Disclaimer: None of the below is official policy. I'm describing things that I believe go on in the heads of interviewers where I work.
I'm in academia, and I've been involved search committees. Before we bring someone on site, we do skype interviews and thoroughly scrutinize their CV. We select the best CVs regardless of gender or ethnicity or anything, and the skype interview is to assess their communication skills. A slightly smarter person who can't be understood is going to be less effective than someone who is perhaps a little less creative but communicates well. Teaching is very important, and being understood is very important in teaching. Gender does not factor into this, and people from other countries vary massively in how intelligible their accent is, so ethnicity doesn't directly enter into it either. (I'm pretty good at pronouncing other languages, because I have studied phonetics extensively. To some degree, a thick accent occurs due to a lack of talent -- they just have a really hard time understanding how to produce the sounds, and grammars are a chalelnge too. Most people don't have a talent for learning languages as an adult. However, I also think there's a laziness factor. Some people work harder than others and develop explicit compensating strategies. For instance, several of my colleagues from China have learned to *just slow down*, which helps like you wouldn't believe.)
With regard to female faculty candidates in in-person interviews, we tend to make two major assumptions:
1. They are a priori no more or less competent than the men.
2. Various cultures (including our own) make them less up-your-nose about their accomplishments.
3. Since women are generally perceived as less competent, a woman making it through a PhD program at a good school is often an indicator of superior "grit" (courage and resolve).
So when we interview, I think we tend to work a little harder at making sure we aren't missing any sparks of creativity, good ideas, or important accomplishments that the women may be unnecessarily humble about.
There are also some other factors:
4. Although we'd like to have stellar candidates, the main thing we evaluate is just whether or not they will be succcessful in research and bring positive attention to the university. While we certainly like the rock stars, there are many people who fit into the "very good" category, whom we would be very happy to make an offer to. Very few of the people we interview *aren't* in the very good category, independent of gender and background.
5. We're not Cal Tech or Harvard. The rock stars will go to the higher-ranked schools. With limited hiring slots and limited time to make decisions, we often choose "very good" and "likely to an accept an offer" over "rock star" but "likely to go somewhere else."
6. With programmed lower esteem, a more competent female candidate is slightly more likely to accept an offer than an equivalent male. We're not exactly taking advantage, because they decide whether or not they want to accept the offer. It's just a female applicant is likely to be more competent than they appear, and we're happy to factor that into deciding on the limited number of offers we give out.
So if we have a female candidate who has done good research and but was so-so in the interview, although we don't give slack on the quality of their publications, not bowling us over with how awesome they are in the interview is not going to hurt their case perhaps as much as it might for the men.
In my time here, two male faculty in engineering have washed out. No female faculty have. The thing is, the men who washed out were clearly not meeting standards, while all the female faculty have objectively strong publication records. It's not like we have much in the way of borderline cases where we let a woman get tenure with the same level of accomplishment as a man who didn't. The recoil effect of #2 the direct effect of #3 up there is that women in academia often work harder than their male colleages because (again culture) they are worried they will be perceived as less competent, so they over-compensate. They may generally be shy about telling you how great they are, but they're not shy about doing good research and publishing papers.
The conclusion is that gender does play a role in the interview process. To say otherwise would be disingenuous. But the effects are subtle, and the main reason we'd prefer a seemingly less competent female is that the seemingly more competent male isn't likely to accept the offer; if we were to make that offer, we would not only lose that candidate but also all of the others we were considering for the same position (because it's important to make offers in a time manner). So a *given* female candidate who is objectively competent is probably more likely to be made an offer than a similar male candidate. We haven't done the statistics on this, but it's probably true, but probably much more pragmatic than it is discriminatory. It's betting on the horse with the lower standard deviation rather than the higher mean. Despite this possible bias in favor of women, we still hire mostly men, because the vast majority of applicants are male. And that is disappointing to us.