Let’s analyze the problem using Bayes Net terminology. Let A={suspect=Knox’s demographic profile}, B={suspect lives with victim} and C={suspect guilty}. Then your claim is that the net is structured as A->B->C, or that the demographic evidence is conditionally independent of guilt given co-habitation.
No.
My claim is that the net is structured as A->C<-B; both A and B affect the likelihood of guilt, and in particular A substantially reduces the likelihood of guilt as James_Miller points out (Note that I’m not saying B is irrelevant, obviously this is wrong).
No.
I agree.