This is an intriguing article. Author Peter A. Lawrence argues that by not acknowledging that on average men and women have different characteristics, we are doing science a disservice. The current system selects for aggressiveness, assertiveness, self-confidence, self-aggrandization, and so on. Men, on average, tend to have these qualities in higher quantity than women. Women on the other hand, tend to have nurturing, understanding, and social skills that will encourage their peers and mentees. Such characteristics are sorely lacking in the field of science currently, and their lack continues to add to the attrition of "feminine women and feminine men."
The argument of the paper is that where men and women have different characteristics we need the feminine, and where they are the same (originality, creativity, insight), the current trend makes us select against desirable women and for less desirable men. In addition, the social taboo against acknowledging these differences in averages perpetuates the sexism inherent in the system. It's a very interesting thesis, and one that treads a fine line between trying to abolish bias, and perpetuating bias.