In case you missed it earlier this week, Christie Aschwanden from fivethirtyeight.com published a superb article challenging the perceived reproducibility and retraction problems of current science. It’s a long piece about the scientific method and statistics, complete with wonderful interactive tools for p-value hacking. In the post, the Center for Open Science’s Brian Nosek describes science as operating, “as a procedure of uncertainty reduction. The goal is to get less wrong over time.”
That reminded me of the video below, a 2012 conversation between two favorites here at The Scholarly Kitchen, author Neil Gaiman and astrophysicist Neil deGrasse Tyson. In this segment, they discuss the differences between science and religion, and failed notion of “the God of the Gaps” (the idea that anything science can’t currently explain must be supernatural). Plus you’ll learn why the end of the rainbow is a really good place to hide a pot of gold.
3 Thoughts on "Science Versus Religion, The God of the Gaps, Being "Less Wrong" and Two Favorite Neils (Gaiman and deGrasse Tyson)"
What a pleasant–and smart–collection of present and past writings and ruminations. The notion of science working to reduce uncertainty is sensible and in no way invalidates the scientific process. Rather the process is validated. Because science can’t explain everything now doesn’t make science bad. Faith can’t explain everything, either. That’s part of the faith process. Why these two pursuits can’t be mutually respected has baffled the grown-up me. And I think is so because, when a child, I attended a church that respected both faith and science. And so I think it should go now for scientists and faith believers. Thank you, David Crotty, for this posting! And thanks to those who pursue science and pursue faith. (No thanks to those who fight about it.)
For people who are not as intellectually analytical as these panelists, faith can be a quite firm foundation for constructing a workable worldview, and a modus operandi for a functional life. The rights of those simple people should not be obstructed by smart people who can see rainbows clearly as the optical events they actually are. Furthermore, the faith-based person’s right to construct a worldview on religious foundations instead of scientific or data-based evidence is pragmatically legitimate, as well as being Constitutionally protected. The rights of those believers should, and will be, always defended by thinking people who truly understand the value of intellectual liberty.
Thank you, David, for linking us to this colorful exchange.
When we’re discussing the practice and philosophy of science, I think we tend to fall into the false dichotomy of either pronouncing science broken or defending it in such a way that discounts criticism.
I prefer to take a third view, as the article does, that science has a great history of success and we owe all of modern technology and medicine to it but there are some issues that need to be addressed. At this point, I think that we all know the biggest problem is P-hacking in the biomedical sciences. Before I knew that this was a widely discussed problem, I remember seeing the misapplication of ANOVAs and T-tests in some biology labs and being really concerned. The problem, as the article suggests, is that scientists in some fields are so strongly incentivised and so personally motivated to find positive results (wouldn’t you be if you were trying to invent treatments for pediatric brain tumors) that they misuse statistics without consciously knowing that they’re even doing it.
I do think that biology could benefit from a little bit of a course correction when it comes to attitudes to statistics. I sometimes felt that those who wanted to do thorough analyses and explore variable spaces properly, as the article suggests, were consider by their colleagues to be overly fussy. I remember once expressing concern about a data set that a colleague had produced because I felt that there was a serious systematic problem with it and being rebuffed with the answer that they’d done the negative controls and gotten a positive P-value, so it must be right. That’s just an anecdote, but reflective of an attitude that the most important thing is to find a stat to support a meaningful conclusion. That doesn’t mean that science is broken, but it could do with one or two bolts tightening here and there.