70% of published scientific findings cannot be replicated. The one field that kept replicating results got shut down for it.
On the reproducibility crisis, what it actually means, and the specific punishment reserved for research that keeps showing up.
In 2011, a team at Bayer AG attempted to reproduce the findings of 67 published cancer biology studies that their drug development work depended on. They could reproduce fewer than 25%.
In 2012, a similar effort by Amgen scientists targeting 53 landmark oncology studies replicated 6 of them. John Ioannidis at Stanford published the paper “Why Most Published Research Findings Are False” in 2005, and the citation count now exceeds 7,000 - not because scientists disagreed with it, but because they agreed with it enough to keep citing it. The reproducibility crisis is not some fringe claim. It is a technical consensus about the state of the technical literature.
The standard response is that this is a failure of incentive structures like publish-or-perish culture, underpowered studies, p-hacking, or selective reporting. All of which are real and documented. But there is a more interesting question beneath this diagnosis. If those are the mechanisms that explain why the literature fails to replicate, what does it mean that one research program kept replicating across decades, across independent labs, across methodological variations… and was still terminated?
The Princeton Engineering Anomalies Research laboratory was founded in 1979 by Robert Jahn, Dean of Engineering at Princeton. Brenda Dunne served as the laboratory manager. The PEAR Lab’s research program investigated a single question for 28 years - does human consciousness interact with random physical systems at a statistically significant rate? The methodology was rigorous. The random event generators were calibrated and independently audited. The data was accumulated across 2.5 million individual trials. The effect was small by most physical standards - on the order of one part in ten thousand. It was also consistent, replicable across independent operators and across different random event generator designs, and statistically significant at levels that would be considered definitive in any clinical drug trial.
The PEAR Lab closed in 2007. Jahn said, in an interview afterward, that the scientific community had decided not to engage with the data. Several major journals had returned papers without review. The response from establishment physics was not to attempt replication, which would have been the scientific response, but to decline participation in the conversation.
This is a distinct pattern from the general reproducibility crisis. In that crisis, results fail to hold up when repeated. In the PEAR case, the results held up repeatedly and the field declined to repeat them. The scientific consensus that dismissed the work was not based on failed replication. It was based on prior assumptions, that effects of this kind cannot exist, therefore results of this kind cannot be valid. The evidence was filtered at the model level before it reached the data level.
This is worth calling out directly because it exemplifies the filter that has been running on a very large category of human experience for a very long time. The evidence for this is documented, replicated, and archived. The decision not to include it in the official account was institutional, not empirical.
Career Reading closes tonight at midnight. Three questions, full divination suite, direct read of your specific field. $50, regular price $144.
Further Reading
Begley, C. G., & Ellis, L. M. (2012). Raise standards for preclinical cancer research. Nature, 483(7391), 531-533. [online] Available here: https://doi.org/10.1038/483531a
Ioannidis, J. P. A. (2005). Why most published research findings are false. PLOS Medicine, 2(8), e124. [online] Available here: https://doi.org/10.1371/journal.pmed.0020124
Jahn, R. G., & Dunne, B. J. (1987). Margins of Reality: The Role of Consciousness in the Physical World. Harcourt Brace Jovanovich.
Jahn, R. G., & Dunne, B. J. (1997). Science of the subjective. Journal of Scientific Exploration, 11(2), 201-224.
Nelson, R. D., Bradish, G. J., Dobyns, Y. H., Dunne, B. J., & Jahn, R. G. (1996). FieldREG anomalies in group situations. Journal of Scientific Exploration, 10(1), 111-141.
Open Science Collaboration. (2015). Estimating the reproducibility of psychological science. Science, 349(6251). [online] Available here: https://doi.org/10.1126/science.aac4716
Radin, D., Michel, L., Galdamez, K., Wendland, P., Rickenbach, R., & Delorme, A. (2012). Consciousness and the double-slit interference pattern. Physics Essays, 25(2), 157-171.