[tungstenfurnace]

Unfortunately there's no explanation which means there's no science here. (The world seems to be full of studies which measure things that are easy to measure and say things that people want to hear.)

Here's an explanation as to why exercise might be good for you: exercise transfers attention away from repetitive thoughts and into one's body and one's surroundings. It thus confers benefits similar to those of meditation.

This theory may well be totally false. But, hey, at least it offers an explanation. There's something there to be false.

The next step would be to criticise the theory and then, if it stands up, to test it against current theories.

Any testing though, would be highly unlikely to resemble the observations referred to in the linked article. It would instead depend on the relevant details of the rival theories.

[xiaoma]

>"Unfortunately there's no explanation which means there's no science here."

Experimental science has a long and colorful history of answering the what before the how or why. Consider the question of black body radiation. Without the experimental research results and Plank's desperate attempts to fit them to a mathematical model, Einstein wouldn't have had a reason to think of light existing as discrete particles and theoretical quantum mechanics would not have possible.

Creating frameworks for why things work is satisfying, but gathering data is just as vital to science.

[tungstenfurnace]

Yeah, sometimes novel phenomena are stumbled upon. Where they are deliberately stumbled upon I would call it exploration rather than science.

I don't think that exercise falls into that category. There is nothing surprising about exercise per se.

Popper once challenged an audience to go out and 'observe'; I think the point of his joke was that without knowing what you're trying to observe and why then the instruction is absurd.

In order to claim that 'more exercise is good' you need to start by guessing why some exercise might be good, whilst being open to the possibility that it may not be, and then start testing the best guesses. Otherwise you literally won't be able to interpret your 'more exercise' data, even if it is relevant.