I wasn't clear. I couldn't figure out how to express my full point concisely, so I left it short and potentially misleading. I had hoped the meaning was implied, but I see now it wasn't. Thanks for pointing that out – I definitely didn't mean to say attractive women are less intelligent!

When I said "less likely" I meant "less likely than if you were looking for intelligence specifically".

So maybe, odds of intelligence when looking for it = 25% Odds of intelligence in general population = 10% Odds of intelligence when searching by attractiveness = 10%

I didn't mean attractive women are less intelligent. I meant that by searching for attraction he's less likely to find intelligence than if he were searching specifically for intelligence.

Strictly speaking, "If you are selecting based on looks (as opposed to picking randomly)" simply implies that you believe fewer attractive women are intelligent, proportionally. This actually could be reasonable without thinking less of attractive women, if you assume that attractive and intelligent women are likely to already be taken, but it's still certainly an assumption.

It gets even less damning if we say "If you are selecting based on looks (as opposed to picking based on intelligence)". It should be uncontroversial that a group of women selected for intelligence is going to be smarter than one selected based on an uncorrelated (or even weakly correlated) feature.

Thank you, that's what I meant. I just wrote a paragraph in another comment, but you expressed it much more concisely.

Searching based on looks rather than (traits likely to signify) intelligence is likely to lead to less intelligence found, even if the subset (attractive women) is equally likely to be intelligent.