Based on the evidence presented in the study, the conclusion that faculty have a lower regard for female job applicants than for male job applicants is at best an unsupported assumption and at worst a misinterpretation of the evidence. Furthermore, the study results are consistent with faculty holding beliefs favorably biased toward women and against men.
To rate job applicants on the basis of jobs applications one must hypothesize a relationship between the application and the applicant. One must explicitly, or implicitly through action, supply a general answer to the question: Given an application, how well will the applicant perform on the job?.
The only definite conclusion which can be reached from this study is that faculty hypothesize different relationships between application and applicant for male than for female applicants. But here is the kicker: The "bias" exhibited in this study is consistent with a belief among hirers that women job applicatns tend to look better on paper, not worse, than male applicants. Faculty offering lower salaries to women could be operating in the belief that women are better than men at presenting themselves.
If Professor Jane Doe believes the following to be true:
"Women are usually awesomely fantastic at presenting themselves, so if this female applicant looks looks like a 10/10 on paper, she is really probably an 8/10"
"Men are terrible ignoramuses at presenting themselves, so if he looks like an 8/10 on paper, he's probably a 10/10".
Those statements 1) Evidently display belief favorably biased toward female and against male applicants 2) Are consistent with the study results.
So, the traditional interpretation is flawed, because it is not a conclusion, but an assumption; there is no reason whatsoever to favor it over a handicapping explanation.
Someone should study what are the assumption of the faculty about the relationship between jobs applicants and job applications. And separately, if there is a difference in those assumptions between male and female job applicants, how accurate are they?