A very interesting point, Cathy! I seem to remember that somewhere in the Ginther & Kahn paper there is a brief aside to the effect that things might be tougher for women at top-tier institutions, but I'm not sure they looked at the issue in detail. You might ask them (I don't know Ginther, but have met Shulamit Kahn briefly, and she seems like she'd be amenable to such questions). I would speculate that the impact of children would be the greatest at institutions that are the most demanding of people's time and energy. However, family issues don't really provide a very good explanation for your observations about Berkeley and Harvard.

Ginther and Kahn are looking at data across all institutions and disciplines; I could certainly see that there might be some interesting non-uniformity once you get to the level of individual departments. I think that people tend to hire / associate with people similar to themselves, so perhaps all-male or almost-all-male places need a bit of leadership (or a kick in the pants) to get with the program.

If it is in fact the case that there is a bit of an old-boys network still at some places, one upside is that as the ranks of women increase, these problems will presumably decline and the resulting virtuous cycle will further accelerate the pace of hiring of women faculty members.