They are mathematically bases guesses, or assumptions if you will.
-Stoic Joker
Actually, they're a lot more than that. -40hz
Only to the person that created them. Because only they know the (true) reasoning behind whatever deviation(s) has/have been used to achieve said target number(s). Statistics are projections of what might be if ... All of the other formulaic assertions used fall into place in accordance with the known factual parameters used, and nothing odd happens. To anyone else there simply an educated guess, that leverages their level of trust in whom ever ran-the-numbers.
Much unlike facts, which require empirical evidence/reproducible results, "valid" statistics merely have to avoid being dead wrong.
Discussion that happens frequently at the office (usually on rush jobs):
Brass: So, how's the testing going?
Me: Well, the
first test went well.
Brass: So it works then...
Me: No. The first test didn't fail... But a statical sampling of one doesn't prove shit to anyone (or at least it shouldn't...)
Brass: Oh... Okay, I'll come back later.