Second, most couples in my experience aren't of equal levels of attractiveness (and I'd like to see the research you quote because I don't know how levels of attractiveness could be gauged effectively enough to conduct such research). In fact, I'd say just the opposite in that there is legitimate research that says women are prone to date/marry men older than they are which would mean other traits rank higher than attractiveness (which of course fades as we age).
Beyond that, looking at the extreme, have you ever looked at the husbands of SuperModels?
Like Paulina Porizkova: http://img208.imageshack.us/img208/9066/paulinaqv6.jpg
Or Heidi Klum: http://www.clevelandleader.com/files/heidisealbaby.jpg (I love ya Seal but that stuff on your face is weird)
This is all a question of degree. I do think that physical attractiveness does play a very big role in both directions. But clearly women do place a somewhat lower premium on it than men do.
Just to add to the anecdotal useleness, rich men do attract women, but women swoon over Brad Pitt or that twilight star waaaaaay more than they every will over "the donald".
At 46 years old, Brad Pitt is also about twice the age of the women who get the most attention from men. In order to make a fairer comparison, imagine a man who's very handsome, but whose career and overall signals of success are on par with the physical attractiveness of "the donald".
EDIT: Looking at the spouses of supermodels or the super rich is probably unhelpful in gauging the behavior of most of the population.
The question has never been "is attractiveness a factor" but "how big of a factor is it". So a study that artificially boosts the importance of one factor is useless is answering that question
Also, why aren't supermodels relevant? I think they're the most relevant in that their physical beauty makes their choices almost limitless so when they choose an unattractive man it shows how little that aspect means to them.
Supermodels aren't relevant because they lie so far outside the norm, and can't be used to predict the behavior of a typical woman (even a very attractive woman).
Are women more flexible about their partner's attractiveness based on other factors? Maybe, but without any kind of research, it's all conjecture and confirmation bias.