I can't remember in which book I read it, but it said that the biggest single predictor for language success amongst those IIRC entering or trying to enter the US FSI programmes was "auditory memory". So (making this up now/extrapolating here) the type of candidate who'd respond "Huh? Whaddya say? Doobidoo...bidoo?" when asked to repeat 'Duibuqi' (ooh, some Mandarin for ya, LSK!
) would, in having not even caught the correct number of syllables, be less likely to succeed than the ones who responded "Dweypoochee" (we'll let them off the -p- they substituted for the -b- LOL). Mind you, thinking back on it now, that smacks a little of Audiolingualism, but never mind, as the basic point is doubtless valid.
"she also happens to study much harder--possibly as a result of her success" - Hmm, it's a bit chicken and egg though. A large part of her superior pronunciation etc must be due to practice making perfect, but hey, I'm sure there are some "innate" attributes at work there too. (That being said, these males are still nothing like the stones, rabbits or telephone exchanges beloved by Chomsky when arguing for innateness in humans - they do possess the power of hearing and ears, albeit made of cloth, apparently!).
By the way, perhaps also take a look if you haven't already at e.g. Jennifer Jenkins' (author of
The Phonology of EIL etc) research on NNS interactions and 'mutual intelligibility', as the native standard isn't always the one most well-suited to this type of communication it would seem.