At RealClearScience, David Randall, Director of Research of the National Association of Scholars offers a look at potential reforms in light of the growing reproducibility crisis in science.

First, a synopsis of the problem:

A large part of modern science and social science already is tainted by the irreproducibility crisis of modern science—the combination of politicized groupthink, academic publish-or-perish culture, and statistical carelessness that means more than one half of modern scientific and social scientific research may be wrong. The same incentives that produce sloppy research, and which prevent any checks to see if research has been done badly, also provide space for actual research fraud. If you don’t provide open access to your data, if you don’t preregister your work, if you shove negative results into the file drawer—you allow the deadwood to get tenure, but you also allow the actual cheats to prosper.

Modern science tolerates fools, and that means it has no way to tell who are knaves.

“Modern Science Tolerates Fools and Knaves.” Here’s a Solution, November 4, 2024

NAS offers a reform prop[osal based on changing federal science grant incentives:

The National Association of Scholars (NAS) is working for reform within the system. We have published our Model Science Policy Code as a way to improve the system, but not tear it down. What we most want to improve is the role of federal government in America’s system of scientific research.

The federal government is the largest single funder of scientific research in the world, and federal funds not only distort American regulatory policy but also subsidize the wholesale production of irreproducible research – some outright fraudulent – in American universities. Federal regulations, in turn, can have real effect in changing the practice of scientists in America—and, indeed, worldwide. If scientists won’t get federal research funds unless they act honestly, then they will (with bad grace) act honestly. “Fools and Knaves.”

One example is the NAS Model Science Policy Code: “ten individual pieces of model legislation are meant to inspire policymakers and the public to work for concerted reform of federal science policy.”

Two questions going forward: Can science even survive our era of private truth in which truth becomes what we need to believe about ourselves and our world? Related question: If money for science becomes a way of awarding public funds on a basis other than achievement, do research results even matter? The inability to reproduce research results matters only if we assume that there is some kind of truth that we must find out there. Many influencers do not assume that any more.