A little while back I read a Language Log post on this, one of Mark Liberman's breakfast experiments. He looks at differences in switch timing, which I think is the same as what you're calling LSV, between male and female speakers in a large corpus of telephone conversations.