In: Economics
The question is really a philosophical one with some obvious relevance to economic modelling. Let me start with Plato's cave.
Plato's Cave is a thought experiment which came to be known by the name in remembrance of Greek philosopher Plato who presented it in his treatise The Republic. He holds its plausibility as sufficient to conclude that we could not realize reality as it truly is because this Reality (capitalized R intentional) exists on a metaphysical plane impenetratable by the physical senses.
Instead, what we (mistakenly) perceive as the real world are the shadows caused by this true reality. This true reality is the real causal world where accurate relationships between cause and effect can be drawn; but unfortunately it is inaccessible and whatever we learn from what we perceive ourselves are mistaken inferences kinda like giving undue importance to spurious correlations.
As such all models and modelling is flawed since, by Plato's argument, we may not be able to observe all the relevant antecedent causes.
-----
Unlike his younger contemporary Plato, Hippocrates was a physician and had the benefit of trial and error in the real world. He recognized that at least for some phenomena (viz. some diseases) cause and effect could be deduce; theoretical frameworks could explain observed physical phenomena in terms of simple materialistic - or in other words, observable - causes.
The stress on observability is important because a model like Plato's - which poses a question but proposes no solution - is dangerous on humanistic grounds. "Divine intervention" from above has been exercised as an excuse for exploitation since time immemorial.
As such, if (and especially if) Plato's argument is correct then all human models / opinions are equally flawed. Thus an observability-oriented model is at least as good as any other. And it's definitely easier to come to group conclusions - which is what all government policy is really about - when the decision-making is grounded in observability (e.g. numbers and metrics).
-----
Thus models may not be be "just" in an Absolute sense, going by Plato's argument e.g a particular homicide somewhere may indeed be the outcome of supernatural activity.
However, societal laws (e.g. punishing homicide) exist to preserve the integrity of the fabric of human existance.
As a societal collective we've decided to outlaw (and thus disincentivize) certain activities judged to be harmful to this integrity. This decision making is crucial, otherwise we cannot coordinate and derive comparative advantage.
----
If the true reality is indeed inaccessible, perhaps models of this reality are an acceptable Second Best
. . . . and so are conclusions derived from this conscientious (honest) modelling.