Our quants tend to line up with the telephone polling that Nielsen, Morgan and Newspoll have been publishing. Not that the absolute numbers fall in line, but the theoretical movement since the last election is similar.We had 282 responses, which is a little small for me to do the same manipulations that I do with larger samples, so I've generally treated this one as three separate focus groups with each of Labor, Liberal and Greens voters.

What is the composition of the sample?

On the left are our figures, and on the right are percentages taken from Newspoll's survey of the 15th November. Newspoll doesn't add up to 100 and the balance is made up of an amalgamated category called "Others". It is impossible for me to know how the individual parties in our "Others" matches up to those in Newspoll's, although their total is 5% and ours is 12%. It reflects the fact that our sample is much less likely to say that it will vote for the major parties than the community at large.

Our sample is also much more likely to vote Greens. Three times more likely, in fact, if the Newspoll figures are accurate. On the other side Labor and Liberal are around half what they are in the community at large. While the Greens component of our polls is generally higher than in the community at large, this degree of over-representation means that the Greens vote is enthusiastic and will probably do much better than expected tomorrow. They may win Melbourne despite the Liberals' decision to preference against them.

Is the state heading in the right direction?

As usual the answer depends on voting intention. Labor voters are extremely positive, while Liberal voters are negative. There is a net positive balance between Liberal and Labor, so allowing for the fact that the Greens vote is three times too high, it is probable that the direction question is actually pretty finely balanced between those who think it is heading in the right direction and those who don't.

Distribution of preferences

No surprises here (apart from the small percentage of Liberal and Labor voters who say they will preference the other side). Most significant is the Greens preference allocation as it will probably determine the result in a number of seats. Will be interested on the night to see whether it matches up with reality. at only 23% to the Coalition it is lower than it used to be, but that is probably a reflection of the fact that the increase in the Greens vote has come mostly from Labor voters, as you can see in the table below.

Vote last election versus vote this election

So, out of those who say they will vote Greens this time, 46% voted Greens last time and 52% Labor. As the Greens vote has doubled this essentially suggests it is all from disaffected Labor voters. At the same time there appears to have been a swing back to Labor from some Greens. The increase in the Liberals vote would appear to be straight from Labor, with a small contribution from Greens and Independents.

Brumby approval

Again, on party lines, with Liberals being less inclined to disapprove than Labor voters to approve. Just over half of Greens voters disapprove of Brumby, providing a possible reason for the large number of defections from Labor to the Greens.

Baillieu approval

This table tells us a lot about the two leaders. Their approval rating is very similar, which should cheer Ted Baillieu. It is also obvious that Brumby is a polarising personality. While 48% disapprove of Brumby, only 38% disapprove of Baillieu. A similar mathematical relationship holds true for approvals where it is 33% for Brumby and only 21% for Baillieu. Similar relationships hold true across the three major parties.

In some ways this is a benefit for Brumby. His team is energised behind him. In other ways it is a benefit for Baillieu because no-one hates him, although this might also mean they believe he is less effective than Brumby too.

Preferred premier

This shows the weakness of Brumby's position. While Labor voters are unanimous that he is the best only 61% of Greens voters agree. If it comes down to a personality competition on the day, despite what the aggregated figures say, Baillieu is well-positioned. He will get all of the Liberal vote but he could sway a few more Greens, a quarter of whom are not sure who is the better premier.

Expectations

Expectations are important not because they tell you who will win, but because they give you clues to likely voter behaviour when you combine the results with who voters want to win. This table says that a hung parliament is viewed as likely by many voters, while a Liberal win is not even viewed as likely by Liberal voters. The largest constituency for a hung parliament is amongst Greens voters. This table says that Greens voters are not likely to be scared back into voting Labor, and that they will be emboldened to protest against Labor as they don't expect the Liberals to even come close to winning.

Desires

As would be expected, Liberal and Labor voters both want their side to win, with some small aberrations. (Four times as many Labor voters as Liberals want a hung parliament). Greens voters are most inclined to want a hung parliament, and only 23% of them want a Labor win. The split is similar amongst those Greens voters who voted Labor last election. So Greens voters are looking to engineer a situation where they will be able to exercise power to rein in a Labor government. They don't show a high propensity to want to vote Liberal to try to do the same thing.

Conclusion

Our quants point to a large Greens turnout, with most of these voters wanting to change the direction of government. They believe they can do this through a hung parliament, and will vote to try to make that happen. They dislike John Brumby more than they dislike Ted Baillieu, and there is therefore a possibility that in situations where they don't think they can elect a Green some of them may stray to the Liberals for Labor to lose some additional seats. The Liberals problem from this point of view is that Baillieu may not be seen as sufficiently forceful to be an effective way of bringing Brumby to account.

Share this article on your favourite social bookmarking sites: