Keith I think you are getting to the nub of the problem. Graham, seems to have used data from vote that was gained in each electorate. I see a few problems with this - the redistribution after 2007 meant that the profile of some electorates had changed significantly.
One way to compensate for this is to focus on booth results (this is also the strategy parties use in planning their campaigns.) the trouble is all we can get from the AEC website (or more precisely all I could find) is two party preferred by polling booth.
Graham - I take your point that you and David are talking about different things. However, one may wonder whether or not the protest vote in the way you describe it gives us any relevant information. Indeed one wonders if it is possible to draw any firm conclusions simply by interrogating the poll results.
(as an aside why is it that thread seems to have lead to an attempt at a rational discussion? We are on page 2 and there has been no vitriol)