...[A series of studies] provided respondents with a portfolio, a job application, an individual essay, or other information that was attributed to a male or a female. Whether the task is to admit someone to a graduate program, to select someone for tenure, or to assign a grade to an essay, the studies demonstrate that documents associated with a male name consistently get a higher rating than the same documents associated with a female name. For example, Elizabeth Spelke and Ariel Grace report on a study of a tenure decision for a candidate with an average record. When the dossier was associated with a male name, 70% of the reviewers recommended tenure; when it was attributed to a female name, only 45% recommended tenure.
From this review by Marcia C. Linn Why Aren't More Women in Science? Top Researchers Debate the Evidence .
Thanks to Peggy (at the excellent Women in Science blog) for the pointer. Peggy also offers longer excerpts with a commentary.