A random dataset could appear to be ordered to us, and we would be
none the wiser.


That's why experiments must be reproducable.


Wally wrote:
Still doesn't address the underlying problem that the notion of causality is
no more than an inference wrought of the empire of our experience.


That's true enough. But here's where the difference between "science"
and philosophy creeps in... when it's science, it happens every time.

For example, the phenomenon of water flowing downhill is essentially a
random event, the illusion caused by trillions of odd-shaped molecules
bouncing around any way they want. But somehow, water *always* flows
downhill.

Fresh Breezes- Doug King