Reproducibility Project

Battling Bad Science

  • Prosperity
  • 2017

By 2017, the Laura and John Arnold Foundation had already donated more than $80 million to fix a problem the rest of the world was just becoming aware of: a large fraction of all scientific research today is badly flawed, impossible to reproduce, and inaccurate. When it comes to improving the quality of science, says science watchdog John Ioannidis, “the Arnold Foundation has been the Medici.”

One beneficiary has been the Reproducibility Project, launched by a University of Virginia professor to test how many of the studies published in top psychology journals could be repeated with the same experimental result by other scientists. Only about four out of ten, it turned out. With Arnold funding, this has led to new efforts to improve the quality and integrity of research by helping, and pressing, scientists to post their raw data for public study and otherwise be more open about their procedures and assumptions.

Similar critiques and reform projects supported by the foundation helped expose the arbitrary and incomplete nature of many of today’s scientific pronouncements on nutrition, and flaws in much of the research that produces pharmaceutical drugs. This single-handed work by the Arnolds helped convince a majority of scientists themselves that current research is plagued with biases and “reproducibility” flaws. A 2016 investigation by the journal Nature found that more than 70 percent of researchers have tried and failed to reproduce another scientist’s experiments, and more than half have failed to reproduce their own experiments.

The Arnold Foundation has made long-term commitments to continue airing this problem, and funding potential solutions built on better and more open research methods, for decades to come.