Science, Uncertainty, and Statistics
Just yesterday, Allyson and I had a great discussion in the car on the way to dinner about the general lack of scientific understanding here in America. My contention after reading through this article on science journalism was that very few people in America have an understanding of the role statistics plays in science and the scientific method. Frankly, scientists don’t exactly prove anything. Instead, experiments are largely designed to assess the statistical likelihood that some combination of factors is correlated in some scientifically significant way. Each scientist must lay bare methodology and raw data in a peer-reviewed scientific paper. Then that scientist’s peers are expected to challenge or confirm those results with new experiments to test the relationship of those factors using a different method. Science is a conversation, and statistics is the language in which that conversation takes place.
Coming out of our discussion yesterday, I did some digging online and found one of the best examples I’ve found of explaining the role that statistics plays in science. This quote from Gavin Schmidt comes from a recent public debate on whether or not climate change is a crisis.
I want to talk to you a little about the nature of this public debate. And I want to give you some background to what you’ve been hearing so far, and what you’ll hear a little bit later on. The issue of global warming and whether it’s a crisis or not, is in fact a scientific decision, it’s a scientific issue. It’s not a political one. On the other hand, deciding what to do about it is obviously political. Science can inform those decisions, but it can’t determine what decisions society makes. But we’re here to debate the existence of the problem and whether it is a crisis. That’s something that the scientists on this side are eminently suited to do. You’ve all seen or heard about the CSI police drama, where high tech forensic scientists try and work out who done it when they come across the scene of a crime. Well think of climate scientists as CSI planet Earth, we’re try-, we see a climate change and we try and work out what’s done it. Just like on CSI we have a range of high tech instruments to give us clues, satellites, ocean probes, radar, a worldwide network of weather stations and sophisticated computer programs to help us make sense of it all. The aim is to come to the most likely explanation of all the facts fully anticipating that in the real world there are always going to be anomalies, there are always going to be uncertainties. Conclusions will be preliminary and always open to revision in the light of new evidence. If this all sounds familiar, it’s because it’s exactly the same approach that doctors take when examining a patient. They don’t know everything about the human body, but they can still make a pretty accurate diagnosis of your illness. We end up then with a hierarchy of knowledge. Some things that are extremely likely, some things we’re pretty sure of, and some things that we think might be true, but really could go either way. There isn’t a division into things that are completely proven and things which are completely unknown. Instead, you have a sliding scale of increasing confidence. Let me give you a few examples. We’re highly confident that the sun is gonna rise tomorrow, it might not, it might go nova. But it’s likely that it will happen. It’s quite likely that you’ll be able to get a cab home from this event, unless it’s raining of course. [LAUGHTER] But, but those two things have different levels of certainty. You’re used to the idea that different kinds of knowledge come with different levels of certainty, and that’s exactly what we’re talking about when we talk about the impacts of climate change.
If more people understood this simple truth about science, I think that we would be in a much better position to have intelligent discussions about scientific topics in the media. If you’re a science educator at any level, I implore you to stress this point before all others. Hypotheses aren’t proven. Science is an on-going conversation.
Technorati Tags: Education, Science, Statistics, Gavin Schmidt