CounterCurrent: Week of 5/23
The National Association of Scholars has long been concerned with a lack of reproducibility in the sciences, that is to say, the ever-increasing prevalence of peer-reviewed scientific findings that cannot be reproduced, and that are therefore not real science. This work culminated in NAS Director of Research David Randall’s 2018 report, The Irreproducibility Crisis of Modern Science, co-written with Christopher Welser.
To be sure, irreproducible science is as old as science itself, as there have always been scientists who draw faulty conclusions from spurious research. But the irreproducibility crisis as such is indeed a new phenomenon, given both the scope of fields it has infected and the amount of work within each field that has been compromised. What happened? A combination of shoddy statistical analysis; a scientific culture focused chiefly on producing positive results rather than seeking the truth; political conformity bent toward drawing specific conclusions; and the lack of professional accountability, among other factors.
Unfortunately, the irreproducibility crisis does not stay in the test tube—it has had and will continue to have substantial real-world effects on American environmental and health regulations, and thereby on the American people. The Environmental Protection Agency (EPA) is primarily responsible for this travesty, as it frequently relies on irreproducible, unreliable scientific research when crafting regulations.
How many untold billions, even trillions, of dollars have been spent to comply with these dubious regulations? Nobody really knows, and herein lies the problem. As long as the irreproducibility crisis goes on unchecked, the EPA and other regulatory bodies will continue to operate on an unstable foundation—in other words, on shifting sands.
To get to the bottom of this issue, David Randall has joined forces with an all-star team of researchers to produce a series of in-depth reports, aptly titled Shifting Sands. Last Thursday, NAS launched the first of these reports, Keeping Count of Government Science: P-Value Plotting, P-Hacking, and PM2.5 Regulation, written by Randall, Adjunct Professor in the School of Public Health at the University of Alberta-Edmonton Warren Kindzierski, and the Director of the Shifting Sands Project, Stanley Young.
These three analyze a wide range of scientific malpractice that contributes to the irreproducibility crisis—including opaque methodology and data sets, P-hacking, HARKing, and political groupthink—specifically as they relate to the field of environmental epidemiology. The results are striking:
EPA regulations rely on environmental epidemiological literature, without applying rigorous tests for reproducibility ... Such rigorous tests are needed not least because earlier generations of environmental epidemiologists have already identified the low-hanging fruit.
These include massive statistical correlations between risk factors and health outcomes—e.g., the connection between smoking and lung cancer. Modern environmental epidemiologists habitually seek out small but (nominally) significant risk factors and health outcome associations. These practices render their research susceptible to registering false positives as real results, and to risk mistaking an improperly controlled covariable for a positive association.
Randall, Kindzierski, and Young also offer eleven recommendations to the EPA, designed to curb the irreproducibility crisis and bring regulations back to a foundation of reliable science. Does this all sound a bit too technical to you, my non-scientist readers? Fear not—as one who is a musician by training, I can assure you that the authors of Shifting Sands have worked hard to make sure that the core of these reports is accessible to all. I would recommend you read this first installment.
CounterCurrent is the National Association of Scholars’ weekly newsletter, written by Communications & Research Associate David Acevedo. To subscribe, update your email preferences here.
Image: Keith Hardy, Public Domain