From New Scientist:
Unconscious biases and data-torturing are weakening our knowledge base – but unlike politicians and bankers, scientists aren’t covering up their failings
I started hearing in the mid-90s, about how scientists were going to fix all that stuff, but they didn’t actually do much better than politicians and bankers.
Science cannot afford to be complacent. Over the past few years there has been a creeping realisation that while bad apples are few and far between, there is a deeper problem. The barrel itself may be rotten.
Take heart! Bad apples are rarely the problem. They rot. It’s the good apples we need to watch.
What about the problem of reliable knowledge? On this front, things might also not be as bad as they seem. Last month, Science published a follow-up to the reproducibility paper arguing – ironically – that it used flawed statistics. Correct for these, and almost all 100 studies were reproducible, its authors claimed.
That, of course, may be just another case of torturing the data; the authors of the original paper have accused those of the new one of selectively interpreting the numbers. And it doesn’t absolve other problematic branches of science. More.
What’s mainly nded here is intellectual freedom and watchfulness.
Follow UD News at Twitter!