Cameron, I don't think the VB model allows you to predict magnetisation. VB is good for the structures, but the problem is, it's far too complicated and ungainly to deal effectively with - ie its just not general enough...(just an example: sp3d hybridisation - trigonal bipyramidal structure - can also turn out ot be dsp3 depending on the metal and the available orbitals, and it gets a lot worse from there. It's difficult to form a consistent model in this way that is meaningful)
Magnetisation is best observed through Crystal Field Theory, in which ligands are approximated by negative point charges, and you observe the shifts this causes to the d orbitals of the metal (formation of t2g and eg subsets for an octahedral site), which in turn, depending on the nature of the ligand, gives you info abt the number of unpaired electrons in the ligand and hence whether it is para- or diamagnetic. Furthermore the energy differences produced as a result of the d-orbital splitting give you the colours or lack of them.
I'm not entirely sure what the dot-point wants, but I think it's fairly safe to talk firstly about the simplest of all models- the ball and stick - followed by lewis dot diagrams/lewis acid base theory, then VB and the basics of crystal field theory...no doubt Lewis would be the focus of analysis. For each of these you would assess the positives and the negatives and come to a conclusion I'd say.