Courtesy of GhostWhoVotes, the full results of Galaxy’s marginal seat polling, which will be published in tomorrow’s News Limited tabloids, can be found here.
Let me make a number of points first, which I’ll illustrate with reference to the Queensland seats:
(a) It’s not “the biggest poll ever” – it’s a series of polls of marginal seats in different states, each with a sample of 800 covering a number of electorates. The aggregate numbers at state and national level would be meaningless.
(b) What’s crucial in interpreting the findings is which seats have been surveyed, and the fact that the sample for each of the 4 seats surveyed in each state is only 200. Additionally, if the seats are geographically and demographically diverse, reading too much into the aggregate of the marginals can be misleading.
Let’s take a look at Queensland, by way of illustration.
Galaxy surveyed 800 voters in Bowman, Dawson, Dickson and Flynn.
Dickson and Bowman are on Brisbane’s northern and southwestern fringes respectively. Dickson is nominally Labor, but with a Coalition incumbent, and Andrew Laming holds Bowman for the LNP, but on the smallest margin in Queensland. Both almost fell to the Kevin07 surge.
Dawson and Flynn are huge Central to North Queensland seats which Labor holds, but with no incumbent MP in Dawson where James Bidgood has stepped down.
As I’ve been saying for about a week or so, Labor is behind in Dickson, where Peter Dutton looks to have tightened his grip. In Bowman, Labor is barely running a campaign, with reports appearing for weeks in the Brisbane Times that their candidate is invisible, and the local papers can’t get hold of her for an interview. In short, these are two seats in outer Brisbane which Labor has little chance of winning, though Dickson is more competitive, and a lively race continues there, probably designed to pin Dutton down as much as anything else.
Dawson is probably the only Queensland seat where the mining tax had a significant impact. Chris Trevor’s seat of Flynn‘s a better shot for Labor, although it’s being contested by one of the few non-incumbent LNP candidate’s who’s much chop. Worth noting, though, that Dawson was the subject of a previous Newspoll survey finding it tied at 50-50.
But, basically, these four seats have little in common as a group, and it’s been repeatedly observed that both the LNP and Labor are agreed (and other evidence confirms) that seats in Queensland are swinging in both directions, and that swings are highly variable not just regionally (as is common) but also seat by seat.
In short, if I had to pick four seats likely to be won by the LNP in Queensland, these would be among the contenders, though I’d expect that the LNP won’t win all of them. I’d place no credence whatsoever in extrapolating the 5.4% 2PP swing against the ALP to the rest of the state, and in any case, I suspect most of it’s coming from Bowman and Dickson.
If Herbert, Leichardt and Brisbane had been thrown into the mix with, say, Dickson, Labor would have probably have been comfortably ahead (because it appears to be in three of these four seats), and we’d have known as little about what will happen in the state as a whole.
But the message would have been quite different.
I won’t do the exercise for the rest of the country, but the logic should be clear.
Update: Antony Green has now written a blog post explaining how the “national” result of this poll of 200 voters in a number of marginal seats has been misreported wildly due to a series of egregious errors.
Update: William Bowe at The Poll Bludger:
Note that for all the vastness of Galaxy’s total national sample, as far as all-important Queensland is concerned the results are less sturdy than yesterday’s Newspoll, which targeted eight Queensland seats rather than four and 1600 respondents rather than 800. That poll produced a swing of 3.4 per cent against Labor compared with Galaxy’s 5.4 per cent, which in uniform swing terms would mean a difference of no fewer than four seats.
Update: I’ve posted an overview of the interpretation of the poll over at The Drumroll, where I’ve also had a look at some of the other questions asked by Galaxy.
Elsewhere: Possum makes an important point about marginal seat polling:
So the aggregated marginal seat polling tells us no more about what is happening in any particular seat than the large national polling breakdowns do – the expected variation in the swing between 4 marginal seats isn’t fundamentally different from the variation we would expect to see from any arbitrarily larger number of aggregated seats at the state level, be they marginal or not. Essentially, these marginal seat polls just verify what we already know, rather than telling us anything particularly new.
The 2007 figure quoted is 50.9% the latest 48.5 for Labor in Queensland. Sample size was 800. Margin of error is 3.4% and the decrease comes to 2.4%.
Possum’s poll cruncher comes in very handy in times like these.
http://blogs.crikey.com.au/pollytics/thepollcruncher/
There are some other ‘strange’ elements in the whole set ie Qld plus the other states.
Gillard is almost unanimously preferred as PM.
The Rudd impact is minimal.
Most people, by a large margin about 2:1 do not believe Abbott will keep his election promises.
Most people, by a small but significant margin, reckon they are better off now compared to 3 years ago.
The only element that appears to be working against the ALP is the impression of ‘disunity’.
Quite frankly the media can take the credit for that.
It has managed to paper over the gaping cracks in the COALition, the Minchin takeover, the 41:40 party room split, the climate believers vs the deniers, the considerable disquiet that the Nats have with the Libs over broadband and paid parntal leave and focus on Rudd.
Which is now of minimal impact.
I’m sceptical about this poll and about the timing of its release.
these polls findings have the same ring as 14 labor candidates are still holding govn sponsored positions aka Robb and thus should be disendorsed a last ditch effort to save a doomed campaign hold the faith labor to gain 7 seats my prediction same as a few weeks ago
If the way this poll is spun results in a perception that Labor is an underdog going into the last week, that’s a good thing for the ALP and bad for the Coalition. “Momentum” is a media fiction.
“I won’t do the exercise for the rest of the country, but the logic should be clear.”
What’s this logic thing you’ve mentioned? Surely it’s a bit late in the campaign to start bringing that into play?
Andrew – heh! But the results, I’d argue, are somewhat independent of the way the campaign is reported and analysed.
SA’s in particular appear skewed. I wouldn’t consider Boothby or Grey to be marginal at all, especially when Hindmarsh and Adelaide are left off the list, while Sturt was I think an unusual case in the 2007 election that was always going to bounce back once Howard was no longer a factor. Different breed of Lib in the eastern suburbs of Adelaide.
WA – Cowan sounds a bit like Bowman, demographic change up in Wanneroo had taken it well out of Labor’s reach anyway and it’s probably the most affected seat by the mining tax as a lot of fly-in-fly-outers live there. I’d have actually surveyed Canning instead, which has flip-flopped between parties about as often as Stirling has. But not too bad I guess when one is only talking one seat.
VIC – Actually seems about right.
NSW – Macarthur’s been Lib-held for a very long time, and the boundaries of Macquarie would in any usual case be a Lib seat; it’s gained bits from Lindsay and Greenway which have typically been Lib areas.
SkyNews is promoting this as the ‘Sunday Telegraph SuperPoll’, not as a Galaxy Poll—despite the fact Galaxy means more to a national TV audience than the Terrorgraph does.
That pretty much sums it up.
Galaxy doesn’t want to be associated too closely with this ‘research’.
Hannah’s Dad: net +8% of QLDers saying ‘more likely to vote ALP if Rudd has high profile’ is not ‘minimal’!! It might just have saved this govt its hide.
And note that its +6 in SA, and +5 in NSW, and +4 elsewhere.
In QLD – he remains preferred PM over Gillard by a country mile. This hasnt been MSM spin: apparently, QLDers really do feel that way.
Any Quincelander here will not be surprised: Im one, and I can assure you no one does parochialism quite like it, (though maybe WA is close). This shit is for real!
Overall, where’s the kapow? Coalition ahead in QLD selected marginals, a bit ahead in NSW selected marginals, and behind in VIC. Tell me something I dont know.
@7 – It might have been more useful, Andrew, to have surveyed only seats with Labor incumbents in NSW, Qld and WA.
@8 – You could well be right, Nickws. The seats to be polled might have been picked by the editor, not by Galaxy.
I’m with you on this – at least for Qld it seems like a funny bunch of seats. People on Twitter seem to be suggesting that the Sunday Tele will be making a fearless prediction of a Coalition win on the basis of this polling. Make of that what you will, I guess.
@12 – May well be so, Jason. But, like I said, not a bad thing for Labor to appear to be the underdog.
Oh Dear, the poor old Curious Snail has fallen for its own propaganda.
http://www.couriermail.com.au/news/national/huge-swing-against-labor-puts-tony-abbott-on-bring-of-stunning-election-win/story-fn5z3z83-1225905307127
@14 – that story is complete rubbish!
Ps – @9 – Lefty E, I think there’s no question at all that Kevin’s coming on board has been a big help to the ALP in Queensland.
Antony Green says there is serious problem with the figures and the Labor figure is really 51%. Egg on faces is Antony’s tip.
http://www.abc.net.au/elections/federal/2010/
No, nothing like a prediction along those lines to focus voters’ minds on the propsect of an Abbott government.
There will be a lot of noise in the next week. Your post does well to caution against jumping at shadows, in either direction.
If you believe in Polls then you believe that there a very few people living in Australia who are allowed to vote.
Also if you believe The Candidates especially the sitting members who seems to come out of the woodwork in the lead up to an Election, well keep their promise then you believe in giving false hopes to yourself and others within the community.
Hi Mark,
I agree with your post and offer a few additional comments. Flynn and Dawson have both been bit with CPRS fear and mining tax fear. Both – even more so Flynn – are mining dependent.
Half the survey being from Flynn and Dawson was always going to warp the result. The remaining two having incumbent Liberals also warps the mix.
At the moment, it would appear from Newspoll that Forde (3.4%), Petrie (4.2%) and Bonner (4.5%) are scraping through, but Leichhardt (4.1%) and Brisbane (3.8%) are at risk.
Alas, a uniform 5.4% swing is, to borrow Tony Abbott’s climate change statement, crap.
Another nail in Galaxy’s credibility.
LOL, what a classic Curious Snail beat up. Arch Bevis didnt even lose in 1996; he aint going anywhere.
Meanwhile, I’m getting my first exposure to round the clock LNP attack ads while watching James Bond on channel 7 – they’re really bad.
@18 – It’d be worth following Possum’s lead, Jason, and tuning out of the media over the last week, I suspect!
Its also worth noting that Galaxy has a 4-election track record of underestimating ALPs primary by 1%-1.5%.
You can practically bank on it.
And here’s antony @ Poll Bludger:
All that said – doesnt hurt at all for punters to be hearing its really tight.
Here’s more Antony:
http://blogs.crikey.com.au/pollbludger/2010/08/14/morgan-phone-poll-51-49-to-labor/comment-page-14/#comment-569388
Lefty E, Arch is in a tight spot. It’s more to do with the redistribution than this variable polling. He has been thrown into some tiger country…
@22 – Worth observing, too, that this one overlaps with the Nielsen 53-47 poll.
I still don’t think the national or state 2PPs have any meaning at all, because the marginals surveyed are not representative samples of either the states or the nation.
But if News Limited want to run their “Coalition in stunning victory” rubbish, good luck to them. Probably do a good job in consolidating Labor’s win.
Ok Stuart – didnt know about the redistribution.
My bet: Arch will hang on.
Yep he’s picked up Hamilton and Ascot and lost some good Labor areas. Margin is still 4.6 % though and I don’t think he will be too much troubled.
Update: Antony Green has now written a blog post explaining how the “national” result of this poll of 200 voters in a number of marginal seats has been misreported wildly due to a series of egregious errors.
There goes Galaxy’s credibility. Gone!
Who’d have thought the first casualty of the election would be a pollster?
On the News Ltd treatment, worth having a look at the front page of the last Sunday Telegraph before the election in 2007 (Thank you Malcolm Farnsworth.)
While you’re there, have a good chuckle at the sight of Glenn Milne as political editor. Those were the days.
Oh, and that story in the top right corner may also interest.
Note that the Queensland part of the poll, which provides 10 of the 17 seats projected as Labor losses, covers half as many seats and half as many respondents as yesterday’s Queensland marginals poll from Newspoll, which had the swing 2% lower. In uniform terms, that would make a difference of four seats.
Which ones? I’ve got them 0.7% too high in SA 2010, 1.2% low in Queensland 2009, 3.2% high in WA 2008, 0.9% low in federal 2007, 1.0% high in NSW 2007, 1.1% low in Victoria 2006, 1.1% high in Queensland 2006, 1.4% high in federal 2004.
Note that the dud national vote figure doesn’t appear in the PDF for the poll, which is presumably what Galaxy provided. I would say the 51.4% figure was knocked together by News Ltd editors who decided they needed a national figure for the headline, and did it the stupid way.
Will probably be good PR for Labor in QLD – might sway a few back from the Greens. Hope not. The poll is just the media raising uncertainty under the assumption that people are more interested in the media when needing resolution. I wouldn’t read the Courier Mail to find out what’s on TV.
Thewetmale:
Milne has columns published by their ABC these days.
The Sunday Mail stuff is completely delusional. Talk about projecting the wishes of your aging readership onto the data.
It sounds like it was written by the guy who led th punch up on Morayfield Road on Wyatt Roy’s behalf:
http://www.couriermail.com.au/news/sunday-mail/campaign-violence-caught-on-video-as-wyatt-roy-talked-on-his-phone-in-the-background/story-e6frep2f-1225905296726
Steve@14 says:
‘Oh Dear, the poor old Curious Snail has fallen for its own propaganda.
http://www.couriermail.com.au/news/national/huge-swing-against-labor-puts-tony-abbott-on-bring-of-stunning-election-win/story-fn5z3z83-1225905307127‘
The typo in its link makes me think the poll is as valid as their proofreading.
The CEO of Gallaxy is on ABC Newsradio agreeing with the simplest interpretation, the one that gives Abbott the keys to the Lodge.
Anthony Green made a point in an interview about two weeks ago where he said that the newspapers are under pressure as there are so many other news sources these days, but the one thing they still have a virtual monopoly on is polling.
Is this what we’re seeing I wonder?
The credibility or otherwise of a poll is only of interest to psephological tragics such as hang out here. The rest of the population only reads the headlines and bits of articles. We’re stuffed. [checks passport]
@Helen – I don’t see how the poll means much – How much does wanting to be on the winning team influence voting, and if it does why does every party want to be the underdog?
Polling showing a Tony Abbott win keep Greens voters focused on their second preferences. Evidence that the Greens are flocking back to Labor has been one of the untold stories of the last fortnight. Now that Labor will always have to fight elections on two fronts, this becomes an important part of its strategy.
Can’t we yawn at the latest attempt at tabloid push polling.
Some have questioned the methodology and one wonders in how the questions fed torespondentswas framed.
Oh no, now Abbot has turned up in “Backsliders”; (hits mute).
It’s probably true that the one thing the labor leadership would loath even beyond a coalition victory would be a qualitive improvement in Greens representation in our parliamentary houses.
Update: William Bowe at The Poll Bludger:
As a betting man I am keeping my money in my pocket.
Update: I’ve posted an overview of the interpretation of the poll over at The Drumroll, where I’ve also had a look at some of the other questions asked by Galaxy.
I cant believe the number of people buying into this guff 10 minutes after the Nielsen 53-47.
Calm down! This 3-card trick from Galaxy and News isnt going to do anything but form up swingers against Abbott.
Remember it actually show ALP ahead 51-49 nationally – its just that they are misrepresenting it. deliberately, id say at this point.
Have at look at the front page of the OZ – yhou’ll see two stories: http://www.theaustralian.com.au/
Poll show swings to coalittion in marginals
ALP fights backs in marginals
The latter being Newspoll.
The ALP are odds on to win.
After the Galaxy fiasco of the Rooty Hill debate stacked with Liberal Party shills I wouldn’t use a Galaxy Poll for toilet paper.
@51 – I don’t think it’s the actual results that are the problem – but the selection of seats to poll, and the dodgy headline numbers, which as William points out, are down to News Limited not Galaxy.
William at 34: I was shamelessly referencing this unsourced comment on PB! http://blogs.crikey.com.au/pollbludger/2010/08/14/morgan-phone-poll-51-49-to-labor/all-comments/#comment-569309
Of course, the issue here is the 2PP headline figure. It does look like a News rather than Galaxy job.
It wouldn’t surprise me to find the cold hand of Murdoch in these poll results. He has form as I’ve noted before.
I read the comments on PB about the sample size from each electorate and the commenters reckoned it would only have been about 200 people polled per seat which gives a very high MOE. Some speculated it could be as high as 10%, 5% at the least.
I don’t know enough about how these things are calculated, but even to my untrained eye, 200 seems an awfully small number to sample in electorates the size of the marginals in 2 of the most populous states in the country.
Antony Green and Andrew Catsaras have questioned the results.
Hi Jane
Proportional marginal error is 1 divided by the square root of the sample size.
e.g. sample size 100, square root of 100 is 10; 1 divided by 10 is 0.1. So the proportional error is 10%. So you might say, “the sampled result was 48% but I need to keep in mind that the true figure could be as low as 38% and as high as 58%”. [i.e. 48% +/- 10%]
Now with sample size 200, the square root of 200 is approximately 14.14, and 1 divided by that is about 0.07071,
so there’s a margin of error of about 7.1%.
So in the example above, you’d say: “The sampled result is 48%, but the true figure could be as low as 41% and as high as 55%.”
A wide range of possibilities.
And the “margin of error” itself is just an estimate.
At least you can see why the results in this poll must be treated with caution (to say the least)!
Antony Green provides a proper analysis of this poll(s) on his blog, which has been highlighted on this thread.
One blogger’s response takes the cake
(Arthur): Whether or not your analysis is correct in this instance , I don’t care. Watching you comment on elections over the years, it is so obvious you are anti-Coalition, it makes me wonder why some part of my salary still goes to pay yours.
(Antony’s)COMMENT: I’m intrigued that you don’t care whether my analysis is correct or not. I’ve just used the part of your salary to publish your comment. And it’s my day off!
Say it all really.
Say it all really…
Should be Says it all really.
So it couldn’t be that the voters are sick and tired of the absolute incompetence of the ALP?
Gillard, from the same people who brought you the NSW government.
Thanks, Ambigulous @55.
Second that, ETR @57.
“the cold hand of Murdoch in these poll results”
Yep, its a three card trick: release op-eds backing JG (except in QLD_ which, incidentally, all have the same verbatim paragraphs in them – and then deliberately misrepresent the Galaxy polls figures across all News Ltd papers.
I think its quite clearly orchestrated to give a false impression of momentum.
Do it with Galaxy – and you cant be blamed.
I predict a 53-47 newspoll tonight to put pay to today’s nonsense.
Elsewhere: Possum makes an important point about marginal seat polling:
Speaking of News Ltd the Courier Mail & Sky are trying to get another ‘People’s Forum’ happening in Queensland for this Wednesday night.
What? So they can rig it for Abbott again?
As mentioned before, it smells like push-polling by News Ltd and Galaxy.
Saturday’s paper has a front-page picture of that nice Mr Rabbit, Sunday’s paper has a front-page headline that Labor is rooned, and a story on the inside that some young lass is gunna that for that nice Mr Rabbit, and an editorial sighing in sorrow that this newspaper is gunna support that nice Mr Rabbit.
It’s all designed to salve the conscience of anyone who voted for Labor last election that it’s okay to vote for that nice Mr Rabbit coz wouldn’t you know everyone else is doing it too (and here’s our poll to prove it).
Egg on their face? As if they care. They’ve done their job.
Paul, I’m with you. I think Labor would be mad to agree to it. IMO Galaxy/News Ltd did their dash with the last one letting Joe Scalzi’s son ask a question and I think a couple of other Young Liberals got to ask questions as well. “Undecided voters” my arse.
Indeed Jacques. Normally my attitude to the media is a bit like water off a ducks back. But they’vce actually made me very angry this election, and, indeed in the months running up to it. This is all too reminiscent of the anti-Whitlam bloodlust for my liking. I want Labor to win, just to rub egg in the bastards’ faces – the media that is.
The media is, in my opinion, the single most important political issue in this country, dwarfing all others because it controls how all others are perceived, or not perceived, and directing discussion along its own desired path and that of the interests which it reresents and belongs to.
Until there is fundamental change in the ownership and function of media in this country then we will never have anything representing a true democratic process.
If the COALition wins the election the situation will become even lots and lots worserer, if the ALP wins, well then ……?
Paul, As you know I lean more towards Labor than the Liberal Party but I’m finding it very difficult in pumping myself up for a Labor win. I almost don’t care who wins, not quite but almost. The mandatory internet filter, the disgraceful welfare quarantining in the NT and soon to come in for all unemployment & single parenting benefit recipients and the Howard-esque border protection stuff makes my blood boil.
Watching an ALP Senator (Trish Crossin) on Sky today outright lie in saying that in the NT during the “intervention” school attendance has gone up when the former editor of the National Indigenous Times refuted that to her face with the findings in a FAHCSIA report that it in fact has gone down just go off onto a different subject is appalling.
You’d expect this stuff from the Libs not Labor. They’re essentially giving people on the Left the middle finger so in that sense fuck ‘em. The only thing keeping me from being up in arms about an Abbott government is that at least the Greens will have the balance of power in the Senate.
“As mentioned before, it smells like push-polling by News Ltd and Galaxy….Sunday’s paper has …an editorial sighing in sorrow that this newspaper is gunna support that nice Mr Rabbit.”
Our News Ltd Sunday paper (i.e. the Herald Sun) had an editorial for that nice Ms Gillard, saying that on balance, they believe it’s in the best interests of Australians to reelect a Labor Government and that Mr Rabbit has not made the case for a change.
http://www.smh.com.au/opinion/politics/she-should-have-waited-instead-she-bolted-20100814-123zb.html
A fine example of make-work “journalism”. Apparently the author is:
“a Canberra-based writer and an award-winning political journalist”
Jeebus H, the guy writes like he’s on acid.