There's not as much reason to pay attention to statistical reasoning when we have insight into causal mechanisms. Particularly when our knowledge of the causal mechanisms suggests that the statistical results are very susceptible to misleading interpretations.

Incidentally we have essentially perfect insight into the causal mechanisms of what makes a number prime, and yet this sort of reasoning is spectacularly successful:

Cramer's random model of the primes asserts, roughly speaking, that the primes behave as if every large integer n had an independent probability of 1 / log(n) of being prime (as predicted by the prime number theorem).