Map of life expectancy at birth from Global Education Project.

Thursday, July 22, 2021

On the bias of science: a digression

This series has elicited a couple of what I consider idiotic and unpublishable comments to the effect that I have just proved why people don't trust science. This is a basic problem in science communication to the general public that is much discussed. Here's my two cents.


The problem is that scientists are generally reluctant to make highly definitive statements. Science is a continuously progressive endeavor and it's not uncommon for what we thought we knew yesterday to turn out to be not quite right today. Many conclusions are probabilistic - we're 90% or 95% sure of something. Sometimes an association holds under the circumstances we've tested so far, but we can't be sure it holds under other circumstances. Sometimes conclusions are expressed in terms of a confidence interval -- we're 95% sure that the true value lies between x and y. Scientists will even be oddly shy about assertions they are 100% sure are correct. For example, they'll tend to say things like "There is no convincing evidence for the effectiveness of homeopathy," when they know damn well that homeopathy is not just fake, but completely insane.


People generally hear this as unconvincing. They think it gives them license to believe something else. In the first place, I wish that scientists would be more direct in how they phrase conclusions when addressing the general public. In the abstract, you could say that we don't know anything for sure, maybe we're all living in a computer simulation or the world is an illusion. But if you accept the basic premise of intersubjective reality, there are some things we damn well know. These include evolution, anthropogenic climate change, tobacco causes cancer and heart disease, and the benefits far outweigh the risk of authorized vaccines for Covid-19. There is uncertainty about some of  the details and exact numbers. We can't precisely predict all the effects of climate change; or precisely say what the numerical chances are of some side effect of a vaccine and who might be most at risk, vs. the precise reduction in your risk of severe illness, again perhaps depending on some personal characteristics.


But the general conclusions in these cases are absolutely definitive. If you are willing to believe that the course of the sun across the sky is an illusion created by the earth rotating, you should believe this. There have been widely disseminated conclusions that have turned out to be not quite right, such as the relationship between consuming various kinds of fats and heart disease risk, or the risks of hormone replacement therapy. But what scientific authorities are telling you today is the best available understanding, and your best bet is to believe it. It may get even tighter and more definitive in the future -- that's most likely -- or it might turn out to be a little off. But you should bet the odds.

1 comment:

Don Quixote said...

At the end of the day, the comments of your critics are moot. They trust science, indeed, when they end up in the hospital and need a broken bone set, or radiation therapy for their cancer. If they have an accident and the EMT crew shows up, they don’t send them away, except perhaps if they’re Christian Scientists, saying, “I don’t trust science! Leave me to die!” People talk and talk and talk, but when it comes time to save their ass, they don’t worry about saving face.