In an interview with John Horgan at Scientific American:
Horgan: The “reproducibility crisis” in research has raised questions about science’s reliability. Do scientists deserve some blame for widespread debate over climate change, evolution and vaccines?
Dawkins: It is a real worry, perhaps especially acute in medical research. Part of the problem is the tendency for results to be simplified in order to make a neat, easily summed-up story. And this is exacerbated when recent research results hit the newspapers or other media.
Another problem is the “file drawer effect” whereby papers that fail to disprove the null hypothesis are never published, because authors or editors think they’re too boring. This could theoretically lead to falsehoods being propagated: If enough studies are done, a minority will yield statistical significance even if the null hypothesis is true.
Despite the “reproducibility crisis” there are some scientific conclusions that really are robust and become progressively more so as time goes by. The fact of evolution is one such. More.
We all know about the file drawer effect but Dawkins’s colleagues can make a living off failure. So why should they care? Any reform needs teeth.
And what does it mean to say that the fact of evolution is “robust”? The problem is that evolution is a history and there is currently a huge mess around what exactly happened in many areas and how.
See also: Teaching evolution to creationist students: Why would anyone who was embarking on teaching evolution as a serious project try to involve a virulently anti-religious figure like Dawkins in the argument?
and
What the fossils told us in their own words