Idle Speculation: Bryan overboard edition

After a gruelling two-and-a-half years refereeing Australia’s rowdiest non-parliamentary slanging match, Bryan Palmer has exited the battlefield. That just leaves one psephological punter mug enough to operate a comments facility. For anyone out there who wants to argue the toss about recent opinion polls, the following have come through since the last episode of Idle Speculation two very long weeks ago:

Roy Morgan yesterday released a survey of 556 voters which has Labor at 46 per cent on the primary vote and 55 per cent on two-party preferred. These are the softest figures Morgan has produced for Labor since January. There’s also a bunch of attitudinal guff that has the Prime Minister doing quite a bit worse on every measure than when these questions were last asked a month before the 2004 election. A possibly interesting exception is, “Who would be best for minimising the tax you pay?”

• Speaking of attitudinal guff, Galaxy‘s venture on to this turf last week raised a few eyebrows, including those of polling mavens Irving Saulwick and Denis Muller. Writing in Crikey on Tuesday, Saulwick and Muller described the wording of one of Galaxy’s questions – “Which of the following concerns, if any, do you have about the prospect of Kevin Rudd and the Labor Party running the Australian economy?” – as a “disgrace”. Nonetheless, there does not seem any particular reason to be dubious about its finding that Labor had very modest leads of 1 per cent on the primary vote and 4 per cent on two-party preferred.

• Monday brought another poll from Galaxy, this time of 800 voters in Queensland. It showed that Labor’s lead had narrowed from 7 per cent to 1 per cent since its last such poll in February, and from 55-45 to 52-48 on two-party preferred.

• A Westpoll survey of 408 voters in Saturday’s West Australian produced a very encouraging result for the Coalition, who led 50 per cent to 38 per cent on the primary vote and 56.3-43.7 on two-party preferred. That amounts to a 0.9 per cent swing to the Coalition from 2004.

• Since Westpoll escapes notice on the eastern seaboard, I might also note Tuesday’s poll of state voting intention, which I gather was from the same sample as the federal poll. It has Labor up to 42 per cent from 41 per cent in April, the Coalition up from 38 per cent to 40 per cent, and Labor’s two-party lead narrowing from 54.5-45.5 to 52.3-47.7.

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.

40 comments on “Idle Speculation: Bryan overboard edition”

  1. I’ll argue the toss. Kevin Rudd has set the agenda since the start of the year, which gave him a commanding lead. He has not finished setting the agenda. Labor will maintain its lead until the election. The Liberals have wasted six months in trivial attacks on him and his party, and they have not landed a blow. The polls will move up and down and the future is not set in stone, but I am confident of a Labor victory. The only advice I would give him is to be clear and stop jumping all over the place on IR. He should take my advice, just like Frank McManus should have in 1974.

  2. I agree with Marcus, it is sad about the Oz Politics website.

    As for the polling it has narrowed. The question now is whether it will continue to narrow enough so that the Coalition will win in Nov/Dec or will it stabilise around 45/7 – 53/5 TPP making it too difficult to peg back come the election.

  3. It is sad about OzPol, I signed up to comments this year and was just starting to read the comments and so on. I was only just starting to get into it too.

  4. My understanding from the latest post on OzPolitics is that the blog will go on. Just without commentary.

    In a sense that blog was a victim of its own success, some comments went to the 500 mark, which is phenomenal.

    In regards of polls it is amazing how much they have become news in recent time. Every release is commented and analysed by the media with the same reverence as some mystical revelation.

    As a Labor supporter I do hope that voters will think that Howard is a shoe in the the polls indicate that he will win. I am a Victorian and I remember the 1999 Victorian election 🙂

  5. It’s a pity about OzPolitics but I symathise with Bryan. But why does it need to be moderated, he asked in sweet innocence ? Things did run hot but surely a filter could automatically remove offending language. My view was that if you couldn’t stand the heat, get out of the kitchen. Eventually the slanging matches die out and there were enough new threads starting.

    And I was enjoying it so much too. I logged on to say how hopeless Shrek was on 7.30 report last night and had just missed the close !

    Anyway, I’ll post here about the Howard/McKew stoush if it seems appropriate.

  6. It was getting frustrating to have to wade through Steven Kaye et al, and people endlessly, repetitively rebutting him, to get to the good analysis, of which there was rather a lot. Bryan’s own stuff is top-notch, and I hope that free from the need to moderate pillocks, he will have time to write more.

  7. I’ll reserve my opinion on the Galaxy results till after the next round of polls, next week. However, any organisation that has a conservative cheerleader as it’s head, can’t get the past election figures correct in its analysis, shows a swing of 5% against Labor’s primary vote over a one week period (the other polls all showed Labor around 50% a week or so previously) and produces this figure over the same weekend another poll has Labor on 51% primary vote, surely has to be treated with some caution.

  8. Wow.. no comments at OzPolitics. I started reading the ozpolitics blog back in.. 2003/2004 ? Before the 2004 federal election anyway, so it’s sad to see the comments go. I remember when it only received 0 or single digit comments per thread. Them was the days…

    Ozpolitics was on the verge of overtaking Larvatus Prodeo as #2 popular political blog (Tim Blair #1 in traffic by a country mile).

    However I admit I stopped reading it about six months ago due to the constant partisan bickering.

    Realistically, why do people bother ? Are readers going to suddenly change political philosophy based on some comment area point scoring ?

    Hopefully the worst offenders don’t migrate here.

    I have my speaker’s robes and trusty gavel ready to bring them to order if they do!

    (note: last sentence not endorsed by the bludger, I just made it up).

  9. I would have thought that there would be more comment about some of the concerns surrounding the latest galaxy results in Qld – especially given the comments made by Saulwick – who, i would think, would have some cred. The latest Morgan – which also shows a tightening – is also a bit of a worry (maybe the govt scare campaigns are working). Not being a psephologoist (sp?) – is a sample of just over 500 for the latest Morgan an issue? The rest seem to be around 1,000 generally.

    Thanks for providing a much needed outlet in this amazing election year. I know there’ll be a lot of the old ozpolitics lot joining in – given the fever that has already been generated for election 07

  10. Nature abhors a vacuum and the lack of policy launches means people, media and the blogs have focussed on polls. With the absence of newspoll, everyone has been going mental jumping on any data they can find.

    Hopefully the next newspoll will slake everyone’s thirst and things can get back to normal.

  11. It’s a trifle petulant of Bryan to close down comments on his blog, but hopefully he’ll rethink his position as the election nears. Regarding the polls, it should be obvious to everyone that they’ve narrowed considerably recently and that any future poll that has the Government’s primary vote below 40% and Labor’s above 46% should be automatically discarded. That 6 point range can be easily traversed by the Coalition in time for the election, which they’ll win handily thanks to the unions.

  12. I’ll bite, but only once Steven.

    Take out Galaxy, and where is this miraculous narrowing?

    Go look at Bryan’s graphs – they are generally flat as a tack.

    Galaxy may be an outlier or the harbinger of doom for Rudd, but there is little evidence from any other source to support your statement.

    Od course, we could just do what you suggest, and discard any poll or set of polls unless they get your approval first.

  13. I think William urgently needs to equip himself with the means to suspend or ban people from his blog, either for a fixed period in the sinbin or permanently, otherwise this site will go the same way Bryan’s did. No doubt I would be at risk myself from time to time, although I try not to be provoked by idiots like… oops.

  14. Gary Bruce

    Re Galaxy being a “conservative cheerleader”: if this were so, I’d have thought it would do the opposite and pump up Labor’s support as much as possible to try to maximise the possibility of a protest vote.

    Maybe it’s coincidence, but I’ve noticed Newspoll tends to overstate the support for Labor state governments on the eve of state elections. Newspoll also predicted a much closer contest federally in 2004 than what eventuated, whereas AC Nielsen predicted a bigger winning margin for the Coalition.

  15. Steven Kaye – I’ll bite too. So from now on Steven no other poll counts? Where’s this trend you seem to be referring to? Oh, I know, you’re joking right? Ha. Good one.

  16. On a non federal note, has any meaningful polling been done for state politics? Would be interesting to see if Bracks has been hurt by the Tatersalls scandal and if Baillieu has managed to consolidate further from the good campaigning last year. I’m aware of Omodei being up a certain creek but that’s it in state terms. Anyone have more to add?

  17. BigBob and Gary Bruce – the turnaround in the polls started with the last Nielsen, and has since then been picked up by Newspoll, Galaxy, Westpoll and the most recent Morgan phone poll. I do not count Morgan’s face-to-face polling because I believe that methodology is fundamentally flawed and understates support for the Coalition.

    By the way, today’s Age has an article about the latest Westpoll, which shows the Liberals’ vote strengthening in their seats of Stirling and Hasluck, and ahead in Labor’s Cowan.

  18. From Patterson Research that produces WestPoll

    “For the marginal seat analysis Patterson had to limit the surveys to a sample size of N=200 per seat, which provided a survey error of ?6.5% at the 95% level of confidence, or about ?3.8% at the 70% level of confidence.”

    6.5% margin of error for a marginal seats which could change hands with a 2-3% deviation or even less – its about as useless as tits on a bull.

  19. I think Bryan should have considered expanding into a vBulletin forum, and hired a few moderators. then he could have disconnected the comments on the front page, had a thread in the forums on each thread.

    thus keeping the large traffic flow, but the front page keeps the appearance opf balance he wanted. The volunteer moderators on the forum could have stopped him needing to spend so much time controlling things.

  20. Steven,

    I’m not disputing that there may be a drift back to the Coalition. 58-60% of TPP for the ALP was not sustainable.

    As to a substantial movement – it isn’t anywhere except in Galaxy.

    As I have said, Galaxy may be right or it may be wrong, but there is little other evidence to point to any substantial trend to the government.

    More polls will show the picture.

    And as Stephen Hill points out, Westpoll’s border on the unusable.

  21. Stephen, correct me if I’m wrong here, but I think Patterson is referring to the 200-sample polls of specific marginal seats they often do during election campaigns. This poll had a sample of 400 and would thus have a smaller margin of error. I admit that 400 is still pretty small, but I will make this point: it appears these were the same people they surveyed for this week’s state poll (the sample for each was 408), and there doesn’t seem anything roguish about that one. They’re basically the same figures as for other recent Newspoll and Westpoll surveys.

  22. yup 400 per seat, 1275 all up

    the strange bit was Swan wasn’t polled. Given its one of two labor marginals (the other cowan) why was this not polled as well. ANother installment ?

  23. Latest Morgan here. Has ALP 55:45 TPP, but it’s a very small phone poll with a sample error of around 4.5%. It isn’t directly comparable to the previous Morgan face-to-face, IMO.

  24. Steven Kaye,

    The last Newspoll was 60/40 to ALP. What trend back to government?

    I hope William does not let the partisan comments get out of hand. This has been a great blog without the drivel of partisan comments at Ozpolitics.

  25. IMHO, Steven Kaye earned himself a reputation on ozpolitics of being excessively (ie. unreasonably and repetitively) partisan. I call on him to be much more moderate on this site.

  26. Panadawn asks: I think Bryan should have considered expanding into a vBulletin forum

    I’d be happy to put up a set of forums (though not vBulletin ones, I don’t like vBulletin) if there was sufficient interest, and there were some people prepared to be the moderators. This isn’t really the place to discuss it, so if you’re interested, register on this board, and post your interest in the only board that currently exists. If there’s sufficient interest, I’ll go ahead with it.

    If there’s interest, this board will metamorphose into the live one.

  27. Excessively partisan posts are not to be encouraged, so I agree, Amber Dekstris, with your call for moderation and add that applies to posters of all political persuasions. Being “excessive” usually reflects extremism on one or other side of the political spectrum and does not contribute to an informed discussion on particular matter(s). I will be voting for the party which, in my opinion, is less affected by extremist policy positions on a range of issues including workplace relations, education, the environment and health.

  28. I think you’re right William, I was just going by the limited info from the Age, and from the Patterson Research sight. A 400 people poll would reduce the margin of error by a few percept, mind you marginal seat polls in their nature are going to be more volatile. I can remember at the last election various seats being claimed to have a Labor lead in the early weaks, that actually swung further away from the opposiiton.

  29. Stephen, we were both wrong. I thought we were talking about last week’s 400-sample poll of the entire state. We were in fact talking about this week’s 400-sample polls of three different electorates, which you assumed were 200-sample.

  30. Has anybody here seen my old friend Bryan,
    Can you tell me where his gone, oh,
    He freeded alot of people but it seems the good die young, yeah,
    I just looked around and he was gone.

Comments are closed.