Pineapple Party Time

Just another Crikey Blogs weblog

Cards on the table II

Saturday, March 21, 2009 – 10:20 am, by Mark Bahnisch

I’m going to chance my hand on a seat by seat prediction. I tip Labor to win 45 seats, the LNP to win 40 and Independents to win 4.

Details of the seats I think are worth watching are below the fold, categorised into likely changeovers (the basis for the numbers above), and possibles. They’re listed by party and region and in ascending order of margin. Seats with an asterisk are those where the nominal party has changed subsequent to the redistribution, and Indooroopilly where Ronan Lee shifted from the ALP to The Greens.

This has been a very hard campaign to read, and I wouldn’t be surprised to be wrong, and I’d be very surprised if all the seats I’ve nominated toppled. No doubt some will be held, and others will change hands.

Up until Wednesday or Thursday, I thought the most likely outcome was a very narrow LNP win, followed by a hung parliament. I think a large number of voters have only focused on the choice in the last few days, and the ALP’s last minute blitz has had some effect, combined with Lawrence Springborg’s own negatives. I also think that the release of the LNP costings very late in the campaign may have been a mistake, highlighting their implausibility just when folk are watching. If I’m right and there’s a narrow ALP win, it will be despite not because of the Labor campaign. It will also be in the face of the “time for a change” factor and the ALP’s position on almost all the issues. Anna Bligh will deserve all the credit, and the apparat should collectively be exiled to whatever purgatory or circle of hell where old operatives go to rest uneasily.

If I’m wrong, the LNP will have won despite and not because of Lawrence Springborg.

Read More »

Comments (17) | Permalink

Graham Young’s new poll

Saturday, March 21, 2009 – 6:23 am, by Mark Bahnisch

Graham Young has written up his last lot of polling for the state campaign at What The People Want. Of interest is the number of respondents in his sample who’ve switched their vote during the campaign, the association between switching and particular issues, and the direction the vote’s headed:

The purpose of any election campaign is to win votes. It doesn’t matter how many awards your ads win, or how the journalists score every day, if you don’t end up with more votes at the end of the campaign than you started with, then you have failed.

Queensland Labor has failed.

15% of our respondents claim to have changed their vote since the campaign began. Those who have favour the Liberal National Party by 54% to 46%. That means that while the government has won some voters, the LNP has won slightly more. Politics is all about percentages, so this is pretty good.

It confirms that if anything the swing has intensified.

On the basis of our results I suspect that the LNP is on about 52% of the vote with a greater than 7% swing. This is better for them than Galaxy or Newspoll, but historically the National Party vote has tended to be underestimated in polls. Galaxy and Newspoll also don’t tend to catch the late behaviour until it is too late.

Voters’ reasons for changing their votes are a microcosm of the issues driving the whole sample.

Labor is losing because voters believe they haven’t performed and that they are all spin and no substance. Health policy, and delivery of services generally, are their key weaknesses.

The Liberal Nationals aren’t winning because voters perceive them as insubstantial. They’re not losing, because most believe they wouldn’t be any worse than the government.

The Labor Party has tried to make jobs and the economy the issues of the campaign, but they have failed.

Comments (10) | Permalink

Newspoll 49.9/50.1

Friday, March 20, 2009 – 11:48 pm, by Mark Bahnisch

Newspoll [pdf] is out [hat tip to Oz in comments on a previous thread]. The poll, taken on Wednesday and Thursday with a weighted sample of 1328, suggests a very evenly poised contest.

Labor’s primary is 41.7 to the LNP’s 42.1. The Greens are on 7.4 and Others on 8.3. In the capital, Labor is on 42.5, the LNP on 40.1, The Greens 8.6 and Others on 8.1. Outside Brisbane, Labor has 40.9, the LNP 43.9, The Greens 6.4 and Others 8.5.

Bligh’s satisfaction/dissatisfaction/uncommitted split is 46:44:10, Springborg’s numbers are 39:49:12 and Bligh leads The Borg 53/33 on preferred premier with 14 uncommitted.

Labor has a 61% decided vote and the LNP 70%. Win expectations are 56/24 with 20 uncommittted.

UPDATE (William): Steven Wardill of the Courier-Mail offers intelligence from both parties’ tracking polling, which (assuming the reports are accurate) are telling somewhat different stories. Wardill says LNP insiders believe Broadwater, Southport and Burleigh on the Gold Coast are “too tough to claim”. On Labor polling:

Sources last night said Labor tracking showed the party’s support fought back on Wednesday night during the final TV advertising blitz but then dropped away again. The tracking apparently found the LNP was likely to win with a 60 per cent chance of a minority victory and a 40 per cent chance of a majority. Swings away from Labor were found in Gold Coast seats, including Burleigh, as well as in Mundingburra, Mansfield, Pumicestone and Everton.

However, Wardill appears to believe Labor will win Gladstone from independent Liz Cunningham. Seats he tips to fall include, but are not exclusive to, Barron River, Whitsunday, Keppel, Pumicestone, Indooroopilly, Cleveland, Chatsworth and Aspley. This comes in an article that doesn’t seem to be online in which six Courier-Mail journalists lay out their predictions, which echo the consensus by ranging from LNP minority to narrow Labor majority government (which is Wardill’s prediction, despite what Labor are telling him).

UPDATE: (Possum)

Graphporn and Monte Carlo sim time.

Adding this rather spiffy Newspoll addition to the Newspoll time series, we get the following charts of all the metrics (they’re thumbs, so just click on them to expand and place your mouse over them to see what they are)

npprims npqldtpp npbetterprem premiersat Opposition Satisfaction

Leader Satisfactions Net Satisfactions

One problem I have with these Newspoll numbers is that I can’t get the primaries to match the Two Party Preferred estimate without having 60% of Greens preferences exhaust and the remaining flow to the ALP at a rate of 70%, while the broad “Others” category exhausts at a much smaller rate of between 40% and 50% (leaving 50-60% flowing through as preferences) and giving the preference flow to the ALP from the “others” at only 40%. Not massively out, but I’d be surprised if that’s what actually happens. You can play around with the figures yourself using the Editgrid widget at the bottom of this post.

Plugging this into our Monte Carlo simulation then becomes a bit tricky. Using the breakdowns of both Galaxy and Newspoll weighted by sample size, and using the Newspoll exhaustion and flow rates, the ALP two party preferred aggregation comes in at  51.7% for Brisbane and 47.8% for non-Brisbane. If we plug that into our simulation, we get an implied probability of 66% that the ALP will win at least 45 seats, and where the most likely outcome is of the ALP winning 46 seats. That simulation includes the sampling error uncertainty of the polling aggregations. If we assume that the polls are actually 100% correct, the implied probability of the ALP winning at least 45 seats with a uniform swing and standard deviation of 4% increases to 87% and where the most likely outcome is the ALP winning 48 seats.

One of the more peculiar results in this Newspoll is the rather low Greens vote in Brisbane, coming in at only 8.6% – which suggests a swing of 2.1% away from them in Brisbane since the last election. I find that hard to believe. In terms of its impact on seats won or lost – assuming a 50% exhaust rate and a 70/30 pref flow to Labor, the difference between an average Green vote of 10.7% at the last election and 8.6% this election is to effectively knock one half a percent from the ALP TPP in Brisbane, which tonight will probably equate to 3 or 4 seats.


Comments (24) | Permalink

Revenge of the sub-editors

Friday, March 20, 2009 – 11:20 pm, by Mark Bahnisch

Andrew Bartlett’s review of the Queensland campaign at New Matilda is graced (?) by this tagline:

Andrew Bartlett was sceptical that the Libs and Nats could even pull off a merger, let alone win an election as the newly formed Liberal National Party. It looks like he got it very wrong.

I’m not sure I’d be very happy if that para appeared beneath my byline. But perhaps there’s a reasonable query lurking in that text, one that was also implied on an LP comments thread:

A considerable swing was obviously on from months out from the election. It is widespread and (i) the LNP merger and (ii) the calling of an early election consolidated it. Labor knew this. It’s not clear why so many commentators ignored the Peter Brent school of electoral analysis and assumed Labor would win despite its very tired look, or assumed that Queenslanders were old enough to see politics only through the lens of the Joh years.

That might be making the case a tad too strongly, but I think it is reasonable to ask why so many commentators assumed before the campaign began that Labor would be returned with no great difficulty, albeit with a reduced majority. For my part, I think there’s a simple answer. Most commentators, myself included:

(a) overestimated the competence of the ALP machine;

(b) overestimated the bar Lawrence Springborg had to clear to be a (just) credible alternative;

(c) underestimated the degree to which fracture lines inherent in the LNP could be papered over – particularly when the ALP hardly got the wedge out.

That Labor was on the nose from before Peter Beattie’s last victory, and only saved by a combination of the Coalition implosion and a superb campaign is not, I think, in any doubt. Both factors have been absent this year, as have Beattie’s political skills.

Elsewhere: John Quiggin:

I’m not at all impressed by the apparent quality of the alternative government, but it’s more convincing as a united party than as the chaotic rabble that went under the name “Coalition”. The Bligh government has some reasonably strong performers (Bligh herself and Paul Lucas for example), but there have been plenty of duds or worse. And the involvement of proven disasters like Mike Kaiser in the campaign was a big mistake.

Comments (7) | Permalink

The Official Crikey Election Tipping Competition!

Friday, March 20, 2009 – 5:14 pm, by Possum Comitatus

OK Sportsfans, here it is – the official Crikey Qld election tipping comp. The winner will not only receive the resentful opprobrium of the losers, not only will they receive bragging rights to belt other wannabe psephologists around the head with until the next election, not only will they receive a lucky dip of products from the bowels of the Crikey merchandise empire (rumored by certain editorial powers to be First Dog On The Moon T-Shirts), but the winner will also receive either:

a) a free Crikey sub for you slackers out there that haven’t subscribed

or

b) for you responsible folks that have, the gift that keeps on giving – a free Crikey sub for that someone special.

The rules are simple and must be OBEYED. The competition requires three answers.

1. The number of seats the ALP will win

 2. The number of seats the LNP will win

3. Pauline Hanson’s primary vote to two decimal places.

The way it works is by process of elimination – if there are people tied after correctly predicting the number of ALP seats won, they then move on to the LNP round. If there are still people that predicted both the ALP and LNP seat numbers, they next move on to the Hanson round where the person closest to predicting Hanson’s ultimate official primary vote to two decimal places wins. If there’s still a tie after that, William, Mark and I will come up with some excruciatingly painful and humiliating tie breaker which we’ll broadcast to the world  for our own cruel enjoyment.

I’ll notify the winner by email once the final official results are delivered.

There are two other very important rules that MUST BE OBEYED.

1. Only 1 prediction is allowed per person – it cannot be updated. Once you make it, you’re stuck with it.

2. The competition closes Saturday 21st March at 3.00pm QLD time (that’s 4.00pm to you Mexicans and whatever that is for you South Australians and Sandgropers with your funny clocks). Exit poll results have generally started to leak after about 4.00pm, so to stop smarty pants latecomers with good contacts from stealing the show, we’ll shut down the comp before the exit polls get a wide enough sample to be utilised.

So, let the predictions flow. ALP seats won, followed by LNP seats won followed by the ultimate tie breaker – Hanson’s primary vote to 2 decimal places.

UPDATE:

Thread Closed – we’ll announce the winner when the results are official.

Comments (135) | Permalink

Queensland Election Preview Liveblog: chat with Crikey

Friday, March 20, 2009 – 1:40 pm, by Mark Bahnisch

Join William Bowe, Possum and Crikey political correspondent Bernard Keane from 2pm today for a lively chat about tomorrow’s election.

Comments (0) | Permalink

Borg can’t be sold to Brisbane; nor Bligh to Beattie’s battlers

Friday, March 20, 2009 – 1:30 pm, by Mark Bahnisch

From today’s Crikey email:

I don’t often agree with former ALP Senator John Black’s political analysis – he places too much weight on demographics for my liking – but I did nod my head vigorously when reading his piece on the Queensland election in yesterday’s Financial Review.

Black argued that the same cohort of voters who Peter Beattie held through three successive landslides, who swung to Kevin Rudd and the federal ALP in 2007 after a long incarnation as “Howard’s battlers”, are in danger of deserting state Labor tomorrow. These voters are on low to middle incomes, socially conservative and worried about services and jobs. Many of them are the same voters, or coming from the same place, as the folk who gave One Nation 25% of the state vote in 1998. They’re also the same voters who swung some regional seats further in Labor’s direction in 2006 out of anger at WorkChoices.

While there’s often been an assumption in speculation about this campaign that Lawrence Springborg couldn’t easily be sold in Brisbane, few seem to have asked – until the swing became obvious – whether Anna Bligh could hold onto Beattie’s vote on the urban fringes and in the regions. Under Beattie, the very model of a blokey Queensland populist, a complex backflip could be executed where he surrounded himself with urbane Ministers – think Matt Foley and one Anna Bligh, talked up biotech and creative industries and cosmopolitanism, but still projected down home Canberra hating parochialism and praised Joh to the skies on well timed occasions. But all the hard hats in the world haven’t dispelled Bligh’s urbane image.

There’s more to the puzzle than this. Lawrence Springborg is no Barnaby Joyce, and doesn’t bang the populist drum like Nationals past. He’s more attuned to the monotonous tone of moderation. So neither leader has really convinced – anywhere much.

The other side of the coin is that a big chunk of the urban base Labor relied on is fleeing at a frightening rate of knots – parked with the Greens, not so much out of conviction, but out of disappointment at a Labor party that appears directionless and managerialist yet without the competence. Two party preferred votes are almost meaningless in Queensland’s optional preferential system. The key is the primaries. If there’s a high rate of exhaustion, and Labor is far enough behind in a lot of city seats, all the greenwash and social policy stuff from the last week won’t lure those voters back. They’re disengaged and not in a mood to be wooed. Outside Brisbane, it looks like a wasteland.

As I’ve been arguing throughout the campaign, neither party has succeeded in reviving Beattie’s state wide appeal, or even looked convincing talking of “plans” which appear too nakedly and opportunistically targeted to particular regions and interest groups. But while the campaign has been boredom incarnated, the electorate is exceptionally unpredictable. The outcome could be anything from a very narrow Labor win through a hung parliament to a small LNP majority. The latter appears most probable, but really we’re all guessing – locals wouldn’t make the same mistake as Peter Van Onselen and write of a swag of Labor seats in “southwest Queensland” — but all the same this is one of the most difficult of recent elections to read.

Aside from the obvious early election angle, there will be a motto here for the Rudd government. It’s this – sooner or later, maybe sooner when the economy’s in strife, it becomes impossible to hold together a coalition of urban denizens and outer-urban and regional voters. You need quite exceptional political skills – a la Beattie – to pull it off. You can’t do it with the political strategist’s soundbite laden focus grouped book of oh so clever tricks.

Cross-posted at Larvatus Prodeo.

Comments (27) | Permalink

Vote 1: Dave Zwolenski

Friday, March 20, 2009 – 1:29 pm, by Possum Comitatus

Most of you have probably seen or heard about this already, but for those that haven’t – campaigning via brutal honesty.

YouTube Preview Image
Comments (0) | Permalink

Galaxy Part 2

Friday, March 20, 2009 – 8:37 am, by Possum Comitatus

Another Galaxy Poll and it’s just like the last 3 – a remarkable consistency for an unremarkable election campaign. It comes off a sample of 800 for an MoE or around the 3.5% mark

galaxyfinal

There’s no ‘Better Premier’ beauty contest result this morning, but there’s a few other things worth looking at. First up, the question on campaign performance.

campperformance

Next we have that brave Galaxy question where they put “deserve” in the same sentence as politicians – with predictable results.

alpwin

lnpwin

The cynicism would appear to be rather bipartisan on all fronts in statistically equal measures.It reminds me of the intro to Borgs primary 30 second ad where he said “people are telling me they deserve better“. About 57% of people by the looks of it.

The only regular Galaxy question which has undergone any real movement over the campaign is on win expectations.The question asks:

In the State election on Saturday, which of the two major parties do you expect to win the most seats, the ALP or the Liberal National Party?

winexpall

partywinexps

The other part of the poll was a breakdown of the primary vote in terms of Brisbane seats vs. non-Brisbane seats – but with small subsamples from an already smallish total sample of 800, it adds far too much uncertainty to even bother with. For instance, the MoE on a sample of 400 is 4.9% which makes the results pretty much meaningless in an election that appears to be close.

Update:(Possum)

For a bit of fun, turn primaries and exhaustion rates into TPP results. Just click the “click to edit” button below and enter the primary votes for each of the parties (except Others as that’s calculated automatically).

Next to “Exhaustion Rates”, enter the proportion of votes for the Greens and Others that you think will exhaust e.g. if you think 60% of Greens voters will “Just Vote 1″, then give that a value of 0.6.

Then enter the preference flow to the ALP for Greens and Others. So if you think 80% of Greens preferences that actually flow will go to the ALP then enter 0.8.

Once you’ve done that, it will calculate the Two Party Preferred.

Comments (2) | Permalink

Campaign updates: election minus 1 day

Friday, March 20, 2009 – 3:57 am, by William Bowe

As our clapped-out nags go neck-and-neck into the home strait:

Pumicestone (Labor 5.4%), Cleveland (Labor 1.3%) and Redcliffe (Labor 6.0%): Steven Wardill of the Courier-Mail reports that “Labor sources believe the recent oil spill has cost the party crucial support” in these coastal metropolitan seats. Highlight of LNP 2009 Election Financial Commitments: “Bee Gees Redcliffe Fire Station Tourist Attraction”. Projected expenditure over four years: $0. How deep is their love?

Toowoomba North (Labor 7.6%): The LNP has sought to make hay from Attorney-General Kerry Shine’s portfolio area by promising to boost councils’ powers to veto brothel applications, which is apparently an issue of particular local concern. Local newspaper The Chronicle has helpfully outlined the parties’ promises relevant to the electorate. Labor: $1 million for a birth centre at Toowoomba Hospital; new $1.7 million kindergarten at Fairview Heights State School; $2 million upgrade for Toowoomba Hospital’s emergency department; $3.78 million to upgrade the Toowoomba Young Women’s Christian Association; $3.23 million to build 12 two-bedroom social housing units at Newtown; $2.5 million to secure land for a new high school at Highfields. LNP: “Protection of farming areas like Felton and Haystack Plain from mining”; a maternity centre (one of 12 across regional Queensland at a cost of $27 million); $5 million to extend the Wilsonton campus of Toowoomba State High School to years 11 and 12; $2 million to secure land for a new high school at Highfields.

Beaudesert (Nationals 5.9%): The Courier-Mail reports “a late rush of bets” for Pauline Hanson, with SportingBet cutting her from $6.50 to $3.50. The LNP’s Aidan McLindon remains firm favourite at $1.65, although this is slightly higher than the $1.40 and $1.45 being offered by Centrebet and Sports Acumen a week ago.

Comments (47) | Permalink