Biologist at Monash University interested in size, energy and life history
Dustin Marshall
Loading...
the big problem is that in biology, we tend not only conflate prediction with explanation, we also assess the virtue of theories based solely on their fit to data. Numerous debates for many biological theories focused on relative fit but we think this is the wrong focus.
What if your own theory was tricking you? What if the theory your research was based on was just a “Clever Hans”? Out today in TREE, we explore this problem and more. In this 🧵 I will elaborate the Clever Hans problem in biological theory.
What if your own theory was tricking you? What if the theory your research was based on was just a “Clever Hans”? Out today in TREE, we explore this problem and more. In this 🧵 I will elaborate the Clever Hans problem in biological theory.
That a theory predict data is a necessary but insufficient condition for a theory to be sound. So rather than just focusing on fit we recommend that biologists spend more time scrutinising the assumptions of competing theories. some theories are merely Clever Hans, the challenge is to find them.
We usually build theories based on a series of biological principles and an expectation of what output seems reasonable from a theory. Biological theories aren’t derived out of thin air, we want them to have certain dynamics. e.g theories of growth should predict growth to be fast as first then slow
Of course Hans couldn’t actually do math, he was just trained to stop tapping when the human asking the question gave inadvertent clues that he’d arrived at the same answer. Theories can be like that. They can give the right answer (predictions match data), tempting us to infer they EXPLAIN BUT!
For example, several different models all predict growth trajectories equally well, but because they have different underlying mechanics, they can’t all explain growth equally well www.sciencedirect.com/science/arti...
