> I do not agree with the first statement - it supposes that men and women will, given an option and without external pressures, always choose the same thing in equal amounts.
That's not quite what I'm supposing.
I'm supposing that the choices of individual people without external pressures will always choose things unpredictably. In other words, correlation between individual choice and any particular demographic facet of a person should be pure chance.
This is how random number generators work. You want to provide as even a distribution as possible. That doesn't change the fact that, given a certain seed, the generator will always return the same "random" result.
The analogy comes back around like this: if you can reliably predict that a random number generator will return a number divisible by 3 if the input is odd, there's something weird going on even if some results are not divisible by 3. Chances are, if you look at the RNG's algorithm, you'll find something that creates that bias.
Do you care? Is that an issue? Is that a vulnerability? That's an entirely different question. As software engineers, we'd call it a business or design decision. In the wider world, we call it morality or ethics.
Edit to add: One of the consequences of this perspective is that it factors in any demographic variety, not just "men and women". If someone has a chromosomal set of XXY, then it still makes sense to consider their representation. That it's vanishingly small, such that they rarely register a blip on populations smaller than "the entire world", is part of the same analysis.