Right off the bat let's simplify things and say that human attractiveness as a mate has only two components: "Looks", by which I mean looks, and "charm", by which I mean all the rest - not only what is commonly called charm. The dependent variable is mating opportunities. The hypothesis refers to both short- and long-term mating, although the coefficients for the variables will differ between the two. Here's the hypothesis:
Charm has the largest influence on the mating opportunities of people with about average looks.
The basic idea here is that if you look like Naomi Campbell, you can afford to behave like Naomi Campbell, people are still going to want to mate with you. Conversely, if you look like Jabba the Hut, you may well be able to impress people with your wit, but that won't make people want to mate with you ("I'm seeing us more, like, friends."). When you look o.k., people are going to be sort of indifferent on the basis of your looks, so it's charm that tips the scales.
That isn't that original an idea, so I would be kinda surprised if it weren't in the research literature. But only kinda. That's because the hypothesis posits a nonlinear moderator effect. Just sticking the two as controls into your regression won't make you find such an effect (if it exists). Even if you go through the trouble of looking at scatterplots: positive linear effect for looks; positive linear effect for charm (I presume). Linear interaction between charm and looks? Probably exists, but that's not the effect we're talking about. You're not going to find this effect unless you go down The Popper Path: thinking first, testing later. It's sorta demanding.
In case it hasn't actually been tested, I'm giving away this hypothesis for free. All you'll have to do is cite LemmusLemmus (2011). Yeah, I know. Some people can't even bring themselves to cite people with real names, such as Steve Sailer (2006). It's sorta demanding.