Where is the data to support your claim? It's possible that strangers are more dangerous on a per opportunity basis, but I suspect it's untrue. For example, children spend about the same time at school as at home (at least on weekdays). Yet abuse is much more frequent at home.