[anileated]

There are no tests that can prove causality due to reasons I outlined in my previous comment. Correlation can be reliable enough to be considered causation for practical purposes, but unless you possess some sort of transcendental sacred knowledge about the nature of the world you’re in, strictly speaking any “causation” is mostly correlation with a pinch of unfalsifiable pixie dust (or, more charitably, metaphysics).

A model, especially when described to a layperson, can offer various simplifications (such as electrons flying around their atoms like planets around a star, or X causing Z), but no one seriously considers a model as literally depicting reality.

An explanation with unfalsifiable claims is always slightly metaphysics.

Metaphysics, and philosophy in general, is how we have natural sciences, scientific method, reasoning, all those things. Dismissing it is not even wrong, it’s nonsensical because it’s not a sibling, it’s a parent.