Please enter your name here