Mixed messages

As the election campaign in Queensland gets under way, a lot is being made of Saturday’s opinion poll in the Courier-Mail showing the Coalition leading 51-49 on two-party preferred. This is partly because it is the nearest poll to hand, and partly because its finding that the government is in danger makes good copy. However, there are a number of reasons to be cautious. The poll seems to be the first Queensland state survey ever conducted by Galaxy Research, whose emergence during the 2004 election campaign was noted here. Given Queensland’s complicated electoral terrain, it seems reasonable to suggest that a new agency would face a steep learning curve. Secondly, unless there’s something in the print version that I’ve missed, the Courier-Mail provides no indication of the sample size (UPDATE: Gary Bruce in comments notes that the sample was 800, as reported here). Presumably it would be in the ballpark of the surveys conducted for the paper until recently by TNS, which polled about 500 people – well below half the samples used by Newspoll and Roy Morgan. Last but not least, the result is at odds with every poll published in the current term, except for the TNS/Courier-Mail poll published on February 3.

August 12 (Galaxy) 42 43 49 51
June 3 (TNS) 40 38 52 48
February 3 (TNS) 42 43 49 51

Newspoll and Roy Morgan tell different stories, both from the Courier-Mail and each other. Newspoll’s has the greater ring of truth. Despite the fairly constant pace of upheaval during the Beattie government’s past term, it tells of a game of two halves – of a government maintaining its supremacy in the first half and facing a stiff challenge in the second. The obvious dividing line is the politically calamitous Dr Death scandal, which erupted early in April 2005.

The numbers indicate the timing of 1. the "Winegate" episode involving Indigenous Affairs Minister Liddy Clark, 2. Clark’s resignation as minister, 3. the emergence of the Dr Death scandal, 4. Labor’s defeat in the Chatsworth and Redcliffe by-elections, 5. Gordon Nuttall’s resignation as Health Minister, 6. Labor’s defeat in the Gaven by-election, 7. the Mary River dam announcement, and 8. the failure of the proposed merger of the Coalition parties.

Roy Morgan, whose most recent poll came out yesterday, paints a rosy picture for Labor even by its own world-famous standards. Indeed, every single poll it has conducted since the 2004 election – and they have come at regular bi-monthly intervals, with samples of around 1200 – has shown Labor set to increase its primary vote, and the Coalition parties struggling to break even.

At best, a very mild Dr Death effect can be determined through a narrowing of the two-marty margin from October 2005, but even Labor’s worst result had them in a better position than at the 2004 election. However, care should be taken even with plausible-sounding two-party figures for Queensland, which has optional preferential voting and a tradition of strongly performing independents and minor parties.

Roy Morgan at least has the advantage of being up-to-date. Events since Newspoll’s last survey include the state budget, the Toowoomba recycled water referendum, Beattie assuming the role of Water Minister and Bruce Flegg’s rise to the Liberal leadership – not to mention the first day of campaigning which, as a quick look at The Australian makes clear, has not gone well for the Coalition. Among the articles featured is an assessment from Malcolm Mackerras that Beattie will survive with a five-seat majority.

Author: William Bowe

William Bowe is a Perth-based election analyst and occasional teacher of political science. His blog, The Poll Bludger, has existed in one form or another since 2004, and is one of the most heavily trafficked websites on Australian politics.

7 comments on “Mixed messages”

  1. Ah, the sweet smell of elections is in the air!

    Not being a Queenslander I find it interesting to gaze into the state’s political pantomime from beyond. The state which brought us One Nation and is scattered with distrait independents and rugged Nationals types will always be worth watching when it comes to ballots, exponentially more interesting that the grey political characters in NSW and Vic.

    I liked Mackerras’ article in the Australian this morning. It was crisp and clear and set out his early predictions with decent explanations for his decisions. Like Mackerras I reckon Nangano will fall back to the Nationals on a huge swing, simply because of the Belke-Petersen name, while Gympie is an obvious return to the National camp.

    In relation to Mackerras’ predictions I’d like to ask why in The Australian he mentioned that the Libs would pick up Springwood. With a 9.7% Labor majority it is well beyond the Coalition hopes in Mundingburra (6.2%), Pumicestone (5.4%) and Townsville (5.3%). Even the over-inflated margins in Toowoomba North (7.3%) and Glasshouse (8.9) should be wiped well before Springwood … or are there localised factors that I’m missing?

    One other early prediction. I reckon Libs will hold Redcliff without any bother but will really struggle with Chatsworth. As a Liberal supporter but a detestor of the factions I wouldn’t mind seeing ‘kingpin’ Michael Caltabiano fall, but if he does it’ll be by only a couple of hundred votes.

  2. I think Malcolm Mackerras is way off in his predictions that the Liberals will have 17 seats (up from their current 7).

    Of the Liberal gains I predict: Kawana, Mudgeeraba and Clayfield would be easy wins for the Liberals.

    But Indooroopilly, Broadwater and Burleigh are all very much still lineball. I wouldn’t write Labor off in holding any of these because of good local MPs.

    Then you have the third layer: Aspley, Barron River, Cairns, Hervey Bay and Southport – which all still seem FAR more Labor than Liberal.

    I don’t think the Liberals have ANY chance at all of gaining Mt Ommaney, Greenslopes, Mansfield, Cleveland, Springwood or Pumicestone.

    Infact it is not outside the realm of possibility that the Liberals’ themselves could lose either Chatsworth or Redcliffe to Labor. Neither is Noosa a foregone Liberal gain.

    So based on all this – I would say the Liberals will gain maybe 4 or 5 seats at the most.

    Mackerras prediction of a Liberal gain of 10 seats, seems quite extreme if unbelievable.

    I do predict Beattie Labor will see a moderate swing against it. But with an opposition who can’t even decide who will be premier – overall I think it would be far more realistic to assume a Labor loss of only perhaps 6-8 seats (almost all of them in the regions).

  3. Regarding Nanango – “John Bjelke-Petersen” (the son) already ran for parliament back in the 1996 federal election (in the seat of Fisher for the National party).

    He totally bombed. He finished 3rd place behind the Liberals and Labor with only 20% of the vote.

    So don’t assume that the Nationals will win Nanago automatically. (The sitting Independent MP has 62% of the 2PP vote).

  4. I think Mackerras has it about right…Labor looks vulnerable this time and there should be some bigs swings..even a repeat of 1995.

    But I cant see any way that the Libs can hold Chatsworth. This was a comfortable Labor seat from 1977 and they will get it back – even if they lose the election.

    I’d also watch Redcliffe & Currumbin. Expect the Libs to win both but I’m not sure either are a foregone conclusion. Currumbin is the most Labor of the Gold Coast seats and swung 18% in 92 because of the backlash against Merri Rose..if the Libs hold it it could be because of the personal vote of the new MP.

    I’d disagree with Malcolm about Keppel, I think the Nats have more chance in either Mulgrave or Toowoomba North but thats just a hunch.

    As for Springwood, the Libs could just as easily get Mansfield, Mt Gravatt or Mt Ommaney.. I think his idea is that there should be a large correction in these once Liberal southern seats and while not enough to topple them, on the balance of probabilities it should be enough to prevail somewhere.

  5. Does a large majority protect against defeat?

    In theory, no. In practice, yes.

    It says here in my spreadsheet of about 200 Australian elections since 1901, The larger the majority, the bigger the swing against- but no party with more than 68% of the seats has ever been defeated.

    Cute. Peter Beatty has 67.4% of the seats.

  6. I think that the ‘predictions’ of what seats will fall to the Liberals is far more complicated.

    There will be some seats that are expected to fall like Indooroopilly that will not. Ronan Lee is a very solid candidate who has done about as much as he can to hold the seat. He also has a very high profile in the community in combination with a unique set of local issues that do not necessarily match with ‘wider queensland’. Clayfield may also fall into this category with Liddy Clarke a popular member against Local Councillor Tim Nicholls. The problem for Nicholls is a) Liddy Clarke is popular and b) a Nicholls win will mean the voters of Clayfield go to another election in the next month or two.

    Seats that are not expected or highly unlikely to fall like Springwood and Mt Ommaney may actually come through despite more illustrious seats not. The Springwood Liberal candidate is a very popular local councillor who is held in very high esteem by the local community. Barbara Stone lacks profile, has accomplished little for her electorate and had no oppostion at the last election. This means the margin in this seat and Mt Ommaney where the same applies gives people a ‘false sense of security’. Mt Ommaney is also a high chance with the Liberal candidate very very high profile- his campaign has been running for many many months now, he has had a lot of local coverage through the local rags and letters to voters. Julie Attwood has been non-existant, lazy and seems to be waiting out her three term super-payout moreso than doing anything for her electorate.

    Seats such as the above are examples of localised voting trends that I expect to see happen at the upcoming election.

Comments are closed.