This is an blog posting that's "just too late to be useful". The polls for the Melbourne District Byelection are closing. (This means I don't need to worry about whether to comply with election comment laws and post my name and address.)

Since I had to vote today, and since I don't tend to slavishly follow tickets, I decided I needed to have a summary view of where each candidate stood. There are 16 people standing, so there's no way I will instinctively recognise them all.

Almost all the candidates have produced how to vote cards that rank the other 15 candidates in their preferred order. I figure there is a good graphic way of showing the interrelationships. Here's my first attempt.

I've used IBM's ManyEyes software to generate a visualisation.

Image:Melbourne Byelection Preferences
Click here to see the full sized visualisation.

I've listed each candidate's top three preferences (2, 3, 4 on their ticket). A nomination in either direction creates a line. So, for example, O'Connor (Socialist Equality Party) gave no preference, so he's linked to "none". But one of Borland (Public Housing's) top three preferences went to O'Connor.

The two top "Networked" candidates are Ahmed and Collyer (Australian Democrats). In a sense these are candidates that are generally liked by other candidates.

Green candidate Oke gets quite a bit of love from some of the minor left parties. A lot of the pre-election commentary said that Kanis (ALP) had done a better job of gaining preferences from other candidates. That isn't especially obvious in this picture, so either she didn't do as well as it was reported, or, more likely, my diagram doesn't show this - if she was fifth preference for a minor candidate, but everyone from 1 - 4 has been eliminated, then she might pick up a preference ahead of the Greens Candidate Oke.

All will be revealed when counting starts in a few minutes!

Rolling Commentary on results (Press F5 or Refresh to see updates during the evening.)

The Victorian Electoral Commission kindly makes progressive results available on its Tally Room web site using a special standard for election results.

At 6:12pm the first results have been posted.

... each candidate has received 0 votes. It's too close to call. :-)

.... waiting for the next set of results to be posted.

As I wait for some meaningful results to be posted, I'm thinking about how to improve my preferences diagram in the future. It works OK for 'clumping' like minded candidates together (the FF, DLP, Aus Christian cluster), but doesn't help in deciding which of the likely candidates will pick up preferences. Maybe for each Candidate I ought to show the order in which they preference the top three likely candidates (in this case, ALP, The Greens and either Mayne or Nolte) - whether they are at 2, 3, 4 or maybe 6, 9, 12... or whatever.

While I sit here waiting for some results to be posted, I've been contemplating the campaign and the four opinion poll surveys I received (and answered) during the campaign, including the "push poll" that tried to persuade me that, amongst other things, The Greens wanted to close down Melbourne Zoo. There was also a lot of commentary which suggested it was important to send votes to the ALP to send a message re Baillieu or Gillard. The implication (from the left and the right) is that if the ALP vote is down, it is either a failure to show an opinion against Baillieu's government and/or a comment against the Gillard government. I don't think that's a reasonable comment. If there were a Liberal Candidate, confidence in the Baillieu government would be shown mostly through that candidate's primary vote. In the absence of that, I think it can best be seen by the number of first preferences given to the "right" bunch of candidates in my diagram (probably Schorell-Hjavka belongs in that group too.) The fourth of my opinion polls, The Australian's Reachtel poll, indicated that the ALP vote is unaffected by thoughts about the Federal Gillard government.

I sing to myself: "Why are we waiting..." No results from the VEC so far. There's a tweet that says that the Greens won the RMIT booth by 300 to 489. But what does that mean? Is it a first preference vote? If so, how many people voted in that booth? (Have the Greens reached 50%? And I guess you'd expect them to do well near RMIT.)  Twitter seems to be using the #melbvotes hashtag.

Ah: They've posted some results.
The Greens 40%
ALP 31%
I need to think about whether this is strictly (mathematically) accurate, but I think that this means that if The Greens pick up only 1/3 of preferences and 2/3 go to ALP, they would win. Could end up being fairly tight.
The results came from 2 booths. There are 14 in total.
After ALP and The Greens, three of the four highest candidates (Patten, Ahmed and Nolte) all preference the ALP. Mayne goes to The Greens.

8:23pm Woo Hoo! More results.
ALP and The Greens are now closer: 32% to 39%.

The Greens need to rise 11%, the ALP needs to pick up 18%. A whole bunch of Liberal voters will presumably send their votes towards the ALP, so it's not necessarily going to be easy for The Greens to reach 50%. Unless, of course, the 9% of informal votes are overwhelmingly from Liberal voters?

Pollbludger is doing good things in adjusting preferences to predict the way the votes have been seen to be going. At 8:38pm, the Pollbludger prediction is for a line ball result.

Based on the 8:31 results, the only booth where the ALP got more first preference votes than the Greens is East Melbourne. (ALP 466, The Greens 436).
Oh: And Hotham Hill: ALP 375, The Greens 213.

More results posted. ALP 32% Greens 38%.

At the risk of being a bit premature: it looks like the ALP might come from behind and cross the finishing line ahead of The Greens. (But I wish I had a spreadsheet like Pollbludger's so I make predictions based on something sounder than a bunch of observations and gut feelings.)
(On the other hand: Pollbludger's spreadsheet just crashed. Oops!)

It looks like it's too close to call tonight.

The VEC snuck in another set of results at 11pm. Early votes went slightly towards The Greens. Postal Votes went strongly to the ALP (presumably as a result of their experience in this area).

The VEC says there are about another 1000 Postal Votes to come. Based on the current party split seen with Postal Votes, it's looking pretty certain that the ALP has won the Melbourne District Byelection.

Comments (0)
Anthony Holmes July 21st, 2012 05:31:55 PM

No Comments Found
 Add a Comment
Web Site:
Comment:  (No HTML - Links will be converted if prefixed http://)
Remember Me?