One of the features of our Federal Senate voting system in Australia is the option to vote below the line. Although there has been some substantial (and well deserved) complaint regarding flaws in the system, it actually forms a pretty good basis for selection that works well while the number of candidates remains approachable.
However, the number of candidates in the Senate, in most states, has long since become unapproachable in terms of the effort required to allocate preferences – in NSW for 110 candidates this time around, and even in Tasmania 54.
In anticipation of this issue, the system was designed with a short cut – you could nominate to vote above the line and pick one of the provided pre-arranged preference lists. Each party is able to provide one such list.
This however has lead to some problems in that these lists are not necessarily — nor commonly — ordered in the way that party members would select if they were to allocate themselves. This leads to a lot of the fuss that has occurred with micro parties managing to ‘snowball’ a very small vote up to the point where they pick up a seat through a combination of good fortune and good analysis.
Despite the ease of voting above the line, some people choose to take the effort below the line. In Tasmania this time around, around 10.3% chose to do so. After the election is counted, the AEC publishes the preferences selected by those who vote below the line, naturally without any identifying details. This allows a look at what order people who voted for a particular party actually did use, and thus a comparison with the party’s above the line selection.
A couple notes before looking at some of the data. Firstly, people have different reasons for voting below the line. A significant reason for doing so is because they don’t like some aspect of the way the party has arranged their above the line list. As such, the arrangement of votes sampled from below the line should not be confused with how those who voted for that party above the line would have voted below the line. Secondly, although the data is nominally by individual, the difference between individual and party results is almost exclusively determined by the order in the party and thus the results can be simplified to party results without loss.
With that said, lets have a look at the results for the Liberal Party.
So what can we get out of this? Apart from the order of preference which you can muse over yourself, there are a couple of interesting features. One is the shape of the curve; the other the deep spike in the 20 percentile line for Labor.
The shape of the curve illustrates the consistency of response; where there is a consistent response to another party, their point lies close to the diagonal. In this case it is obvious that the only really consistent responses are for the Liberal Party and the Greens. The other parties are either not well enough known or opinion is split upon them. The Liberal Democrats and Palmer United also show a lesser consistency.
Labor’s lower line shows a deep spike down; this reflects the substantial portion of voters who preference Labor very early in their votes, before most of the minor parties. Its not enough to outweigh the much larger grouping who put them last or second last, (see the spike up in the 80 percentile line), but it does reveal the distinct split in votes that occurs.
In my next piece I’ll look at the division that is revealed within Liberal Party voters with a little cutting and slicing of the data.