John Kay of all people is surely not getting things quite correct. He starts the article with a complaint about the Economics profession (well, who wouldn’t) but then states:
“And I have learnt that such bias is almost as common in academia as among the viewers of Fox News: the work of John Ioannidis has shown how few scientific studies can be replicated successfully.”
What John Ioannidis studied was largely the field of RCT or observational epidemiology in medicine — not science. RCT are, at best, a form of technology assessment usually with no scientific hypothesis in sight; and observational epidemiology is known to be largely bunk, because of confounding. Much pharma animal work is similarly devoid of a meaningful hypothesis. The real issue is, in a world where technology is usually so effortless, and where much or most of this technology relies on science, so much medicine (and some other branches of natural science) are so unreliable and insecure. To understand why this is the case, is the relevant question. The Mertonian norms have ended, and the currency is being actively debased. As John Kay knows, institutions matter, and it is to these we should look. A statistical hypothesis is not the same as a scientific hypothesis.