> And yet, people who exercise more are overwhelmingly more likely to be a healthy weight
This is what can be called a “correlation not causation.” Control for eating, and the picture would likely look different.
In the same way as “people who ride subway are more healthy than those driving cars.” It isn’t that riding subway on its own produces some miracle health benefits, riding subway usually means that you walk more and sit less (compared to driving a car as your primary commute method). If you control for that and start walking the same amount even if you drive, you will likely find the difference in health outcomes between driving and using subway disappear.
Yes, if you control for eating the picture will look different, because you're creating a different picture.
Modern, Western scientific thinking always tries to isolate variables, our medicine is always trying to isolate that one specific molecule, etc. This has given us fantastic insights, and allowed us to see the trees for the forest, but we also then start to miss the forest I feel.
The forest is a complex and difficult to quantify system of interrelation and interconnection. We may never fully understand the forest, but that doesn't mean the route to effective knowledge about navigating this landscape is to ignore forests, and concentrate on individual trees. I hope the metaphor is clear enough.
There are different ways of knowing, and to jarringly switch metaphors, just because you've built a hammer, doesn't make every form of knowledge a nail.
This is what can be called a “correlation not causation.” Control for eating, and the picture would likely look different.
In the same way as “people who ride subway are more healthy than those driving cars.” It isn’t that riding subway on its own produces some miracle health benefits, riding subway usually means that you walk more and sit less (compared to driving a car as your primary commute method). If you control for that and start walking the same amount even if you drive, you will likely find the difference in health outcomes between driving and using subway disappear.